Non-alcohol fatty liver disease: Balancing supply and utilization of triglycerides.

Abstract

Purpose of review:

Non-alcoholic fatty liver disease (NAFLD) is defined as the abnormal accumulation of lipids in the liver, called hepatic steatosis, which occurs most often as a concomitant of the metabolic syndrome. Its incidence has surged significantly in recent decades concomitant with the obesity pandemic and increasing consumption of refined carbohydrates and saturated fats. This makes a review of the origins of NAFLD timely and relevant.

Recent findings:

This disorder, which shares histologic markers found in alcoholic fatty liver disease, was named NAFLD to distinguish it from the latter. Recently however, the term Metabolic Associated Fatty Liver Disease, or MAFLD, has been suggested as a refinement of NAFLD that should highlight the central, etiologic role of insulin resistance, obesity, and diabetes mellitus. The complexity of the pathways involved in the regulation of hepatic TG synthesis and utilization have become obvious over the past 10 years, including the recent identification of monogenic causes of MAFLD. These include PNPLA3, TM6SF2, GCKR, MBOAT7 suggest targets for new therapies for hepatic steatosis.

Summary:

This review can serve as a guide to the complex pathways involved in the maintenance of hepatic TG levels as well as an introduction to the most recent discoveries, including those of key genes that have provided opportunities for new and novel therapeutics.

INTRODUCTION

Non-alcoholic fatty liver disease (NAFLD) is characterized by the presence of neutral lipids, mostly triglyceride (TG), in lipid droplets (LD) within the liver [^1–2]. This is best demonstrated by a biopsy showing more than 5% of hepatocytes with LD, or demonstrating>5% proton density fat fraction by magnetic resonance imaging (MRI) or similar imaging methods. A diagnosis of NAFLD requires the absence of other causes of steatosis, including alcohol intake and use of a hepatotoxic drugs. Monogenic disorders resulting in reduced ability to break down LD and/or secrete very low density lipoproteins (VLDL) can act as modifiers of NAFLD, worsening the degree of steatosis.[¹]

NAFLD is the leading cause of liver disease in western countries.[³] Earlier studies reported a NAFLD prevalence of 6-35% worldwide, and 10-46% in North America,[^4–5] but these numbers are increasing. This epidemic is due to the essential, pathophysiologic link between NAFLD and related metabolic disorders, including insulin resistance, type 2 diabetes mellitus (T2DM), and dyslipidemia, which are also increasing. The impact of metabolic disorders on NAFLD has led to a proposal that the term Metabolic Associated Fatty Liver Disease (MAFLD) be used to re-define the disease. [^6*] Numerous studies have identified ethnicity as a risk factor to develop liver steatosis: NAFLD is considerably more prevalent in Hispanics than in Caucasians or African Americans.[^4,7–11] These ethnic differences parallel differences in the prevalence of the Metabolic Syndrome in the groups as well as differences in the frequency of key genetic variants [^12*,13].

NAFLD encompasses a wide spectrum of pathologic changes in the liver, including non-alcoholic fatty liver (NAFL), defined as hepatic steatosis without evidence of hepatocellular damage, and non-alcoholic steatohepatitis (NASH), defined as steatosis with evidence hepatocyte injury. The latter includes hepatocyte ballooning, lobular inflammation, apoptosis, with or without fibrosis.[^{1, 14,15}] Once fibrosis is present, NASH can progress to cirrhosis and hepatocellular carcinoma.[^{14, 16}]

In the past decade, there have been major efforts to develop therapies for NAFLD. At present, most target NASH, but without NAFL, the incidence of NASH would be dramatically decreased. In this review, we focus on the abnormalities that drive the accumulation of TG in the liver as well as the potential for decreasing TG synthesis or increasing TG utilization to reduce hepatic steatosis.

PATHOPHYSIOLOGY OF NAFL AND POTENTIAL THERAPEUTIC TARGETS

Hepatic steatosis results when TG synthesis, whether from esterification of plasma-derived fatty acids (FA) or FA derived from de novo lipogenesis (DNL), is greater than TG utilization or disposal, whether by mitochondrial oxidation of FA or secretion of VLDL-TG.[^17–19] Reduced utilization or disposal of hepatic TG are usually consequences of intrinsic hepatic abnormalities, often associated with loss of function (LOF) of genes involved in those pathways. We will address each of these pathophysiologic drivers of steatosis.

Pathways of TG synthesis:

Increased hepatic uptake of plasma non-esterified FA:

Plasma non-esterified FA, hereafter referred to as plasma FA, are bound to albumin and are the main source of FA for hepatic TG synthesis. Plasma FA account for most of the hepatic TG content used for VLDL-TG synthesis and secretion in NAFLD patients.[^20–22] Hepatic extraction of circulating FA is proportional to their plasma concentration and, when insulin-mediated suppression of adipose tissue lipolysis is diminished, as in the metabolic syndrome, FA release from adipose tissue is increased. Even modest increases in plasma FA, whose flux is high, can significantly increase hepatic uptake. Direct links between hepatic uptake of plasma FA and increased VLDL secretion have been demonstrated in both rodents and humans. [^23–24]. It is clear that plasma FA flux and uptake by the liver are critical drivers of TG synthesis and steatosis, but how this translates to targets for therapy is less certain. FA uptake by the liver occurs via several pathways, making uptake an unlikely target. [^25–27*] Furthermore, once FA are taken up from plasma, complex intracellular trafficking of FA, including their activation by a family of long-chain fatty acyl CoA synthetases and targeting to different cell organelles, adds additional barriers to the development of targeted therapies prior to FA entry into specific metabolic pathways [^28*].

Increased hepatic uptake of TG-rich lipoprotein (TRL)-remnant TG-FA:

In addition to plasma non-esterified FA, TG-rich lipoprotein-remnants, products of lipolysis of nascent VLDL and chylomicrons, can enter the liver via several pathways [²⁹], contributing a modest amount of FA after lysosomal degradation of their TG [³⁰]. Insulin deficiency or insulin resistance, as well as polygenic causes of decreased lipoprotein lipase (LpL) activity, can result in incomplete lipolysis of TRL, leading to TG-enriched remnants that will contribute more TG-FA after hepatic uptake and lysosomal processing. A small, indirect source of TRL-FA is a pool of FA that escapes local uptake in adipose and muscle capillary beds during lipolysis. These “spill-over” FA can bind to albumin and merge with the general plasma FA pool coming from adipose tissue [³¹].

Increased de novo lipogenesis (DNL):

De novo lipogenesis, which occurs almost exclusively in the liver, is the pathway whereby acetyl-CoA, the product of carbohydrate metabolism in mitochondria, is converted to FA. Glucose and fructose, via glycolysis, generate pyruvate, which is converted to acetyl-CoA by the pyruvate dehydrogenase complex. Acetyl-CoA enters the tricarboxylic acid cycle (TCA or citric acid cycle) for energy generation or is transferred to the cytosol for FA synthesis when the cell is in positive energy balance. Under these circumstances, the synthesized FA are used mainly for TG synthesis.

DNL normally contributes 5-10% of TG that is stored in the liver and secreted as VLDL-TG in healthy humans. However, in individuals with obesity, insulin resistance, diabetes, or very high sucrose consumption, de novo lipogenesis can contribute as much as 30% or more to hepatic and VLDL-TG [^20,32,33*].

DNL is regulated by several hormones, substrates, products of metabolism, and transcription factors. [²⁷] Hormonal control is mainly via insulin’s effects on sterol regulatory element-binding protein 1 (SREBP-1c), probably through liver X receptors (LXRs) and mammalian target of rapamycin (mTOR), whereas glucagon affects Acetyl-CoA carboxylase (ACC) and Malonyl-CoA activities. Key substrates are glucose and fructose. Products of metabolism are the ratio of adenosine triphosphate (ATP) to adenosine monophosphate (AMP), which controls AMP-activated protein kinase (AMPK), an inhibitor of ACC activity. Transcription factors include SREBP1-c, the master regulator of fatty acid synthesis mentioned above, and carbohydrate response element binding protein (ChREBP), which responds to increased products of glucose metabolism.

There is abundant literature in rodent models and humans showing associations between consumption of fructose-sweetened beverages, increased de novo lipogenesis, and steatosis risk.[^34,35] Fructose is a stronger inducer of DNL than glucose because the conversion of fructose to fructose-1-phosphate by fructose kinase, a necessary step for the further conversion of fructose to pyruvate, is not regulated by phosphofructokinase as is the case for glucose entry into the glycolysis pathway[^36.37]. Fructose and glucose may also activate SREBP-1c, thereby stimulating the genes involved in DNL.[³⁸]

Several enzymes, including ATP citrate lyase (ACLY), ACC, fatty acid synthase (FASN), and Stearoyl-CoA desaturase-1 (SCD1) are involved in the synthesis of the long chain FAs, palmitate (C16:0) and oleate (C18:1,) from acetyl-coA [³⁹]. ACC1 and ACC2 seemed to be excellent targets for inhibition, which would decrease hepatic TG content by both inhibiting de novo lipogenesis and increasing mitochondrial oxidation of FA. Despite the demonstration of reduce hepatic TG content when ACC is inhibited, there was also an unexpected and significant elevation of plasma TG due to compensatory activation of Glycerol-3-phosphate-acyltransferase (GPAT1) resulting from deficiency of polyunsaturated fats. The latter resulted SREBP-1c gene activation and increased synthesis and secretion of VLDL-TG. A reduction in TG clearance by lipoprotein lipase has also been reported.[^40,41] Thus monotherapy with ACC inhibitors seems unlikely.

Hepatic FA, irrespective or their origin, are the substrate for synthesis of TG, as well as phospholipids and cholesterol esters. TG synthesis results from a series of enzymatic steps beginning with the addition of a FA to glycerol-3 phosphate by GPAT to generate lysophosphatidic acid (LPA). This is followed by the addition of a second FA to LPA by 1-acylglycerol-3-phosphate O-acyltransferase (AGPAT) to form phosphatidic acid (PA), that loses its phosphate group by the action of phosphatidic acid phosphatase (PAP1) also known as Lipin1 to form diacylglyceride (DG or DAG). The latter receives the third FA via the action of diacylglycerol transferase (DGAT) to produce TG.

DGAT exists in two isoforms: DGAT 1 and DGAT2. The former is primarily expressed in intestine and plays a central role in fat absorption by esterification of exogenous FA, while DGAT2 is expressed in liver and adipose tissue and incorporates endogenous FA into newly formed TG.[⁴²] In animal studies, DGAT1 knockdown proved to be ineffective as a treatment for NAFLD. Additionally, human studies of DGAT1 inhibitors demonstrated gastrointestinal side effects, including diarrhea, abdominal pain, and nausea.[⁴³] DGAT2 knockdown both reduces TG synthesis and downregulates SREBP1 activity, leading to decreased lipogenic gene expression and increased FA beta-oxidation; DGAT2 inhibition significantly improves hepatic steatosis in mice [^44,45*]; clinical trials in humans have begun.

Common genetic variants that affect de novo lipogenesis include those within the genes encoding glucokinase regulatory gene (GCKR) and membrane-bound O-acyltransferase 7 (MBOAT7), also known as lysophosphoinositol transferase 1 (Lpiat1). [^46*,47] GCKR encodes the GCKR protein, and a GCKRP variant with proline to leucine (P446L) change results in an unregulated increase in the activity of GCK leading to increased glycolytic flux and activation of carbohydrate-responsive element-binding protein (ChREBP), a potent stimulant of de novo lipogenesis. This variant, which has been reported to have a frequency of 40%, is associated with the atypical combination of lower levels of blood glucose and hypertriglyceridemia, along with NAFLD due to increased DNL.[⁴⁸]

MBOAT7 is a six-transmembrane protein found in several organelles, including the ER, mitochondria, and LDs. It is required to maintain normal levels of 1,2-diacyl-sn-glycero-3-phosphoinositol, or PI, in cells. MBOAT7 is involved in the re-acylation of phospholipids with polyunsaturated fatty acids in the Lands cycle. Recent studies demonstrated that targeted deletion of the gene results in hepatic steatosis with increased TG synthesis, a finding supported by meta-analysis of several large cohorts demonstrating an association between a single-nucleotide polymorphism (SNP) near the MBOAT7 gene and a range of NAFLD abnormalities. [^49*,50]

Decreased lipid droplet lipolysis:

Hepatic steatosis is a storage disease, with cytosolic LDs acting as the depots. However, numerous studies have identified LD as active sites of synthesis and degradation of lipids, particularly TG, as well as a home to molecules involved in signaling and even gene transcription [^51*]. LD TG accumulation results excess TG synthesis, regardless of the source ; how newly synthesized TG moves from sites of synthesis to the LD is, however, beyond the scope of this review, but our understanding of this pathway has increased greatly in the past 25 years [⁵²]. Characterization of the degradation of hepatic LD lipids, usually called LD turnover has also evolved, heralded by the important contrasting rolls of perilipin 2 (PLIN2), also known as adipose differentiation related protein (ADRP), and adipose triglyceride lipase (ATGL), encoded by patatin-like phospholipase domain‐containing protein 2 (PNPLA2). The latter is activated by CGI-58 (comparative gene identification-58) and inhibited by G0S2 (G0/G1 switch gene 2) and CIDEC (Cell Death Inducing DFFA Like Effector C) [⁵¹]. A number of investigators have made a case for lipophagy, either a macro–form where autophagosomes envelop a piece of a LD and transport to the lysosome, or a micro–form where LDs fuse directly with lysosomes, as other pathways for LD-TG turnover. The quantitative importance of lipophagy, relative to lipolysis of LD lipids, in the whole liver under physiologic conditions, remains to be determined.

In contrast, there is no doubt that the LD surface protein PNPLA3, is a critical regulator of LD turnover. [^53*] Indeed, a common single amino-acid change in PNPLA3 (i.e. isoleucine to methionine at position 148 or I148M) is the most clinically relevant genetic modifier of NAFLD [⁹] PNPLA3 expression is regulated by sterol receptor element–binding protein-1c (SREBP-1c),[^53,54] and, although there has been some uncertainty about the exact mechanism, appears to cause steatosis through resistance to proteasomal degradation, leading to its accumulation on the LD surface and inhibition of lipolysis by ATGL and CGI-58 [⁵⁵]. Importantly, the degree of hepatic steatosis in people with the PNPLA3 variant is closely linked to the presence of components of the metabolic syndrome [¹³]. Additionally, there are significant ethnic differences in the frequency of this variant: it was much higher in Hispanics (mainly Mexican Americans in this study (49%) compared to European Americans (23%) and African Americans (17%).[⁹] PNPLA3 has already become a target for innovative therapies of NALFD. [⁵⁶]

Mitochondrial oxidation of cell FA:

The liver generates energy from FA oxidation, which occurs within the mitochondria. Carnitine palmitoyl transferase (CPT1) converts fatty acyl-CoA to fatty acyl carnitine in the outer mitochondrial membrane, and this is transferred, by carnitine translocase, to CPT2, which regenerates fatty acyl CoA and free carnitine within the mitochondrial membrane.[⁵⁷] This process allows for beta-oxidation to begin and is indicative of the importance of the rate of FA flux into the mitochondrial matrix that is influenced by FA availability and CPT1 activity, and which can be in inhibited by malonyCoA.[⁵⁸] Although a significant literature demonstrates mitochondrial dysfunction in NASH [⁵⁹], the question of whether the development of simple hepatic steatosis is associated with, or exacerbated by, defective mitochondrial FA oxidation has not been definitively answered in humans, mainly because of the lack of completely validated methods. Measurement of plasma ketones as a surrogate of the rate of FA oxidation (high rates generate incomplete utilization of acetyl-CoA leading to ketone formation) actually indicate increased or normal rates of FA oxidation in people with NAFL versus normal subjects [⁵⁸]. Studies with multiple stable isotopes to interrogate the TCA cycle and ketogenesis indicated that increasing severity of hepatic steatosis was associated with reduced ketogenesis and increased acetyl-CoA oxidation in the TCA cycle. [^60,61] However, the methodology used requires a number of assumptions with varying views of the validity of those assumptions [^62,63]. This key issue seems, at present, to require further study, although there is no evidence (except for rare genetic disorders of mitochondria metabolism) [⁶⁴] for defective mitochondrial FA oxidation in livers of individuals with NAFL.

Assembly and secretion of VLDL:

Non-oxidized FA in the liver are esterified into TG, which can then be stored in LDs or used for VLDL assembly and secretion. In response to excess hepatic fat, NAFLD patients actually secrete more VLDL-TG than do subjects without NAFLD.[⁶⁵] The largest source of the increased TG secreted in NAFLD is not derived directly from systemic FA, but from lipolysis of LD-TG or from DNL.[^20,65] VLDL-apolipoprotein B (ApoB) secretion rates in individuals with NAFLD are typically also greater than rates in non-NAFLD subjects,[⁶⁶] although rates were similar between the two groups in one study. [⁶⁵] In all of these studies, VLDL was TG-enriched and larger in the NAFLD subjects. A question that might be asked is “Why doesn’t VLDL secretion keep up with the liver’s need to export TG and avoid steatosis?” In a study where excess FA were delivered to hepatoma cells or mice, ER stress resulted, leading to less than maximal secretion of ApoB. [⁶⁷] These data suggest that in humans with both NAFLD and higher than normal VLDL secretion rates, increased VLDL secretion is less than maximal, resulting in hepatic steatosis.

Rare pathogenic variants in the genes encoding ApoB and microsomal triglyceride transfer protein (MTP), two proteins critical for the assembly of VLDL and chylomicrons, were identified several decades ago as causing low levels of circulating ApoB-lipoproteins and hepatic steatosis [⁶⁸]. Familial hypobetalipoproteinemia (FBHL) is a rare autosomal disorder caused by rare pathogenic LOF variants in the APOB gene. In heterozygous FBHL, there is modest to moderate hepatic steatosis and reduced levels of TG and low density lipoproteins (LDL) in the circulation. However, homozygous or compound heterozygous forms of FBHL present with fat malabsorption, significant hepatic steatosis, and extremely low or undetectable levels of plasma TG and LDL [⁶⁸].

Abetalipoproteinemia (ABL) is an extremely rare autosomal recessive disorder caused by bi-allelic rare pathogenic variants of the MTTP gene. Homozygous or compound heterozygous LOF of MTTP presents with fat malabsorption and undetectable levels of ApoB in the circulation []. In both ABL and FHBL, patients can develop steatosis in the absence of modifying factors such as insulin resistance, although the presence of that defect, along with obesity, diabetes, and excessive alcohol intake, will increase the likelihood of hepatic steatosis and its severity. As might be expected, variants that lead to loss of function of either MTP or ApoB are associated with reduced risk for cardiovascular disease.

Anderson’s Disease, also known as Chylomicron Retention Disease, is an extremely rare autosomal recessive disorder resulting from bi-allelic rare pathogenic variants in the secretion-associated ras related GTPase 1B (SAR1B) gene and complete loss of function of the protein, SAR1B GTPase, which is required for the generation of very large coat protein complex II (COPII) vesicle necessary to carry chylomicrons from the ER to the Golgi in small intestine enterocytes [⁷⁶⁹]. This results in malabsorption of dietary fat and fat soluble vitamins, with failure to thrive. Hepatic steatosis is also present in patients with this disorder along with decreased secretion of VLDL as, similar to chylomicrons, large VLDL also require SAR1B coated COPII vesicles. [⁷⁰].

The identification of a common variant in transmembrane 6 superfamily member 2 (TM6SF2), a gene of unknown function at the time of its discovery, significantly extended understanding of the assembly and secretion of VLDL. [⁷¹], TM6SF2 is a transmembrane protein thought to reside in the endoplasmic reticulum (ER) and endoplasmic-reticulum–Golgi intermediate compartment (ERGIC) compartments of cells that, in ways incompletely defined, plays a role in the conversion (via addition of TG) of nascent, relatively TG-poor VLDL to a TG-enriched VLDL. The presence of a mutant TM6SF2 protein with a glutamic acid to a lysine at position 167 (E167K9) impairs lipidation, resulting in secretion of less TG from the liver and hepatic steatosis. [^71–73] The effect of E167K in the secretion of VLDL apoB is less clear, with no change, decreased secretion, or increased secretion all being observed. [^72,73*,*74] The prevalence of the common E167K variant is between 5 and 10% in the population, with higher prevalence in Caucasians than in African-Americans or Hispanic-Americans (mainly Mexican Americans) [⁷¹]. Concordant with reduced secretion of VLDL TG, the TM6SF2 variant is associated with reduced cardiovascular disease.

CONCLUSION

NAFLD is the outcome of complex combinations of altered metabolic pathways and both common and rare genetic variants that result in an imbalance between TG synthesis and utilization. The many steps involved in the ultimate outcome, hepatic steatosis, offer both a broad range of approaches to prevention or reversal of TG accumulation, but also increase the difficulty of choosing the optimal target for achieving those goals. Increasing knowledge of major genes that significantly increase the risk of developing NAFLD suggest that successful development of precise treatments are not far off. Meanwhile, lifestyle changes that diminish insulin resistance and its associated metabolic abnormalities must remain central to the approach to patients.

KEY BULLET POINTS.

• NAFLD is the leading cause of liver disease with a reported prevalence of 6–35% worldwide, and 10–46% in North America

• This epidemic is due to the essential, pathophysiologic link between NAFLD and related metabolic disorders, including insulin resistance, type 2 diabetes mellitus (T2DM), and dyslipidemia.

• Hepatic steatosis results when TG synthesis, whether from esterification of plasma-derived fatty acids (FA) or FA derived from de novo lipogenesis (DNL), is greater than TG utilization or disposal, whether by mitochondrial oxidation of FA or secretion of VLDL-TG.

• DNL normally contributes 5-10% of hepatic or secreted TG in healthy humans, but can contribute as much as 30% or more in obese, insulin resistant individuals or those with T2DM.

• PNPLA3, TM6SF2, MBOAT7, and GCKR, are key genes involved in hepatic lipid homeostasis and variants in these genes, in concert with the common metabolic disorders mentioned above, can cause significant NAFLD.

ABBREVIATIONS

ABL

Abetalipoproteinemia

ACC

Acetyl-CoA carboxylase

ACLY

ATP citrate lyase

ADRP

Adipose differentiation related protein

AGPAT

Acylglycerol-3-phosphate O-acyltransferase

AMP

Adenosine monophosphate

AMPK

AMP-activated protein kinase

ApoB

Apolipoprotein-B

ATGL

Adipose triglyceride lipase

ATP

Adenosine triphosphate

CGI-58

comparative gene identification-58

ChREBP

Carbohydrate response element binding protein

CIDEC

Cell death inducing DFFA like effector C

COPII

Coat protein complex II

CTP1

Carnitine palmitoyl transferase

DAG

Diacylglyceride

DGAT1

Diacylglycerol transferase 1

DGAT2

Diacylglycerol transferase 2

DG

Diacylglyceride

DNL

De novo lipogenesis

ER

Endoplasmic reticulum

ERGIC

Endoplasmic-reticulum–Golgi intermediate compartment

FA

Fatty acids

FASN

Fatty acid synthase

FBHL

Familial Hypobeta Lipoproteinemia

G0S2

G0/G1 switch gene 2

GCKR

Glucokinase regulatory gene

GPAT

Glycerol-3-phosphate-acyltransferase

LD

Lipid droplets

LDL

Low density lipoproteins

LOF

Loss of function

LPA

Lysophosphatidic acid

LPCAT1

Lysophosphatidylcholine acyltransferase 1

LpL

Lipoprotein lipase

LXR

Liver X receptors

MAFLD

Metabolic Associated Fatty Liver Disease

MBOAT7

Membrane-bound O-acyltransferase 7

MRI

Magnetic resonance imaging

mTOR

Mammalian target of rapamycin

MTP

Microsomal triglyceride transfer protein

NAFL

Non-alcoholic fatty liver

NAFLD

Non-alcoholic fatty liver disease

NASH

Non-alcoholic steatohepatitis

PA

Phosphatidic acid

PAP1

Phosphatidic acid phosphatase

PLIN2

Perilipin 2

PNPLA

Patatin-like phospholipase domain-containing protein

SAR1B

Secretion associated ras related GTPase 1B

SCD1

Stearoyl-CoA desaturase-1 (SCD1)

SNP

Single-nucleotide polymorphism

SREBP-1c

Sterol regulatory element-binding protein 1

T2DM

Type 2diabetes mellitus

TCA

Tricarboxylic acid cycle

TG

Triglyceride

TM6SF2

Transmembrane 6 superfamily member 2

TRL

Triglyceride-rich lipoprotein

VLDL

Very low density lipoproteins