Excessive consumption of fructose can result in non-alcoholic fatty liver disease (NAFLD)


Excessive consumption of fructose – a sweetener ubiquitous in the American diet – can result in non-alcoholic fatty liver disease (NAFLD), which is comparably abundant in the United States. But contrary to previous understanding, researchers at University of California San Diego School of Medicine report that fructose only adversely affects the liver after it reaches the intestines, where the sugar disrupts the epithelial barrier protecting internal organs from bacterial toxins in the gut.

Developing treatments that prevent intestinal barrier disruption, the authors conclude in a study published August 24, 2020 in Nature Metabolism, could protect the liver from NAFLD, a condition that affects one in three Americans.

“NAFLD is the most common cause of chronic liver disease in the world. It can progress to more serious conditions, such as cirrhosis, liver cancer, liver failure and death,” said senior author Michael Karin, Ph.D., Distinguished Professor of Pharmacology and Pathology at UC San Diego School of Medicine.

“These findings point to an approach that could prevent liver damage from occurring in the first place.”

Fructose consumption in the U.S. has skyrocketed since the 1970s and the introduction of high fructose corn syrup (HFCS), a cheaper sugar substitute that is broadly used in processed and packaged foods, from cereals and baked goods to soft drinks.

Multiple studies in animals and humans have linked increased HFCS consumption with the nation’s obesity epidemic and numerous inflammatory conditions, such as diabetes, heart disease and cancer.

The U.S. Food and Drug Administration, however, currently regulates it similar to other sweeteners, such as sucrose or honey, and advises only moderation of intake.

The new study, however, defines a specific role and risk for HFCS in the development of fatty liver disease.

The ability of fructose, which is plentiful in dried figs and dates, to induce fatty liver was known to the ancient Egyptians, who fed ducks and geese dried fruit to make their version of foie gras,” said Karin.

“With the advent of modern biochemistry and metabolic analysis, it became obvious that fructose is two to three times more potent than glucose in increasing liver fat, a condition that triggers NAFLD. And the increased consumption of soft drinks containing HFCS corresponds with the explosive growth in NAFLD incidence.”

Fructose is broken down in the human digestive tract by an enzyme called fructokinase, which is produced both by the liver and the gut. Using mouse models, researchers found that excessive fructose metabolism in intestinal cells reduces production of proteins that maintain the gut barrier – a layer of tightly packed epithelial cells covered with mucus that prevent bacteria and microbial products, such as endotoxins, from leaking out of the intestines and into the blood.

“Thus, by deteriorating the barrier and increasing its permeability, excessive fructose consumption can result in a chronic inflammatory condition called endotoxemia, which has been documented in both experimental animals and pediatric NAFLD patients,” said the study’s first author Jelena Todoric, MD, Ph.D., a visiting scholar in Karin’s lab.

In their study, Karin, Todoric and colleagues from universities and institutions around the world, found that leaked endotoxins reaching the liver provoked increased production of inflammatory cytokines and stimulated the conversion of fructose and glucose into fatty acid deposits.

“It is very clear that fructose does its dirty work in the intestine,” said Karin, “and if intestinal barrier deterioration is prevented, the fructose does little harm to the liver.”

The scientists noted that feeding mice with high amounts of fructose and fat results in particularly severe adverse health effects.

“That’s a condition that mimics the 95th percentile of relative fructose intake by American adolescents, who get up to 21.5 percent of their daily calories from fructose, often in combination with calorie-dense foods like hamburgers and French fries,” Karin said.

Interestingly, the research team found that when fructose intake was reduced below a certain threshold, no adverse effects were observed in mice, suggesting only excessive and long-term fructose consumption represents a health risk. Moderate fructose intake through normal consumption of fruits is well-tolerated.

“Unfortunately, many processed foods contain HFCS and most people cannot estimate how much fructose they actually consume,” said Karin. “Although education and increased awareness are the best solutions to this problem, for those individuals who had progressed to the severe form of NAFLD known as nonalcoholic steatohepatitis, these findings offer some hope of a future therapy based on gut barrier restoration.”

Nonalcoholic fatty liver disease (NAFLD) consists of a benign form – hepatic steatosis (defined by the presence of lipid droplets within hepatocytes on histopathologic examination) – and the more perilous nonalcoholic steatohepatitis (nasH), which may result in cirrhosis and hepatic failure. NAFLDis currently the most common liver disease worldwide, both in adults and children.1

Considering that NAFLD was first reported in adults in 1980,2 and in children in 1983,3 the secular trend of NAFLD prevalence is both astounding and alarming. in the past 30 years, the prevalence and severity of NAFLDhas paralleled that of obesity, type 2 diabetes mellitus (t2Dm) and the metabolic syndrome, and a mech­ anistic association between NAFLD and these disorders has been proposed.4

Despite concerted research into the etiology and pathogenesis of naFlD, the causes of the disease remain unknown. understanding the pathol­ ogy of NAFLD and other chronic metabolic diseases, and how they relate to environmental changes that have occurred in the past 30 years is essential to preventing and treating NAFLDand the metabolic syndrome in the future.

NAFLD and the metabolic syndrome

NAFLD is a condition of aberrant lipid storage in hepatocytes.5 although intrahepatic lipid accumulation is related to lipid accumulation in adipocytes (that is, obesity),6,7 not all intrahepatic lipid accumulation can be explained by obesity alone. For instance, patients with lipodystrophy are not obese, but have high levels of intrahepatic lipid.8 intrahepatic lipid accumulation is, in fact, independent of Bmi,9 but more accurately reflects the existence of metabolic complications.10

The preva­ lence of the metabolic syndrome is 34% in american adults overall, but 53% in patients with hepatic steato­ sis, and even higher (88%) in those with nasH.11

The prevalence of both the metabolic syndrome and NAFLDincreases with age and with Bmi; at autopsy, intra­ hepatic lipid is identified in 36% of lean adults and 72% of adults with obesity,6 versus 5% of normal weight chil­ dren, 16% of overweight children, and 38% of children with obesity.12

Patients with NAFLD are more likely to have dys­ lipidemia and increased blood pressure than individuals of a similar weight without NAFLD.13,14

Furthermore, the degree of intrahepatic lipid predicts metabolic dys­ function even better than does the degree of visceral adipose tissue,15 and hepatic steatosis may precede the onset of the metabolic syndrome and its complications.16 NAFLD more accurately predicts the existence of insulin resistance than do the ATP III criteria, which were devel­ oped to diagnose the metabolic syndrome.17 in children18 and in adults,19 alanine aminotransferase (ALT) corre­ lates with both intrahepatic lipid, insulin resistance, and other components of the metabolic syndrome.18

The congruence of NAFLD and other manifestations of the metabolic syndrome have led the Japanese govern­ment to include a threshold level of alt as part of their definition of the metabolic syndrome.20

Risk factors for NAFLD

genetic risk factors

in a study of the correlation of alt with the metabolic syndrome, elevated levels of this enzyme were posi­ tively correlated with elevated serum lipid and glucose concentrations, blood pressure, and waist circumference in most geographic regions except asia,21 suggesting the existence of hereditary modulators and/or alternate pathogeneses of NAFLD in specific populations. NAFLD affects specific racial and/or ethnic groups preferentially22 and efforts have been made to find specific genetic pre­ dispositions for developing the disease.

An associa­ tion between the presence of the rs738409 allele of the patatin­like phospholipase domain­containing protein 3 (PNPLA3), which encodes a protein under metabolic control, and intrahepatic lipid has been documented. this allele is particularly frequent in latino individuals, who have the highest prevalence of NAFLD in the usa.22

Another allele in the same gene is associated with low hepatic fat content in african americans, who have the lowest risk of developing NAFLD . the role of PnPla3 in lipid processing is not known, but this protein may also affect other ectopic lipid depots, as visceral adipose tissue is related to intrahepatic lipid accumulation, irrespective of race or ethnicity.23

Metabolic risk factors

obesity­associated insulin resistance and subsequent hyperinsulinemia seem to be necessary though not suffi­ cient to precipitate the development of NAFLD .24  in a retrospective study of children with biopsy­confirmed nasH, 75% of them had fasting hyperinsulinemia, which predicted steatosis, inflammation and fibrosis.25 adolescents with both obesity and NAFLD had higher liver and skeletal muscle insulin resistance, as measured with hyperinsulinemic–euglycemic clamp techniques, than those who were obese without NAFLD .13 

 Changes in the concentrations of adipocytokines derived from visceral adipose tissue, including increases in leptin, resistin, tumor necrosis factor (tnF), and interleukin 6 concentrations, and decreases in adiponectin concentra­ tions, are all associated with NAFLD and insulin resis­ tance.26 whether these changes in cytokine levels are markers or causes of intrahepatic lipid, or whether all of these factors are driven by an additional, yet unidentified cause is not clear.

in addition, hyperglycemia may also exacerbate NAFLD . in a Japanese study, the prevalence of steatosis among lean adults was 62% in those who had been newly diagnosed as having t2Dm and 43% in those with impaired fasting glucose levels, but only 27% in individuals with normal fasting glucose levels.27

Dietary risk factors

exposure to environmental factors, especially dietary factors, is also likely to contribute to the generation of intrahepatic lipid.28,29 some studies have suggested that specific dietary fats, such as trans­unsaturated fats, contribute to hepatic steatosis.30,31

Conversely, mono­ unsaturated lipids such as oleic acid (the primary com­ ponent of olive oil),32 linoleic acid,33 or n–3 fatty acids34 decrease accumulation of intrahepatic lipid and improve postprandial triglyceride levels, possibly by increasing peroxisomal activity, which reduces damage by reac­ tive oxygen species (ros).35

Another dietary factor that probably contributes to hepatic steatosis, and the focus of this review, is the monosaccharide fructose. in case­ controlled studies, sugar­sweetened beverage consump­ tion was associated with hepatic steatosis, and this association was independent of the degree of obesity.36,37

In other case­controlled studies, total fructose consump­ tion was associated with NAFLD in general, and nasH in particular.38,39 micronutrient insufficiencies associated with the consumption of sugar­sweetened beverages40 may aggravate their toxicity in producing NAFLD .

The ‘two-hit’ theory of NAFLD

Histologically, NAFLD is similar to alcoholic fatty liver disease (aFlD). mechanistically, although NAFLD is thought to represent a continuum of hepatic insult, it is probably the result of two distinct but related ‘hits’ to the hepatocyte.41  These two steps of hepatic injury are similar to those that are caused by ethanol and seen in AFLD.42  

The first ‘hit’ is the development of intra­ hepatic lipid and hepatic steatosis owing to an imbal­ ance of normal hepatic lipid metabolism, which results in either excessive lipid influx, decreased lipid clearance, or both. at this point, steatosis is potentially reversible and does not necessarily lead to permanent hepatic injury.43

The second, less common, but more virulent ‘hit’, which occurs in 5% of individuals with steatosis, is a concomi­ tant inflammatory process that presumably results from oxidative stress, lipid peroxidation and cytokine action. the resulting lobular inflammation leads to ballooning degeneration and perisinusoidal fibrosis, which promote apoptosis and hepatocellular death, with resulting scar­ ring and progression to NASH.44

Approximately 25% of patients with nasH will develop portal fibrosis, complet­ ing the process that leads to cirrhosis.45 once inflamma­ tion has started, progression to cirrhosis may only take a few years.46

Understanding the etiology of these two ‘hits’ is essential in the prevention and treatment of NAFLD .

The first ‘hit’: hepatic steatosis

Intrahepatic lipid accrues when the rate of hepatic lipid influx, by fatty acid import or de novo synthesis of fatty acids, exceeds the rate of hepatic lipid clearance, by fatty acid catabolism or lipoprotein export (Figure 1).47,48 the following five mechanisms that influence lipid influx or clearance may lead to hepatic steatosis.

Increased ingestion of dietary fat

the role of dietary fat in the pathogenesis of NAFLD remains controversial.

Ad lib highfat liquid feeding to rats generates intrahepatic lipid,49 whereas ad lib feeding of lipidrich chow does not.50 in humans, ingestion of dietary fat influences the accumulation of intrahepatic lipid,51 but stable isotope studies have demonstrated that up to only 15% of intrahepatic lipid is derived from dietary fat.52 lastly, lowcarbohydrate diets, which are
otherwise rich in protein and fat, are frequently used as treatment for NAFLD.53

Increased influx of free fatty acids

Free fatty acids that result from lipolysis of subcutaneous or visceral adipose tissue depots circulate to the liver, and may contribute to intrahepatic lipid accumula­ tion.54,55  For instance, in t2Dm complicated by obesity and chronic inflammation, overexpression of adipose tissue cytokines can foment insulin resistance in the liver and adipose tissue, in part by serine phosphorylation of insulin receptor substrate 1 (irs­1).

Insulin resistance in adipocytes leads to failure of insulin­mediated suppres­ sion of hormone­sensitive lipase (Hsl) and release of free fatty acids from adipose tissue (especially visceral adipose) into the circulation.56 release of free fatty acids from visceral adipose tissue is particularly problem­ atic, as their first circulatory pass is through the liver.

Inflammation also increases tnF­mediated upregula­ tion of hepatic fatty acid translocase,57 which, coupled with increased influx of free fatty acids, leads to steato­ sis.58 other conditions characterized by increased lipo­ lysis and insulin resistance, however, do not necessarily result in steatosis.

For instance, patients with poorly con­ trolled type 1 diabetes mellitus have increased lipolysis and insulin resistance, but have minimal intrahepatic lipid accumulation,59 presumably because of enhanced β­oxidation of fatty acids to ketones, which accelerates hepatic lipid clearance for energy usage by the rest of the body.

Figure 1 | Pathways of hepatic lipid metabolism. Three main pathways contribute to production of intrahepatic lipid (lipid input): import of lipoprotein triglyceride by the LDL receptor; lipolysis, which is the source of free fatty acids that are imported    into the liver; and use of carbohydrate as an energy source by de novo lipogenesis. Two pathways contribute to clearance of intrahepatic lipid (lipid output): complete β-oxidation of intrahepatic lipid, which leads to carbon dioxide (CO2) production, or incomplete β-oxidation, which leads to production of ketones (which are then exported for energy use by the rest of the body); and lipid export, usually by conjugating intrahepatic lipids with apolipoproteins to form VLDLs. Perturbation of any of these pathways, such as in T2DM, kwashiorkor, Reye syndrome, or hypobetalipoproteinemia, can result in accumulation of intrahepatic lipid and lead to hepatic steatosis. Abbreviation: T2DM, type 2 diabetes mellitus.

Increased de novo lipogenesis

De novo lipogenesis is accentuated more by excess dietary carbohydrate than by excess dietary fat.60 For example, when total energy intake from carbohydrate exceeds total energy expenditure, hepatic de novo lipogenesis is increased 10­fold.61 similarly, the rate of de novo lipogenesis associated with a high­carbohydrate diet is 27­fold greater than that associated with a low­ carbohydrate diet in the fasting state, and fourfold greater in the fed state.62

For example, patients with kwashiorkor (protein mal­ nutrition despite adequate caloric intake) who are fed a maize diet, rich in carbohydrate,63 manifest severe hepatic steatosis, more so than patients with other malnutrition syndromes.64

Excess accumulation of metabolites generated byde novo lipogenesis is seen in human and rat models of hepatic steatosis.65,66 For instance, labeled isotope studies in individuals with obesity and steatosis have shown that 26.1% of the intrahepatic lipid pool is gen­ erated by de novo lipogenesis.52

On a typical high­fat diet, lean individuals exhibited less than 3% (1–2 g per day) of carbohydrate converted to free fatty acids by de novo lipogenesis,67,68 but this percentage was >10% in individuals with obesity and insulin resistance.69

Impaired hepatic β-oxidation of fatty acids

β­oxidation, (the sequential removal of two­carbon fragments for ketone production or energy genera­ tion within mitochondria) is the main route for the metabolism of long­chain fatty acids under normal physiological conditions.70  

This process is accomp­ lished by mitochondrial trifunctional protein, which catalyzes three separate reactions to liberate these two­ carbon fragments from acyl­Coa, shortening them in the process.

This process, however, may be disrupted at several key enzymatic stages. Dysfunction of this protein may cause abrupt and massive hepatic failure with steatosis,71 similar to the pathophysiology and liver pathology in patients with reye syndrome.72  

The transesterification and import of fatty acids into the mitochondrial matrix for the process of β­oxidation is achieved through covalent binding to carnitine, a mitochondrial carrier molecule. each fatty acid that is transported requires the cleavage of the carnitine–fatty acid complex, mediated by the enzyme carnitine O­palmitoyltransferase 1 (CPt1).

Regeneration of carni­ tine by CPt1 is the key rate­limiting and regulatory step in the process of β­oxidation. malonyl­Coa, formed by dimerization and subsequent decarboxylation of acetyl­ Coa (the first step in fatty acid synthesis) is also a steric

inhibitor of CPt1.73,74  experimental suppression of malonyl­Coa formation in rats counteracts the adverse effects of NAFLD and insulin resistance.75 malonyl­Coa concentrations are increased when excessive citrate, the primary substrate of de novo lipogenesis, is generated beyond the oxidative capacity or needs of the liver; thus, de novo lipogenesis and defective β­oxidation are linked, and both promote intrahepatic lipid accumulation.

in lipid­engorged hepatocytes, several vicious cycles involving tnF, ros, peroxynitrite, and lipid peroxida­ tion products partially block the flow of electrons in the respiratory chain. the over­reduction of upstream respiratory chain complexes increases mitochondrial ros and peroxynitrite formation. oxidative stress increases the release of lipid peroxidation products and cytokines, which contribute to the development of nasH.76 impaired lipid β­oxidation seems to exert only a minor influence in the development of NAFLD in humans,77 but might arguably be enhanced in persons with inadequate diet­derived and endogenous antioxidant defense mechanisms.78

Impaired triglyceride export

Hepatocytes esterify excess free fatty acids into tri­ glycerides, which are packaged with apolipoprotein B­100 (apo B­100) by microsomal triglyceride transfer protein (mttP), and exported as VLDLS. VLDL pro­ duction in patients with nasH is decreased compared with that in healthy controls,79 and this decrease could accentuate the accumulation of triglycerides in the liver. However, whether decreased VLDL production is a cause or a result of nasH is not clear.

Patients with hypo­ betalipoproteinemia—an autosomal recessive disease caused by a defect in mttP and characterized by fat malabsorption, low plasma cholesterol levels at a young age, and progressive neurologic degenerative disease— have severe hepatic steatosis.

Hepatic lipid export is diminished in this condition, yet peripheral clearance is normal, so affected individuals have markedly dimin­ ished serum triglyceride levels.80 Patients with nasH, however, have hypertriglyceridemia due to impaired clearance as well as defective lipid export.

Dietary factors and hepatic steatosis

each of these five processes that control hepatic lipid influx and/or clearance can be perturbed sufficiently in humans to increase intrahepatic lipid and promote hepatic steatosis, contributing to—but not fully explain­ ing—the global increase in prevalence and severity of NAFLD . The inexorable rise in daily calorie intake seen in the past 30 years in industrialized countries may also have a role in this increase. in the usa, adult men have increased caloric intake by 187 kcal per day, women by 335 kcal per day, and adolescents of both sexes by 275 kcal per day.81  

The prevalence of both NAFLD and the metabolic syndrome have increased dramatically worldwide with the global export of the western diet. whether the excess of energy intake is the main factor that underlies the NAFLD epidemic, whether specific macronutrient and/or micronutrient components of the diet are involved, or whether both factors have a role has not been fully elucidated.

Dietary fat consumption

Despite documented increases in daily caloric intake, total dietary fat intake has remained stable in the last 30 years (5.3 g per day decrease in men, 5 g per day increase in women), while the percentage of calories ingested from saturated fat has decreased.82 although high­fat feeding can induce overnutrition and intra­ hepatic lipid in experimental animal and human models, it does so only with concomitant carbohydrate inges­ tion.50,51,83

Furthermore, long­term voluntary consump­ tion of an increased percentage of fat, as is seen in the atkins diet, did not result in increased caloric intake, and did not increase the risk of obesity or NAFLD .84,85 in a 2­year intervention study, participants on either the fat­ liberalized mediterranean (monounsaturated fats) or the atkins (all fats) diet without imposed caloric restriction, demonstrated comparable weight loss and decreased cir­ culating fasting insulin and triglyceride concentrations.86

Indeed, high­fat, low­carbohydrate diets seem to amelio­ rate elevated liver enzyme levels and decrease metabolic disease risk.53 therefore, the quantity of ingested dietary fat does not seem to be the cause of NAFLD , although qualitative fat composition may be of concern.

Trans-unsaturated fat consumption

with the advent of Crisco, the first commercial trans­ unsaturated or ‘partially hydrogenated’ fat, in 1911, various processed food products, especially margarines and baked goods, have included trans­unsaturated fat in order to improve shelf life.

Due to the trans­isomerization of the double bond in these fatty acids, bacteria cannot digest them, which prevents food from becoming rancid. unfortunately, our mitochondria (derived evolutionarily from bacteria) also cannot digest trans­unsaturated fats, which, therefore, do not undergo mitochondrial β­oxidation,87 and increase the risk of hepatic steatosis. experimental feeding of trans­unsaturated fats (termed the ‘western lifestyle Diet’) to animals is associated with rapid accumulation of intrahepatic lipid.31

Consumption of trans­unsaturated fats peaked in the 1960s, but since 1988, owing to the recognized association of trans­ unsaturated fat with cardiovascular disease, the percent­ age of calories from trans­unsaturated fat consumed in the western diet has been gradually decreasing. a reduc­ tion from 3.0% in 1982 to 2.2% in 2002 in total energy derived from trans­unsaturated fats was documented in the minnesota Heart survey.88 thus, while consumption of trans­unsaturated fats could mechanistically underlie the development of NAFLD and the metabolic syndrome, consumption trends are temporally disparate with the current epidemic of these diseases.

Carbohydrate consumption

when it comes to which macronutrient promotes intra­ hepatic lipid, who knows better than the French? in recognition that dietary carbohydrate is more steato­ genic than fat, the prototypical fatty liver, ‘paté de foie

gras’, is made by force­feeding ducks or geese with a diet of maize (corn), wheat, and soya cake—a macronutrient breakdown that favors carbohydrate (47.9%) over fat (2.1%).89 in contrast to the results of dietary fat intake, carbohydrate overfeeding in humans results in exces­ sive weight gain and hepatic steatosis in a short period of time.

For example, in the ‘Guru walla’ overfeeding para­ digm, consumption of a diet of 7,000 kcal per day that consists of 70% carbohydrate and 15% fat, and is mainly composed by sorghum (a kind of maize), leads to hepatic steatosis within a few weeks.90 Food consumption survey data from the us Department of agriculture, reported by the national Health and nutrition examination survey program between 1971 and 2004, also indicate that the observed increase in total energy intake in the us popula­ tion is accounted for almost completely by carbohydrate consumption, with a 67.7 g increase per day in men and a 62.4 g increase per day in women within that time frame.82 an excess in dietary carbohydrate consumption has been reported specifically in patients with nasH.91

Secular trends in fructose intake

the predominant carbohydrate responsible for the rise in caloric consumption associated with the typical western diet is the monosaccharide fructose, which is consumed either as sucrose (50% fructose) or as high­fructose corn syrup (42% or 55% fructose).

Before 1900, us americans consumed approximately 15 g of fructose per day (4% of total calories), mainly through fruits and vegetables. Before world war ii, fructose intake had increased to 24 g per day (5% of total calories); by 1977, it was 37 g per day (7% of total calories); and by 1994, 55 g per day (10% of total calories). adolescents today consume over

72.8 g per day (12.1% of total calories) of fructose;92 20% of teenagers consume 25% or more of their total calories as fructose.93  thus, fructose consumption has increased fivefold over the last century and more than doubled in the last 30 years. Food disappearance data from the economic research service (ers) of the us Department of agriculture support this secular trend. although the ers documents partially decreased sucrose intake per capita, the total annual consumption of caloric sweeteners per capita has increased from 33 kg to 43 kg in 30 years.94

Although the presence of high­fructose corn syrup in soft drinks has received most of the atten­ tion,95,96 high fruit juice intake has also been associated with childhood obesity.97 Currently, americans consume sugar at a rate of 66.8 kg per year (180 g per day), half of which is fructose.

Fructose and the metabolic syndrome

many investigators have implicated fructose in the pathogenesis of the metabolic syndrome40,98–105 and NAFLD .39,106,107 the liver is the principal site of fructose metabolism, as it possesses the fructose­specific Glut5 transporter.108  

Although adipocytes possess Glut­5 mrna and protein, the level of this transporter in adipose tissue is quite low.109 the kidney and small intestine also possess Glut­5 transporters, but their function is to transport fructose molecules across their lumena, either

for urinary excretion (to eliminate any systemic fructose that escapes hepatic clearance) or for release into the portal circulation, which passes directly to the liver. the hepatic metabolism of fructose is very different to that of glucose in that it is insulin independent, bypasses the process of glycolysis, and increases de novo lipogenesis to a greater extent. indeed, the hepatic metabolism of fructose is more reminiscent of that of ethanol.110 similar to ethanol, fructose can induce each of the phenomena associated with the metabolic syndrome (Figure 2).


Fructose is phosphorylated by fructokinase, which uses atP as the phosphate donor, depleting the hepatocyte of intracellular atP. the scavenger enzyme amP deami­ nase 1 reclaims additional phosphates from aDP, and in the process generates the waste product uric acid. uric acid acts within vascular smooth muscle to inhibit endothelial nitric oxide synthase and resultant nitric oxide production, which promotes hypertension.100

Our group has shown that sugar­sweetened beverage consumption positively correlates with uric acid and blood pressure levels in children,111 while others have documented this association in adults.112 Furthermore, the uric acid inhibitor allopurinol can reduce blood pressure in adolescents113 and adults with obesity.114

Hepatic steatosis

owing to the excess substrate load, excess mitochondrial acetyl­Coa is formed, exceeding the ability of the tri­ carboxylic acid (tCa) cycle to metabolize it. the excess acetyl­Coa is converted to citrate, exits into the cytosol via the citrate shuttle, and serves as the substrate for de novo lipogenesis.

Acetyl­Coa dimerizes and is decarboxy­ lated to form malonyl­Coa, which inhibits mitochon­ drial β­oxidation. triglycerides newly formed by de novo lipogenesis115 can overwhelm the lipid export machin­ ery and precipitate in the liver, forming intrahepatic lipid and leading to hepatic steatosis.

Hepatic insulin resistance

Fructose­1­phosphate activates dual­specificity mitogen­ activated protein kinase kinase 7 (mKK7),116 which stimulates the hepatic enzyme mitogen­activated protein kinase 8 (maPK8).117  this kinase is thought to be the bridge between hepatic metabolism and inflamma­ tion.118 Furthermore, the intermediate diacylglycerol, which accumulates during de novo lipogenesis acti­ vates hepatic protein kinase C ε type (PKCε).119 Both maPK8 and PKCε trigger serine phosphorylation and subsequent inactivation of irs­1, which leads to hepatic insulin resistance.120–123

Figure 2 | Hepatic fructose metabolism. Fructose induces: substrate-dependent phosphate depletion, which increases uric acid and contributes to hypertension through inhibition of endothelial nitric oxide synthase and reduction of NO (green); excess formation of citrate, which serves as the substrate for de novo lipogenesis (orange); excess formation of malonyl- CoA, which inhibits β-oxidation (red); hepatic lipid droplet formation and steatosis; activation of MAPK8 and PKCε, which contributes to serine phosphorylation of IRS-1 and hepatic insulin resistance, which in turn promotes hyperinsulinemia and influences substrate deposition into fat (yellow); export of free fatty acids, which leads to VLDL formation and muscle insulin resistance (light blue); increased synthesis of FOXO1, which promotes gluconeogenesis and hyperglycemia (pink); and central nervous system hyperinsulinemia, which antagonizes central leptin signaling and promotes continued energy intake. Abbreviations: ACL, ATP-citrate lyase; ACC1, acetyl-CoA carboxylase 1; apo B, apolipoprotein B-100; ChReBP, carbohydrate response element binding protein; CPT-1, carnitine O-palmitoyl transferase 1; FAS, fatty acid synthase; DAG, diacylglycerol; FOXO1, forkhead box protein O1; GLUT-5, solute carrier family 2, facilitated glucose transporter member 5; IRS-1, insulin receptor substrate 1; LPL, lipoprotein lipase; MAPK8, mitogen-activated protein kinase 8; MKK7, mitogen- activated protein kinase kinase 7; MTTP, microsomal triglyceride transfer protein; NO, nitric oxide; PFK, 6-phosphofructokinase; PGC-1β, peroxisome proliferator-activated receptor γ coactivator 1β; Pi, inorganic phosphate; PKCε, protein kinase C ε type; PP2A, protein phosphatase 2a; pSer-IRS-1, serine phosphorylated IRS-1; SReBP-1c, sterol regulatory element binding protein-1c; TCA, tricarboxylic acid

Dyslipidemia and muscle insulin resistance

Free fatty acids are also formed, which, when packaged as triglycerides into heavily fat­laden VLDLs, are cleared with low efficiency, causing dyslipidemia and augmenting the risk of cardiovascular disease.115,124 excess circulating lipid is also taken up by skeletal muscle to form intramyocellular lipid, which leads to muscle insulin resistance.125,126

Hyperglycemia and t2DM

Fructose, a gluconeogenic precursor, increases synthe­ sis of the forkhead box protein o1 (FoXo1).127 Hepatic insulin resistance, made worse by elevated fructose concentrations, prevents the phosphorylation of FoXo1, which allows this protein to enter the nucleus and induce the transcription of enzymes that promote gluconeo­ genesis. increased hepatic glucose output foments hyper­ glycemia, and is likely to contribute to the development of t2Dm.


Fructose also contributes to increased food consump­ tion and obesity. Direct effects of fructose on the central nervous system (Cns) include stimulation of hormones

that stimulate appetite and reward,128–130 and reduction of hypothalamic malonyl­Coa levels, which results in increased amP kinase concentrations, driving further food intake.131 indirect effects of fructose on the Cns include hypertriglyceridemia, which reduces leptin transport across the blood–brain barrier,132 and hyper­ insulinemia, which blocks the leptin signal transduction pathway, resulting in a sense of starvation, again driving further food intake.133


Fructose consumption may also contribute to bacterial overgrowth and increased intestinal permeability.39,105 animal and human studies suggest that NAFLD is associ­ ated with small intestinal bacterial overgrowth,134 which is linked to circulating endotoxinemia and inflamma­ tory cytokines. the presence of these factors in the blood has also been reported in patients with aFlD.135

Translocation of bacterial endotoxins through a leaky gut lumen to the portal blood increases the exposure of the liver to inflammation and injury. the gut microflora, therefore, may be a key link between fructose feeding, plasma endotoxin levels, and systemic and hepatic inflammation associated with the metabolic syndrome.

Similarities between fructose and ethanol

Promotion of de novo lipogenesis

in contrast to the hepatic conversion of glucose pri­ marily to glycogen, a fructose bolus is metabolized by glycolysis directly to pyruvate (Figure 2). this process is accentuated in individuals with insulin resistance and/or obesity.52,69,136 thus, in response to a fructose bolus, a large volume of acetyl­Coa is generated to enter the hepatic mitochondrial tCa cycle.

This cycle, however, has a rela­ tively fixed maximum velocity to deal with its substrate, modifiable by exercise, cold, altitude, and concentrations of thyroid hormone. when the liver mitochondria are not able to metabolize the entire fructose­derived acetyl­ Coa substrate excess, any extra exits the mitochondria into the cytosol in the form of citrate via the citrate shuttle.137

In the cytosol, carbohydrate response element binding protein (ChreBP) activates the enzymes respon­ sible for de novo lipogenesis,138,139 which transforms the citrate, through malonyl­Coa, to generate fatty acyl­Coa. Furthermore, fructose­1­P stimulates per­ oxisome proliferator­activated receptor γ coactivator 1β (PGC­1β), a transcriptional co­activator for sterol regu­ latory element binding protein 1c (sreBP­1c).

sreBP­1c further increases the activity of the enzymes involved in de novo lipogenesis, which leads to the generation of even more fatty acyl­Coa.140  the majority of the fatty acyl­ Coa is packaged into VLDL for export, but a proportion accumulates instead as lipid droplets in the hepatocyte.29

This double activation of de novo lipogenesis, mediated through both the carbohydrate (ChreBP) and lipogenic (sreBP­1c) transcription factor pathways, seems to be exclusive to fructose, and probably explains its ability to generate intrahepatic lipid very rapidly, as it relies upon a substrate­driven effect that does not need insulin stimulation.

Qualitatively and quantitatively, the genera­ tion of acetyl­Coa, increased de novo lipogenesis,141 and increased intrahepatic lipid formation142 that result from excessive fructose intake are similar to those seen with ethanol consumption.

Figure 3 | Association between sugar-sweetened beverage consumption and serum alanine aminotransferase in a population of children seeking obesity treatment at the University of California, San Francisco.164 These two variables are more strongly correlated in white children than in African American children.11

Inhibition of β-oxidation

Fructose increases the intracytosolic citrate available for de novo lipogenesis. the process of de novo lipogenesis involves three enzymes: atP­citrate lyase, which recon­ verts citrate to acetyl­Coa; acetyl­Coa carboxylase 1 (aCC1), which dimerizes and decarboxylates acetyl­ Coa into malonyl­Coa; and fatty acid synthase, which adds further two­carbon fragments to malonyl­Coa to elongate it to form fatty acyl­Coa. these two­carbon fragments are generated in the mitochondria, after

regeneration of carnitine by CPt1 after transesterifica­ tion and import of fatty acids into the mitochondrial matrix. malonyl­Coa, which is generated by de novo lipogenesis from carbohydrate, inhibits CPt174 and pre­ vents β­oxidation,143 which contributes to intrahepatic lipid accumulation. in animal models, defective fatty acid oxidation and hepatic insulin resistance result in steatosis.144  

These same phenomena occur in animals when ethanol is the energy substrate.145,146 in ethanol­ fed rodents, β­oxidation is blocked by increased activity of aCC1, which leads to increased levels of malonyl­Coa and decreased activity of CPt1.147

Suppression of hepatic lipid export

the principal exit strategy of intrahepatic lipid is its export from the liver as VLDL. the synthesis of VLDL depends on mttP for transfer of lipids to the apo B­100 protein, which is necessary for its correct folding before export. Peroxisome proliferator­activated receptor α (PParα) regulates mitochondrial and peroxisomal fatty acid oxidation, and stimulates hepatic mttP expression, coordinating energy stores and facilitating the normal excretion of hepatic lipid for peripheral processing.148 Both fructose149 and ethanol148,150 reduce hepatic PParα activity, which results in downregulation of mttP.151

Hepatic triglyceride availability is the major determi­ nant of VLDLsecretion rate, but mttP activity seems to determine VLDL size, which determines the rate of clearance of these molecules in the plasma.152 elevated circulating VLDL concentrations in a hamster model of insulin resistance induced by high fructose feeding was associated with VLDL overproduction;153,154 while decreased clearance of triglyceride­rich VLDL has been demonstrated in fructose­fed rat models.151,155 similarly, suppression of mttP activity by ethanol increases VLDL production but reduces lipid export machin­ ery.156  triglyceride­enriched VLDL levels are typically elevated in patients with alcoholism.157

Treatment with a PParα agonist leads to upregulation of mttP and increased vlDl export and turnover, and alleviates the intrahepatic lipid accumulation of both naFlD158 and aFlD.148,150

Figure 4 | Molecular renditions of glucose and fructose. a | Glucose in the linear, chair (Haworth), and space-occupying projections. b | Fructose in the linear, chair (Haworth), and space-occupying projections. In the linear form, both substrates possess a reactive aldehyde or ketal moiety, which can bind nonenzymatically to freely available amino groups of proteins. At body temperature and pH, however, the chair form of glucose predominates. This conformation is a glucopyranose (6-membered ring), with equatorial hydroxyl groups and is molecularly stable, which limits its protein reactivity. The chair form of fructose is a fructofuranose (5-membered ring) with two axial hydroxymethyl groups that exert allosteric and ionic forces on the unstable furanose ring, which favors the linear form. Thus, at body temperature and pH, the majority of fructose exists in the linear form and is more reactive to proteins than glucose.

Chronic ethanol159 and fructose160 feeding in rat and hamster models respectively, are also associated with increased intestinal delivery of apolipoprotein B­48 (apo B­48). apo B­48 is a protein unique to chylomicrons. it is only generated in the small intestine, where an alter­ natively spliced stop codon truncates the apo B­100 mrna transcript.

Chylomicrons containing apo B­48 delivered to the liver are metabolized into triglyceride­ rich remnant lipoproteins that, unlike VLDLs (which contain apo B­100) and their LDL metabolites, cannot be cleared by the LDL cell­surface receptor that mediates endocytosis of cholesterol­rich apo-­B-­100 ­containing lipoproteins in all nucleated cells.

This process contrib­ utes to enhanced circulation time of triglyceride­rich proteins, hypertriglyceridemia, and atherogenicity. while the specific role of fructose or ethanol in human apo B­48 synthesis has not been examined, increased production of intestinally derived particles containing apo B­48 occurs in humans with hyperinsulinemia and insulin resistance.161

Intrahepatic lipid accumulation

In mouse models, fructose overfeeding results in rapid development of intrahepatic lipid, hypertriglyceridemia, and insulin resistance.162,163 our group has shown similar correlative and causative phenomena in humans. analysis of data from the weight assessment for teen and Child Health (watCH) Clinic at the university of

California, san Francisco, demonstrated that daily sugar­ sweetened beverage consumption assessed by recall (in kcal per day) correlated with alt concentrations in white children, although this correlation was smaller in african american children164 perhaps owing to genetic differences in modulators of lipid metabolism, such as PNPLA3165 or PPARα (Figure 3).166

Although alt concentrations are an imperfect measure of NAFLD ,167 serum levels of this protein correlate, nonetheless, with intrahepatic lipid accumulation, especially in children.168 a similar relationship between consumption of sugar­ sweetened beverages and NAFLD has also been shown in adults.33,36

Moreover, in a crossover, isocaloric feeding study in adults performed by our group, in which diets rich in fructose or complex carbohydrates were com­ pared, fructose feeding increased intrahepatic lipid accumulation by 38% within 8 days, as measured by magnetic resonance spectroscopy.169 

These results indi­ cate that, even in an isocaloric state, fructose is more likely to overwhelm the hepatic lipoprotein packaging machinery than other dietary factors, which results in accumulation of intrahepatic lipid, hepatic steatosis, and elevation of alt.

These associative and mechanistic data from animals and humans point to excessive fruc­ tose consumption as a proximate cause of NAFLD . once fructose­induced obesity and insulin resistance are estab­ lished, other factors, such as adipocytokines and intesti­ nal endotoxins, may also increase intrahepatic lipid levels and exacerbate hepatic dysfunction.

The second ‘hit’: inflammation

Fructose is different from glucose in yet another way, in that it can promote hepatocellular damage. molecularly,

Figure 5 | Generation of reactive oxygen species by fructose or ethanol. Fructose first forms an intermediate Schiff base with the ε-amino group of lysine, which then spontaneously hydrogenates to form an irreversible Heyns rearrangement product (hydroxyamide linkage or fructose adduct), through the ‘Maillard reaction’. each protein fructosylation generates
one superoxide radical (O2 –), which must be quenched by an antioxidant, such as glutathione. Alternatively, ethanol is metabolized by alcohol dehydrogenase 1B (ADH1B) to NADH and acetaldehyde. Acetaldehyde, through a similar Maillard reaction, forms acetaldehyde adducts, with generation of superoxide radicals, which must also be quenched by antioxidants. In the absence of adequate antioxidant capacity, production of reactive oxygen species leads to peroxidation, hepatocellular damage, necroinflammation (nonalcoholic steatohepatosis), fibrosis, and ultimately cirrhosis.

glucose is found in two steroisomeric forms: the linear aldehyde form, and the glucopyranose (6­membered ring) form (Figure 4). the aldehyde form of glucose is highly reactive with ε­amino groups of lysine. the nonenzymatic exothermic reaction between these factors leads to protein glycation,170 which is termed the maillard or ‘browning’ reaction; this is the reason why bananas become brown with time.

However, at 37 °C and pH 7.4, the ring form is molecularly stable and nonreactive, owing to its 6­membered gluco­ pyranose and equatorial hydroxyl groups. in these conditions, 80% of glucose is thought to remain in the ring form. Fructose is also found in two stereoisomeric forms: the linear ketal form, and the fructofuranose

(5­membered ring) form. the latter has two axial (abutting) hydroxymethyl groups, which exert allo­ steric and ionic forces to the unstable furanose ring and drive it toward the linear form. thus, at body tem­ perature and pH, 80% of fructose is thought to exist in the linear form, with a reactive ketal group. this difference explains why nonenzymatic fructosyla­ tion is seven times more rapid than protein glyca­ tion with glucose as the substrate.171,172

Each protein fructosylation reaction releases a superoxide radical (Figure 5).173 Fructose generates 100 times more ros than glucose,174,175 which, if not quenched by an anti­ oxidant (in the case of liver, glutathione), can promote hepatocellular damage.

Although difficult to demonstrate in humans, this phenomenon is easily recapitulated in vitro. in one study in mouse lymphoma cells, fructose induced non­ enzymatic fructosylation and Dna damage, which resulted in internucleosomal cleavage and apoptosis.176

In a study of hepatocytes in monolayer culture, incuba­ tion with fructose yielded no direct damage.177 However, when these hepatocytes were preincubated with sub­ lethal doses of hydrogen peroxide, their ros­quenching ability was extinguished.

Incubation with fructose was as toxic as that with other organic aldehydes and caused hepatocellular death. these experiments suggest that in a susceptible redox environment fructose can act as a direct hepatotoxin.

a similar phenomenon was also demonstrated in vivo. the methionine­choline­deficient diet in rats generates hepatic steatosis and nasH within 3 weeks.178 methionine is the source of the methyl group for gluta­ thione synthesis, and choline is a building block of phosphatidylcholine, a required component of VLDL particle assembly.179

Deficiency in methionine and choline promotes intrahepatic lipid accumulation and ros­mediated damage. interestingly, if the energy substrate for the methionine­choline­deficient diet is starch (glucose), no intrahepatic lipid formation or hepatocelluar death ensues; however, if the substrate is sucrose (glucose plus fructose), the liver undergoes massive steatosis, apoptosis, necrosis, and fibrosis.180

these data suggest that large volumes of fructose, in combination with micronutrient insufficiencies, can impair glutathione and possibly other hepatic anti­ oxidant reserves,181 contributing to the second ‘hit’ of hepatic damage in NAFLD , and the evolution of hepatic steatosis to nasH.

Although not yet demonstrated in humans, this hypothesis is particularly attractive, as several studies have shown that children with obesity, especially sugar­sweetened beverage drinkers, despite being overnourished are also micronutrient mal­ nourished.182,183 in this sense, hepatic inflammation caused by fructose is again reminiscent of that caused by alcoholism, in which micronutrient deficiencies are also commonplace.184 

The hepatic conversion of ethanol to acetaldehyde by the enzyme alcohol dehydrogenase 1B, which results in the generation of glycoaldehyde radi­ cals, is akin to the conversion of fructose to carbo­ nyl metabolites.146

Glutathione­dependent hepatic detoxification processes are impaired under conditions of nutritional deficiency,185 and improved with provi­ sion of nutrients that serve as methyl donors.186 in the absence of adequate nutritional substrate, we surmise that these toxins generate a flux of ros that can exceed antioxidant reserves.

Hepatic ros may be generated through other mechanisms as well, such as oxidation of saturated fat.187 ros formation, either directly from fructose, ethanol or fat, or indirectly from mitochon­ drial dysfunction caused by defective β­oxidation, is a likely initiator of endoplasmic reticulum stress106 and the unfolded protein response,123,188 both of which may also contribute to the pathogenesis of NAFLD and the metabolic syndrome.189

REFERENCE LINK : DOI: 10.1038/nrgastro.2010.41 · Source: PubMed

More information: Jelena Todoric et al. Fructose stimulated de novo lipogenesis is promoted by inflammation, Nature Metabolism (2020). DOI: 10.1038/s42255-020-0261-2


Please enter your comment!
Please enter your name here

Questo sito usa Akismet per ridurre lo spam. Scopri come i tuoi dati vengono elaborati.