High-protein Diet More Effectively Reduces Hepatic Fat Than Low-protein Diet Despite Lower Autophagy and FGF21 Levels

Chenchen Xu; Mariya Markova; Nicole Seebeck; Anne Loft; Silke Hornemann; Thomas Gantert; Stefan Kabisch; Kathleen Herz; Jennifer Loske; Mario Ost; Verena Coleman; Frederick Klauschen; Anke Rosenthal; Volker Lange; Jürgen Machann; Susanne Klaus; Tilman Grune; Stephan Herzig; Olga Pivovarova-Ramich; Andreas F. H. Pfeiffer


Liver International. 2020;40(12):2982-2997. 

In This Article

Materials and Methods

For a detailed description of methods and materials, please refer to the SI Methods.

Study Design, Dietary Intervention, Blood and Tissue Samples, and Proton Magnetic Resonance Spectroscopy

In cohort I, 19 participants with morbid obesity undergoing bariatric surgery were divided into two hypocaloric (1500–1600 kcal/d) diet groups (HP: 30 E% protein, 25–30 E% fat, 35–45 E% carbohydrates, n = 9; LP: 10 E% protein, 25–35 E% fat, 55–65 E% carbohydrates, n = 10), for three weeks prior to surgery (Table S1). Groups were matched for age, gender and body mass index (BMI). Study participants obtained food plans given as 10-d rotating menus (Table S2) and included recipes. HP food plans were rich in low-fat dairy products (cheese, milk, yoghurt), meat, eggs, fruits and vegetables, whereas LP (low-methionine) food plans were rich in bread, rice, potatoes, soy products, fruits and vegetables. Sweets, cookies and sweet beverages were excluded in both groups. Part of the food (protein shakes in the HP and vegan spreads in the LP group) was provided to the participants.

The primary outcome of the study was change in liver fat content after three weeks of dietary intervention; secondary outcomes were anthropometric measurements, routine biochemistry of liver parameters, serum markers of glucose metabolism and liver fat (FGF21), autophagy flux, gene expression and mitochondrial activity in the liver.

At the beginning and at the end of the intervention (3–5 days before surgery), anthropometric measurements, fasting blood samples, body composition determination via BOD POD (Cosmed, Rome, Italy) were performed. Assessment of intrahepatic lipid (IHL) content, measured by proton magnetic resonance spectroscopy ([1]H-MRS) was performed in the early afternoon, at the end of each clinical investigation day. On the day of surgery, blood, liver and adipose tissue samples were collected. Study participants were advised to follow the diets during this gap and not to change their physical activity.

In cohort II, 10 obese patients, who consumed a 3-week hypocaloric (1200–1500 kcal/d) diet with a moderately increased protein intake (20–22 E% protein) and did not undergo bariatric surgery, were selected as a reference protein (RP) group (Table 1, Table S1). The serum data and IHL content from these patients were used as reference. The register and ethics approval of the study are shown in the SI Methods.


After haematoxylin and eosin (H&E) staining, liver biopsies were assessed for hepatocellular ballooning, lobular inflammation, fibrosis and hepatic triglyceride content. Study subjects were divided into three groups according to the steatosis-activity-fibrosis (SAF) score[22,23] into "normal", "steatosis" and "NASH" groups. The histological diagnoses by SAF score in the LP and HP groups are shown in Table S3.

Ex Vivo Examinations, Protein Extraction and Western Blotting for Autophagy Flux

Fresh liver tissue was incubated with or without 0.1 μM bafilomycin A1 and 10 μg/mL leupeptin (both from Sigma Aldrich Chemie) for 2 hours as described.[24] Protein extraction was performed by lysing frozen liver tissue in RIPA buffer. Western blotting and chemiluminescent detection were carried out on Fusion SL (PeqLab Biotechnologie GmbH).

RNA Extraction and Gene Expression Analysis

Followed by RNA extraction from liver samples, gene expression profiling was performed by RNA-Seq as described in SI Methods. Expression of genes involved in autophagy, ER-stress, FGF21 pathway and lipid metabolism was assessed by quantitative real-time polymerase chain reaction (qRT-PCR) as described before.[25] Primer sequences are shown in Table S4.

Hepatic Triglyceride and Citrate Synthase Activity Analysis

Frozen liver tissue was powdered, lysed and homogenized for further analyses. Triglyceride content analysis was performed according to Triglyceride Determination Kit (Sigma Aldrich Chemie) and the absorbance changes were detected at 540 nm by spectrophotometry. Citrate synthase (CS) activity analysis was completed by adding 10 μL of 1:5 diluted supernatant, 215 μL reaction buffer, 25 μL acetyl CoA (3.6 mM) and 50 μL oxaloacetate (3 mM) in each well of the 96-well plates and the absorbance changes were monitored at 412 nm for 7 minutes by spectrophotometry.

Hepatic Mitochondrial Respiration Analysis

Mitochondrial respiration analysis was performed in the homogenized liver using the high-resolution respirometry OROBOROS O2k-Core (Oroboros Instruments) at 37°C in a hyperoxygenated environment. Substrates were added according to a multiple substrate-uncoupler-inhibitor titration (SUIT) protocol and shown in the online supplemental methods.[26–28]

Power Calculation and Statistical Analysis

Power calculation was completed using the G-Power software v.3.1 for the primary end point Δ liver fat content. Power calculation was based on the difference of liver fat after 6 weeks of the high-protein diet observed in [ref. 3] where the mean reduction in the liver fat content was −7.0% with standard deviation 5.4%. For the paired parametric design and α < 0.05, the calculated sample size was seven subjects in each treatment group for statistical power of 80%, and six subjects in each treatment group for statistical power of 70%.

Statistical analysis was performed via SPSS Statistics (IBM) and figures were plotted in GraphPad Prism (GraphPad Prism Inc). Data are reported as mean ± standard error of the mean (SEM). T test and Mann-Whitney-U test were used to compare two groups, while ANOVA with subsequent post-hoc analysis (LSD test) was applied to compare more than two groups. Correlation analysis was performed through Pearson's coefficient or Spearman's rank correlation coefficient. Statistical significance was defined as P < .05.