Novel AGXT2-PYCR3 macrophage subtypes identified in fatty liver disease

In a discovery that could reshape how scientists understand and potentially treat one of the world’s fastest-growing liver diseases, researchers in China have identified a previously unrecognized population of immune cells that appears to drive inflammation and scarring in metabolic dysfunction-associated steatotic liver disease, or MASLD. The findings, published in Genome Medicine, reveal that macrophages lacking two key amino acid–metabolizing enzymes, AGXT2 and PYCR3, accumulate in diseased livers and behave in ways that actively promote the disease process—and, remarkably, their harmful behavior can be reversed in laboratory models simply by restoring the amino acids those cells can no longer properly process.

MASLD, formerly known as non-alcoholic fatty liver disease, affects a substantial and growing proportion of the global population, closely tracking the worldwide rise in obesity, type 2 diabetes, and metabolic syndrome. In its early and middle stages, the condition is still reversible with timely clinical intervention. Left unmanaged, however, it can progress to inflammation, fibrosis, cirrhosis, and ultimately liver failure or cancer. Precisely because the window for intervention is widest early in the disease, identifying new molecular targets has become a major priority for hepatology researchers. The new study, led by Tiansu Lv, Hongshan Dai, Shihu Zhang, and colleagues under the co-corresponding authorship of Feng Zhang and Xiqiao Zhou at Jiangsu Province Hospital of Chinese Medicine and collaborating institutions in Nanjing, offers one of the most detailed multi-scale portraits to date of what goes wrong inside the liver microenvironment during MASLD—and introduces an entirely new cell type into the picture.

What makes the study technically striking is the layered, multi-platform strategy the team employed. Rather than relying on a single analytical technique, the researchers combined high-dimensional single-cell immunophenotyping with mass spectrometry–based proteomics, phosphoproteomics, and spatial proteomics, followed by mechanistic validation in cell models. The first stage used cytometry by time of flight, or CyTOF, a technology that tags cells with heavy-metal-conjugated antibodies and measures dozens of protein markers simultaneously in each individual cell by mass spectrometry. This allowed the team to map the immune landscape of the MASLD liver in unprecedented detail, distinguishing cell populations that conventional flow cytometry would collapse into indistinguishable groups.

The CyTOF analysis produced a clear and consequential signal: myeloid-derived cells—the broad family of innate immune cells that includes monocytes, macrophages, and dendritic cells—were significantly expanded in MASLD liver tissue. That expansion made the myeloid compartment the obvious next target. The team therefore sorted these key cell populations and subjected them to liquid chromatography–tandem mass spectrometry (LC–MS/MS) with label-free quantification, probing both the total proteome and the phosphoproteome—the complete set of phosphorylated proteins that reveals which signaling pathways are switched on or off inside the cells. Phosphoproteomics is particularly powerful here because phosphorylation events are the molecular currency of cellular communication; mapping them provides a direct readout of pathway activity rather than mere protein abundance.

The proteomic and phosphoproteomic analyses converged on a surprising culprit: amino acid metabolism. Two metabolic pathways emerged as severely impaired in the myeloid cells of MASLD patients. The first was the glycine metabolic pathway, regulated by the enzyme alanine-glyoxylate aminotransferase 2, or AGXT2. The second was the proline metabolic pathway, regulated by pyrroline-5-carboxylate reductase 3, or PYCR3, an enzyme that catalyzes the final step in proline biosynthesis, converting Δ1-pyrroline-5-carboxylate into proline. Glycine and proline may sound like obscure biochemical players, but both are deeply intertwined with cellular health: glycine feeds glutathione synthesis, the cell’s master antioxidant defense, while proline is essential for protein synthesis, redox balance, and—critically for the liver—collagen production by fibrotic cells.

To find out where in the diseased liver these metabolic defects were concentrated, the researchers turned to imaging mass cytometry, or IMC. This spatial proteomics technique combines the multiplexing power of mass cytometry with high-resolution tissue imaging: tissue sections, including formalin-fixed paraffin-embedded clinical samples, are stained with panels of metal-tagged antibodies, and a laser ablates the tissue pixel by pixel while a mass spectrometer records the metal signal at each position. The result is a map showing, at single-cell resolution, which cells express which dozens of proteins—and, crucially, which cells sit next to which. Applying IMC to liver biopsies from MASLD patients and healthy controls, the team homed in on the two metabolic enzymes and made their central discovery: a subset of macrophages that were negative for both AGXT2 and PYCR3.

These AGXT2−PYCR3− macrophages were significantly enriched in MASLD livers compared with healthy tissue. But abundance alone was not the striking part. The spatial analysis showed that these cells exhibited high colocalization with inflammatory cells and fibrotic cells—they were physically clustered in the exact neighborhoods where inflammation and scarring unfold. Within the macrophage compartment, the team compared different dysregulated subsets and found that the M2-type dysregulated cluster (designated M2-C1), which encompasses the AGXT2−PYCR3− population, displayed even stronger pro-inflammatory and pro-fibrotic potential than the dysregulated M1 subset (M1-C3). This is notable because M2 macrophages are classically considered the “reparative,” anti-inflammatory arm of the macrophage family; the finding that a dysregulated M2-like subset could be more inflammatory and fibrogenic than its M1 counterpart underscores how profoundly amino acid metabolic failure rewires immune cell identity.

To move beyond correlation, the researchers built in vitro models using both human THP-1-derived macrophages and murine RAW264.7 macrophages, in which AGXT2 and PYCR3 expression was knocked down using siRNA and shRNA approaches, recapitulating the metabolic defect seen in patient tissue. The results were unambiguous. Macrophages lacking AGXT2 and PYCR3 showed enhanced proliferation and migration—behaviors consistent with aggressive tissue infiltration. They secreted higher levels of inflammatory cytokines and chemokines, the signaling molecules that recruit further immune cells to sites of damage. They also released elevated amounts of classic fibrotic proteins and exerted a strong inductive effect on hepatic fibrotic cells, essentially coaching other cells in the liver to adopt a scar-producing phenotype. A key biochemical clue accompanied these observations: intracellular glutathione, or GSH, was downregulated in the defective macrophages, linking the metabolic lesion to a collapse in antioxidant capacity and the oxidative stress that drives inflammation.

The mechanistic dissection revealed which signaling circuits were responsible. The heightened inflammatory output traced to activation of the NF-κB pathway and the MAPK/AP-1 pathway—two of the most important transcriptional programs governing inflammatory gene expression. The pro-fibrotic behavior, meanwhile, was driven by phosphorylation of SMAD3 at threonine 8 within the TGFβ signaling axis, the canonical pathway that instructs cells to produce collagen and other extracellular matrix components. In other words, losing two amino acid metabolic enzymes in macrophages was sufficient to switch on the master regulators of both inflammation and fibrosis—the twin engines of MASLD progression.

Perhaps the most clinically tantalizing result came next. When the researchers supplemented the defective macrophage cultures with the corresponding amino acids—restoring the glycine and proline supply that the broken metabolic pathways could no longer adequately generate—the aberrant phenotypes were effectively rescued. Proliferation, migration, cytokine secretion, and fibrotic signaling all receded, accompanied by reversal of the abnormal NF-κB, MAPK/AP-1, and p-SMAD3/TGFβ pathway activation. While amino acid supplementation in a culture dish is a very long way from a therapy in a patient—the study’s in vitro findings will require extensive validation, including animal studies and ultimately clinical trials—the result establishes an initial, mechanistic link between amino acid metabolism and early-to-middle-stage MASLD, and it suggests a conceptual framework in which metabolic support of immune cells might blunt disease progression.

The work also carries methodological significance for the field. By integrating CyTOF, quantitative proteomics, phosphoproteomics, and IMC within a single study design, the researchers demonstrated a pipeline that moves fluidly from unbiased discovery of cellular changes to spatial localization in actual patient tissue to mechanistic confirmation in controlled models. This end-to-end approach addresses a persistent weakness in single-cell biology, where discoveries made in dissociated cell suspensions often fail to be anchored in their true tissue context. Here, the spatial data were essential: without IMC, the intimate physical association between AGXT2−PYCR3− macrophages, inflammatory cells, and fibrotic cells would have remained invisible.

The study, conducted with ethical approval from Jiangsu Province Hospital of Chinese Medicine and published open access, was funded by the National Natural Science Foundation of China and provincial research programs. The authors note that the term MASLD is used throughout to avoid stigmatizing patients, and that no animal experiments were involved in the research. As MASLD prevalence continues to climb worldwide, the identification of AGXT2−PYCR3− macrophages offers researchers a new cellular target, a new biomarker candidate, and a fresh biochemical hypothesis—all pointing toward the possibility that the road to liver fibrosis runs, at least in part, through broken amino acid metabolism in the immune cells that patrol the hepatic frontier.