Lactobacillus plantarum ameliorates NASH-related inflammation by upregulating L-arginine production

Nonalcoholic fatty liver disease (NAFLD) affects about one-quarter of adults and encompasses a spectrum from simple steatosis (NAFL) to nonalcoholic steatohepatitis (NASH), cirrhosis, and hepatocellular carcinoma. NASH can progress to advanced liver disease and is a major and growing indication for liver transplantation. Effective pharmacologic treatments are lacking; lifestyle interventions remain standard of care. With the gut-liver axis concept, the gut microbiota has emerged as an environmental factor influencing steatosis, inflammation, and fibrosis. Probiotics may benefit NASH by altering microbiome diversity, producing metabolites that signal via host receptors, modulating bile acid signaling, and strengthening the gut barrier. Lactobacillus plantarum is a safe dietary supplement with immune-regulating properties, reported to maintain gut homeostasis, restore barrier function, regulate immune responses, and reduce liver enzymes. Despite these potential benefits, its role and mechanisms in treating NASH are unclear. This study aimed to determine whether L. plantarum can alleviate steatohepatitis, and to define the extent and mechanisms by which it affects NASH, hypothesizing actions via gut microbiome-mediated pathways.

Study design and animal models: Six-week-old male C57BL/6N mice were housed under controlled temperature (23 ± 2°C), humidity (60% ± 10%), and a 12-h light/dark cycle with ad libitum access to diet and water. After a 7-day acclimation, all mice except chow controls were fed a choline-deficient high-fat diet (CD-HFD; D05010402, Research Diets; 22.6% protein, 23.5% fat (43% kcal from fat), no choline) for 30 weeks to induce NASH. Mice were then randomized (n=8/group) to CD-HFD plus vehicle, CD-HFD plus L. plantarum, or CD-HFD plus empagliflozin for 12 weeks; chow controls received chow plus vehicle. L. plantarum (supplied by Ildong Bioscience; administered once daily by oral gavage) and empagliflozin (10 mg/kg/day by oral gavage) were given for 12 weeks. Vehicle groups received equal-volume PBS. At study end (week 42), mice were anesthetized; blood was collected by cardiac puncture; tissues were harvested for snap-freezing or formalin fixation. An additional NAFL prevention cohort received CD-HFD for 8 weeks followed by 8 weeks of L. plantarum or vehicle.

L-arginine intervention model: To test sufficiency of L-arginine, mice were fed a methionine-choline-deficient (MCD) diet (A02082002BR; 16.9% protein, 9.9% fat, 64.9% carbohydrate). After 2 weeks on MCD, mice received daily oral gavage of L-arginine (Sigma-Aldrich A5131) or vehicle for 4 weeks (total MCD duration ~6 weeks), followed by tissue and serum analyses.

Human samples: Thirty-three human liver samples (28 NAFLD patients by biopsy; 5 healthy donor controls) from Severance Hospital were subjected to RNA-seq if RNA quality was sufficient. Serum L-arginine was measured by ELISA in available individuals. IRB approval (#4-2018-0537) and informed consent were obtained per the Declaration of Helsinki.

Metabolic phenotyping: Body and liver weights and liver-to-body weight ratios were recorded. Intraperitoneal insulin tolerance tests (IPITT; 0.75 U/kg insulin after 6-h fast) and oral glucose tolerance tests (OGTT; 1 mg/g glucose after overnight fast) were conducted with tail-vein glucose measured at 0–120 min. Metabolic cage studies (Oxymax/CLAMS) measured VO2, VCO2, RER, heat production, activity, and food intake over 72 h following 1-day acclimation.

Biochemistry: Plasma AST, ALT, total cholesterol, and triglycerides were quantified (Fujifilm DRI-CHEM). Hepatic lipids (total cholesterol, triglycerides, free fatty acids) were measured in homogenates using BioVision kits. Serum endotoxin was determined with a Pierce LAL chromogenic kit.

Histology and immunohistochemistry: Formalin-fixed paraffin-embedded liver sections were stained with H&E and Sirius Red. An expert liver pathologist, blinded to groups, scored histology using NASH CRN and FLIP criteria. IHC used antibodies against F4/80 (macrophages) and MPO (neutrophils), with DAB detection and hematoxylin counterstain; images were quantified using ImageJ.

Gut barrier and permeability: Intestinal permeability was assessed by oral gavage of FITC-dextran 4 kDa (60 mg/100 g); serum FITC fluorescence was measured 4 h later. Serum endotoxin levels were measured as above. Ileal RNA-seq assessed gene expression related to barrier function.

Cell culture and in vitro assays: Caco-2 cells were cultured in DMEM with 20% FBS; LPS challenge (to disrupt barrier) was followed by L. plantarum treatment; immunoblotting for E-cadherin and occludin was performed. Primary Kupffer cells (DMEM + 10% FBS) were cultured and cocultured with mouse liver organoids (STEMCELL #70932) grown in Matrigel with HepatiCult medium in Transwells. After differentiation, LPS (100 ng/mL, 24 h) was applied to induce a NASH-like microenvironment, followed by treatment with DMSO or L-arginine (1 mM, 24 h). Inflammatory gene expression (e.g., IL-1β, IL-12) was measured by RT-qPCR.

Molecular assays: RNA isolation from mouse liver and ileum, Kupffer cells, and organoids used TRIzol or RiboEx; cDNA was synthesized and qPCR performed with SYBR Green; GAPDH served as housekeeping control. Immunofluorescence on organoids used standard fixation, blocking, and staining with Alexa-conjugated secondaries and DAPI; confocal imaging by Zeiss LSM780.

Microbiome analyses: Cecal DNA was extracted (QIAamp DNA Stool Mini Kit). Whole metagenome shotgun libraries were prepared and sequenced on Illumina HiSeq 4000. Processing used bioBakery workflows: KneadData (QC), MetaPhlAn (taxonomic profiling), HUMANN (functional profiling via UniRef90 and KEGG), and StrainPhlAn (strain profiling). 16S rRNA amplicon sequencing also assessed community diversity (ACE, Chao1 richness; Shannon, Simpson diversity) and beta diversity (Bray–Curtis PCoA). Taxonomic compositions were summarized from phylum to species, including detection of L. plantarum in treated mice.

Serum L-arginine: Quantified by ELISA (ImmuSmol #IS 1-0400) per manufacturer’s protocol.

RNA sequencing and bioinformatics: Mouse and human libraries were prepared (Illumina TruSeq) and sequenced (NovaSeq/HiSeq). Reads were mapped to mouse (GRCm38/mm10) or human (GRCh38/hg38) genomes using CLC Genomics Workbench. Count data were variance-stabilized and analyzed with DESeq2. Differential expression criteria: adjusted P < 0.05 and |log2FC| ≥ 1. Gene set enrichment analysis (GSEA v4.3.1) used MSigDB GO, hallmark, and KEGG sets. Cross-species comparisons correlated pathway dysregulation between murine and human NAFLD.

Statistics: Unpaired nonparametric tests (including Mann–Whitney) and ANOVA with Tukey’s multiple comparisons were applied; P < 0.05 was considered significant. Analyses were conducted in SAS 9.4; visualization in GraphPad Prism and BioRender.

• In CD-HFD-induced NASH mice, L. plantarum reduced liver weight and liver-to-body weight ratio compared with vehicle; macroscopic livers were smaller.
• L. plantarum did not significantly improve insulin resistance or glucose tolerance in the NASH model; empagliflozin did. L. plantarum nonetheless reduced hepatic lipid content and improved NAFLD-related metabolic phenotypes without altering energy expenditure, activity, or food intake.
• Histology and IHC showed that L. plantarum decreased steatosis-associated inflammation: reduced F4/80+ macrophage and MPO+ neutrophil areas; lowered NAFLD activity score; and decreased serum ALT.
• Liver RNA-seq in L. plantarum-treated mice showed significant downregulation of inflammatory processes (e.g., regulation of T-cell activation/proliferation). Hallmark pathways such as inflammatory response, interferon-gamma response, interferon-alpha response, IL6–JAK–STAT3, and TNFα signaling via NF-κB were negatively enriched.
• Cross-species analysis confirmed that inflammatory KEGG pathways upregulated in human NAFLD were also up in murine NASH and were downregulated by L. plantarum treatment. Murine and human NAFLD transcriptomes were positively correlated (Pearson r ≈ 0.71 for comparable contrasts).
• Metagenomics revealed enrichment of microbial functional genes for L-arginine biosynthesis in L. plantarum-treated mice, specifically pathways III (via N-acetyl-L-citrulline) and IV (via archaeal-like pathway). Serum L-arginine concentrations were significantly higher in L. plantarum-treated mice than in vehicle or chow groups.
• L. plantarum improved gut barrier integrity: reduced serum FITC-dextran permeability and endotoxin (LPS) levels; ileal transcriptomics showed rescue of barrier-related genes suppressed by CD-HFD; in LPS-challenged Caco-2 cells, L. plantarum restored E-cadherin and occludin levels.
• Preventive NAFL model: L. plantarum given during early disease inhibited hepatic fat accumulation and lowered AST/ALT; it also reduced insulin resistance in this prevention context.
• L-arginine sufficiency: In MCD diet-induced NASH, oral L-arginine for 4 weeks reduced liver inflammation (lower neutrophil infiltration) and decreased serum AST/ALT without significantly changing hepatic triglycerides, cholesterol, or FFA. Liver transcriptomics showed broad downregulation of inflammatory genes and pathways, with TNFα–NF-κB signaling most significantly suppressed, mirroring effects seen with L. plantarum.
• In vitro, L-arginine reduced IL-1β and IL-12 mRNA in LPS-exposed liver organoids cocultured with Kupffer cells.
• Human data: In a limited cohort (n=19 with serum), circulating L-arginine levels tended to decline with NAFLD progression from steatosis to NASH with fibrosis, suggesting a potential association with disease severity.

The study addressed whether L. plantarum can ameliorate NASH and through what mechanisms. In a CD-HFD murine model that recapitulates human NAFLD transcriptomes, L. plantarum markedly reduced hepatic inflammation, immune cell infiltration, injury biomarkers (ALT), and inflammation-related transcriptional programs (e.g., interferon responses and TNFα–NF-κB). Whole-metagenome profiling implicated microbiome-mediated augmentation of L-arginine biosynthesis (pathways III and IV) in L. plantarum-treated mice, which corresponded with elevated circulating L-arginine. Functional validation using MCD-induced NASH and liver organoid–Kupffer cocultures showed that L-arginine alone is sufficient to reduce inflammatory markers and downregulate NF-κB signaling, aligning with the anti-inflammatory signature observed with L. plantarum. L. plantarum also strengthened gut barrier integrity and reduced endotoxemia, a known driver of NASH progression, suggesting a dual mechanism: increased L-arginine availability and improved intestinal barrier function that together dampen hepatic inflammation. These findings support a gut-liver axis mechanism in which L. plantarum colonization enhances microbial arginine biosynthesis, raises systemic L-arginine, and suppresses inflammatory signaling in the liver, highlighting a plausible therapeutic strategy for NASH.

L. plantarum administration ameliorated NASH-associated liver inflammation in mice, decreasing histologic inflammation, immune infiltration, and ALT levels while downregulating key inflammatory pathways such as TNFα–NF-κB and interferon responses. Shotgun metagenomics and serum assays indicated that L. plantarum enriches microbial L-arginine biosynthesis pathways and increases circulating L-arginine. Independent L-arginine supplementation reproduced anti-inflammatory effects and suppressed NF-κB signaling in vivo and in liver organoids, supporting L-arginine as a mediator of L. plantarum’s benefits. L. plantarum additionally improved intestinal barrier function and reduced endotoxemia, further contributing to hepatoprotection. Future research should validate these mechanisms in human NAFLD/NASH, quantify L. plantarum-driven arginine production in humans, assess clinical efficacy and dosing, and explore long-term safety and durability of microbiome modulation strategies.

• Translational gap: Findings from CD-HFD and MCD murine models may not fully generalize to human NASH due to species differences, including gut mycobiome and microbiome ecology.
• Human validation is limited: Elevated L-arginine production and the therapeutic effect of L-arginine or L. plantarum require confirmation in human interventional studies.
• The precise L. plantarum dose-response and long-term colonization dynamics were not fully characterized.
• While both increased L-arginine and improved barrier function were implicated, the relative contribution of each mechanism was not dissected with causal interference (e.g., arginase inhibition/NO pathway blockers, barrier-specific modulation).
• Dietary and first-pass intestinal metabolism may limit systemic availability of exogenous L-arginine, complicating translation and dosing strategies.
• Some metabolic outcomes (insulin/glucose tolerance) did not improve with L. plantarum in the NASH treatment setting, indicating a primarily anti-inflammatory rather than insulin-sensitizing effect in that model.