Skip to main content
NCBI home page
Scientific Reports logoLink to Scientific Reports
. 2026 Apr 9;16:16174. doi: 10.1038/s41598-026-47841-8

Remote regulation of hepatic lipid secretion in the intestine by metabolic interaction of dietary arginine with ornithine

Hiroki Nishi 1,2,✉, Sena Nakanishi 2, Lin Xie 3, Masayuki Fujinaga 3, Yiding Zhang 3, Daisuke Yamanaka 4, Suguru Fujita 2, Junji Saruwatari 5, Kentaro Oniki 5, Ming-Rong Zhang 3, Fumihiko Hakuno 2,✉
PMCID: PMC13201593  PMID: 41957409

Abstract

The quantity and quality of dietary protein profoundly influence satiety, body growth, and systemic metabolism. Among the 20 major proteinogenic amino acids, arginine (Arg) is considerably associated with hepatic steatosis; when animals fed an Arg-deficient diet (ΔArg), triacylglyceride (TAG) dramatically accumulates in the liver. To explore the underlying mechanism, we first investigated the role of ornithine (Orn), as Orn is the primary metabolite of Arg and it is reportedly involved in the regulation of liver metabolism. While male Wistar rats fed a ΔArg diet exhibited a significant increase in liver TAG levels due to an attenuated TAG secretion, and consistently marked reduction of TAG-rich lipoproteins in the circulation, Orn addition to the diet completely abolished all these metabolic changes. Orn was only effective when taken orally, but not through intraperitoneal administration, suggesting that the intestine plays an essential role for Orn to regulate liver metabolism. The metabolic features similar to those of our rat model was also observed in the analyses of clinical samples, implying the common mechanism in humans. Conclusively, dietary Arg deficiency lowers local Arg-to-Orn conversion in the intestine, which in turn inhibits hepatic lipid secretion remotely via gut-liver axis.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-47841-8.

Keywords: Fatty liver, Dyslipidemia, Very-low-density lipoprotein, Arginine, Ornithine

Subject terms: Biochemistry, Gastroenterology, Physiology

Introduction

Although the contribution of dietary carbohydrates and fats to the progression of metabolic diseases such as dyslipidemia and metabolic dysfunction-associated fatty liver disease (MAFLD) were extensively studied, knowledge of the role of dietary proteins remains limited1–4. Emerging evidence demonstrated that the total protein intake and the source of dietary protein is closely associated with the risk of metabolic syndrome5,6; hence, studies should focus on the amino acid composition of proteins, which is associated with protein quality. The excess or reduced levels of several major proteinogenic amino acids in a diet (e.g., leucine, valine, serine, and tyrosine) impact the maintenance of metabolic health7–10. Previously, we demonstrated that arginine (Arg) intake is an important determinant of fatty liver development in rats11,12. To explore the underlying mechanism, we focused on the role of the non-proteinogenic amino acid ornithine (Orn), a primary metabolite of Arg, which is produced by an arginase-catalyzed reaction13 and is involved in various metabolic regulations in the liver14–16. Here, using this fatty liver model, we further investigated the physiological function of dietary Arg and its impact on the hepatic lipid metabolism, focusing on the role of Orn.

Results and discussion

A ΔArg diet containing 0.17% Arg (Supplementary Data Table 1) fed for 1 week induced considerably increased triacylglyceride (TAG) accumulation (almost eight times) in the liver of male Wistar rats than that in the control diet (CN, containing 0.52% Arg); however, body growth, food intake, and calorie intake were comparable in these groups (Fig. 1A–E). The Total cholesterol (TCho) levels in the liver did not significantly differ between the two groups, indicating that a ΔArg diet selectively induced a TAG-dominant fatty liver (Fig. 1F). Contrastingly, serum TAG concentrations were significantly reduced in ΔArg diet-fed rats (Fig. 1G,H). As the majority of circulating TAG is associated with very-low-density lipoprotein (VLDL) and chylomicron, derived from the liver and intestine, respectively17, these results suggested that the release of TAG from the liver to the circulating blood is interrupted by the ΔArg diet. A liver TAG secretion assay, performed by injecting liquid surfactant tyloxapol inhibiting lipolysis of lipoproteins, confirmed that TAG secretion from the liver was significantly slowed down in ΔArg diet-fed rats than that in CN diet-fed rats (Fig. 1I).

Fig. 1.

Fig. 1

Open in a new tab

Low-arginine diet promotes the accumulation of triacylglyceride (TAG) in liver by decreasing its secretion from the liver. Six-week-old male Wistar rats were reared on either a control (CN) or a low-arginine diet (ΔArg) for 7 days. (A,B) Changes in the body weight and total food intake during the period of experimental diet feeding. (n = 6). (C) Representative photographic images of the liver of the rats and representative images of H&E-stained sections of the liver. (D–F) Liver weight of the rats, and the total triacylglycerol (TAG)- and cholesterol (TCho)-contents. (n = 4). (G,H) Concentrations of total TAG and TCho in the rat serum. (n = 3). (I) The rats were administered 200 mg/kg of tyloxapol via the tail vein, and tail blood was collected 0, 1, 2, and 4 h after administration (n = 5). The serum level of TAG was measured (left). The slope of the linear regression lines obtained based on the individual plasma TAG data is presented as hepatic TAG secretion rate (right). Bar: mean ± S.E.M., *p < 0.05 (Student’s T test).

To explore the underlying mechanism, we focused on the role of Orn. Orn supplementation to a ΔArg diet completely restored fatty liver development (Fig. 2A,B). Consistently, attenuated TAG secretion from the liver and the reduced circulating TAG levels were completely restored (Fig. 2C,D). Furthermore, while the concentration of TAG-scarce high-density lipoproteins (HDLs) in the serum was comparable between CN-fed and ΔArg-fed rats, TAG-rich fraction of lipoproteins (non-HDL-cholesterols [non-HDL-C]; calculated as the total lipoprotein cholesterols —HDL-cholesterols [HDL-C]) was correspondingly restored by Orn supplementation (Fig. 2E,F). Surprisingly, in rats previously fed a ΔArg diet, the Orn supplementation immediately increased the reduced levels of serum TAG and non-HDL-C, achieving comparable levels with the CN group within one night only (Supplementary Fig. 1).

Fig. 2.

Fig. 2

Open in a new tab

Ornithine supplementation restores low-arginine diet-induced liver TAG accumulation. Wistar rats were fed with the CN or ΔArg diet, supplemented with or without 1% Orn, for 7 days as described in Fig. 1. (A,B) TAG concentrations in the liver and serum were measured. (n = 5). (C) Representative images of H&E-stained section of the liver. (D) Liver TAG secretion assay conducted as described in Fig. 1I. (n = 6). (E,F) HDL-C and non-HDL-C concentrations in the serum. (n = 5). (G) Western blotting assay using sera and liver lysates. (n = 5). (H) Band intensity of G was quantified. (I) TAG transfer activity of crude liver homogenates. In the presence of liver homogenates, the amount of labeled-TAGs transferred from donor vesicles to acceptor vesicles by MTP was monitored by measuring fluorescent intensity. Bar: mean ± S.E.M., *p < 0.05 (CN vs. ΔArg), #p < 0.05 (ΔArg vs. ΔArg + Orn) (Tukey–Kramer test).

Next, we assessed lipoprotein assembly in the liver. The primary TAG-carrying species of non-HDL is VLDL, which is assembled in hepatocytes carrying one molecule of apolipoprotein B (ApoB) per particle and undergoes TAG loading concomitantly mediated by microsomal triglyceride-transfer protein (MTP) and protein disulfide isomerase (PDI), and is secreted from the cell into blood circulation through the SAR1B-SURF4-COPII cargo trafficking system17–20. Expectedly, ApoB protein levels were enhanced in the livers of ΔArg-fed rats, whereas those in the serum were decreased; however, Orn supplementation reversed these changes (Fig. 2G,H). Furthermore, SURF4 expression in the liver tended to decrease after feeding the ΔArg diet, further supporting the notion that a ΔArg diet attenuates the VLDL export from the liver. MTP and PDI proteins were upregulated in the livers of ΔArg-fed rats, implying that VLDL assembly was enhanced; it was further validated by the in vitro MTP assay, reflecting TAG transfer activity of crude liver extract of each rat, which revealed that MTP activity in the liver was increased by a ΔArg diet and reversed by Orn supplementation (Fig. 2I).

Orn supplement, administered orally rather than intraperitoneally, exhibited the aforementioned counteracting effects (Fig. 3A). Consistently, serum ApoB concentrations, which were reduced in ΔArg-fed rats, were not recovered by Orn intraperitoneal injection (Fig. 3B). Additionally, the Orn concentration was elevated locally in the portal blood immediately after oral ingestion; however, at the equilibrium state, its concentration in the systemic circulation and liver was almost comparable among the different diet groups, regardless of Orn administration (Supplementary Fig. 2). Hence, it was unlikely that the ingested Orn per se traveled to and was captured by the liver, where it functioned as a metabolic modulator; passing through the gastrointestinal tract was potentially crucial for Orn to display metabo-regulatory effects on the liver. To clarify this, we performed a biodistribution assay using radiolabeled Orn (11C-Orn). In this assay, 11C-Orn was administered to the stomach of rats through oral gavage and traced using positron emission tomography (PET) coupled with live computed tomography (CT); the massive signal, initially observed in the stomach, gradually decreased for 90 min after the administration (Supplementary Figs. 3, 4). Signals in the intestinal tract increased over time and its accumulation in the bladder was detected within 30 min. Based on this observation, we conducted the biodistribution assay during 30 min after the administration of 11C-Orn. Results indicated that ΔArg-fed rats exhibited faster expulsion of 11C-Orn from the stomach than that in the CN-fed rats. Contrastingly, more 11C-Orn was retained in the small intestine and urinal excretion was lessened in ΔArg-fed rats, however, no significant difference was detected (Fig. 3C). The oral gavage of an indigestible and unabsorbable dye, Evans Blue, revealed that the gastrointestinal motility was comparable between ΔArg- and CN-fed rats (Fig. 3D), which excludes the possibility that the retainment of Orn in the intestine was attributed to the impaired gastrointestinal transit of luminal content. Instead, the active trapping of Orn by intestinal tissue is possible.

Fig. 3.

Fig. 3

Open in a new tab

Orally ingested ornithine remotely regulates liver lipid metabolism through the gut-liver axis. Wistar rats were fed with the CN or ΔArg diet for 7 days, as described in Fig. 1. (A,B) Rats were intraperitoneally injected with Arg, Orn (0.5 mmol/rat/day), or vehicle (saline) daily during the period of experimental diet feeding, and the liver TAG-content was measured (A). The amount of ApoB in the serum was analyzed by immunoblotting (B). (n = 5). (C) Biodistribution assay using 11C-Orn. After the experimental diet feeding, rats were starved of food and water, 9.9–10.36 MBq of 11C-Orn was administered through oral gavage, and kept free for 30 min; subsequently, the rats were sacrificed, and their tissues were collected immediately. Data are expressed as the mean percentage of the injected radioactivity dose per gram of tissue weight (%ID/g). (n = 3). (D) The gastrointestinal transit of luminal contents was assessed. An indigestible and unabsorbable dye, Evans Blue, was put in the stomach through oral gavage, and the distance it spread over 50 min was expressed as a percentage of the total length of the small intestine. (n = 4). (E,F) Rats were treated with rapamycin (2 mg/kg) via intraperitoneal injection daily during the period of experimental diet feeding. The liver TAG content and secretion were measured. (n = 5). Bar: mean ± S.E.M., *p < 0.05 (CN vs. ΔArg), #p < 0.05 (ΔArg vs. ΔArg + Orn) (C, Student’s T test; others, Tukey–Kramer test).

The gut-liver axis plays an important role in organismal nutrient sensing and associated metabolic regulation in the liver21–23. Considering the above results, we hypothesized that ingested Arg reaches the small intestine, where it is metabolized into Orn, which transfers signals to the liver via the portal vein. To test this hypothesis, we initially investigated whether Orn is further metabolized to produce a signal-transmitting metabolite. Since the major metabolic pathway of Orn in mammalian tissue yields polyamines (putrescine, spermidine, and spermine), citrulline, and proline, we measured relative concentrations of these metabolites in the portal blood plasma of rats (CN: ΔArg: ΔArg + Orn); the results (putrescine, 1.00:0.93:0.71; spermidine, 1.00:1.19:1.12; spermine, not detected; citrulline, 1.00:1.05:0.84; and proline, 1.00:0.73:0.95) reflected no apparent correlation with hepatic TAG accumulation. Next, the contribution of the gut microbiota was evaluated, because it plays an indispensable role in gut-liver communication by converting luminal content into beneficial products, such as bile acids, required for the host23. A bile acid species derived from microbiota was reported to ameliorate diet-induced fatty liver in mice24. The total bile acid levels in the portal blood plasma were significantly higher in rats fed a ΔArg diet than those in CN-fed rats (Supplementary Fig. 5B). Therefore, we fed rats the bile acid sequestrant cholestyramine as a dietary supplement to inhibit bile acid uptake from the intestinal lumen, and the increase in bile acids in the portal blood of ΔArg-fed rats was depleted; however, significant hepatic TAG accumulation by a ΔArg diet was still observed (Supplementary Fig. 5A-C). Additionally, after the depletion of gut microbiota by an antibiotic cocktail added to the drinking water, a significant amount of ΔArg diet-induced hepatic TAG accumulation was continued; however, the extent was slightly smaller compared with that of the drug-untreated control counterparts (Supplementary Fig. 5D,E). Furthermore, 16 S ribosomal DNA metagenomic sequencing analysis of stool samples revealed that the composition and abundance of the gut microbial population were not affected by dietary Arg and Orn (Supplementary Fig. 5F–H). Summarily, these results suggest that the intestinal tissue, rather than the luminal microbiota, receives Arg/Orn and transmits the information to the liver.

To identify the molecules receiving Arg/Orn and evoking downstream signal transduction, we tested the involvement of the mechanistic target of rapamycin complex 1 (mTORC1), which is a key player in cellular nutrient sensing and is involved in diverse metabolic regulations in response to nutritional status25,26. We focused on Arg, considering that it is one of the best-known activators of mTORC127. Pharmacological inhibition of mTORC1 by intraperitoneal administration of rapamycin totally prevented ΔArg diet-induced attenuation of hepatic TAG secretion, and resultant fatty liver development (Fig. 3E,F), suggesting that ΔArg diet-induced mTORC1 activation, induced anywhere in the body, underlay the mechanism.

Fatty liver and dyslipidemia are generally associated with obesity and excessive energy intake; for instance, a high-fat diet is considered its primary cause27–30. However, the current results show that some cases exhibit a trade-off between fatty liver and dyslipidemia without consuming massive amounts of energy. To validate these results, we investigated amino acids and lipids in the blood and fatty liver using human clinical samples, and investigated whether individuals having such characteristics existed. We sampled the data of amino acids, TAG, and LDL-C levels in the blood, as well as parameters related to the liver condition of 678 individuals, which were collected during physical examinations at Kumamoto University in Japan, and canonical correlation analysis was performed (Fig. 4A–C). Overall, the common component 1 (CC1) reflected a positive correlation of fatty liver with TAG/LDL-C concentrations, BMI, and ALT, validating the general assumption. Furthermore, a remarkable correlation between fatty liver and blood amino acid concentrations was detected, which was comparable with that of the earlier findings in a rat model11,31. Using those data, we searched for individuals having fatty livers with low blood levels of LDL-C and TAG (referred to as ‘FL_LLLT’). First, the data were divided into two groups based on the presence of fatty liver, and subsequently categorized into nine subgroups based on LDL-C and TAG levels in the blood (Fig. 4D). Here, we considered LDL-C levels ‘high,’ ‘norm,’’, or ‘low’ at ≥ 140 mg/dL, 101–140 mg/dL, and < 101 mg/dL, respectively, whereas the corresponding ranges of TAG levels were ≥ 150 mg/dL, 61–150 mg/dL, < 61 mg/dL, respectively32. The individuals having both fatty liver and hyperlipidemia (‘FL_HLHT’) tended to show overall higher levels of plasma amino acids than those of the other categories, which was consistent with that of the results of the canonical correlation analysis discussed earlier (Fig. 4D,E). Finally, we found that two out of 678 individuals were classified as ‘FL_LLLT’ (Fig. 4E,F). Their blood amino acid profiles compared with those of individuals with no fatty liver, normal blood LDL-C, and normal blood TAG (‘NFL_NLNT’) revealed a tendency to exhibit lower Arg and Orn, and higher Gln and Met concentrations; however, statistical significance remained unclear due to the small sample size (Fig. 4E, Supplementary Fig. 6). This pattern of amino acid profile was similar to that of ΔArg diet-fed rats31. Furthermore, the blood amino acid profile of the individuals with no fatty liver and hyperlipidemia (‘NFL_HLHT,’ 18 individuals) did not exhibit such patterns, and sometimes was contrasting to that of ‘FL_LLLT’ (Fig. 4E). Therefore, similar to those of the outcomes in the rat model, the distribution of lipids between the liver and circulation in humans can be perturbed by a certain amino acid composition of dairy food, which results in hepatic lipid accumulation regardless of excessive energy intake.

Fig. 4.

Fig. 4

Open in a new tab

Blood amino acid levels correlate with the lipid translocation between the liver and circulation in humans. (A) Conceptual diagram of canonical correlation analysis (see methods). (B) Heatmap reflecting the weight of blood amino acids, ALT and AST levels, and clinical items. CCx represents a common component created from blood amino acids, ALT and AST levels, and clinical items. (C) Heatmap of the blood LDL and TG levels, and fatty liver. Darker purple and orange indicate higher negative and positive correlations, respectively. CCy represents the common component created from blood LDL, TG levels, and labels for fatty liver. (D) Classification of subgroups in this study. Numerals in the table represent the number of individuals in each group. (E) Scatter plot of fatty liver, blood LDL-C, and blood TAG concentrations. The dots indicate each individual. Dotted lines represent the border of subgroups according to (D). The color of the dots highlights the subgroups of our interest (see D). (F) The focused average amino acid profile of the subgroups. Data are expressed as relative values normalized against the average value of the ‘NFL_NLNT’ group. Definition of the terms: ‘NFL_NLNT’ (blue), individuals with no fatty liver, normal blood LDL-C, and normal blood TAG levels; ‘NFL_HLHT’ (purple), individuals with no fatty liver and hyperlipidemia; ‘FL_HLHT’ (cyan), individuals with fatty liver and hyperlipidemia; ‘FL_LLLT’ (magenta), individuals with fatty liver, low blood LDL-C, and low blood TAG levels.

Conclusion

The current rat model-based study revealed that dietary Arg is sensed by the intestinal tissue, where it is metabolized into Orn, and it remotely regulates the expression of PDI/SURF4 proteins in the liver via an mTORC1-mediated mechanism needing further clarification, through which the hepatic lipid flow is properly maintained. Earlier reports demonstrate that mice lacking SURF4 in the liver and cultured hepatocytes overexpressing PDI phenocopies the current results18,20, and phenotypes resembling the ΔArg-fed rat model were detected in humans (Fig. 4, Supplementary Fig. 6), underpinning the conclusion. This also implies that humans and rodents share a common mechanism associated with their metabolic responses to the nutritional status, at least in part. Simultaneously, the current study has some limitations. Since the current study has included only male animals, sexually dimorphic traits might possibly lie in the phenotype/mechanism discussed here. Additionally, further study can clarify the following questions: ‘what receives Arg/Orn?’, ‘where it functions?’, and ‘how it undertakes distant communication between the gut and liver?’. Here, we demonstrated that the generally accepted Arg-sensing machinery, mTORC1, was also involved in this mechanism (Fig. 3E,F). However, this result seemed surprising as well as confusing because the fact that Arg deficiency may activate mTORC1 contradicted the well-described canonical effect of Arg on mTORC1 activation33. Actually, in tissues of ΔArg-fed rats tested here, activation of mTORC1 was not more than CN-fed counterparts, at least when whole tissue lysates were analyzed (Supplementary Fig. 7). The results imply that mTORC1 of a minor population of the specific cell type may be activated in response to a ΔArg diet rather than that of a whole organ/tissue. This idea was proven to be applicable in some cases, such as the intestine that has undergone calorie restriction; for instance, mTORC1 in most cell types is inactivated by calorie restriction, whereas only intestinal stem cells exhibit a remarkable upregulation of mTORC1 activity34. Further exploration to identify these cell types is warranted to determine the molecules communicating nutritional information to the liver, which can deepen our understanding of organismal nutrient-sensing systems and metabolic responses to nutrients and provide a new insight into the pathophysiology of diet-induced metabolic dysfunction-associated diseases.

Materials and methods

Materials

Vitamin mixture (OYC formula), mineral mixture (OYC formula), cellulose powder, and corn starch used here were purchased from Oriental Yeast Co. (Tokyo, Japan); soybean oil was supplied by Nacalai Tesque (Tokyo, Japan); Rapamycin, Cholestyramine, and Tyloxapol were obtained from LC Laboratories (MA, USA), MedChemExpress (NJ, USA), and Sigma Aldrich (MO, USA); total Bile Acid Assay Kit was purchased from CELL BIOLABS (CA, USA). All other reagent-grade chemicals used here were commercially available.

Animal experiments

All experimental diets used here were prepared by mixing ingredients evenly (Supplementary Data Table 1). All diets contained an amino acid mixture as the sole nitrogen source, whose formula was determined based on that of bovine casein. The control diet (CN) contained 15.3%, (w/w) 64.8%, and 5.0% amino acids, carbohydrates, and lipids, respectively, with a total energy of 3.7 kcal/g; the ΔArg diet with a similar composition, differed only by the reduced amount of Arg compensated with the same amount of starch, making the diets isocaloric. For the ΔArg + Orn diet, 10 g/kg of Orn (FUJIFILM Wako Pure Chemical, Osaka, Japan) was added to a ΔArg diet. All the diets were prepared immediately before the experiment and stored at − 30 °C until use. In all experiments, male animals were subjected to exclude the influence of estrous cycle.

Five-week-old male Wistar rats were purchased from The Jackson Laboratory (Kanagawa, Japan) and Japan SLC (Shizuoka Japan), caged individually, and maintained under a controlled environment with 50–60% humidity and a 12 h light/dark cycle (8:00–20:00/20:00–8:00) at 24 ± 1 °C; free access to food and water was provided throughout the experiment. Initially, the rats were reared on a normal chow diet (CE-2; CLEA Japan, Tokyo, Japan) for at least 3 days, and subsequently, fed the CN diet for 4 days as an acclimation period. Next, the animals were divided into experimental groups; the average and variance of body weight were approximately the same across the groups. Each group was fed one of the experimental diets and maintained for 7 days. The body weight and food intake of all rats were measured daily at 10:00 a.m.. On day 7, the rats were fasted for 1 h in the morning, followed by anesthetization using isoflurane (DS Pharma Animal Health, Tokyo, Japan) and decapitation. Blood samples were collected from the carotid arteries, and livers were isolated from the carcass. The blood samples were kept on ice for 1–2 h to induce clotting; subsequently, they were centrifuged at 1200×g, 4 °C for 15 min to collect sera. For portal blood collection, the abdomen was opened under isoflurane anesthesia, the portal vein was exposed, and blood was drawn using a syringe and a heparinized needle. The blood samples were immediately centrifuged at 1,200×g, 4 °C for 15 min to collect plasma. The isolated tissues were weighed and immediately frozen in liquid nitrogen for biochemical analyses or immersed in a 10% paraformaldehyde/PBS solution to prepare paraffin sections. Paraffin-embedded livers were sliced into 4 μm-thick sections and stained using hematoxylin and eosin (H&E) following the standard protocol. Frozen samples were stored at -80 °C for further use.

For intraperitoneal injection of amino acids related to Fig. 3A,B, 0.5 mmol of Arg or Orn dissolved in saline was injected once a day, daily after the acclimation period; rats in the control group were injected with the same volume of saline. The injection dosage was determined based on the daily Arg intake; the oral Arg intake differed by ~ 0.5 mmol/rat/day between the CN and ΔArg groups. For radio-imaging and biodistribution assays, rats were fasted overnight and deprived of drinking water for 6 h before the experiment. For rapamycin treatment, 2 mg/kg rapamycin was injected intraperitoneally daily after the acclimation period. A solution containing 0.9% NaCl and 2% ethanol was used as the vehicle. For the cholestyramine treatment, 5% cholestyramine was added to the diet and administered after the acclimation period. As the addition of cholestyramine tended to reduce food intake, pair feeding was conducted during the cholestyramine experiment.

All animal care and experiments conformed to the Guidelines for Animal Experiments of the University of Tokyo and National Institutes for Quantum Science and Technology, and were approved by the Animal Research Committee of the University of Tokyo and National Institutes for Quantum Science and Technology.

Measurement of lipids and lipoproteins

Total lipid in the liver was extracted following Folch’s method with minor modifications35. The frozen tissue pieces (weighed in advance) were homogenized in methanol: chloroform solution (1:2, v/v), followed by the addition of 20% volume of 0.8% KCl solution and centrifugation at 13,000×g, 4 °C for 10 min. Subsequently, the organic (chloroform) layer was collected and the solvent was evaporated; the remaining lipids were dissolved in isopropanol. The TAG and TCho contents in the lipid extracts were measured using the Triglyceride E-test Wako and Cholesterol E-test Wako (Fujifilm Wako Pure Chemical Corporation, Osaka, Japan), respectively. The concentrations of serum/plasma TAG, TCho, and HDL-C were measured using the Triglyceride E-test Wako, Cholesterol E-test Wako, and HDL Cholesterol E-test Wako, respectively, according to the manufacturer’s protocol. Non-HDL-C levels were calculated by subtracting the HDL-C concentration from that of TCho.

Western blotting

Frozen tissue pieces were homogenized in RIPA buffer. The protein concentration of the lysates was measured using the Pierce BCA protein assay kit (Thermo Fisher Scientific, MO, USA). A fixed amount of protein was subjected to SDS-PAGE. Serum samples were diluted 20 times with distilled water, and a fixed volume was used for SDS-PAGE. Band intensity was quantified using ImageJ software (https://imagej.net/ij/). Primary antibodies used in this study were as follows: anti-APOB antibody (#20578-1-AP, Proteintech, IL, USA), anti-MTP antibody (#sc-135994, Santa Cruz Biotechnology, TX, USA), anti-PDI antibody (#2446, Cell Signaling Technology, MA, USA), anti-SURF4 antibody (#PA5-69676, Thermo Fisher Scientific), anti-S6 ribosomal protein antibody (54D2, #2317, Cell Signaling Technology), anti-phospho-S6 ribosomal protein antibody (Ser240/244, #2215, Cell Signaling Technology), anti-4EBP1 antibody (#9452, Cell Signaling Technology), anti-phospho-4EBP1 antibody (Thr37/46, #9459, Cell Signaling Technology), anti-β-actin antibody (#A5441, Sigma Aldrich), and anti-α-tubulin antibody (#T9026, Sigma Aldrich).

Liver TAG secretion assay

The rats, fed experimental diets for 7 days, were fasted for 6 h (08:00–14:00) and 200 mg/kg of tyloxapol was infused through the tail vein. Blood samples were collected 0, 1, 2, and 4 h after drug administration, and the plasma TAG levels were monitored. The slope of the linear regression lines obtained based on the individual plasma TAG data was considered as the hepatic TAG secretion rate.

Lipid transfer activity assay

Hepatic triglyceride transfer activity was assessed based on earlier methods36,37, with minor modifications. For this assay, crude liver homogenates containing MTP were incubated with fluorescent TAG-containing (donor vesicles) and unlabeled (acceptor vesicles) liposomes. When a labeled TAG is packaged into a donor vesicle, its fluorescence is masked. Contrastingly, labeled TAG exhibits detectable fluorescence when it is removed from the vesicles and carried by MTP. Hence, the TAG transfer from the donor to the acceptor can be evaluated by measuring the fluorescence.

The rats, fed experimental diets for 7 days, were fasted overnight; in the next morning, the liver was exposed under isoflurane anesthesia and was immediately perfused using saline through the portal vein to remove the blood. Liver pieces (each ~ 10 mg) were collected and homogenized with homogenization buffer (10 mM Tris-HCl, 1 mM MgCl2, 1mM EGTA, proteinase inhibitor cocktail, pH 7.4). After centrifugation at 4 °C, 13,000×g for 10 min, the supernatant was collected as a crude homogenate containing MTP.

For preparing liposomes, lipids (4.5 µmol Phosphatidylcholine [Egg PC, Avanti Polar Lipids, Birmingham, England] and 140 nmol fluorescent TAG [16:0-LR/18:1/18:1 TG-lissamine rhodamine, Avanti Polar Lipids] for the donor vesicle; 24 µmol Phosphatidylcholine for the acceptor vesicle) were dissolved in chloroform and put in test tubes. After evaporating the solvent, 1 mL of liposome buffer (40 mM NaCl, 1 mM EDTA, 0.02% NaN3, 15 mM Tris, pH 7.4) was added and sonicated for 30 min in a prewarmed water bath (55 °C), which was immediately followed by vigorous vortexing. This sonication/mixing cycle was repeated until the solution turned cloudy; finally, the specimens were sonicated for 40 min on ice, yielding donor and acceptor vesicles.

In the transfer assay, 7 µL each of donor and acceptor vesicles, 20 µL of reaction buffer (5 mM Tris-HCl, 1 mM EDTA, 0.75 M NaCl, 5 mg/ml BSA, pH 7.4), and 16 µL of water were mixed. Next, 50 µL of liver homogenate were added and the specimens were incubated at 37 °C; its fluorescence (531 nm/590 nm) was monitored. In the blank sample, the homogenate was replaced with 50 µL of water. Total fluorescence in the reaction was determined by adding 7 µL of the donor vesicle in 93 µL of isopropanol. Lipid transfer activity is indicated as the percentage of blank-subtracted fluorescence intensity of each well over blank-subtracted total fluorescence, normalized against the amount of protein.

Synthesis of 11C-labeled ornithine (11C-Orn)

11C-Orn was synthesized using a fully automated synthesis system38. Before irradiation, the mixture of 18-crown-6 (8 mg) in CH3CN (900 µL) and Cs2CO3 (3 mg) in water (150 µL) were azeotropically dried in the reaction vial and then CH3CN (300 µL) was added. 11C-CO2 was generated by the nuclear reaction of 14N(p, α)11C in the target gas (N2 gas containing 0.01% O2 gas) under a cyclotron (CYPRIS HM-18; Sumitomo Heavy Industries, Tokyo, Japan). 11C-CO2 in N2 gas from the cyclotron was concentrated in a coiled stainless tube at -196 °C using liquid N2. By heating the coiled stainless tube at 50℃, concentrated 11C-CO2 gas was released under a N2 stream at a flow rate of 10 mL/min, at which it mixed with H2 gas. The mixed gas was passed through a nickel wire tube at 400 °C using a methanizer (GL Sciences, Tokyo, Japan) and a mixture of 11C-CH4 in the carrier (N2 + H2) gas was obtained. Next, the mixture of 11C-CH4 in the carrier gas was mixed with 5% NH3 in N2 gas and passed through a heated platinum furnace (950 °C) at a flow rate of 400 mL/min to produce 11C-HCN. The 11C-HCN was bubbled into a solution of Cs2CO3 and 18-crown-6 in anhydrous CH3CN (300 µL). The labeling precursor (trimethylbenzyl L-2-[carbobenzyloxy)amino]-4-bromobutyrate)39(5.0 mg) in CH3CN (200 µL) was added to the solution and the reaction mixture was heated at 90 °C for 8 min. After the 11C-cyanation was completed, the reaction solvent was completely removed at 100 °C and the residue was dissolved in MeOH (500 µL). This reaction mixture was transferred into another reaction vessel containing CoCl2·6H2O (7 mg) and NaBH4 (10 mg) at room temperature and was incubated for 10 min. After the reduction, the reaction mixture was returned to the original reaction vessel, treated with 6 M HCl (100 µL), and heated at 120 °C for 10 min. The reaction mixture was neutralized using 1 M NaOH (500 µL) and transferred onto a column (CAPCELLPAK SCX, 10 mm internal diameter × 250 mm; Osaka Soda, Osaka, Japan) fitted to a Jasco HPLC system. Elution with 50 mM phosphorus buffer at a flow rate of 5.0 mL/min yielded pure 11C-Orn by passing the radioactive fraction (retention time = 11–12 min) through a 0.22 μm Millipore filter. Finally, 11C-Orn was obtained in 6 ± 3% (n = 6) radiochemical yield at the end of irradiation and > 90% radiochemical purity. Starting from 33 to 37 GBq of 11C-CO2, 0.15–0.29 GBq of 11C-Orn was produced within an average synthesis time of 60 min from the end of irradiation.

PET/CT imaging and 11C-Orn biodistribution assay

PET was performed using a small-animal Inveon PET scanner (Siemens, Knoxville, TN, USA) after oral gavage of 11C-Orn (74–81.4 MBq/0.5–1 mL/rat)38. The scanner provided 159 transaxial slices with 0.796 mm (centre-to-centre) spacing, a 10 cm and 12.7 cm transaxial field of view (FOV), and axial FOV, respectively. Emission scans were acquired in three-dimensional list mode with an energy window of 350–650 keV under isoflurane anesthesia. Scans were performed at intervals of 20–30, 50–60, and 80–90 min after 11C-Orn administration in ΔArg-fed and CN-fed rats. After imaging, the rats were awakened and allowed to move freely, except during scanning. All list-mode acquisition data were sorted into three-dimensional sinograms, which were then subjected to Fourier rebinning into two-dimensional sonograms. Corrections for scanner dead time, randoms, and decay of the administered 11C-Orn were applied. The dynamic images were reconstructed using filtered back-projection with a Hanning’s filter and a Nyquist cutoff of 0.5 cycles/pixel. Immediately after 90 min of the PET scans, non-enhanced CT scans were performed for 4 min using the breath-holding model. The scan conditions included the following radiation parameters: 200 µA, 90 kV, and an FOV of 60 mm using a small-animal CT system (R_mCT2; Rigaku, Tokyo, Japan). Averaged CT attenuation and dynamic PET images were reconstructed and fused using Siemens Inveon Research Workplace (IRW) software (version 4.0). The average radioactive signal in each VOI was quantified as standardized uptake value (SUV), representing the injected dose per cm3 volume normalized by body weight.

In the biodistribution assay, the rats were euthanized by cervical dislocation under isoflurane anesthesia 30 min after oral gavage of 11C-Orn (9.9–10.36 MBq/0.5 mL/rat). Organs and tissues, including the blood, heart, lungs, liver, pancreas, spleen, kidneys, small intestine, colon, rectum, muscle, bone, testis, bladder, and brain, were promptly removed, collected, and weighed. The radioactivity in each tissue sample was immediately measured using a 2480 Wizard auto-γ scintillation counter (PerkinElmer, Waltham, MA, USA), and expressed as a percentage of the injected dose per gram of wet tissue weight (%ID/g). All the radioactivity measurements were corrected for decay. The same experiment was independently conducted twice and reproducibility of the result was confirmed.

Gastrointestinal transit assay

Gastrointestinal transit was assessed following a previously reported protocol40. Experimental diet-fed rats were fasted overnight; in the next morning, the corresponding diet was provided for 20 min. Subsequently, the diets were removed, and simultaneously, 200 µL of Evans Blue solution (5% Evans Blue, 40% glucose in PBS) was administered through oral gavage, followed by keeping the animals free for 50 min without food and water. The rats were then anesthetized using isoflurane and the entire gastrointestinal tract was removed. The distances from the pylorus to the ileocecal junction and from the pylorus to the forefront of the transiting dye were measured, and the gastrointestinal transit rate was calculated as the percentage of the traveling distance of the dye over the total length of the small intestine.

Metabolite analysis

For preparing serum/plasma metabolites, 50 µL of serum/plasma was mixed with 120 µL of methanol containing internal control substances (25 µM of 2-morpholino-ethanesulfonic acid and 100 µM of methionine sulfone) on ice. After centrifugation (16,000×g for 10 min at 4 °C), 130 µL of supernatant was mixed with 250 µL of ultrapure water and subjected to ultrafiltration using 3 kDa cutoff filters (Amicon Ultra 3 K device, Merck, Darmstadt, Germany), followed by evaporation for 30 min and lyophilization for 6 h. For liver metabolite extraction, liver pieces (~ 100 mg) were homogenized in 500 µL of methanol containing internal controls. The homogenates were diluted with 250 µL of ultrapure water, and 600 µL of each homogenate was mixed with 400 µL chloroform, followed by centrifugation (16,000×g for 5 min at 4 °C); next, 800 µL of the upper phase was collected and subjected to evaporation for 30 min. Subsequently, 300 µL of ultrapure water was added, followed by ultrafiltration and lyophilization. The lyophilized specimens were resuspended in ultrapure water and subjected to LC-MS/MS (LCMS-8030, Shimadzu, Kyoto, Japan) using the Method Package for Primary Metabolites ver. 2 (Shimadzu), following the protocol provided by the manufacturer. The standard solution consisting of a fixed concentration of the target metabolites, was simultaneously analyzed and the absolute amounts of the metabolites were calculated.

Antibiotic-induced microbiome depletion

To deplete the gut microbiota of rats, drinking water supplemented with an antibiotic cocktail (Abx) was provided, as previously reported41,42. Half of the rats (12 animals) were given water containing 1 g/L ampicillin-Na, 500 mg/L vancomycin, 1 g/L neomycin sulfate, 1 g/L metronidazole, and 10 mg/L amphotericin B, whereas the other half was given water only as a control. Abx treatment was initiated at the age of 4 weeks, and the rats were reared on a normal chow diet for two weeks. Next, they underwent a 4-day acclimation period followed by 7 days of experimental diet feeding, as described for other experiments, with the Abx treatment continued during the entire period. The growth rate, food intake, and water intake were monitored to ensure these factors are comparable among all groups throughout the experiment. Immediately before sacrificing the rats, freshly discharged feces were collected from the anus and subjected to DNA extraction using the FavorPrep stool DNA isolation kit (FAVORGEN, Pingtung, Taiwan).

16 S ribosomal DNA metagenomic sequencing analysis of the gut microbiome

The rats, fed experimental diets for 7 days, were moved to new cages, and fresh feces dropped into the cages within 1–2 h were collected. Six rats reared in each group were randomly divided into two sub-groups, and the feces from three rats in each sub-group were pooled into one sample, yielding two samples per dietary group used in further analysis. The stool samples were stored at − 80 °C until use. Genomic DNA was extracted and purified using the MagPure Stool DNA KF Kit (Magen Biotechnology Co., Guangdong, China), and 30 ng of DNA extracted from each sample was subjected to polymerase chain reaction to amplify the V3–V4 region of the 16 S ribosomal RNA gene using the following primers: 338 F-ACTCCTACGGGAGGCAGCAG, and 806R-GGACTACHVGGGTWTCTAAT. The amplicon library was sequenced using a DNBSEQ-G400_PE300_150 bp paired-end platform (BGI, Shenzhen, China). Paired-end reads were merged using FLASH (v1.2.11) tool to generate consensus sequence tags, which were clustered into operational taxonomic units (OTU) with USEARCH (v7.0.1090) software. Chimera sequences were filtered using the UCHIME (v4.2.40) algorithm. OTU representative sequences were aligned against the Greengene (V201305) database for taxonomic annotation, and taxonomic analysis was performed using RDP classifier (v2.2) software. Library production and sequencing were performed at BGI Japan (Hyogo, Japan).

Human data collection

In the human study, we recruited 678 Japanese subjects from the participants in a health screening program conducted at the Japanese Red Cross Kumamoto Health Care Center, Kumamoto, Japan. Peripheral blood samples were collected from the subjects after fasting for ~ 12 h. Lipid profiles and other biomedical parameters were measured using the standard methods recommended by the Japan Society of Clinical Chemistry at the Japanese Red Cross Kumamoto Hospital Health Center. Hepatic ultrasound was used to diagnose fatty liver based on four criteria: diffuse high-echo texture, increased echo texture compared with that of the kidney, vascular opacity, and deep attenuation43.

Processing of human clinical data

We applied canonical correlation analysis (CCA) to human data44. CCA was designed to maximize the correlation between two common components (Inline graphic and Inline graphic). The dataset of Inline graphic and Inline graphicwas used to extract common components (CC). Here, Inline graphic includes plasma LDL-C and TAG levels, and labels for fatty liver (1 or 0). Inline graphicincludes plasma amino acid, ALT, and AST levels, and clinical items including BMI, gender (‘GEN’), and average daily alcohol intake (‘ALC’). The canonical coefficients can be determined through CCA by solving the following optimization problem:

graphic file with name d33e810.gif 1
graphic file with name d33e814.gif

where, Inline graphic Inline graphic indicate weights. Inline graphic and Inline graphic are the functions for covariance and variance, respectively. We used scikit-learn, a Python library, to extract common components. Here, we focused on the common components up to the number of variables in Inline graphic (i.e., 3).

We adopted the following criteria to characterize dyslipidemia: HDL-C ≥ 140 mg/dL, TG ≥ 150 mg/dL indicated ‘high’ and HDL-C < 101 mg/dL, TG < 61 mg/dL indicated ‘low.’33 ‘Normal’ was defined based on values between ‘high’ and ‘low’ levels (Fig. 4). IQR indicates the difference between the first and third quartile points. We calculated the median IQR in our analysis using the median (Supplementary Fig. 6).

Ethics statement

The human study complied with the principles of the Declaration of Helsinki, and the study protocol was approved by the Ethics Committee of the Faculty of Life Science, Kumamoto University (No. 169). All study participants provided written informed consent before participating in this study. Also, all experiments involving live animals and human samples were performed in accordance with the ARRIVE guidelines (https://arriveguidelines.org).

Statistical analysis

Comparisons between the two groups were performed using the Student’s t-test. Comparisons among more than two groups were performed through analysis of variance (ANOVA) and the Tukey–Kramer post-hoc test. For all analyses, statistical significance was considered at p < 0.05. All statistical calculations and graphs were generated using GraphPad Prism 10 (Dottmatic, CA, USA).

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material 1 (636.8KB, pdf)
Supplementary Material 2 (2.5MB, docx)

Acknowledgements

We thank Dr. Minoru Yoshida and Dr. Yasuhiro Ogata (Japanese Red Cross Kumamoto Health Care Center) for the human data collection. We also thank Ms. Tomomi Ueda and Ms. Noyumi Nakamura (Department of Veterinary Medical Sciences, the University of Tokyo) for their technical support. Further, we thank Dr. Shin-Ichiro Takahashi (The University of Tokyo) for his contribution in supervising our experiments. We would like to thank Editage (www.editage.com) for the English language editing.

Author contributions

H.N. designed the study; H.N. and S.N. conducted most experiments; L.X., M.F., Y.Z., and M.R.Z. contributed to 11 C-Orn synthesis and related experiments; D.Y. contributed to LC-MS-based metabolite analyses; J.S. and K.O. collected human blood samples, performed human data curation, examined patients, and collected clinical samples; S.F. contributed to the clinical data processing and informatics analyses; F.H. supervised the study; H.N., S.F., and F.H. wrote the manuscript. All the authors have read and approved the final version of the manuscript.

Funding

This work was supported in part by the Japan Society for the Promotion of Science (JSPS) KAKENHI, #23K13923 granted to H.N., 24K21295 and 23K27558 granted to M.R.Z.; 15H04583, 23H00358 and 23K23793 granted to F.H.; and the Moonshot Research and Development Program (##21zf0127003h001) granted to M.R.Z.; and the Cross-ministerial Moonshot Agriculture, Forestry, and Fishers Research, from the Bio-oriented Technology Research Advancement Institution, BRAIN (#20350956) granted to N.H. and F.H.

Data availability

Raw 16 S ribosomal DNA amplicon sequencing data related to Supplementary Fig 5 are available in the DDBJ BioProject database with links to BioProject accession number #PRJDB40630. The human datasets generated and/or analyzed in this study are not publicly available for protecting personal privacy, but are available upon reasonable request. The other data described in this manuscript and the analytical code will be made available upon reasonable request. Further information and requests should be directed to and will be fulfilled by the lead contact, Hiroki Nishi ([shwest@mail.saitama-u.ac.jp](mailto: shwest@mail.saitama-u.ac.jp)).

Declarations

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Hiroki Nishi, Email: shwest@mail.saitama-u.ac.jp.

Fumihiko Hakuno, Email: hakuno@g.ecc.u-tokyo.ac.jp.

References

  • 1.Hyde, P. N. et al. Dietary carbohydrate restriction improves metabolic syndrome independent of weight loss. JCI Insight4, (2019). [DOI] [PMC free article] [PubMed]
  • 2.Mardinoglu, A. et al. An integrated understanding of the rapid metabolic benefits of a carbohydrate-restricted diet on hepatic steatosis in humans. Cell. Metab.27, 559–571e5 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Gao, Y., Hua, R., Hu, K. & Wang, Z. Carbohydrates deteriorate fatty liver by activating the inflammatory response. Nutr. Res. Rev.35, 252–267 (2022). [DOI] [PubMed] [Google Scholar]
  • 4.Yki-Järvinen, H., Luukkonen, P. K., Hodson, L. & Moore, J. B. Dietary carbohydrates and fats in nonalcoholic fatty liver disease. Nat. Rev. Gastroenterol. Hepatol.18, 770–786 (2021). [DOI] [PubMed] [Google Scholar]
  • 5.Lyu, J. et al. Protein source associated with risk of metabolic syndrome in adults with low and adequate protein intake: A prospective cohort study of middle-aged and older adults. J. Nutr. Health Aging. 28, (2024). [DOI] [PMC free article] [PubMed]
  • 6.Nishi, H. et al. Metabolic effects of short-term high-fat intake vary depending on dietary amino acid composition. Curr. Dev. Nutr.8, 103768 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Hu, X. & Guo, F. Amino acid sensing in metabolic homeostasis and health. Endocr. Rev.42, 56–76 (2021). [DOI] [PubMed] [Google Scholar]
  • 8.Taya, Y. et al. Depleting dietary valine permits nonmyeloablative mouse hematopoietic stem cell transplantation. Sci. (1979). 354, 1152–1155 (2016). [DOI] [PubMed] [Google Scholar]
  • 9.He, L., Ding, Y., Zhou, X., Li, T. & Yin, Y. Serine signaling governs metabolic homeostasis and health. Trends Endocrinol. Metab. 34, 361–372 (2023). [DOI] [PubMed] [Google Scholar]
  • 10.Kosakamoto, H. et al. Sensing of the non-essential amino acid tyrosine governs the response to protein restriction in Drosophila. Nat. Metab.4, 944–959 (2022). [DOI] [PubMed] [Google Scholar]
  • 11.Nishi, H. et al. Importance of serum amino acid profile for induction of hepatic steatosis under protein malnutrition. Sci. Rep.8, 1–11 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Otani, L. et al. Low-arginine and low-protein diets induce hepatic lipid accumulation through different mechanisms in growing rats. Nutr. Metab. (Lond). 17, 1–13 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Jenkinson, C. P., Grody, W. W. & Cederbaum, S. D. Comparative properties of arginases. Biochem. Physiol.114 (1996). [DOI] [PubMed]
  • 14.Canbay, A. & Sowa, J. P. l-Ornithine l-Aspartate (LOLA) as a novel approach for therapy of non-alcoholic fatty liver disease. Drugs. 79, 39–44. 10.1007/s40265-018-1020-5 (2019). [DOI] [PMC free article] [PubMed]
  • 15.Vela, C. I. B. & Ramírez, J. L. P. Efficacy of oral L-ornithine L-aspartate in cirrhotic patients with hyperammonemic hepatic encephalopathy. Ann. Hepatol.10, (2011). [PubMed]
  • 16.Fukushima, S. et al. A novel amino acid signaling process governs glucose-6-phosphatase transcription. iScience24, 102778 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Olofsson, S. O., Boren, J. & Apolipoprotein, B. A clinically important apolipoprotein which assembles atherogenic lipoproteins and promotes the development of atherosclerosis. J. Intern. Med.258, 395–410 (2005). [DOI] [PubMed] [Google Scholar]
  • 18.Wang, S. et al. IRE1α-XBP1s induces PDI expression to increase MTP activity for hepatic VLDL assembly and lipid homeostasis. Cell. Metab.16, 473–486 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Wetterau, J. R., Combs, K. A., McLean, L. R., Spinner, S. N. & Aggerbeck, L. P. Protein disulfide isomerase appears necessary to maintain the catalytically active structure of the microsomal triglyceride transfer protein. Biochemistry30, 9728–9735 (1991). [DOI] [PubMed] [Google Scholar]
  • 20.Wang, X. et al. Receptor-mediated ER export of lipoproteins controls lipid homeostasis in mice and humans. Cell. Metab.33, 350–366e7 (2021). [DOI] [PubMed] [Google Scholar]
  • 21.Albillos, A., De Gottardi, A. & Rescigno, M. The gut-liver axis in liver disease: Pathophysiological basis for therapy. J. Hepatol.72, 558–577 (2020). [DOI] [PubMed] [Google Scholar]
  • 22.Pabst, O. et al. Gut–liver axis: Barriers and functional circuits. Nat. Rev. Gastroenterol. Hepatol.20, 447–461 (2023). [DOI] [PubMed] [Google Scholar]
  • 23.Hsu, C. L. & Schnabl, B. The gut–liver axis and gut microbiota in health and liver disease. Nat. Rev. Microbiol.21, 719–733 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Zhong, J. et al. Hyodeoxycholic acid ameliorates nonalcoholic fatty liver disease by inhibiting RAN-mediated PPARα nucleus-cytoplasm shuttling. Nat. Commun.14, (2023). [DOI] [PMC free article] [PubMed]
  • 25.Saxton, R. A. & Sabatini, D. M. mTOR signaling in growth, metabolism, and disease. Cell. 168, 960–976. 10.1016/j.cell.2017.02.004 (2017). [DOI] [PMC free article] [PubMed]
  • 26.Goul, C., Peruzzo, R. & Zoncu, R. The molecular basis of nutrient sensing and signalling by mTORC1 in metabolism regulation and disease. Nat. Rev. Mol. Cell Biol.24, 857–875. 10.1038/s41580-023-00641-8 (2023). [DOI] [PubMed]
  • 27.Rodríguez-Correa, E., González-Pérez, I., Clavel-Pérez, P. I., Contreras-Vargas, Y. & Carvajal, K. Biochemical and nutritional overview of diet-induced metabolic syndrome models in rats: What is the best choice? Nutr. Diabetes10, (2020). [DOI] [PMC free article] [PubMed]
  • 28.Bays, H. E. et al. Obesity, dyslipidemia, and cardiovascular disease: A joint expert review from the Obesity Medicine Association and the National Lipid Association 2024. J. Clin. Lipidol.18, e320–e350 (2024). [DOI] [PubMed] [Google Scholar]
  • 29.Yanai, H., Adachi, H., Hakoshima, M., Iida, S. & Katsuyama, H. Metabolic-dysfunction-associated steatotic liver disease—Its pathophysiology, association with atherosclerosis and cardiovascular disease, and treatments. Int J. Mol. Sci.24, (2023). [DOI] [PMC free article] [PubMed]
  • 30.Deprince, A., Haas, J. T. & Staels, B. Dysregulated lipid metabolism links NAFLD to cardiovascular disease. Mol. Metab42, (2020). [DOI] [PMC free article] [PubMed]
  • 31.Nishi, H. et al. Alteration of serum amino acid profiles by dietary adenine supplementation inhibits fatty liver development in rats. Sci. Rep.10, 1–10 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Yamakado, M. et al. Plasma free amino acid profiles predict four-year risk of developing diabetes, metabolic syndrome, dyslipidemia, and hypertension in Japanese population. Sci. Rep.5, (2015). [DOI] [PMC free article] [PubMed]
  • 33.Chantranupong, L. et al. The CASTOR proteins are arginine sensors for the mTORC1 pathway. Cell. 165, 153–164 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Igarashi, M. & Guarente, L. mTORC1 and SIRT1 cooperate to foster expansion of gut adult stem cells during calorie restriction. Cell166, 436–450 (2016). [DOI] [PubMed] [Google Scholar]
  • 35.Folch, J., Lees, M. & Sloane Stanley, G. H. A simple method for the isolation and purification of total lipides from animal tissues. J. Biol. Chem.55, 999–1033 (1957). [PubMed] [Google Scholar]
  • 36.Athar, H., Iqbal, J., Jiang, X. C. & Hussain, M. M. A simple, rapid, and sensitive fluorescence assay for microsomal triglyceride transfer protein. J. Lipid Res.45, 764–772 (2004). [DOI] [PubMed] [Google Scholar]
  • 37.Anaganti, N., Rajan, S. & Hussain, M. M. An improved assay to measure the phospholipid transfer activity of microsomal triglyceride transport protein. J. Lipid Res.62, (2021). [DOI] [PMC free article] [PubMed]
  • 38.Zhang, Y. et al. L-[5-11 C]Glutamine PET imaging noninvasively tracks dynamic responses of glutaminolysis in non-alcoholic steatohepatitis. Acta Pharm. Sin. B. 10.1016/j.apsb.2024.07.023 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Ding, Y. S., Antoni, G., Fowler, J. S., Wolf, A. P. & Langstrom, B. Synthesis of L‐[5‐11 C]ornithine. J. Label. Comp. Radiopharm. 27, 1079–1090 (1989). [Google Scholar]
  • 40.Jarry, A. C. et al. Neuromedin U is a gut peptide that alters oral glucose tolerance by delaying gastric emptying via direct contraction of the pylorus and vagal-dependent mechanisms. FASEB J.33, 5377–5388 (2019). [DOI] [PubMed] [Google Scholar]
  • 41.Fagarasan, S. et al. Critical roles of activation-induced cytidine deaminase in the homeostasis of gut flora. Sci. (1979). 298, 1421–1424 (2002). [DOI] [PubMed] [Google Scholar]
  • 42.Rakoff-Nahoum, S., Paglino, J., Eslami-Varzaneh, F., Edberg, S. & Medzhitov, R. Recognition of commensal microflora by toll-like receptors is required for intestinal homeostasis. Cell118, 229–241 (2004). [DOI] [PubMed] [Google Scholar]
  • 43.Jimba, S. et al. Prevalence of non-alcoholic fatty liver disease and its association with impaired glucose metabolism in Japanese adults. Diabet. Med.22, 1141–1145 (2005). [DOI] [PubMed] [Google Scholar]
  • 44.Zhuang, X., Yang, Z. & Cordes, D. A technical review of canonical correlation analysis for neuroscience applications. Hum. Brain Mapp.41, 3807–3833. 10.1002/hbm.25090 (2020). [DOI] [PMC free article] [PubMed]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1 (636.8KB, pdf)
Supplementary Material 2 (2.5MB, docx)

Data Availability Statement

Raw 16 S ribosomal DNA amplicon sequencing data related to Supplementary Fig 5 are available in the DDBJ BioProject database with links to BioProject accession number #PRJDB40630. The human datasets generated and/or analyzed in this study are not publicly available for protecting personal privacy, but are available upon reasonable request. The other data described in this manuscript and the analytical code will be made available upon reasonable request. Further information and requests should be directed to and will be fulfilled by the lead contact, Hiroki Nishi ([shwest@mail.saitama-u.ac.jp](mailto: shwest@mail.saitama-u.ac.jp)).

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group

RESOURCES