Rifaximin Protects Against Inflammation and Fibrosis in MASH: Any Role for Ethanol-Producing Bacteria?

TL;DR

This study explores whether rifaximin can protect against liver inflammation and fibrosis in MASH by targeting ethanol-producing gut bacteria.

Contribution

The study shows that rifaximin's protective effects in MASH are not due to reducing ethanol-producing bacteria but may be due to reducing endotoxemia and inflammation.

Findings

Rifaximin reduced liver inflammation and fibrosis in mice fed a Western diet.

Rifaximin did not reduce ethanol-producing bacteria or faecal ethanol levels.

Bacteria isolated from rifaximin-treated mice developed resistance to the drug.

Abstract

Metabolic Dysfunction-Associated Steato-Hepatitis (MASH) is a multiple-hit disease. Endotoxins, ethanol, and other metabolites of certain gut microbiota can reach the liver and accelerate inflammation and disease progression. Targeting ethanol-producing colonic bacteria with rifaximin could affect the progress of MASH. In the present study, thirty mice were assigned to three groups (n = 10 mice per group). Mice received either a normal diet, a Western diet, or a Western diet with oral rifaximin. After 12 weeks, liver function, serum levels of TNF-α, interleukin (IL)-1β, IL-6, and lipopolysaccharides (LPS) were measured. Liver specimens were assessed for pathological changes, lipid deposition, and fibrosis. Expression of p53, GFAP, CD68, and TLR-4 in the liver was also assessed. Faecal samples were evaluated for ethanol contents. Lactobacillus acidophilus, in addition to ethanol-producing Klebsiella pneumoniae and Escherichia coli, were isolated, quantified, and tested for sensitivity to rifaximin. Rifaximin was able to ameliorate Western diet-induced biochemical changes and elevated TNF-α, IL-1β, IL-6, and LPS levels. Changes in liver histology, fibrosis, and lipid content were attenuated. Expressions of p53, GFAP, CD68, and TLR-4 in the liver were all reduced. The Western diet-induced increases in faecal ethanol or ethanol-producing bacteria were not corrected by rifaximin. After 12 weeks, isolated bacteria from the rifaximin group were rifaximin-resistant. Our findings imply that the protective impact of rifaximin in the MASH model is unlikely to be mediated by alteration of ethanol-producing colonic bacteria because of acquired rifaximin resistance. Rifaximin-induced reduction in endotoxemia and inflammation in the liver appears to be a more relevant explanation.

Figures

Full text

1. Introduction

Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD), formerly known as non-alcoholic fatty liver disease (NAFLD), is a chronic liver disease in which excess triglycerides are accumulated in the liver (steatosis) in association with metabolic disturbances such as obesity, dyslipidemia, diabetes, or high blood pressure. MASLD could progress from hepatic steatosis to more complicated hepatocellular inflammation, injury, and fibrosis, which is known as Metabolic Dysfunction-Associated Steato-Hepatitis (MASH), which was formerly known as non-alcoholic steatohepatitis (NASH) [1]. The progression of MASLD to MASH can then lead to subsequent and more serious complications, such as liver cirrhosis and hepatocellular carcinoma [2]. Failure and limited success of many clinical trials made the current treatment options for MASLD and MASH limited, and MASH has become one of the major indications for liver transplantation [3].

The pathophysiological process through which MASLD/MASH develops is so complicated. Oxidative stress, mitochondrial and immunological dysfunction, insulin resistance, genetic susceptibility, and gut endotoxins are different factors that could collectively lead to MASLD/MASH [4]. Gut microbiota (GM) plays a pivotal role in the development and progress of MASH. GM’s lipopolysaccharides (LPS) endotoxins increased intestinal permeability and dysregulation of metabolism, which can all accelerate hepatic injury and the ongoing fibrotic process [5]. More recently, it was suggested that the dysbiotic GM of MASLD-predisposed patients produces higher amounts of ethanol, which, in turn, produces further impairment of intestinal permeability and augments liver damage through the generation of reactive oxygen species. Several ethanol-producing strains of Escherichia coli (E. coli) and Klebsiella pneumoniae (K. pneumoniae) were isolated from susceptible individuals and have been linked to the development of MASLD [6,7].

Rifaximin is an antibiotic that has shown some usefulness in liver diseases and hepatic encephalopathy [8]. Rifaximin has also been suggested as a potential treatment for MASLD. Such a suggestion was based on both experimental and clinical studies [9,10,11,12,13]. Rifaximin appears to be an ideal therapy for MASLD for several reasons. First, its systemic absorption is very low, which made its systemic adverse effects minimal. The long-term safe use of this drug has already been documented in patients with chronic liver diseases [8]. In addition, the drug has a broad-spectrum activity against most of the aerobic and anaerobic bacteria in the gut [8]. Moreover, rifaximin showed promising results in several MASH models [11,12,13]. Studies that supported the usefulness of rifaximin in MASLD suggested attenuation of endotoxemia, improvement of insulin resistance, and reduction in proinflammatory cytokines as potential mechanisms by which the drug can ameliorate MASLD [10,11]. However, none of these studies directly investigated the effect of rifaximin on ethanol-producing GM as a potential mechanism by which the drug could slow the progression of MASLD. Hence, the present study aimed to assess the ability of rifaximin to modify some of the ethanol-producing colonic bacteria and reduce ethanol and toxin delivery to the liver using a mouse model of MASH.

2. Materials and Methods

2.1. Animals

The experimental protocol was approved by Mansoura University Animal Care and Use Committee (MU-ACUC; approval number MED.R.24.04.35). The study was conducted according to the guidelines of the Declaration of Helsinki. The study involved a total of 30 pathogen-free male C57BL/6 mice (7–8 weeks of age; 24–30 g of weight). The mice were obtained from Mansoura Experimental Research Centre (MERC), Mansoura University. In the one-week adaptation period, the mice were fed a normal diet and had free access to water. Housing was in an air-conditioned room, and the mice were exposed to a 12 h light/dark cycle. They were then assigned into three groups:

1. Control group (n = 10); mice were kept on normal diet.
2. Western diet group (n = 10), mice were fed on a Western diet.
3. Treated group (n = 10), mice were fed on a Western diet and treated with a daily oral rifaximin (100 mg/kg/day by gavage).

A normal diet provides 3.6 kcal/g, which is distributed as 19.1% from protein, 67.9% from carbohydrates, and 13% from fat. In comparison, the Western diet provides 4.5 kcal/g and has more cholesterol, more fat, and more sucrose. The 4.5 kcal/g distribution is 15.2% from protein, 42.7% from carbohydrates, and 42% from fat. More carbohydrate content increases blood ethanol even in an ethanol-free diet [7]. Thus, fructose (42 g/L) was added to the drinking water in the two Western diet-fed groups to accelerate the model development [14]. The overall duration of induction was 12 weeks. Every week, mice were weighed. Collection of fresh faecal samples, using metabolic cages, took place in the last 7 days of the experiment (from the end of the 11th week to the end of the 12th week). Samples were immediately put at −80 °C until measurements were conducted. Animals were euthanized by the end of the 12th week using halothane overdose. Blood was obtained through cardiac puncture, collected into dry tubes, centrifuged at 2500 g for 15 min at 4 °C, and stored at −80 °C. Livers from all animals were dissected to be prepared for histological and immunohistochemical studies.

2.2. Measrement of Biochemical, Immunological Parameters and Serum Cytokines’ Levels

According to the manufacturers’ instructions, ALT and AST serum levels were measured using specific ELISA kits (Cusabio, Houston, TX, USA; Cat. CSB-E16539m and CSB-E12649m, respectively). Similarly, specific ELISA kits were used to assess TNF-α, interleukin IL-1β, IL-6, and lipopolysaccharides (LPS) levels in the serum (Cusabio, Houston, TX, USA; Cat. CSB-E04741m; CSB-E08054m, CSB-E04639m, and CSB-E13066m, respectively).

2.3. Histopathology

For the histopathological study, liver specimens were fixed in 10% neutral buffered formalin solution for 48 h and processed to prepare five μm-thick paraffin sections. Liver histology was identified using hematoxylin and eosin, while collagen deposition was assessed using Masson’s trichrome stain. Steatosis was assessed in frozen sections using Oil Red O stain [15].

2.4. Immunohistochemical Study

Immunohistochemical staining was conducted for the detection of P53, glial fibrillary acidic protein (GFAP), CD68, and toll-like receptor 4 (TLR4). Five μm sections from paraffin blocks were cut, mounted on charged slides, deparaffinized, and then rehydrated in a graded series of alcohol solutions. For antigen retrieval, sections were boiled with sodium citrate buffer (0.01 mol/L, pH 6) for 10 min. The next step was to block endogenous peroxidase activity by treating the sections with 3% hydrogen peroxide. The avidin–biotin immunoperoxidase method was used for immunohistochemical staining. Rabbit monoclonal anti-CD68 antibody (diluted 1: 200 in PBS; Abcam, UK; cat.no. Ab283667), mouse monoclonal anti-TLR4 antibody (diluted 1: 100 in PBS; Abcam, UK; cat.no. Ab22048), rabbit polyclonal anti-P53 antibody (diluted 1: 400 in PBS; Abcam, UK; cat.no. ab131442) and rabbit monoclonal anti-GFAP antibody (diluted 1:2000 in PBS; eBioscience, ThermoFisher, San Diego, CA, USA; cat.no., 14-9892-82) were applied on the slides. After an overnight incubation at 4 °C, the secondary biotinylated antibody was applied and incubated with streptavidin peroxidase (DAKO Corp.) (Dako K0690; Dako, Carpinteria, California, USA). Three times after each step, phosphate-buffered saline (PBS) was used to wash the sections. The next step was to stain the slide with diaminobenzidine chromogen solution and counterstain with Mayer’s hematoxylin [16]. Sections were dehydrated and mounted. To obtain negative control sections, the primary antibodies were omitted.

2.5. Morphometric Studies

An Olympus digital camera (E24-10 megapixel) built into an Olympus microscope with a 0.5× photo adaptor using the 40× objective was used for photography. Ten microscopic fields/mouse were photographed. The fields photographed were non-overlapping and were randomly selected. Data analysis was performed using Fiji ImageJ, version 20250529-2217 Fiji ImageJ software. All analyses of the area percentage of collagen, CD68, TLR4, P53, and GFAP were conducted on the ten selected fields. Calculation of the mean area percentage (%) of collagen fibres content in the sections stained with Masson’s trichrome stain was performed after the colour threshold was adjusted and the background was excluded. The following formula was used to calculate the selected, blue-stained area: Area percentage (%) = (blue-stained area/total field area) × 100. The same approach was used for oil red O staining and the mean areas % of CD68, TLR4, P53, and GFAP immunostaining.

2.6. Stool Analysis for Ethanol Contents

Faecal samples were collected over the last 7 days before animals were euthanized. Collected samples were immediately frozen at −80 °C until further analysis. At the time of assay, a pre-weighed (125 mg) amount of the samples was suspended in 5 mL of water in Headspace glass vials. Ethanol contents in faecal samples were measured as previously described [17,18]. The procedure included the use of an Agilent 6890 gas chromatography equipped with an Agilent mass spectrometric detector (Agilent Technologies, Santa Clara, CA, USA). Diethyl acetic acid (1.5 mg/L) was used as an internal standard, and quantification of ethanol was done in comparison to the diethyl acetic acid.

2.7. Selective Isolation of Ethanol-Resistant E. coli and K. pneumoniae

250 mg of the previously frozen faeces samples were suspended in one mL of 1X PBS. The suspension was vortexed with glass beads, incubated for two hours at 4 °C, then centrifuged for one minute at 80× g to remove the debris as a pellet [19]. MacConkey agar (Oxoid Ltd., Basingstoke, UK) plates were used to culture 50 μL of the supernatant, which was then incubated aerobically at 37 °C for 48 h. Colony-forming units (CFU) per gram of faeces (expressed as log10 of CFU) were then counted. We also used previously emptied blood culture bottles (Oxoid Ltd., Basingstoke, UK). The bottles contained 20 mL of MacConkey broth (Oxoid Ltd., Basingstoke, UK), which had been supplemented with four mL of sterile rumen juice and four mL of defibrinated sheep blood. Each bottle was inoculated with 200 μL of the supernatant of the original suspension (250 mg of faeces, 1 mL of PBS 1X) and incubated aerobically at 37 °C for 10 days. On days 1, 3, 7, and 10, 500 μL of bottle content was sampled, followed by 10-fold serial dilutions and inoculation onto MacConkey agar plates. To detect alcohol-resistant non-fastidious Gram-negative bacteria, the same procedures were applied, but 5% and 10% ethanol were added to the bottle [20]. Bacterial colony counting was conducted on day three and expressed as log10 of cfu/mL. Culture media preparation in this investigation followed the instructions provided by the manufacturer. Using common microbiological techniques, such as Gram stain and biochemical reactions, each bacterial pure colonies were identified (Gram-negative bacilli, lactose fermenters) and subsequently counted. The Vitek-2 compact system was used to verify E. coli and K. pneumoniae isolates using the Gram-negative identification (GN-ID) cards (BioMérieux, Marcy l’Etoile, France). The American Type Culture Collection (ATCC) E. coli strains (ATCC10536) and K. pneumoniae strains (ATCC10031) were used as positive controls in triplicate testing for each isolate.

2.8. Selective Isolation of Lactobacillus acidophilus

DeMan, Rogosa, and Sharpe (MRS) agar (Oxoid Ltd., Basingstoke, UK) plates were used to culture 50 μL of the above-mentioned supernatant, which was then incubated anaerobically at 37 °C for 48 h [21]. The CFU per gram of faeces (expressed as log10 of CFU) was then counted. We also used previously emptied blood culture bottles (Oxoid Ltd., Basingstoke, UK). The bottles contained 20 mL of MRS broth (Oxoid Ltd., Basingstoke, UK), which had been supplemented with four mL of sterile rumen juice and four mL defibrinated sheep blood. Each bottle was inoculated with 200 μL of the above-mentioned supernatant and incubated anaerobically at 37 °C for 10 days. On days 1, 3, 7, and 10, 500 μL of bottle content was sampled, followed by 10-fold serial dilutions and inoculation onto MRS agar plates. To detect alcohol-resistant Lactobacillus spp., the same procedures were applied, but 5% and 10% ethanol were added to the bottle [20]. Bacterial colony counting was conducted on day three and expressed as log10 of cfu/mL. Culture media preparation in this investigation followed the instructions provided by the manufacturer. Using common microbiological techniques, such as Gram stain and biochemical reactions, Lactobacillus pure colonies were identified (Gram-positive bacilli, catalase-negative isolates) and subsequently counted [22]. The Vitek-2 compact system was used to verify the Lactobacillus isolates using the anaerobic identification (ANC-ID) cards (BioMérieux, Marcy l’Etoile, France). The American Type Culture Collection (ATCC) Lactobacillus acidophilus (ATCC 4356) was used as a positive control in triplicate testing.

2.9. Antimicrobial Sensitivity Testing of Ethanol-Resistant E. coli and K. pneumoniae to Rifaximin

The minimum inhibitory concentration (MIC) of rifaximin (Sigma-Aldrich, St. Louis, MO, USA) was determined by the broth microdilution method according to the guidelines established by the Clinical and Laboratory Standards Institute (CLSI) [23]. Since there is no MIC breakpoint for rifaximin against E. coli and K. pneumoniae set by the European Committee on Antimicrobial Susceptibility Testing (EUCAST) [24], we utilised a breakpoint of 32 µg/mL between sensitive (≤32 µg/mL) and resistant (>32 µg/mL) E. coli and K. pneumoniae strains according to previous studies [25,26]. The tested rifaximin concentrations ranged between 0.125 µg/mL and 256 µg/mL. The E. coli (ATCC10536) and K. pneumoniae (ATCC10031) strains were used as positive controls in triplicate testing for each isolate.

2.10. Amplification of Ethanol-Resistant E. coli and K. pneumoniae adh Gene by PCR

PCR was used to check for the adh gene in the ethanol-resistant isolates that were verified by the Vitek-2 compact system. To obtain DNA templates, E. coli and K. pneumoniae chromosomal DNA were isolated using a previously outlined procedure [27]. A pair of primers (Sigma; F: 5′-ATGAAGTATGTGAATCTGGG-3′, and R: 5′-TTAATAGTTCTGGATCGCTG-3′) chosen in accordance with previous studies was employed for the PCR amplification of the adh gene [7,28]. The PCR reaction used a final volume of 25 μL and included 12.5 μL of Taq PCR Master Mix, which was quickly vortexed to prevent salt concentration discrepancies, one μL of each primer that was thawed on ice and thoroughly mixed, five μL of extracted DNA template, and 5.5 μL of nuclease-free doubly distilled water.

The thermal cycler programme was preceded as follows: initial denaturation for two minutes at 94 °C; thirty cycles of denaturation for 10 s at 98 °C, annealing for 30 s at 60 °C, and extension for three minutes at 68 °C, and final extension for five minutes at 68 °C. The amplified adh gene was electrophoresed on an agarose gel (1.5%) using a 100 bp DNA ladder (Lonza Inc., Rockland, MA, USA) to detect the anticipated (1038 bp) bands that were visible after staining with ethidium bromide [29].

2.11. Data Analysis

SPSS software (ver. 22, SPSS Inc., Chicago, IL, USA) was used for data analysis. The Kolmogorov–Smirnov test was used for the normal distribution of data. One-way ANOVA was used for multiple comparisons and was followed by a Bonferroni post hoc test. Values are expressed as mean ± SD, and a p-value < 0.05 is considered statistically significant.

3. Results

3.1. Animal Weight

In comparison to normal diet-fed mice, the use of a Western diet and fructose in the drinking water resulted in a significant increase in the animals’ weight. Rifaximin treatment did not significantly alter the weight gain which was induced by the Western diet (Figure 1A).

3.2. Biochemical, Immunological Parameters and Serum Cytokines’ Levels

Feeding mice on a Western diet resulted in a significant increase in both ALT and AST. Mice treated with rifaximin showed a significant attenuation of the increased serum levels of both enzymes. (Figure 1B,C). The Western diet also resulted in elevation of the serum levels of all tested proinflammatory cytokines (TNF-α, IL-1β, and IL-6) and a marker of endotoxemia (LPS). Treatment with rifaximin significantly ameliorated these changes, although it could not normalise their levels as compared to the control group, except for LPS (Figure 1D–G).

3.3. Histological Results

3.3.1. Hematoxylin and Eosin Stain

Hematoxylin and eosin-stained sections obtained from the livers of the control group showed branching and anastomosing cords of hepatocytes radiating from the central vein. These cords were lined with endothelial cells with intervening blood sinusoids. Hepatocytes had central, rounded vesicular nuclei and acidophilic cytoplasm. Binucleated hepatocytes were occasionally spotted. Blood sinusoids were lined with flat endothelial cells and Kupffer cells (Figure 2A). Feeding mice on a Western diet resulted in disruption of the arrangement of hepatocytes with a congested central vein surrounded by cellular infiltration. Hepatocytes showed cytoplasmic vacuolations. Some hepatocytes have pyknotic nuclei, and others have fragmented nuclei (Figure 2B). The use of rifaximin with a Western diet significantly attenuated these changes. Vacuolations were still detected in some hepatocytes, while others were normal. Few cells appeared with pyknotic nuclei, while the central vein appeared normal (Figure 2C).

3.3.2. Oil Red O Stain

In the Oil Red O-stained liver sections from the normal diet-fed mice, few lipid droplets were detected in the hepatocytes (Figure 2D). Western diet-fed mice showed variable-sized lipids, either large lipid globules or small lipid droplets (Figure 2E). Treatment with rifaximin in the third group significantly reduced the number of lipid droplets (Figure 2F).

3.3.3. Masson’s Trichrome Stain

In Masson’s Trichrome-stained liver sections from the normal diet-fed mice, fine collagen fibres were detected around central veins (Figure 2G). Sections obtained from the Western diet-fed group, where fatty liver was evident, revealed a significant deposition of collagen fibres around central veins and blood sinusoids. In addition, collagen fibres were also detected between the cords of hepatocytes (Figure 2H). On the other hand, collagen fibre deposition around the central veins and in between the cords of hepatocytes was significantly reduced in the liver sections obtained from the rifaximin-treated, Western diet–fed group (Figure 2I).

3.4. Immunohistochemical Staining

3.4.1. p53 Immunostaining

In the control group, most of the hepatocytes revealed negative immunostaining reactions for the apoptotic marker, p53. However, a few scattered cells exhibited a faint reaction (Figure 3A). The Western diet group revealed a strong positive reaction in most hepatocytes (Figure 3B), while the rifaximin treatment was able to reduce the p53 reaction in the hepatocytes (Figure 3C).

3.4.2. Glial Fibrillary Acidic Protein (GFAP) Immunostaining

Examination of GFAP immunohistochemically stained sections from the liver of the normal diet-fed mice showed few GFAP-positive cells (Figure 4A). Examination of this marker of hepatic stellate cells activation in sections obtained from the Western diet-fed mice revealed augmentation of both the number and reaction intensity of GFAP-positive cells (Figure 4B). The number of GFAP-positive cells was significantly reduced with rifaximin treatment in comparison to the second group (Figure 4C).

3.4.3. CD68 Immunostaining

CD68 immunostaining marked Kupffer cells hanging in the lumen of blood sinusoids. In CD68-immunostained sections obtained from the livers of normal diet-fed mice, few CD68-positive cells were detectable (Figure 5A). Feeding on a Western diet significantly increased the number and size of CD68-positive cells, which was observed in sections obtained from the diet-fed mice (Figure 5B), while treatment with rifaximin was able to attenuate such enhanced reaction (Figure 5C).

3.4.4. Toll-like Receptor 4 (TLR4) Immunostaining

Only a few hepatocytes in the liver of the normal diet-fed mice showed some TLR4 expression (Figure 6A). TLR4 expression was high and intense in liver sections obtained from the Western diet-fed mice (Figure 6B). Rifaximin treatment of mice that were fed on a Western diet significantly reduced TLR4 expression, indicating a reduction in the overall endotoxin delivery to the liver (Figure 6C).

3.5. Stool Analysis for Ethanol Contents

In comparison to diethyl acetic acid, the relative quantification of the ethanol contents in faecal samples was conducted. Ethanol contents were markedly elevated in all the faecal samples obtained from the Western diet-fed mice. Rifaximin did not have any significant effect on the amount of faecal ethanol contents (Figure 7A).

3.6. Selective Isolation of Ethanol-Resistant E. coli and K. pneumoniae

Animal faeces were cultured on MacConkey agar plates, and several lactose-fermenting (LF) colonies were found. The surfaces of K. pneumoniae colonies were mucoid. Each bacterial isolate’s pure colonies (LF, Gram-negative bacilli) were identified using standard microbiological methods like Gram stain and biochemical reactions. The Vitek-2 compact system was used to confirm the isolates of E. coli and K. pneumoniae.

E. coli and K. pneumoniae isolates that were 5% ethanol-resistant were found in all three tested groups’ samples. There was no resistance observed in any isolate at 10% ethanol. The number of E. coli and K. pneumoniae colonies in samples grown directly from faeces increased significantly in both groups that were fed on a Western diet (Figure 7B). Samples that were cultured on MacConkey broth with 5% ethanol to identify ethanol-producing E. coli and K. pneumoniae colonies likewise showed a marked rise in the number of these isolates in the Western diet-fed groups, regardless of rifaximin treatment (Figure 7C).

3.7. Selective Isolation of Lactobacillus acidophilus

Upon culturing animal faeces on DeMan Rogosa and Sharpe (MRS) agar plates, many colonies were suspected. Using common microbiological techniques, such as Gram staining and biochemical reactions, pure colonies of Lactobacillus isolates were identified (Gram-positive bacilli, catalase-negative isolates). Lactobacillus acidophilus isolates were verified using the Vitek-2 compact system. Upon re-culture in the presence of ethanol, no isolate showed any resistance to 10% ethanol. A few samples in the tested three groups were resistant to 5% ethanol. Lactobacillus acidophilus colonies were counted directly from faecal samples. There was a significant reduction in the Lactobacillus acidophilus colony count in the group that received the Western diet. Rifaximin treatment slightly restored the count to normal (Figure 7B).

3.8. Antimicrobial Sensitivity of Ethanol-Resistant E. coli and K. pneumoniae to Rifaximin

In mice that did not receive rifaximin, all bacterial isolates were rifaximin-sensitive (minimum inhibitory concentration (MIC) of 0.25 to 16 µg/mL for E. coli, and MIC of 8 to 16 µg/mL for K. pneumoniae). Bacterial isolates from the rifaximin group after 12 weeks were rifaximin-resistant (MIC of ≥256 µg/mL for E. coli and K. pneumoniae). The E. coli (ATCC10536) and K. pneumoniae (ATCC10031) strains were rifaximin-sensitive (MIC of 0.5 µg/mL for E. coli and MIC of 16 µg/mL for K. pneumoniae).

3.9. Amplification of the Alcohol Dehydrogenase (adh) Gene by PCR

The ability to produce alcohol was validated molecularly by identifying the adh gene in the ethanol-resistant isolates of E. coli and K. pneumoniae in every sample collected from the three tested groups.

3.10. Highlighting the Most Important Findings

Table 1 summarises the important findings regarding the protective impact of rifaximin in the MASH model.

4. Discussion

Considering the complex pathology and “multiple hits” of MASLD/MASH, a long list of drugs, chemicals, and probiotics have been tested [4]. One of the tested drugs is rifaximin. Rifaximin is a poorly absorbable, broad-spectrum, locally acting antibiotic that has been originally used in liver diseases and hepatic encephalopathy [8]. Rifaximin was suggested as a potential treatment for MASLD. Several studies suggested that attenuation of endotoxemia, reduction in proinflammatory cytokines, restoring the intestinal barrier, improvement of insulin resistance, and regulating gut microbiome-related metabolism of bile acids are potential mechanisms by which rifaximin helps in the treatment of MASLD [10,11,12,13].

GM, through their pivotal role in the metabolism of bile acids and the production of short-chain fatty acids and ethanol, are clearly linked to the development of MASH [5]. Much attention has been directed to the role of ethanol-producing bacteria, K. pneumoniae and E. coli, in the development of MASH [6,7,28]. To our knowledge, this is the first report investigating the effect of rifaximin on ethanol-producing colonic bacteria and relating such a potential effect to the progress of MASH. In the present study, we provided evidence that rifaximin was unable to reduce the total faecal ethanol content or modify the ethanol-producing colonic bacteria. Still, it was able to prevent the progress of MASH, possibly through reducing toxin delivery to the liver and inhibiting the inflammatory and fibrotic processes.

First, we hypothesized that the protective role of rifaximin in MASH might be attributed to its direct effect on ethanol-producing GM. We intentionally selected the Western diet/fructose model to induce MASH so that we can magnify the role of ethanol-producing bacteria. Fructose absorption in the small intestine is limited, allowing a significant amount to pass to the colon, where fructose is metabolised by the GM to produce ethanol [30]. The results of the present study did not show that rifaximin can significantly reduce the faecal ethanol content nor reduce the number of ethanol-producing K. pneumoniae and E. coli. To support our hypothesis, the alcohol production ability of the isolated bacteria was further confirmed by identifying the adh gene in the isolates of E. coli and K. pneumoniae in every sample collected from the three tested groups. One explanation of our results is the lack of minimal ability of rifaximin to modify the faecal/colonic bacteria. It was reported that rifaximin could not make a significant modification of the colonic bacterial diversity [31,32]. More specifically, the effect of rifaximin on the GM might be limited to the duodenum and jejunum as reported in CCl4-induced liver fibrosis, where changes in Lactobacillaceae and Bacteroidetes abundance were reversed by rifaximin only in the duodenum and jejunum but not in the ileum, cecum, or stool [32]. Similar results were obtained in a different model of ethanol-induced liver injury [33]. These results were attributed to rifaximin’s water insolubility. To inhibit bacterial growth, rifaximin must undergo micellization with bile acids [34]. However, most of the intestinal bile acids are absorbed in the ileum, making rifaximin that reaches the colon insoluble and less effective as an antibacterial agent [34]. On the opposite side, modification of small intestinal or even colonic microbiota by rifaximin has been reported in some MASH models as well as other disease models [12,13,35,36]. While some studies reported that rifaximin use might increase the abundance of Lactobacillus species [35,36,37], others reported the opposite [32,33]. Overall, the effect of rifaximin on Lactobacilli or other bacteria was minimal, and rifaximin is an eubiotic agent that restores the normal flora rather than changing their relative abundance [38].

Our results suggested another explanation for the contradictory reports about the effect of rifaximin on colonic bacteria. We found that the MIC for rifaximin has significantly increased, indicating that ethanol-producing K. pneumoniae and E. coli have acquired resistance to the drug. The possibility of development of rifaximin resistance in Gram-negative bacteria is more likely than in Gram-positive bacteria, such as Lactobacillaceae [39,40]. Resistance to rifaximin requires chromosomal mutation. This mechanism is usually vertically transmitted and rarely spreads among intestinal microorganisms [41]. Discontinuation of the drug for a few weeks can help to get rid of the resistant bacteria [39]. Since we used rifaximin continuously for a relatively long period, we believe that resistance has blunted most of the changes exerted by rifaximin on the colonic bacteria. It seems that neither modification of the gut microbiota nor reducing the load of ethanol production by the gut microbiota represents a major pathway by which rifaximin benefits MASH patients. The drug was able to ameliorate even alcoholic hepatitis, where mice were fed on ethanol [33]. Moreover, rifaximin cannot affect the faecal ethanol, which comes from yeasts that can produce 10 times the ethanol produced by bacteria [20,30].

Resistance to rifaximin did not prevent the beneficial effects of the drug on the liver. Our results demonstrated that rifaximin was able to ameliorate all the biochemical and histopathological changes in the liver of the tested mice. Accumulation of lipid droplets in the hepatocytes, hepatic stellate cells activation, as measured by GFAP, and the fibrotic changes were all reduced by rifaximin. The drug was also able to reduce apoptotic changes, endotoxemia, and inflammatory processes as demonstrated by the reduction in p53 expression, LPS serum levels, and attenuating TLR-4 expression in the liver. Reduction in CD68-expressing macrophage or Kupffer cells in the liver also indicated a reduced inflammatory process. Inflammation, apoptosis, and enhanced TLR-4 signalling are all involved in the pathophysiological process of MASH [42]. Due to increased intestinal permeability, LPS absorbed from the gut may accelerate the inflammatory process, whereas several mediators, such as TNF-α, TLR-4, IL-6, and IL-1, are usually involved [43]. The activated macrophage plays a central role in the inflammatory and fibrotic process in the liver [44]. Several clinical and experimental studies suggested that the anti-inflammatory effect of rifaximin may help in the treatment of MASLD and other inflammatory models [10,12,36]. At concentrations lower than MIC, rifaximin was reported to have major effects unrelated to pathogen eradication [45]. Inhibition of bacterial motility and virulence, bacterial adherence to epithelial cells, restoration of mucosal integrity, and modulation of the inflammatory process have all been suggested as secondary effects of rifaximin [38,46,47]. In addition, the drug was reported to inhibit endotoxemia and systemic inflammation in a mouse model of ankylosing spondylitis [36]. Via the pregnane X receptor, rifaximin was able to inhibit the activation of the NF-κB and reduce the expression of many pro-inflammatory cytokines [48]. The protective effects of rifaximin may also come from its ability to reverse stress-induced impairment of intestinal barrier function [35]. Upregulation of the expression of occludin and other tight junction proteins by rifaximin was previously reported [35,36].

Gut-derived LPS contributes to liver inflammation, fibrosis, and angiogenesis via TLR4 signalling. A recent Japanese study examined the effects of rifaximin on liver fibrosis and early hepatocarcinogenesis associated with MASH, with a special emphasis on the LPS-epiregulin-IL-8-angiogenesis pathway. The results showed that rifaximin reduced liver inflammation, fibrosis, hydroxyproline levels, and fibrogenic gene expression. Furthermore, rifaximin reduced hepatic epiregulin and IL-8 expression, decreased CD34-positive new blood vessel formation, and inhibited proangiogenic gene activity, while improving intestinal barrier function and decreasing gut permeability. Overall, rifaximin appears to delay MASH progression by repairing the gut barrier, lowering LPS translocation, and inhibiting fibrogenic and angiogenic pathways. These findings emphasise its potential as a chemopreventive agent in MASH-related liver cancer [49].

The conducted study is not free from limitations. Before extrapolating the results of experimental studies on GM to clinical application, two points should be considered. First, human microbiota might differ from that of experimental animals [50]. Moreover, the composition of the intestinal microbiota could vary between mice obtained from different laboratories [51]. It might even vary between mice obtained from the same provider and fed on the same diet [52]. Second, 16s rRNA sequencing of GM would provide a broader mapping of gut bacterial and yeast diversity in comparison to the counting of selected bacterial strains. The present study suggested a new variable to consider before judging the role of rifaximin, or another antibiotic, on the GM. Before judging the effect of these drugs on GM, we need to consider the dosing schedule, the duration of therapy, and the possible development of resistance, especially with prolonged uninterrupted use of the drug.

5. Conclusions

Our data indicate that the protective effect of rifaximin in the mouse model of MASH does not appear to be related to reducing the load of faecal ethanol nor affecting the ethanol-producing colonic bacteria. The development of resistance to rifaximin might explain the drug’s inability to affect the colonic bacteria. However, the drug can still prevent the progression of MASH by other mechanisms, possibly by reducing endotoxemia and the inflammatory process in the liver.