We read with great interest the review article by Kakazu et al., which outlines a compelling framework for understanding the role of amino acids in the pathogenesis of metabolic dysfunction-associated steatotic liver disease (MASLD) and steatohepatitis (MASH) [1]. While the authors consolidate a wealth of information on amino acid dysregulation, several key aspects merit further investigation. Below, we offer three areas where the current synthesis could be substantially advanced.

The review effectively highlights metabolic zonation within the liver lobule—GLS2-mediated glutaminolysis in zone 1 and GS-mediated glutamine synthesis in zone 3—as a key architectural feature influencing amino acid handling. However, this binary spatial model likely oversimplifies a highly dynamic and disease-modified system. In MASLD/MASH, inflammatory remodeling and hypoxia can disrupt classical zonation, resulting in atypical expression patterns of enzymes such as GLS1, GS, and even xCT, particularly in hepatocellular carcinoma-prone microenvironments. Recent spatial transcriptomics studies have shown that lobular organization in non-alcoholic steatohepatitis livers is progressively lost, with fibrotic tracts and regenerative nodules exhibiting ectopic expression of zone-specific markers [2]. In this context, it is insufficient to describe enzyme expression statically; rather, isotope tracer-based flux analysis (e.g., 13C-glutamine or 15N-glutamate infusion) in diseased human livers or advanced models is needed to track functional amino acid routing. Such approaches could reveal whether zone 1 glutaminolysis still fuels gluconeogenesis in MASLD, or whether GLS1-driven glutaminolysis in activated hepatic stellate cells represents a dominant pathogenic node. These distinctions bear therapeutic consequences— whether to target GLS1 inhibition broadly or in a cell-type-specific manner remains an open question.

Kakazu et al. emphasize the role of the gut microbiome in shaping circulating amino acid profiles, particularly branched-chain amino acids (BCAAs), and their association with insulin resistance. However, current interpretations remain largely correlative. For example, while Prevotella copri and Bacteroides vulgatus are noted BCAA producers, their presence does not consistently predict MASLD severity. Conversely, CRISPR-based microbiome editing in recent studies has demonstrated that manipulation of bacterial amino acid metabolism can causally reshape host serum BCAA levels and affect metabolic phenotypes [3]. The key next step is to clarify the directionality and mechanistic mediators of this interaction. Specifically, do microbiota-derived BCAAs act directly on hepatic mTORC1 to promote lipogenesis, or is their pathogenicity indirect via peripheral insulin resistance or macrophage reprogramming? Furthermore, some microbial metabolites such as indolepropionic acid (a tryptophan derivative) or short-chain fatty acids may counteract the harmful effects of BCAA excess by stabilizing gut barrier integrity or modulating hepatic inflammation. Future studies should aim to stratify MASLD patients by microbial metabolic output—not just taxonomic composition—and integrate this with plasma and hepatic amino acid metabolomics. Therapeutically, microbial manipulation (e.g., selective bacteriophage therapy, engineered probiotics) targeting BCAA overproduction could complement dietary restriction strategies in MASLD.

The review rightly identifies mTORC1 and GCN2/ATF4 as nutrient-sensing pathways linking amino acid availability with anabolic or catabolic states in the liver. However, their translational significance in MASLD is vastly underexplored. For example, while hepatic mTORC1 is often activated in MASLD and promotes lipogenesis via SREBP1, systemic mTORC1 inhibition is limited by adverse effects on muscle and immune homeostasis. This presents an opportunity to explore compartment-specific modulation. Liver-targeted mTORC1 inhibitors, or sestrin mimetics that selectively modulate mTORC1 in zone 3 hepatocytes, may offer precision anti-steatotic therapy without systemic toxicity. Conversely, GCN2 activation under amino acid deprivation promotes ATF4 and FGF21 expression, which can enhance lipid oxidation and improve insulin sensitivity. Interestingly, FGF21 analogues such as pegozafermin have already shown promise in MASH treatment [4]. Yet, whether their beneficial effects are downstream of GCN2-ATF4 signaling in hepatocytes or mediated by other endocrine axes (e.g., adiponectin signaling) remains unclear. Preclinical models with hepatocyte-specific GCN2 or ATF4 deletion could illuminate whether activating this pathway pharmacologically (e.g., with leucine mimetics or tRNA charging inhibitors) holds therapeutic value. Furthermore, the review briefly notes that amino acid sensors like mTOR and GCN2 regulate not only hepatocyte metabolism but also influence hepatic stellate cell activation and immune cell phenotypes. Integrating amino acid sensing with hepatic cellular crosstalk models may help uncover synergistic targets that could halt fibrosis progression without compromising metabolic flexibility.

Kakazu et al. have made a valuable contribution in foregrounding amino acid metabolism within the broader pathophysiology of MASLD. To move beyond descriptive biochemistry, future research should integrate spatial and flux-based tools, leverage microbiota-targeted interventions, and explore zonal- and cell-specific nutrient-sensing therapeutics. Such mechanistic depth and translational focus will be essential to unlock the full therapeutic potential of amino acid modulation in MASLD/MASH.

Citations to this article as recorded by