Metabolic dysfunction-associated steatotic liver disease heterogeneity: Need of subtyping

Abstract

Metabolic dysfunction-associated steatotic liver disease (MASLD) is a widespread global disease with significant health burden. Unhealthy lifestyle, obesity, diabetes mellitus (DM), insulin resistance, and genetics have been implicated in the pathogenesis of MASLD. A significant degree of heterogeneity exists among each of above-mentioned risk factors. Heterogeneity of these risk factors translates into the heterogeneity of MASLD. On the other hand, MASLD can itself lead to insulin resistance and DM. Such heterogeneity makes it difficult to assess the natural course of an individual with MASLD in clinical practice. At present MASLD is considered as one disease despite the variability of etiopathogenic processes, and we lack the consensus definitions of unique subtypes of MASLD. In this review, pathogenic processes of MASLD are discussed and a need of subtyping is recommended.

Key Words: Metabolic dysfunctions-associated steatotic liver disease, Visceral obesity, Genetics, Diabetes, Insulin resistance

Core Tip: Metabolic dysfunctions-associated steatotic liver disease (MASLD) is a pathogenically heterogeneous disease with dynamic pathological features. The cardiometabolic risk of MASLD varies based on underlying etiopathogenic processes. Four subtypes of MASLD have been identified: Diabetic, non-diabetic (insulin resistance), genetic and mixed. These subtypes have not been clearly defined yet as separate entities. Given such heterogeneity, the behavior and outcome of MASLD in each patient is expected to be different. Therefore, a single drug is not expected to work for all patients with MASLD. The natural history of MASLD subtypes is unknown. There is an unmet need to categorize its subtypes and prospective studies are needed to characterize each subtype.

INTRODUCTION

Visceral adiposity is defined as ectopic deposition of fat in tissues other than subcutaneous adipose tissue. It is characterized by the excessive accumulation of fat in fat deposits surrounding the organs or fat accumulation inside the organs within the parenchymal or stromal cells (cellular steatosis). It is unclear whether cellular steatosis, such as in hepatocytes, is a consequence of visceral tissue expansion or if it happens concurrently, independent of visceral adipose tissue volume. There is significant data supporting the association between visceral adipose tissue and hepatic steatosis, and both of these conditions have been associated with increased cardiometabolic risk[1]. However, hepatic steatosis independent of mesenteric adiposity has also been described with higher cardiometabolic risk in lean hepatic steatosis patients compared to those with obesity-related hepatic steatosis[2,3].

Epidemiologic heterogeneity of metabolic dysfunction-associated steatotic liver disease

Obesity and being overweight are among the main risk factors for developing metabolic dysfunction-associated steatotic liver disease (MASLD), the most prevalent chronic liver disease worldwide. The most important driver of MASLD is dysfunctional visceral adipose tissue. MASLD developing in the absence of visceral fat accumulation is uncommon and represents a distinct entity. Increased visceral adiposity is an underlying factor in people with MASLD who are considered lean. Overall, people who are overweight or obese have a higher risk of developing MASLD compared with lean people[4]. The global burden of MASLD has grown in parallel with rising rates of diabetes mellitus (DM) type II and obesity. MASLD increases the risk of end-stage liver disease, hepatocellular carcinoma (HCC), death and liver transplantation. It is also associated with extrahepatic consequences, including cardiometabolic disease and cancers. However, diseases other than metabolic syndrome are also associated with MASLD. Such diseases include genetic disorders of lipid and carbohydrate metabolism[5].

Younossi et al[6] recently (2023) reported the global prevalence of MASLD. In a meta-analysis of 92 (n = 9361716) out of 2585 studies that met eligibility criteria, MASLD prevalence estimates and ultrasound-defined MASLD revealed an overall global prevalence of 30.05% and 30.69%, respectively across the study period (1990-2019). Global MASLD prevalence has increased by 50.4% from 25.26% in 1990-2006 to 38.00% in 2016-2019. The ultrasound-defined MASLD prevalence increased by 38.7% from 25.16% in 1990-2006 to 34.59%. The highest MASLD prevalence was in Latin America at 44.37% then the Middle East and North Africa 36.53%, South Asia 33.83%, Southeast Asia 33.07%, North America 31.20%, East Asia 29.71%, Asia Pacific 28.02%, and Western Europe 25.10%. Among the MASLD cohort diagnosed without a liver biopsy, the pooled mortality rate per 1000 patient years was 0.92% for liver-specific mortality; 2.83% for extrahepatic cancer-specific mortality; 4.20% for cardiac-specific mortality and 12.60% for all-cause mortality[6].

In another meta-analysis of 151 studies comprising 101028 patients who had undergone a liver biopsy, the prevalence of MASLD in the overweight population was 69.99%. A similar MASLD prevalence estimate of 75.27% was reported in the obese population. The prevalence of MASLD in the overweight population was the highest in the Americas (75.34%). Clinically significant fibrosis (stages F2-4) was present in 20.27% of overweight individuals with MASLD and in 21.60% of obese patients with MASLD, while 6.65% of overweight individuals with MASLD and 6.85% of obese individuals with MASLD had advanced fibrosis (stages F3-4)[7]. MASLD is not limited to overweight or obese individuals; it affects the normal-weight population as well. In a meta-analysis evaluating normal-weight individuals, the prevalence of MASLD in lean, healthy individuals ranged from 10.2% to 15.7%[8].

In addition to obesity, MASLD is also strongly associated with DM, dyslipidemia, and insulin resistance[9].

MASLD has been reported in all the pathogenetic types of DM, including type II, type I, ketone-prone DM and maturity-onset diabetes of the young (MODY)[10]. The prevalence of MASLD is highest in DM type II, ranging from 55% to 76%. In contrast, the prevalence of MASLD in DM type I ranged from 4% to 20%, though more studies reported around 20%. However, the prevalence of MASLD in ketone-prone diabetics and in those with MODY is lower compared to patients with DM type I. A study compared MASLD prevalence in DM type I, DM type II and healthy individuals who were matched for age and body mass index (BMI). This study showed that only 4.7% (6 out of 128) of DM type I patients had MASLD, vs 13.4% (9 out 67) of healthy individuals and 62.8% (166 out of 264) of DM type II patients. Almost 20% to 25% of MASLD patients have DM independent of type[10-14].

Dyslipidemia is a very common biochemical manifestation in the metabolic profile of patients with obesity, DM type II and insulin resistance, and is strongly associated with MASLD. The prevalence of dyslipidemia varies from 20% to 80% in patients with MASLD[15,16]. The prevalence of dyslipidemia is 63.4% among diabetics as compared to 16.7% among non-diabetics[14]. The global burden of dyslipidemias has increased over the past 30 years. Furthermore, the combination of high triglyceride (TG) levels and low high-density lipoprotein-cholesterol (HDL-C) levels (together with the presence of small, dense low-density lipoprotein (LDL) particles, is referred to as atherogenic dyslipidemia. The presence of such dyslipidemia increases the risk of cardiovascular disease (CVD) and MASLD. Patients with hypertriglyceridemia have a higher risk of developing MASLD than those with a normal TG level[17]. In a large-population study, a high BMI was an independent risk factor for the incidence of MASLD, and a high TG level was a risk factor in the high-BMI group (BMI ≥ 24). Elevated TG contributes about 25% to the incidence of MASLD in people with obesity[18].

The crude hypertriglyceridemia prevalence rates were 29.6% in the global population, 36.9% in men and 23.8% in women. The independent variables that were most associated with elevated TGs were hypercholesterolemia (OR: 4.6), low HDL-C (OR: 4.1), hepatic steatosis (OR: 2.8), diabetes (OR: 2.0), and obesity. (OR: 1.9)[19].

In summary there is a complex overlap of obesity, DM and dyslipidemia in MASLD (Figure 1), which makes it challenging to understand the pathogenesis of MALSD. It raises the question, why only a proportion of obese, diabetic, and lean individuals develop MASLD. It points towards probable genetic predisposition or other unknown factors.

Figure 1 Overlap of metabolic comorbidities. This diagram shows a hypothetical overlap of metabolic risk factors to highlight the heterogeneity of metabolic overlap. Metabolic risk factors are represented with colors: Obesity; yellow, DM; green, dyslipidemia; blue and metabolic dysfunction-associated steatotic liver disease (MASLD); red. The proportion of overlaps were based upon the data from epidemiological studies. Blue arrow represents MASLD patients and green arrow represents non-MASLD patients but has metabolic risk factors. MASLD: Metabolic dysfunction-associated steatotic liver disease; DM: Diabetes mellitus.

Pathogenetic heterogeneity of MASLD

The liver plays a fundamental role in maintaining the homeostasis of both glucose and TG metabolism by regulating systemic glucose and lipid fluxes during feeding and fasting states. The transition from fasting to feeding results in a significant redirection of hepatic glucose and lipid fluxes. It also plays a role in the metabolism of cholesterol needed for hormone synthesis and prevention of atherosclerosis by eliminating excess cholesterol through bile. The key metabolic functions of the liver related to nutrient metabolism are the production of lipoproteins, de novo lipogenesis (DNL) from fructose and monosaccharides, storage of TGs as lipid droplets, the secretion of TGs as very-LDL (VLDL), glycogenosis, glycogenolysis, gluconeogenesis, cholesterol synthesis and the excretion of cholesterol through bile acids. A mass spectrometry approach has identified 564 hepatocyte-secreted proteins (hepatokines) in a normal person. Many of these proteins are affected by fat accumulation and liver damage. The hepatokines that are dysregulated are shown to induce detrimental effects on nutrient metabolism in the liver as well as in other peripheral tissues[20-24].

MASLD is a complex disease that represents a wide spectrum of clinical and pathophysiological sub phenotypes with variable adverse outcomes. The heterogeneity and complexity of this disease stem from multiple contributing factors, which can be categorized into three broad categories: Individual characteristics, metabolic health, and genetics/epigenetics. Individual characteristics make every MASLD patient a unique case. Individual characteristics consist of age, gender, ethnicity, dietary pattern, activity, alcohol consumption, and microbiota composition. These factors affect the regulators and modulators, and result in a state of hepatic steatosis, an adaptive phase. Continued alterations of individual factors and dysfunctions of the regulators and modulators are responsible for transitioning hepatic steatosis into steatohepatitis. Insulin resistance acts as a cause and effect of cellular glucolipotoxicity. Figure 2 describes the role of nutrients, visceral adiposity, insulin resistance, DM, peroxisome proliferator-activated receptor (PPAR), farnesoid X receptor (FXR), bile acids, adiponectin, dysbiosis and genetics in the pathogenesis of MASLD and metabolic dysfunction-associated steatohepatitis (MASH). Metabolic health includes the contribution of obesity (BMI vs visceral adiposity), diabetes and dyslipidemia, which are the manifestations of regulatory dysfunctions. Such interplay of several factors results in significant individual variation of disease phenotype, natural course and response to the therapy. It is important to note that there is significant pathogenetic variation underlying these metabolic health factors. A hypothetical overlap of these metabolic factors based upon epidemiologic data is shown in Figure 1. MASLD is commonly associated with metabolic features, a prevalence of 80% of obesity, 72% dyslipidemia, and 44% of DM type II[6,9,25-27]. Studies have evaluated the impact of metabolic syndrome features in non-diabetic patients. Each individual metabolic factor does not increase the risk, but the presence of at least three factors, meeting the diagnostic criteria of metabolic syndrome, may have an increased propensity to have MASH. Nonetheless, why some patients with diagnosis of metabolic syndrome develop MASLD and aggressive disease and others do not, it remains unknown (Figure 1)[28,29].

Figure 2 The key modulators of carbohydrate and fat metabolism. Red arrow indicates negative relationship and blue arrow indicates positive relationship. FA: Fatty acid; GLP1: Glucagon like peptide 1; GIP1: Gastric inhibitory peptide 1; FXR: Farnesoid X receptor; PPAR: Peroxisome proliferator-activated receptor.

It is well known that genetic abnormalities play two roles: They increase the susceptibility of the person to developing a disease in combination with individual factors or to developing a MASLD independent of other modifiers such as individual factors or metabolic health. In connection with genetic susceptibility, several genetic variants and single nucleotides polymorphism (SNP) have been reported. Genetic susceptibility of course varies from one person to another. An effort has been made to develop a model to estimate the genetic risk in an individual. The polygenic risk score (PRS) can stratify the risk of this condition and of liver-related complications. For example, “high” PRS can predict the evolution to cirrhosis and HCC independently of classical risk factors in individuals with metabolic dysfunction[30,31]. The presence of patatin-like phospholipase domain-containing 3 (PNPLA3) p.I148M and transmembrane 6 superfamily member 2 (TM6SF2) p.E167K has been shown to cause MASLD, independent of other factors such as metabolic factors, individual factors and polygenetic risk. Studies have shown that these SNPs add to the risk of progression when combined with determinants of metabolic health and individual characteristics.

An Italian research group classified the etiopathogenetic spectrum of MASLD into three subtypes: Metabolic, genetic and mixed. The risk of CVD is highest among patients with acquired MASLD as opposed to those with genetic MASLD. In contrast, the risk of aggressive and progressive liver disease is highest among patients with genetic MASLD. The presence of DM or increasing severity of insulin resistance as measured by homeostasis model assessment of insulin resistance (HOMA-IR) adds to the risks of both progressive liver disease and CVD[32]. In a cross-sectional study of 140 biopsy-proven MASH patients, PNPLA3 rs738409 and TM6SF2 rs58542926 polymorphisms were independently associated with advanced liver fibrosis. DM type II increased the risk of advanced fibrosis in addition to the two polymorphisms. In non-diabetic patients, the HOMA-IR ≥ 5.2 increased the risk of advanced fibrosis when the polymorphisms were present[33]. Moreover, the risk of MASH, cirrhosis and HCC is very high among those with homozygous mutation (MM alleles) as opposed to heterozygous mutation (MI alleles)[34].

As discussed above, several factors play a role in the development of hepatic steatosis, and this creates significant heterogeneity that may lead to several clinical subtypes. Due to the heterogeneity of MASLD, the natural history of MASLD is not predictable to the desired level and it is attributed to the higher sensitivity to subtle variations in the associated risk factors. Since it is a common disease, it also overlaps with other primary liver diseases, creating another level of complexity.

Furthermore, the clinical spectrum is quite broad. It varies from simple hepatic steatosis, variable degrees of histological features, MASH, and cirrhosis without features of MASH. MASH is a histological diagnosis and is characterized by hepatocellular steatosis, lobular inflammation and ballooning of hepatocytes. It is defined by MASLD activity (NAS) ≥ 4. Fibrosis stage is determined independent of NAS score and does not have any bearing on the diagnosis of MASH. Liver pathology analyzed at the time of diagnosis only exhibits the severity of the liver damage at a specific point in time but does not display the dynamic changes of the histological findings. Histopathological features may be categorized into three main categories: (1) Simple hepatic steatosis; (2) meeting the diagnostic criteria of MASH; and (3) an indeterminate group with a variable degree of findings (non-MASH). Change in the severity of the histological findings is nonlinear; instead, it is a dynamic process that is closely related to the duration of exposure and the progress (improvement or worsening) of the metabolic risk factors.

The severity of the liver damage in MASLD depends upon the various combinations of multiple risk factors and each combination of the risk factors results in distinct clinical and histological phenotypes which may also have a distinct natural course. The differential contribution of genetic/epigenetic, environmental, and metabolic factors creates significant interpatient variation regarding the major driver of disease. Such a differential contribution of multiple risk factors in any given patient might explain much lower response rates (20% to 30%) in phase II or III MASH clinical trials than those previously reported for other causes of chronic liver disease such as viral or autoimmune liver diseases[35]. As of now we do not have any consensus on defining MASLD subtypes. Thus, the natural course of subtypes of the disease is unknown to the most part. However, there is limited data evaluating the morbidity and mortality of risks related to cardiovascular events and disease progression to cirrhosis and related complications, including HCC[32,33]. The number and the severity of the metabolic abnormalities (obesity, DM, and dyslipidemia) increase the spectrum of MASLD from patients with simple hepatic steatosis to MASH without and with advanced fibrosis[36].

Chronic liver injury with significant fibrosis has much less functional reserve to maintain adequate metabolic functions as described above. In patients with cirrhosis, lipoprotein metabolism is impaired, and the severity of this impairment correlates with the severity of hepatic impairment as measured by total bilirubin, international normalized ratio and albumin. Impaired lipoprotein metabolism in the setting of advanced hepatic fibrosis translates into abnormal plasma concentrations of total cholesterol, HDL-C, LDL cholesterol (LDL-C), and VLDL cholesterol. Lipoprotein metabolic dysfunctions are secondary to abnormalities in plasma enzymes associated with lipoproteins metabolism. These enzymes include lecithin-cholesterol acyl transferase, hepatic lipase, and phospholipid transfer protein. HDL-C and the enzymes that engage in high-density lipoprotein-cholesterol (HDL) maturation and metabolism. Therefore, there is a swing in the composition of HDL in patients with cirrhosis toward the larger HDL2 subclass with a reduction in small HDL3 particles. This is associated with diminished hepatic cholesterol efflux. Such a disproportion in HDL subclasses independently predicts mortality[37].

Thus, it is likely that the observed heterogeneity of MASLD in the clinic, where some patients develop isolated steatosis while others develop MASH and advanced chronic liver disease, stems from a differential contribution of pathogenetic pathways as well as from the activation or deactivation of redundant mechanisms (Figure 3)[38]. In summary, above mentioned findings may indicate that, in some individuals, metabolic dysfunctions start as carbohydrate homeostasis dysfunction due to incretins or pancreatic β-cell dysfunctions resulting into DM as a first manifestation. Among others, adipose tissue dysfunctions take the lead and manifest as visceral adiposity. Nonetheless, a proportion of obese individuals do not develop either metabolic dysfunction. Individual dietary patterns without or with genetic polymorphism may contribute to predilection for carbohydrate or fat metabolic dysfunction. It is known that some individuals are addicted to sugars and others to saturated fat[39-42]. A third dietary pattern is addiction to calories independent of nutrient type[41]. Hepatic steatosis may result from impairment of any one of the biochemical processes involved in hepatic glucose or fat metabolism as described above. Decreased VLDL secretion is apparently cardioprotective but worse for the liver, such pathogenetic mechanism is associated with genetic (PNPLA3 and TM6SF2) MASLD. TM6SF2 activity is required for normal VLDL secretion, and impaired TM6SF2 function causally contributes to nonalcoholic fatty liver disease (NAFLD)[43,44]. TM6SF2 inhibition is associated with reduced secretion of TG-rich lipoproteins and increased cellular TG concentration and lipid droplet content, while TM6SF2 overexpression reduced steatosis[43,45,46]. The exact physiological role of PNLA3 (adiponutrin) is unknown. The PNPLA3 p.I148M protein variant results in a loss-of-function of TG hydrolase and trans-acylase activity in lipid droplets leading to accumulation of polyunsaturated fatty acids, and reduction of secretion of VLDL from hepatocytes[47,48]. DNL is increased in the setting of impaired glucose homeostasis. Exhaustion of adipose tissue results in increased release of fatty acids which are taken up by hepatic cells to accommodate increased fat storage. Part of this fat is released back into circulation in the form of VLDL, thereby increasing CVD risk. Adipokines play a good role in lipid homeostasis. Impaired regulation of adipokines correlates with MASLD. Hepatocellular insulin resistance, which eventually develops as result of glucolipotoxicity, perpetuates the processes of steatosis and fibrosis. Glycemic control, dyslipidemia control, leptin level and severity of insulin resistance have been shown to affect the degree of progression of fibrosis. Moreover, the severity of fibrosis, independent of underlying etiology, also affects hepatic metabolic functions thereby aggravating the insulin resistance. Variability in lifestyle is also responsible for the dynamic behavior of the disease. Other factors contributing to the variable disease course are yet to be explored.

Figure 3 Diversity of pathogenetic mechanisms leading to metabolic dysfunction-associated steatotic liver disease. MASLD: Metabolic dysfunction-associated steatotic liver disease; DNL: De novo lipogenesis.

Given all the heterogeneity and complexity, it is prudent to classify MASLD into several subtypes based on the contributing factors, pathogenetic mechanisms, and histopathological features to better understand its natural history and subsequently develop therapeutic approaches and monitoring guidelines.

Differential cardiometabolic risk of MASLD

CVD risk increases with the presence of hepatic steatosis and stage of fatty liver disease[49,50]. A meta-analysis of observational studies including 34043 participants with diagnosis of MASLD confirmed an associated higher risk of cardiometabolic complications. Another study revealed similar results that individuals with MASLD have a significantly higher risk of 10-year cardiovascular events. These analyses did not differentiate the cohort between MASLD and MASH based on histological findings. Subsequent studies showed that the CVD risk is higher among patients with MASH compared to those with simple hepatic steatosis. This suggests that the presence of simple hepatic steatosis does not mean patients have a higher CVD risk in contrast to MASH patients.

Current evidence has failed to demonstrate a significant association between circulating levels of leptin and CVD. Yang et al[51] showed that serum leptin was not significantly associated with coronary heart disease [odds ratio (OR) = 1.07, 95% confidence interval (CI); 0.96-1.19] and stroke (OR = 0.98, 95%CI; 0.76-1.25)[51]. Similar results concerning coronary heart disease were also obtained from the meta-analysis performed by Chai et al[52]. Researchers of the Multi-Ethnic Study of Atherosclerosis also observed that serum leptin levels did not feature significant prognostic value, as far as the risk for incident CVD is concerned[53]. Thus, the role of leptin in CVD prediction is currently doubted.

Researchers have studied the effect of serum leptin level on MASLD and MASH. Leptin level increases with expansion of adipose tissue and with increasing BMI, thereby reflecting a state of adipose tissue failure which correlates with the onset of MASH. Other studies have shown a positive correlation between serum leptin and MASH as opposed to MASLD and control. Tumor necrosis factor levels were also significantly higher in parallel with serum leptin levels in patients with MASH[54]. Confirmed subsequently in a meta-analysis, circulating leptin levels were higher in patients with MASLD than in controls. Higher levels of circulating leptin were associated with increased severity of MASLD, and the association remained significant after the exclusion of studies involving pediatric or adolescent populations and morbidly obese individuals subjected to bariatric surgery[55].

However, it has also been demonstrated that the severity of liver disease (measured by FibroScan^®) but not simple liver steatosis is associated with higher mortality among patients with MASLD (hazard ratio = 1.69, 95%CI; 1.09-2.63), determined by the MASLD fibrosis score[54]. Increase in mortality was almost entirely driven by cardiovascular causes, constituting liver fibrosis marker panels as predictors of CVD mortality among patients with MASLD[56,57].

It is unclear if hepatic fat quantification is a better tool for determining cardiometabolic risk compared to intra-abdominal visceral fat measurement. A prospective study from Japan demonstrated that hepatic steatosis is a better determinant of insulin resistance and cardiometabolic risk. The study concluded that in nondiabetic and relatively normal-body-mass-index subjects, hepatic steatosis is independently associated with insulin resistance regardless of extrahepatic adiposity and might be the earliest event in the pathogenesis of the metabolic syndrome. The epidemiological studies revealed that almost 20% of MASLD cohorts had a BMI < 25 m². Lean MASLD patients have similar cardiometabolic risk as seen in MASLD patients with obesity. Some studies have demonstrated higher cardiometabolic risk in lean hepatic steatosis patients compared to those with obesity-related hepatic steatosis.

TM6SF2 activity is required for normal VLDL secretion, and impaired TM6SF2 function causally contributes to MASLD[43,44]. TM6SF2 rs58542926 T-allele–mediated hepatic retention of TGs and cholesterol predispose patients to MASLD-related fibrosis, whereas C-allele carriage promotes VLDL excretion, thus increasing the risk of CVD or atherosclerosis while protecting the liver[43,58]. Also, similar to TM6SF2 mutation, PNPLA3 rs738409 G allele is associated with a low risk of cardiovascular events, owing to the tight association with lower serum LDL level, as low LDL-C and high HDL-C are cardioprotective[50,59].

Heterogeneity in natural course of MASLD

Pathological changes in MASLD are dynamic. Paired liver biopsy studies have shown that the severity of the disease may remain stable, progress, or even regress over time, independent of any intervention. Longitudinal observation studies have exhibited progression of MASLD to MASH with bridging in about 25% of the cohorts. In a meta-analysis of biopsy-proven cohorts, the progression rate of fibrosis by one stage was 14 years for patients with hepatic steatosis and seven years for patients with MASH. The differential rate of progression has been attributed to several factors, but the exact mechanism remains unknown. These factors include variation of individuals characteristics, differential contribution and control of metabolic health risk factors and presence or absence of PNPLA3 or TM6SF2 mutations[60]. Additionally, it has been shown that presence of a moderate degree of fibrosis perpetuates the disease course independent of other factors in a proportion of patients with MASLD. Sanyal et al[61] found that one in five patients with MASH with bridging fibrosis or cirrhosis will develop cirrhosis or decompensation respectively over a period of two years[61]. This was confirmed in longitudinal paired biopsy analysis of MASH cohorts that revealed fibrosis progression in only one-third of patients, and it remained stable in about 40%.

Several studies have assessed the predictors of fibrosis progression over time with serial biopsies. As of now, the key predictor of fibrosis progression is baseline fibrosis. Progression of fibrosis based upon NAS score at baseline remains a matter of debate. As of now, all therapeutic trials incorporate two end points to determine the efficacy of studied medications: (1) Complete resolution of MASH; and/or (2) regression of fibrosis by one stage[62]. Thus far, several clinical trials have failed to prove the therapeutic efficacy of investigational drugs. Moreover, despite the fact that glucagon like peptide 1 (GLP1) agonists drugs have shown significant reduction in glycated hemoglobin (HbA1c) and weight, the histological improvement in liver pathology is moderate. Presence of number of pathogenetic factors and their control are the major factors which dictate the severity of the disease and progression over time.

Experimental data have shown that hepatic steatosis is not harmful per se but rather an adaptive mechanism to accommodate the increased free fatty acid influx. On the other hand, cellular lipotoxicity results from metabolic stress characterized by activation of the unfolded protein response/endoplasmic reticulum stress. Decreased adiponectin hormone secretion from adipose tissue due to metabolic stress of adipocytes contributes to the hepatocyte metabolic adaptation failure, which is responsible for the development and progression of hepatocellular injury, inflammation, hepatic stellate cell (HSC) activation and extracellular matrix accumulation, which defines the phenotype of MASH[63].

The inflammation, even when mild in intensity on the baseline histopathology, along with the worsening of the metabolic risk factors could substantially increase the risk of progression when compared to isolated steatosis[64]. In a large cohort of 713 biopsy-confirmed MASLD patients, 49% with DM, 51.3% with advanced fibrosis/cirrhosis, and 69.6% with nonalcoholic steatohepatitis (NASH), glycemic control was assessed through HbA1c measurements performed for several years preceding liver biopsy. Poor glycemic control was associated with ballooned hepatocytes, steatosis, and increased fibrosis[65]. Studies have shown that NAS correlates strongly with disease course. Thus, an increase in the activity grade is associated with fibrosis progression and, conversely, a decrease in the activity grade and MASH resolution are associated with fibrosis regression[66-71]. Moreover, other factors also predict the progression of MASLD such as: Increased HSC activation evidenced by alpha smooth muscle actin staining on histopathology, severity collagen deposition on histopatholgy, and increased serum fibrosis markers.

Subtypes of MASLD (classification)

In the light of epidemiologic, pathogenic, differential cardiometabolic risk and variable course of progression (natural course), MASLD can be classified into three distinct etiopathogenic categories. Based upon the epidemiologic data, the prevalence of each subtype is estimated in Figure 4. Fourth category is mixed type, with variable degrees of overlaps of three distinct types. Each subtype has its own clinical course and outcome. This phenomenon rationalizes the failure to develop a single drug to treat all types of MASLD. Moreover, to further complicate the understanding of the process, liver disease due to other etiologies such as viral or autoimmune disease can lead to secondary hepatic metabolic dysfunction leading to MASLD, whether this should be classified as separate entity, remains matter of debate. The interaction between DM and MASLD is quite complex, bidirectional and connected with insulin resistance (Figure 5). Moreover, adipose tissue dysfunctions because of unhealthy lifestyle may affect pancreatic function independent of hepatic involvement, which may result in the onset of DM without developing MASLD[65,72]. All subtypes of DM have been implicated in the pathogenesis of MASLD with variable degrees of impact (Table 1). A population based screening of MASLD study revealed that insulin resistance (IR) related MASLD constitute 80% and DM type II 20% of the MASLD population. They did not identify the genetic contribution[14].

Figure 4 Estimated prevalence of metabolic dysfunction-associated steatotic liver disease subtypes. MASLD: Metabolic dysfunction-associated Steatotic liver disease, US: Ultrasound: DM: Diabetes mellitus; BMI: Body mass index.

Figure 5  Bidirectional connection between two pathogenetic pathways leading to insulin resistance.

Table 1 Clinical characteristics of metabolic dysfunctions-associated steatotic liver disease subtypes.

	Pancreatic dysfunction			Hepatic
	Diabetes mellitus			Non diabetic	Genetics	Mixed
	Non type II diabetes mellitus		Type II DM	Insulin resistance	Monogenic	Overlap
	DM type I	Primary beta cell dysfunctions (monogenic)	Secondary beta cell dysfunctions	Metabolic syndrome	PNPLA3 or TM6SF2 associated steatohepatitis	DM+/-IR+/-Genetics
BMI	Normal	Normal	Obesity	Obesity	Normal	Unknown
DM	Yes	Yes	Yes	May be	No	May be
Triglyceride level	Normal	Normal	High	High	Normal	Unknown
Total cholesterol	Normal	Normal	Normal or high	Normal or high	Normal	Unknown
HDL	Normal	Low	Low	Low	Normal	Unknown
LDL	Normal	Normal	High	High	Normal	Unknown
Fasting Insulin	Very low (< 5)	8-12 (< 10)	8-10 (< 10)	> 15, commonly > 20	Normal	Unknown
Waist ot hip ratio	Normal	Normal	May be	Reversed	Normal	Unknown
Fasting blood glucose (FBG) (pre-DM)	No	No	No	Yes	No	Unknown
HOMA-IR	Normal or minimally high	Normal or minimally high	> 3.0 < 5.0	> 3.0 usually > 7.0	Normal	Unknown
FBG/insulin ratio	> 25	> 15	> 15	< 10	< 10	Unknown
Adiponectin	Normal or high	Unknown	Low	Low	Normal	Unknown
Leptin	Low	Unknown	No association with high levels after adjusted with confounder	High	Normal	Unknown
Other features	Family and personal history of autoimmunity	Family history of DM and onset at young age			Family history of liver disease: Unknown
Prevalence (population)	0.5% to 0.75 %	0.2 % to 0.3 %	0.1	30% to 40% among obese and overweight	Variable depending upon ethnicity/race; highest in hispanics (45%) and lowest in African American. Also depends upon heterozygosity vs homozygosity	Unknown
Disease burden (proportion developing MASLD)	20% (most studies)		55%-76%	70%-75% among overweight and obese. Based upon liver biopsy	10% to 15%	Unknown
Genetics vs acquired	Acquired	Genetics	Acquired	Acquired	Genetics	Unknown
Hepatic steatosis	Yes	Yes	Yes	Yes	Yes	High
Mechanism	DNL	DNL	DNL	Lipolysis of adipose tissue OR familial hypertriglyceridemia leading to increased hepatic uptake	Decreased excretion of VLDL	Multilevel
MASH risk	Low	Low	High	Moderate		Very high
Fibrosis risk	Unknown	Unknown	Higher than non-diabetics	0.22	Higher than non-diabetics and diabetics	Synergistically rises depending upon the number overlaping factors
CAD risk	Yes	Yes	Yes	Yes	No	Unknown
HCC risk	Unknown	Unknown	Higher than non-diabetics	High	High	Very high

BMI: Body mass index; DM: Diabetes mellitus; CAD: Coronary artery disease; HCC: Hepatocellular carcinoma; MASH: Metabolic dysfunction-associated steatoheptatits; VLDL: Very-low-density lipoprotein; DNL: De novo lipogenesis; HOMA-IR: Homeostasis Model Assessment-Estimated Insulin Resistance; HDL: High-density lipoprotein; LDL: Low-density lipoprotein; MASLD: Metabolic dysfunction-associated steatotic liver disease; IR: Insulin resistance.

Figure 3 summarizes three distinct etiopathogenic mechanisms leading to hepatic steatosis, insulin resistance, and subsequent course of inflammation and/or fibrosis. Hepatic steatosis results from any single or combination of multiple intracellular processes. These include increased fatty acid uptake released as a result of increased lipolysis in adipose tissue, DNL due to uninhibited gluconeogenesis and increased consumption of particularly fructose, and uncontrolled DM (hyperglycemia), decreased excretion of VLDL in particular due to genetic mutations and decreased beta oxidation of fatty acids. On the other hand, hepatic fibrosis is a consequence of either inflammation, apoptosis or insulin resistance without inflammation or apoptosis[73,74]. Neutrophils, Kupffer cells (the resident macrophages of the liver) bone-marrow-derived monocytes and Th17 cells can promote HSC activation by secreting cytokines and growth factors[75,76].

Among all three subtypes, the progression from hepatic steatosis to MASH and cirrhosis occurs by entirely different mechanisms, which may involve inflammation, inflammation with fibrosis, apoptosis with fibrosis or just fibrosis. As of now we know very little about the disease progression from the onset of hepatic steatosis. We also lack the understanding of natural course of each pathogenic process leading to liver damage and cirrhosis which is the ultimate milestone to affect morbidity and mortality. The dynamic control of each pathogenic process, however, may play a role in the dynamic nature of hepatic fibrosis progression and regression. Which makes it difficult to accurately assess the response of a study medication on the degree of hepatic fibrosis regression. In Table 1, author attempted to summarize the features of each subtype based upon limited published literature on this topic[77-86]. Nonetheless, it does not characterize each subtype accurately, but it gives us a perspective to believe that each subtype is a distinct entity with a distinct course. Thus, we need prospectively designed studies to comprehend the natural course of each subtype. Few studies have compared the cardiometabolic risk of genetic MASLD with DM and non-diabetic IR MASLD.

Table 1[4-19,34,79-88] summarizes the differentiating features among MASLD subtypes. Steatosis in PNPLA3-associated MASLD is not accompanied by features of metabolic syndrome. Insulin resistance defined by HOMA-IR differentiates well among three distinct entities. Genetic MASLD characteristically has normal HOMA-IR score as opposed to IR related MASLD where HOMA-IR is significantly high typically more than 5 or even some studies reported > 7. Diabetic patients with HOMA-IR score are abnormal but not as high as seen in IR. Patients with DM have two distinct patterns of insulin level, high levels are consistent with IR and low levels (usually below 10μmol/dL even below 5 μmol/dL) are consistent with pancreatic β cell failure. It is primarily seen in patients with type I DM, monogenic DM and also due to exhaustion phenomenon as a result of severe IR (probably adipose tissue and muscles). Such patients may manifest with DM way before they develop MASLD and may or may not have visceral adiposity[65]. Poor control of hyperglycemia and increased consumptions of fructose may play a role in the causation of MASLD by DNL. The risk of MASH comparably increased in non-diabetic IR MASLD and PNPLA3^148MM/MI groups compared to their respective control groups. The non-diabetic IR MASLD vs PNPLA3^148MM/MI is markedly enriched in saturated and monounsaturated triacylglycerols and free fatty acids, dihydroceramides (markers of de novo ceramide synthesis) and ceramides. Markers of other ceramide synthetic pathways are unchanged. Whereas in PNPLA3^148MM/MI as opposed to PNPLA3^148II, the increase in liver fat is due to polyunsaturated triacylglycerols while other lipids were unchanged. Similar increases in liver fat and MASH are associated with a metabolically harmful saturated, ceramide-enriched liver lipidome in ‘Metabolic MASLD’ but not in ‘PNPLA3 MASLD’. This difference may explain why metabolic but not PNPLA3 MASLD increases the risk of type 2 diabetes and CVD[76]. Each of the three entities have the risk of developing MASH, fibrosis and HCC independent of other entities. A combination of these entities increases the risk MASH, fibrosis and HCC.

Keeping in align with different pathogenic processes, the same treatment approach may not work. Weight loss is not required in patients with genetic MASLD and lean diabetic MASLD. Whether restricting high glycemic food and saturated fat would benefit patients with genetic MASLD remains unknown and yet to be explored. On the contrary, restricting polyunsaturated fatty acids may be beneficial in genetic MASH. GLP1 and/or gastric inhibitory peptide 1 analogs are probably good choice for DM type II related MASLD. Insulin along with a strict healthy diet and good control of hyperglycemia would prevent and treat MASLD in type I DM related MASLD. PPAR agonist could be a good choice for patients with IR related MASLD. FXR agonist might be beneficial in augmenting the response of PPAR. Recently the FDA has approved resmetirom, the first drug for the treatment of MASH with moderate fibrosis. Resmetirom has pleiotropic effects in the liver given the fact that it is an agonist for a nuclear receptor hormone, namely thyroid hormone receptor beta. It significantly decreases intra-hepatic lipids mainly through increased mitochondrial β oxidation and improving hepatocyte mitochondrial function in MASH patients[87]. Though this drug has proved better in both predetermined end points compared to placebo, the response was limited to only 25% of patient of treatment cohort while 75% of the treatment cohort had no or partial response in relation to both primary end points[88].

CONCLUSION

MASLD is a very heterogeneous disease with different etiopathogenic processes playing a key role in the development of hepatic steatosis and fibrosis. Thus far, three distinct etiopathogenic processes lead to MASLD including unhealthy lifestyle with visceral adiposity related insulin resistance, DM, and genetic mutations. The fourth entity is a combination of any of the three entities. These subtypes have not been clearly defined yet as separate sub-entities. Given such heterogeneity, the behavior and outcome of MASLD in each patient is expected to be different. Therefore, a single drug is not expected to work for all patients with MASLD. The natural history of MASLD subtypes is unknown. There is an unmet need to categorize its subtypes and prospective studies are needed to characterize each subtype.

To differentiate and characterize hepatic steatosis and predict the risk of MASH, cirrhosis, HCC and CVD among patients with each subtype of MASLD we ought to quantify hepatic fat, intra-abdominal visceral fat, dyslipidemia and severity of insulin resistance; glycemic control of diabetes; conduct genetic testing for PNPLA3 (adiponutrin gene) and TM6SF2 gene mutations; and obtain PRSs. Among patients with insulin resistance and/or diabetes, the risk is higher among those with low adiponectin and high leptin levels. Routine testing of adiponectin and leptin would help. We are not yet in a position to routinely assess genetic mutation in clinical practice.

ACKNOWLEDGEMENT

I acknowledge the support and contribution of Mr. Andrew Johnson in preparation of manuscript and English language editing.