Lipid metabolism in MASLD and MASH: From mechanism to the clinic

The liver synthesises several lipids, including triacylglycerols (TAGs), diacylglycerols (DAGs) and sterols, which are synthesised at the level of the endoplasmic reticulum (ER) by re-esterification of free fatty acids (FFAs) derived from a) adipose tissue lipolysis, b) dietary chylomicrons taken up from the circulation or c) synthesised through hepatic de novo lipogenesis (DNL) and then stored within lipid droplets (LDs). Intracellular lipid synthesis and storage in LDs LDs are intracellular dynamic organelles that serve to both store neutral lipids for cellular metabolic needs and sequester lipotoxic lipids that would otherwise be toxic for the cell since they can act as membrane detergents or cause organelle dysfunction. 1 LDs are present in many tissues, principally in the liver, adipose tissue and intestine, and enclose a core filled with neutral lipids, most commonly TAGs and sterol esters, surrounded by a phospholipid monolayer incorporating specific proteins from the perilipin family. [2] , [3] , [4] LD assembly is still poorly understood since it involves multiple steps. TAG and sterol ester synthesis, from the esterification of an activated fatty acid to a DAG or a sterol (such as cholesterol), respectively, involve different enzymes located primarily in the ER. 1 TAG synthesis begins with the entry of fatty acids into the ER ( Fig. 1 ) where they are converted to acyl-CoA and used for synthesis of DAGs by diacylglycerol acyltransferase 1 (DGAT1) and 2 (DGAT2). The DGAT1 enzyme is found exclusively on the membrane of the ER and re-esterifies the DAGs produced by the lipolysis of TAGs, while DGAT2 is found both in the ER and on the LD’s surface and synthesises TAGs, which are incorporated into the LDs when the cytosolic concentration of fatty acids rises. 5 Both DGAT1 and DGAT2 prevent the intracellular accumulation of lipotoxic lipids, DAGs and fatty acids. DAGs are not only intermediates in TAG synthesis but they can also activate PKCε and, by impeding insulin signalling, are implicated in the development of hepatic insulin resistance (IR). 6 , 7 Fig. 1 Lipid droplet biogenesis. Non-esterified FFAs from lipolysis or de novo lipogenesis are re-esterified into TAGs by DGAT1 and DGAT2 in the ER. Palmitate is the first FFA synthesised in the DNL by the enzymatic complex FASN; synthesis begins by combining acetyl-CoA with malonyl-CoA. Palmitate is elongated to stearate by ELOV6 and desaturated by SCD1 to palmitolate and oleate. Lipid droplets contain TAGs and sterol esters and are surrounded by a phospholipid monolayer with proteins from PLIN family members. Lipid droplets expand through droplet-droplet fusion or TAG transfer to or synthesis on the lipid droplet surface. DGAT1/2, diacylglycerol acyltransferase 1/2; DNL, de novo lipogenesis; ELOV6, fatty acid elongase-6; ER, endoplasmic reticulum; FASN, fatty acid synthase; FFAs, free fatty acids; LDs, lipid droplets; PLIN, perilipin; SCD1, stearoyl-CoA desaturase-1; TAGs, triacylglycerols. Fig. 1 LDs may expand through droplet–droplet fusion or transfer of TAG to LDs via ER membrane bridges or through TAG synthesis directly on the LD surface. 1 Thus, LDs act as a lipid buffering system by sequestering lipotoxic compounds while maintaining contact with many organelles, facilitating lipid transfer.

The liver contains a high number of mitochondria that provide the energy required to support its many metabolic functions. Hepatic mitochondria are also critical mediators of metabolic flexibility (the ability to adapt to fluctuations in energy demand and supply to maintain whole-body homeostasis) by dynamically modifying the oxidation of glucose or FFAs according to their availability, i.e . switching from FFA oxidation, during the fasting state, to enhanced glucose metabolism during the feeding state. 31 In MASLD, excess lipid accumulation is due not only to excess FFAs but also to insufficient fatty acid oxidation. Whether this is due to mitochondrial dysfunction and/or reduced metabolic flexibility has been debated. Several metabolic diseases, e.g. type 2 diabetes (T2D), obesity, and MASLD, are associated with reduced metabolic flexibility, with higher FFA oxidation even when glucose is the predominant energy supply, e.g. during the euglycaemic hyperinsulinaemic clamp. 31 In the early phases of MASLD, the impaired suppression of lipolysis by insulin is accompanied by increased FFA oxidation, 8 , 32 and mitochondrial activity and biogenesis are increased rather than decreased. 32 , 33 In the later stages of the disease, but not in isolated steatosis, mitochondrial respiration is reduced due to DNA and protein abnormalities, and hepatic IR is associated with reduced electron transport chain capacity; 32 , 33 however, it is not clear whether this is a cause or consequence of MASH(7). In humans and mice, high body weight, steatosis, and lipolysis from adipose tissue lead to an increase in fatty acids in the liver and their use in β-oxidation, the tricarboxylic acid cycle (probably due to excess production of acetyl-CoA) and ketogenesis. [34] , [35] , [36] In animal models, Einer et al. showed that as MASL progresses to MASH, megamitochondria are observed and accompanied by impaired mitophagy and reduced ATP production 33 ; this was recently confirmed in the livers of individuals with obesity and MASH. 37 However, liver size is also increased in patients with MASH, and no study evaluated if the total number of mitochondria, not only their size, is changed since the lower number of mitochondria per g of liver could explain the reduced ATP production. It is also true that in the study by Einer et al. , 33 the megamitochondria had similar protein content, indicating that they probably include more lipids. Einer et al. 33 showed that megamitochondria with reduced mitochondrial oxidation did not increase the production of reactive oxygen species in animals fed a Western diet. In human livers, Sarabhai et al. 37 showed that individuals with obesity and MASH (not with MASL) had megamitochondria whose diameter was associated inversely with fusion/fission biomarkers and with oxidative capacity but positively with H 2 O 2 . Not only size but also structural changes were observed in individuals with MASH, like loss of mitochondrial cristae and paracrystalline inclusions. 38 Fatty oxidation also occurs in mitochondria and produces ketone bodies. However, many studies have shown an increase in fatty acid oxidation and β-hydroxybutyrate in individuals with MASH compared to those without liver steatosis., 38 , 39 likely driven by the high fatty acid flux to the liver due to Adipo-IR. However, sexual dimorphism in lipid handling was observed, as women show high fatty acid oxidation and production of hydroxybutyrate but lower DNL compared to men matched for the same characteristics and liver fat content. 40 In part, this explains why men have a higher prevalence of MASLD.

The postprandial state is an anabolic state with high insulin secretion stimulated by ingested nutrients, which promotes glycogen, protein, and lipid synthesis in the liver, muscle, and adipose tissue. Dietary lipids are assembled as chylomicrons in enterocytes and then taken up by the liver and other peripheral tissues. 4 Once lipids enter the systemic circulation, TAGs in chylomicrons are hydrolysed by circulating lipases into FFAs and taken up by peripheral tissues ( Fig. 2 ). Dietary lipids can also be temporarily stored in intestinal LDs that might be a buffer for excess lipids. 4 Tracer studies showed that dietary lipids remain in the circulation for up to 18 h after a high-fat meal but not after a low-fat meal, 50 raising the question of whether high-fat meals are retained in the intestine and continue to fuel the liver for a significant part of the day. Different dietary lipids have a different effect on hepatic TAG accumulation. Costabile et al. 51 showed that isoenergetic substitution of SFAs with unsaturated fatty acids reduced IHTG. Luukkonen et al. 52 showed that 3 weeks of a hypercaloric diet rich in SFAs induced the highest increase in IHTG (+55%) compared to the diets high in unsaturated fatty acids (+15%) or sugars (+33%, p <0.05). The high-SFA diet increased lipolysis, adipo-IR, and ceramide levels. Studies on alterations in lipid metabolism over the years have focused on fasting, but postprandial lipid metabolism is also altered in MASLD. 53 Postprandial lipidomics highlight the different metabolic responses to food intake. Bonham et al. 54 showed that TAG species varied significantly in response to the type of food ingested, suggesting a different formation/clearance of chylomicrons, despite similar TAG concentrations in chylomicrons of patients with metabolic syndrome and controls. The Mediterranean diet, food with low-glycaemic index and fibre-rich foods have a protective effect against steatosis, while high intake of refined sugars, in particular fructose, promotes MASLD. 17 , 18 Dietary sugars are critical energy substrates ( Fig. 2 ). Luukkonen et al. 52 showed that DNL was high only in individuals consuming a high-sugar diet, rich in fructose, while a high-SFA diet was associated with higher IHTG, but lower DNL. It is well established that only certain sugars ( e.g ., fructose and sucrose, but not glucose) are precursors of newly synthesised palmitate and stimulate DNL, 24 , 55 while glucose is used to produce energy through oxidation or stored as glycogen. Ter Horst et al. showed that fructose, but not glucose, ingestion stimulated DNL in patients with MASLD. 24 Geidl-Flueck et al. also showed that consumption of beverages containing fructose but not glucose for 7 weeks resulted in a 2-fold increase in basal hepatic fractional secretion rates of newly synthesised fatty acids compared to controls but no change in basal synthesis and secretion of very low-density lipoprotein (VLDL)-TAG. 55 DNL occurs throughout the day but mainly in the postprandial state since its main precursors are exogenous sugars (sucrose and fructose, not glucose). 24 , 55 Other substrates may be used to produce the acyl-CoA used for DNL. Postprandial DNL contributes to 15-26% of liver TAGs in individuals with MASLD 49 , 56 vs. 1-6% in the fasting state in healthy individuals. 48 However, the combination of high levels of insulin and substrates could also explain the stimulation of DNL, as also proposed by Ter Horst et al. 24

Reduction in IHTG and resolution of MASH are associated with weight loss achieved with lifestyle intervention, bariatric surgery (in individuals with obesity) or pharmacological treatment, though some drugs act independently of weight loss. Diet and physical activity The main therapeutic approach in the management of MASLD/MASH is represented by lifestyle interventions, as highlighted in the recent guidelines for MASLD(17). Several randomised clinical trials showed how weight loss leads to significant improvements in MASLD. 17 The clinical trial by Vilar-Gomes et al. is one of the largest and involved 293 individuals with histologically proven MASH treated for 52 weeks with a hypocaloric diet (750 kcal/day less than their daily energy need, with carbohydrates, fats, and proteins accounting for 64%, 22% [including <10% of total calories from SFAs], and 14% of total daily calories, respectively) and physical activity (encouraged to walk for at least 200 min/week). 68 The primary outcome, MASH resolution with no fibrosis worsening, was achieved by 25% of participants ( Fig. 2 A); the degree of weight loss was independently associated with improvements in all MASH-related histologic parameters. The study also suggests that in adults with MASLD and overweight/obesity, dietary and behavioural therapy-induced weight loss should aim at a sustained reduction of ≥5% to reduce liver fat, but at least 7–10% to improve liver inflammation and ≥10% to improve fibrosis. 68 Fig. 2 Percentage of participants with resolution of MASH and no worsening of fibrosis (defined as no increase in the fibrosis stage) in clinical trials with different treatments. (A) Diet and physical activity for 1-year 68 ; (B) bariatric surgery for 1 year 69 ; (C) GLP-1 receptor agonist semaglutide for 72-weeks 71 ; (D) pioglitazone for 18 months 80 ; (E) lanifibranor for 24 weeks 83 ; (F) the THRβ agonist resmetirom for 52 weeks 91 ; (G) the dual GLP-1-GIP agonist tirzepatide for 52 weeks 73 ; (H) dual GLP-1/glucagon agonist survodutide for 48 weeks. 74 GIP, gastric inhibitory polypeptide; GLP-1, glucagon-like peptide 1; MASH, metabolic dysfunction-associated steatohepatitis; THR, thyroid hormone receptor. Fig. 2

Several drugs are currently under investigation for the treatment of MASH, but the only approved (under the accelerated approval pathway) option is resmetirom. Semaglutide and other single GLP-1RAs Glucagon-like peptide 1 receptor agonists (GLP-1RAs), like exenatide, liraglutide, dulaglutide and semaglutide, initially developed for the treatment of diabetic hyperglycaemia, have been shown to reduce IHTG. Liraglutide and semaglutide have been investigated in phase II trials for the treatment of MASH, 71 , 72 but only the semaglutide trial has continued to phase III. In the phase II trial, 320 patients (72% with F2 or F3 fibrosis) were randomly assigned to receive semaglutide at a dose of 0.1 mg, 0.2 mg, or 0.4 mg or placebo for 72 weeks. 71 The highest dose of semaglutide was superior to placebo with respect to resolution of MASH (59% vs . 17%, respectively, Fig. 2 C); however, no significant difference between treatment and placebo was observed with regard to improvement in fibrosis. Semaglutide also significantly affects weight loss. Although there are no GLP-1 receptors in the liver or adipose tissue, the reduction in body fat and the improved glycaemic and lipid profiles are undoubtedly critical factors in the regression of MASLD and related cardiometabolic risk factors.

Peroxisome proliferator-activated receptor (PPAR)-γ agonists, like thiazolidinediones, have been approved for the treatment of diabetic hyperglycaemia, and have also been evaluated for the treatment of MASH in small trials, showing improvement in liver histology. [78] , [79] , [80] Cusi et al. 80 investigated the effect of pioglitazone, 45 mg/day or placebo plus hypocaloric diet for 18 months in 101 patients with prediabetes or T2D 80 ; 51% of patients treated with pioglitazone achieved resolution of MASH vs. 19% in the placebo group ( Fig. 2 F). PPAR-γ agonists accomplish a reduction in IHTG [78] , [79] , [80] and in VAT at the expense of SAT 81 , 82 and despite weight gain (+2.5 kg when combined with a hypocaloric diet 80 ). Thus, adipose tissue remodelling, and not just weight loss, is a potential mechanism for the reduction of IHTG and resolution of MASH. Recently, other multiple PPAR agonists, like the PPAR-α/γ agonist saroglitazar and the pan PPAR-α/γ/δ agonist lanifibranor, have been tested for the treatment of MASH. 81 The phase II trial of lanifibranor enrolled 247 individuals randomised to a dose of 800 mg or 1,200 mg vs. placebo. 83 Lanifibranor was superior to placebo for resolving MASH without worsening fibrosis (51% and 49% vs. 22% in placebo, Fig. 2 G), and was also associated with weight gain (+2.7 kg at the highest dose). The improvement in adipose tissue metabolism through PPAR-γ activation is likely crucial for the resolution of MASH. PPAR-γ is mainly expressed in adipose tissue and regulates adipogenesis, lipid storage and fatty acid metabolism. 84 Mechanistically, this has been explained by the enhanced adipogenesis observed in SAT following treatment with pioglitazone. 82 At the same time, all PPAR-γ agonists have a general insulin-sensitising effect by improving adipose tissue insulin sensitivity, 80 , 85 promoting the suppression of peripheral lipolysis, 86 and thus decreasing the concentrations of plasma fatty acids and their efflux to the liver. Moreover, they have been associated with a reduced concentration of circulating pro-inflammatory adipokines and an increase of anti-inflammatory ones like adiponectin, 87 , 88 together with the promotion of anti-inflammatory pathways. 89 In the liver, in addition to the reduction of IHTG, thiazolidinediones promote the reduction of hepatic inflammation 78 , 79 and fibrosis, 90 also preventing the activation of hepatic stellate cells. 84

Dyslipidaemia and hepatic lipid dysfunction are diagnosed in the presence of high plasma concentrations of TAGs and total and LDL cholesterol. The advancement of mass spectrometry technologies has enabled the identification of hundreds (>800) of lipid species (TAGs, DAGs, ceramides, sphingomyelins, PCs, and lysophosphatidylcholines [LPCs]) in the liver and plasma/serum in recent years. These lipids were significantly altered at different stages along the spectrum of MASLD, from MASL to MASH, and we expect that new lipids will be used in the future to characterise dyslipidaemia. The most important studies are reported in Table 1 . Table 1 Changes in lipidomic profile in serum or liver of subjects with MASLD/MASH. Table 1 Study Study population Diagnosis of MASLD Lipidomic analysis Method Matrix Main findings Araya et al. 2004 97 11 CT (BMI 27.8 kg/m 2 ); 10 MASL (BMI 41.7 kg/m 2 ); 9 MASH (BMI 49.9 kg/m 2 ) Liver biopsy GC Liver ↑ MASL and MASH: MUFA, n-6:n-3 ratio, n-6 long-chain PUFA in phospholipids ↓ MASLD and MASH: long-chain PUFA, n-3 PUFA, n-6 PUFA Allard et al. 2008 96 17 CT (BMI 28.5 kg/m 2 , T2D [21%]); 18 MASL (BMI 27.2 kg/m 2 , T2D [20%]); 38 MASH (BMI 31.1 kg/m 2 , T2D [26%]) Liver biopsy GC Liver ↑ MASH: MUFA, palmitoleic acid, and oleic acid ↓ MASH: long-chain PUFA Puri et al. 2007 21 9 CT (BMI 34.5 kg/m 2 ); 9 MASLD (BMI 37.5 kg/m 2 ); 9 MASH (BMI 34 kg/m 2 ) Liver biopsy GC Liver ↑ MASLD and MASH: TAG, DAG, total cholesterol, SFA, n6: n3 ratio ↓ MASLD and MASH: TAG, FA (20:4, n-6), TAG FA (22:6, n-3) ↓ MASLD: PC, PE ↑ MASH: FC, LPC Chiappini et al. 2017 98 7 CT (BMI 21 kg/m 2 ); 39 MASLD (BMI 25 kg/m 2 ); 15 MASH (BMI 31 kg/m 2 ) Liver biopsy GC/LC-MS Liver ↑ MASH: SFA (14:0, 16:0, 18:0) ↓ MASH: PC, PE, PI, PS PC/PE, SM Peng et al. 2018 99 16 CT (BMI 41 kg/m 2 , T2D [6%]); 10 MASLD (BMI 45.5 kg/m 2 , T2D [20%]); 32 MASH (BMI 48.4 kg/m 2 , T2D [34%]) Liver biopsy HPLC-MS Liver ↑ MASL: TAG, DAG, PE (38:4), CE (14:0, 16:0, 16:1, 16:2, 17:1, 18:2, 18:3, 20:4, 20:5, 22:5, 24:6) ↑ MASH: TAG, DAG, acylcarnitine, dihexosylceramide, CER (18:0/24:1), GM1 (d18:1/16:0), SM (38:1), PE (36:1, 38:4, 18:1/22:6), LPE (18:0), CE (14:0, 15:0, 16:0, 16:1, 16:2, 17:0, 17:1, 18:0, 18:1, 18:2, 18:3, 20:3, 20:4, 20:5, 22:5, 24:5, 24:6) ↓ MASL and MASH: PC (35:2, 40:4, 40:7, 40:8), LPC (18:0, 18:1, 20:0, 22:6, 22:0) Apostolopoulou et al. 2018 100 7 CT (BMI 25 kg/m 2 ); 7 MASLD (BMI 51 kg/m 2 ); 7 MASH (BMI 56 kg/m 2 ) Liver biopsy LC-MS/MS Serum ↑ MASLD: dhCER (20:0) ↑ MASH: total dhCER, dhCER (16:0, 22:0, 24:1) Liver ↑ MASH: total CER, LactCER (24:1), HexCER(22:0, 24:0, 24:1), dhCER (16:0, 22:0, 24:1) Ooi et al. 2021 101 50 CT (BMI 45 kg/m 2 , T2D [14%]); 110 MASLD BMI 47 kg/m 2 , T2D [23%]); 16 MASH (BMI 50 kg/m 2 , T2D [33%]) Liver biopsy LC-MS/MS Plasma ↑ MASLD: CER (d18:0/16:0, d18:0/18:0, d18:0/20:0, d18:0/22:0, d18:0/24:0, d18:0/24:1), DAG SFA (16:0, 18:0), MUFA (18:1), PUFA (18:2), TAG SFA (16:0, 17:0, 18:0), MUFA (18:1), PUFA (18:2, 20:3, 20:4) Liver ↑ MASLD and MASH: dhCER, TAG, DAG, CER (d18:0/18:0, d18:0/20:0, d18:0/22:0, d18:0/24:0, d18:0/24:1), LPC (26:0), PI (18:0/22:5), CE (18:0), DAG, TAG ↓ MASLD and MASH: PC (15-MHDA/18:2), PC (15-MHDA/22:6), PC (17:1/18:2, 18:1/22:6), CE (22:5) ↑ MASLD: total CER, CE, THC, CER (d18:1/16:0, d18:1/18:0, d18:1/20:0, d18:1/22:0, d18:1/24:0), GM3 (d18:1/20:0), PC (28:0, 31:0), PC (O-40:7), PS (38:4), CE (18:3), FC ↓ MASLD: SM (37:2, d18:2/20:0), PC (17:0/18:2, 18:1/18:2, 39:5, 17:0/22:6), PC (P-38:5), PE (18:1/22:6), PE (P-18:1/22:4, 20:1/22:6) ↑ MASH: total dhCER, dhCER(d18:1/18:0, d18:1/22:0, d18:1/24:0), SM (d18:0/16:0), PC (36:0) ↓ MASH: PC (16:1/20:4, 38:6), PC (15-MHDA/20:4), PE (16:0/20:4, 38:5), PI (38:5) Puri et al. 2009 110 50 CT (BMI 21.2 kg/m 2 , no T2D); MASLD (BMI 35.2 kg/m 2 , T2D [28%]); 50 MASH (BMI 32.1 kg/m 2 , T2D [31%]) Liver biopsy GC-MS Plasma ↑ MASLD and MASH: TAG, DAG, FFA and CE MUFA, CE, DAG, PC, PE with 18:3 or 20:3 FA ↓ MASLD: DAG, PC, PE, TAG, SFA ↓ MASH: 22:6n-3/22:5n-3 ratio in PC, PE Oresic et al. 2013 112 392 CT (BMI 34.7 kg/m 2 ); 287 MASLD (BMI 34.8 kg/m 2 ) 1 H-MRS/liver biopsy LC-MS Serum ↓ MASLD: LPC (especially C16:0 and C18:0) Alonso et al. 2017 109 353 MASL (BMI 44.5 kg/m 2 ); 182 MASH (BMI 45.2 kg/m 2 ) Liver biopsy LC-MS/MS Serum ↑ MASH: PE (C20:4), CER ↓ MASH: PC (C22:6), PC (20:4)/PE (20:4) ratio Sen et al. 2022 115 206 MASH (BMI 31.3 kg/m 2 , T2D [53%]) Liver biopsy LC-QTOFMS Serum ↑ MASH (F3): deoxyCER (42:0), CER (d18:1/24:0, d18:1/23:0, d18:1/25:0) ↓ MASH (F3): hexCER (d18:1/20:0, d18:1/22:0, d18:1/23:0, d18:1/24:1, d18:1/24:0) Barr et al. 2010 113 9 CT (BMI 47 kg/m 2 , no T2D); 24 MASLD (BMI 44.8 kg/m 2 , no T2D); 9 MASH (BMI 43.2 kg/m 2 , no T2D) Liver biopsy LC-MS Serum ↑ MASLD and MASH: FFA (16:0), SM (18:0/16:0, 18:1/18:0), PC (28:0), LPC (20:2, 20:1), LPE (P-16:0) ↓ MASLD and MASH: FFA (18:3), (20:2, n-6), SM (18:1/12:0, 18:2/14:0, 18:2/16:0, 36:3), PC (18:0/22:6) ↑ MASH: PC (14:0/20:4, 16:0/20:3, P-18:0/20:4),

MASL/MASH are associated with a pro-atherogenic lipid profile with increased TAGs and LDL and reduced HDL. The reduction in IHTG and weight is linked with an improved lipidomic profile. Weight loss either through lifestyle intervention 68 or bariatric surgery (Roux-en-Y gastric bypass or sleeve gastrectomy) 69 , 70 improves the lipid profile by reducing total TAGs, and increasing PC, PE and HDL concentrations. 125 The effect of semaglutide and other single GLP-1RAs on triglycerides and HDL is generally modest, 126 despite the significant weight loss, while it is better with tirzepatide. 127 GLP-1RAs have a positive effect on postprandial TAG and lipoprotein metabolism (mainly VLDL), with a reduction in TAG concentrations in the postprandial state due to the reduction in VLDL assembly and secretion 42 , 128 , 129 but also an action on chylomicron assembly and clearance. [128] , [129] , [130] , [131] Few clinical trials have investigated the effect of GLP-1RAs on lipidomic profiles only in individuals with diabetes and during fasting ( Table 2 ). Zhang et al. 132 studied the effect of exenatide treatment on lipidomic profile in patients with newly diagnosed T2D and obesity, and despite the fact that several lipids were increased at baseline compared to healthy controls, only a few species were decreased after 12 weeks of treatment. Zobel et al. 133 investigated the effect of liraglutide for 26 weeks compared with placebo, showing a significant decrease in 21 lipid species, mainly ceramides, PCs, and TAGs. Similarly, Jendle et al. 134 showed that 18 weeks of treatment with liraglutide compared with glimepiride significantly reduced a more significant number of lipid species, including ceramides, PCs, phosphatidylinositol, cholesterol esters, and sphingomyelins. Considering the limited evidence available, at fasting, GLP-1RAs appear to reduce mainly TAGs, ceramides, PCs, and sphingomyelins, which are lipid species strictly associated with increased risk of cardiovascular disease; interestingly, this lowering effect remained after adjusting for body weight loss. Table 2 Changes in lipidomic profile after pharmacological treatment. Table 2 Study Study population Treatment Lipidomic analysis Method Matrix Main findings Bagheri et al. 125 104 individuals with obesity (SG: n = 77, RYGB: n = 27; BMI 45.6 kg/m 2 ) 12-months after bariatric surgery: SG and RYGB UPLC-MS/MS Serum ↓TAGs , ↓DAGs , ↓CE (16:1, 18:3, 18:4) ↓ CER (18:0, 22:1), ↓PE (O-16:0/20:4) ↑PC , ↑PE , ↑CE (22:1, 22:4), ↑HexCER (18:0, 20:0, 22:0, 24:0, 24:1), ↑LactCER (14:0, 16:0, 20:0, 22:0, 22:1, 24:0, 24:1), ↑SM (16:0), ↑LPC (17:0, 18:1, 18:2, 20:0, 20:1, 22:4, 22:51), ↑LPE (16:0, 18:0) Zhang et al. 132 35 patients with newly diagnosed T2D (BMI 30.9 kg/m 2 ) 12-week GLP-1RA exenatide UPLC-QTOF-MS Serum ↓SM (d18:1/18:0); ↓SM (d18:1/18:1), ↓LPC (16:0), ↓LPE (18:0) Zobel et al. 133 Liraglutide: 51 T2D (BMI 30.5 kg/m 2 ) Placebo: 51 T2D (BMI 29.3 kg/m 2 ) 26-weeks GLP1-RA liraglutide vs . placebo UPLC-QTOF-MS Plasma ↓CER (d38:1, d37:1), ↓HexCER (d34:1), ↓PE (38:6), ↓PC (36:5, 37:5, 40:7), ↓PE (O-34:2, 40:7), ↓TAG (54:6, 55:4, 56:4, 56:7, 58:6, 58:8, 58:10, 58:1, 60:9, 60:10, 60:1, 60:12) Jendle et al. 134 Liraglutide: 33 T2D (BMI 30.5 kg/m 2 ) Glimepiride: 29 T2D (BMI 29.0 kg/m 2 ) 18-weeks GLP1-RA liraglutide vs. glimepiride UPLC-QTOF-MS Plasma ↓CE , ↓CER (d18:1/16:0), ↓HexCER (d18:1/24:0), ↓ PC (35:4, 36:2, 36:4, 38:3, 38:4, 38:5, 38:6, 39:0, 40:4, 40:5, 40:8, 42:8), ↓PC (O-32:0, 34:0, 34:3, 36:3, 36:4, 36:5, 38:4, 38:5, 40:5), ↓PE (O-38:5, 38:6), ↓PI (38:3, 38:7, 44:4), ↓LPC (16:0, 18:0), ↓TAG (48:4, 48:3, 53:3, 54:5, 56:3, 56:4), ↓CE (16:0, 18:0, 18.2, 20:4, 20:5), ↓SM (37:1, d39:1, 39:2 d42:3, d34:2, d32:1, d33:1, d34:2, d36:1, d36:2, d38:1, d40:1, 40:2, d41:1, d41:2, d42:2) Warshauer et al. 135 Pioglitazone: 19 MetS (BMI 30.7 kg/m 2 ) Placebo: 18 MetS (BMI 36.1 kg/m 2 ) 6-month pioglitazone vs. placebo UPLC-MS/MS Plasma ↓CER (C18:0, C20:0, C24:1) ↓dhCER (C18:0, C24:1) ↓LactCER (C16:0), ↓HexCER (C16:0, C18 : 0, C22:0, C24:1). BMI, body mass index; CE, cholesteryl esters; CER, ceramides; DAG, diacylglycerol; dhCER, dihydroceramides; GLP1-RA, glucagon-like peptide 1 receptor agonist; HexCER, hexosylceramides; LactCER, lactosylceramide; LPC, lysophosphatidylcholine; MetS, Metabolic syndrome; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PI, phosphatidylinositol; RYGB, Roux-en-Y Gastric Bypass; SG, sleeve gastrectomy; SM, sphingomyelin; T2D, type 2 diabetes; TAG, triacylglycerol; UPLC-MS/MS, ultra pression liquid chromatography tandem mass spectrometry; UPLC-QTOF-MS, ultra pression liquid chromatography-quadrupole time-of-flight -mass spectrometry. The reduction of serum concentrations of TAGs and the increase in HDL by pioglitazone is well documented 80 and is associated with a decrease in IHTG and similar effects to lanifibranor. 83 Warshauer et al. has investigated the effect of thiazolidinediones on lipidomic profiles 135 in patients with meta

Adipo-IR, adipose tissue IR; ChREBP, carbohydrate-responsive-element-binding-protein; DGAT1/2, diacylglycerol acyltransferase 1/2; DNL, de novo lipogenesis; ER, endoplasmic reticulum; FFAs, free fatty acids; HDL, high-density lipoproteins; IHTGs, intrahepatic triglycerides; IR, insulin resistance; LDs, lipid droplets; LDL, low-density lipoprotein; LPCs, lysophosphatidylcholines; MASLD, metabolic dysfunction-associated steatotic liver disease; PCs, phosphatidylcholines; PE, phosphatidylethanolamine; PNPLA3, patatin-like phospholipase domain-containing protein 3; PPAR, peroxisome proliferator-activated receptor; PUFAs, polyunsaturated fatty acids; SAT, subcutaneous adipose tissue; SFAs, saturated fatty acids; SREBP-1c, sterol regulatory element-binding protein-1c; THR, thyroid hormone receptor; VAT, visceral adipose tissue; VLDL, very low-density lipoprotein.

A.G. has served as a consultant for: Boehringer Ingelheim, Eli Lilly and Company, Metadeq Diagnostics and Fractyl Health; has participated in advisory boards for: Boehringer Ingelheim, Merck Sharp & Dohme, Novo Nordisk, Metadeq Diagnostics and Pfizer; and has received speaker's honorarium and other fees from Eli Lilly and Company, Merck Sharp & Dohme, Novo Nordisk, and Pfizer. The other authors have no conflict of interest to declare regarding this manuscript. Please refer to the accompanying ICMJE disclosure forms for further details.

The following are the supplementary data to this article: Multimedia component 1 Multimedia component 1