PPAR_α Dependent Regulation of Gut Microbiota: Implication for Host Metabolic Homeostasis

ABSTRACT

The peroxisome proliferator‐activated receptor alpha (PPARα) is a key regulator of lipid metabolism and energy homeostasis. However, its role in shaping the gut microbiota requires further investigation. Therefore, the aim of the present study was to investigate whether the lack of PPARα in a mouse model or the presence of a single nucleotide polymorphism (SNP) rs6008259 of PPARα in humans can modulate the gut microbiota and its association with metabolic alterations. PPARα^−/− and PPAR^+/+ male mice were fed an AIN‐93 diet for 10 days, and a human cohort (n = 177) was genotyped for the PPARα SNP rs6008259. PPARα^−/− mice showed less body weight gain, more fat mass and reduced lean mass compared to WT, despite similar food intake. They exhibited elevated hepatic triglycerides and hyperlipidemia. In humans, carriers of the rs6008259 allele of PPARα showed similar findings in blood lipids and body composition. Colonic analysis revealed reduced hypoxia in PPARα^−/− mice, increased inflammatory markers, and compromised barrier function. Gut microbiota analysis in PPARα^−/− mice and humans, carriers of the rs6008259 allele of PPARα showed reduced alpha diversity and altered composition, including increased Blautia, Parabacteroides and Lanchoclostridium. Genetic background is important in the interpretation on the effects of diet on gut microbiota.

The absence or genetic variation of PPARα alters metabolism and gut microbiota in mice and humans. Increased body fat, dyslipidemia, and reduced microbial diversity are observed, along with an association between Blautia and Lanchoclostridium with serum lipids and fat mass. Genetic variability is important to determine the effects of nutrients on gut microbiota.

Abbreviations

AIN

American Institute Nutrition

ATP

adenosine triphosphate

BMI

body mass index

CPT‐1

carnitine palmitoyltransferase I

EE

energy expenditure

HDL

high density lipoprotein

HIF1α

hypoxia‐inducible factor 1‐alpha

LDL

low density lipoprotein

OxPhos

oxidative phosphorylation

CO₂

carbon dioxide

PPARα

peroxisome proliferator‐activated receptor alpha

RXRα

retinoid X receptor alpha

RER

respiratoy exchange ratio

QIIME

quantitative insights into microbial ecology

DS‐PAGE 

sodium dodecyl sulfate–polyacrylamide gel electrophoresis

SNP

single nucleotide polymorphism

VO₂

oxygen consumption

WT

Wild‐type

ZO‐1

zonula occludens‐1.

1. Introduction

Fatty acid metabolism is tightly regulated through multiple interrelated pathways that balance lipid synthesis and catabolism across organs and tissues [1]. These mechanisms are essential to prevent excessive lipid accumulation and to sustain systemic energy homeostasis. Among the transcriptional regulators that coordinate these processes, the peroxisome proliferator‐activated receptor alpha (PPARα) plays a key role. PPARα is a ligand‐activated nuclear receptor that controls the expression of genes involved in mitochondrial and peroxisomal β‐oxidation, fatty acid uptake and binding, and lipoprotein assembly and transport. In addition to its metabolic functions, PPARα also regulates inflammatory responses, demonstrating its pleiotropic role in cellular physiology. Although most early work on PPARα was conducted in liver, heart, and immune cells, there is growing evidence demonstrating the importance of this receptor in the gastrointestinal tract.

In the intestine, PPARα activation promotes fatty acid oxidation, a process that requires heterodimerization with the retinoid X receptor alpha (RXRα) to initiate transcription. Natural and synthetic agonists, as well as microbial metabolites, have been shown to regulate its activity, thereby associating diet and microbiota‐derived signals to epithelial metabolism. Importantly, experimental studies indicate that in wild‐type mice, PPARα‐dependent lipid oxidation consumes substantial amounts of oxygen, creating a relatively hypoxic epithelial environment [2]. On the other hand, PPARα contributes to increasing the expression of occludins and tight junction proteins in the intestinal epithelium, such as claudins and ZO‐1, which ultimately maintain proper epithelial barrier function [3].

Alterations in the composition and function of the gut microbiota have been observed in PPARα‐deficient mice, including dysbiosis and changes in short‐chain fatty acid production, highlighting a bidirectional relationship between PPARα activity and microbial communities [4, 5]. Although human studies have identified associations between PPARα genetic variants and metabolic responses to diet [6], the impact of these polymorphisms, particularly non‐coding ones, on gut microbiota composition and intestinal gene expression remains poorly understood. One such variant, rs6008259 (G→A), located in the 3′ untranslated region (3′ UTR) of the PPARα gene, does not alter the amino acid sequence but may influence gene expression post‐transcriptionally by affecting mRNA stability or microRNA interactions [4]. While other PPARα variants, such as rs4253778, have been linked to lipid traits in Mexicans [5], rs6008259 is a common SNP with a relatively high minor allele frequency (∼0.4–0.5) in individuals of Mexican or Latin American ancestry, as reported in the 1000 Genomes dataset [6], suggesting its potential relevance to population‐specific metabolic phenotypes and gut microbiota profiles.

Recent evidence suggests that rs6008259 participates in genetic–dietary interactions, particularly affecting lipid profiles in response to polyunsaturated fatty acid intake. For example, carriers of the G allele (GG or AG genotypes) have higher total cholesterol and LDL levels compared to AA homozygotes [7]. These interactions may vary by ethnicity and are thought to contribute to differential cardiometabolic risk. Importantly, data on the global frequency of alleles from Ensembl.org indicate that the G allele is found in approximately 73% and the A allele in 27%, with regional differences observed [8].

Together, these findings position PPARα as a critical regulator in the interconnection between lipid metabolism, oxygen homeostasis, epithelial barrier function, and host–microbiota interactions. Elucidating these mechanisms may provide evidence for therapeutic interventions targeting metabolic disorders, intestinal inflammation, and dysbiosis‐associated diseases. Therefore, the objective of the present study was to assess in PPARα^−/− mice the importance of this nuclear receptor in the taxonomy of the gut microbiota and its effects on intestinal barrier function, and also to examine in Mexican participants the impact of the PPARα rs6008259 single nucleotide polymorphism, present in approximately 8.6% of this population in homozygous variant form, on the composition of the gut microbiota and possible metabolic consequences.

2. Material and Methods

2.1. Animal Model

The study included male PPARα^−/− and PPARα^+/+ mice aged 8 weeks with the same genetic background (n = 20 animals per group). Mice were confined in cages with 2 or 3 animals per cage. The mice were housed under controlled conditions with a 12‐hour light/dark cycle at 21°C and had free access to food. Environmental enrichment was provided through nesting material and tunnels to promote natural exploratory behavior and reduce stress in mice. For 10 days, they were fed a diet according to the AIN dietary recommendations for laboratory rodents (AIN93) [9]. Humanitarian endpoints included significant weight loss (>15%), severe lethargy, reduced mobility, or inability to eat or drink; animals were monitored three times per week for these signs throughout the study. At the end of the study period, blood samples were collected from the portal vein. Mice were anesthetized with sevoflurane, and the blood was centrifuged at 3000 rpm for 10 min to obtain serum samples. Liver and intestinal tissues were also collected and stored at ‐70°C for further analysis. Based on SAGER guidelines, we clarified that only male mice were use in this study in order to avoid sex‐related hormonal variability that could alter the composition of the gut microbiota, and all previous models established in our laboratory using mice for this line of research have been developed in males, ensuring consistency and comparability between studies. Finally, considering the timing of the interventions and the fact that female rodents have hormonal cycles, the most appropriate model was male mice to avoid confounding factors. All procedures were approved by the Animal Care Committee of the National Institute of Medical Sciences and Nutrition (FNU‐1828‐16/19‐1). In addition, to assess colonic hypoxia, another group of mice was injected intraperitoneally with pimonidazole (60 mg/kg body weight) 90 min before euthanasia. To calculate the sample size, the formula for comparing means was used [10]. Previously published studies on gut microbiota were also taken into account to refine the final n for this study, which had adequate statistical power, and the minimum number required for each experiment was taken into account.

2.2. Body Composition in Animal Model

The body composition of each mouse was evaluated at the end of the study using magnetic resonance imaging (EchoMRI, Echo Medical Systems, Houston, TX, U.S.A.) to measure lean and fat mass. Mice were placed in a thin‐walled plastic cylinder with a cylindrical plastic insert to restrict movement in a quantitative magnetic resonance imaging system. Inside the cylinder, the animals were briefly exposed to a low‐intensity electromagnetic field (0.05 Tesla) for 2 min (n = 20).

2.3. Energy Expenditure

The energy expenditure of PPARα^+/+ and PPARα^−/− mice was assessed using indirect calorimetry (n = 10). The animals were housed individually for 48 h in Plexiglas cages with an open‐flow system connected to an Oxymax laboratory animal monitoring system (CLAMS, Columbus Instruments, Columbus, OH, USA). The animals were acclimated for 24 h, fasted for 12 h during the light period, and fed during the dark period. Oxygen consumption (VO₂, ml/kg/h) and CO₂ production (VCO₂, ml/kg/h) concentrations were monitored to calculate oxygen consumption and respiratory exchange ratio. Measurements were obtained in each chamber at 22‐min intervals. Energy expenditure (EE) was calculated using the following equation: EE = (3.815 + 1.232 * RER) * VO₂.

2.4. Serum Biochemical Variables

Total cholesterol, triglycerides, urea, and glucose were measured in serum obtained from mice after 16 h of fasting using COBAS c111 photometric equipment (n = 20).

2.5. Hepatic Triglycerides

The Folch Method was used to extract the lipids from biological samples (n = 10). The sample was first mixed with a solvent mixture of chloroform and methanol to disrupt cell membranes and release lipids into the solvent. The sample was then shaken extensively to ensure extraction. Subsequently, methanol was added to induce separation between the chloroform and the aqueous phase to dissolve the lipids in chloroform. Finally, the samples were dried with nitrogen to concentrate the lipid extracts. Total hepatic triglycerides and cholesterol concentrations were measured with triglycerides and Total Cholesterol FS Kits (DiaSys, Diagnostic Systems, Hoizheim Germany). Lipid concentrations were calculated from samples absorbance measured with a spectrophotometer (Bio Rad Laboratories, Hercules, CA, USA).

2.6. Assessment of Colonic Hypoxia

Hypoxia was measured in longitudinal sections of the colon, previously fixed in paraffin, from mice previously injected with pimonidazole (n = 5), using Hypoxyprobe as an antibody as previously described in [10].

2.7. Histological Analysis

Colon samples were fixed in 10% phosphate‐buffered formalin and embedded in melted paraffin blocks (n = 8). After the paraffin solidified, the blocks were cut to a thickness of 4 µm, which were stained with hematoxylin–eosin. The histological sections were observed under a Leica microscope, stained with hematoxylin and eosin (Leica DM750 Wetzlar, Germany).

2.8. Analysis of Bioenergetics in Isolated Mitochondria

Mitochondrial respiration was determined in isolated mitochondria from mice colon (n = 5) using the Seahorse XFe96 Extracellular Flux Analyzer (Agilent Technologies, Santa Clara, CA, USA) as previously described [11]. After euthanasia, colon was dissected, homogenized, and centrifuged at 800 x g for 10 min at 4°C. The supernatant was collected and centrifuged at 8000 xg for 10 min at 4°C. The pellet containing the mitochondria was washed and resuspended on mitochondrial assay solution (MAS1, 220 mM d‐Mannitol, 70 mM sucrose, 10 mM KH₂PO₄, 5 mM MgCl₂, 2 mM HEPES, 1 mM EGTA, 0.2% BSA, pH 7.2) with the addition of substrates (10 mM glutamate and 5 mM malate). Total protein was determined using the Qubit 3.0 Fluorometer. Then 8 µg of isolated mitochondria were added per well in the XFe96 plate. Oxygen consumption rate (OCR) was measured in seventechnical replicates for each mouse using the following compounds: adenosine diphosphate (4 mM), oligomycin (1.5 µM), carbonyl cyanide‐p‐trifluoromethoxyphenylhydrazone (2 µM), and a mixture of rotenone and antimycin A (1.0 µM). Three respiration states were calculated: basal respiration, ATP production or phosphorylating respiration, and proton leak or non‐phosphorylating respiration.

2.9. Transcriptomic Analysis

Total RNA was extracted from intestinal tissue using the guanidinium thiocyanate/cesium chloride gradient method (n = 4). RNA integrity and quality were assessed by capillary electrophoresis using the QIAxcel Advanced System (QIAGEN, Hilden, Germany). RNA libraries were prepared according to the Illumina TruSeq Stranded Total RNA Sample Preparation protocol (Illumina, San Diego, CA, USA). Libraries were sequenced on the Illumina HiSeq 2500 platform at Lifesequencing S.L., generating paired‐end reads (2 × 101 bp) with total read counts ranging from 38,332,247 to 79,514,286 per sample. Raw sequencing data were subjected to quality control using FastQC, and adapter sequences and potential contaminants were removed using in‐house Perl scripts. Clean reads were aligned to the reference genome using Bowtie v1.1.2. Transcript quantification and normalization across replicates were performed with eXpress v1.5.3, and abundance estimates were converted into a count matrix using the abundance_estimates_to_matrix.pl script from the Trinity pipeline. Differential gene expression analysis was conducted using two independent pipelines: limma‐voom v3.38.3, applying log2 counts‐per‐million (CPM) normalization, and DESeq2 v1.22.2. Genes were considered differentially expressed if they met the following criteria: adjusted p‐value (FDR) < 0.05, absolute log2 fold change ≥ 2, and CPM ≥ 1.

2.10. Western Blot Analysis

Colon proteins were extracted by homogenizing the tissue in a RIPA buffer (n = 6). Thirty micrograms of extracted protein were loaded onto SDS‐PAGE gels (10%) and transferred to a polyvinylidene difluoride membrane (Bio‐Rad). Blocking was performed with 5% skim milk powder, followed by antibody incubation. The membranes were incubated with specific antibodies against HIF‐1 (sc‐13515 1:1000) and Claudin‐1 (sc‐365, 0621:1000) (Santa Cruz Biotechnology, Dallas, TX, and Waltham, MA, USA); OxPhos (458094 1:2000) and ZO‐1 (402200 1:10, 000) (Thermo Fisher Scientific Waltham, MA); CPT1 (ab128568 1:2000) Ocludin (ab167161 1:50000) and GAPDH (ab181602 1:20, 000) from Abcam. The proteins were detected with goat anti‐mouse IgG‐H&L antibody (ab67891:15000) or goat anti‐rabbit IgG H&L (ab6721 1:18000) (Abcam, Cambridge, UK). Antibody detection reactions were carried out using Immobilon Western chemiluminescent substrate of HRP (Millipore, Temecula, CA). Chemiluminescence was digitized using a ChemiDoc MP imaging system (Bio‐Rad, Hercules, CA) and analyzed using Image J software.

2.11. 16sRNAr Sequencing of Gut Microbiota

Fresh fecal samples were collected at the end of the intervention, immediately frozen, and stored at −70 °C until further processing (n = 15). Genomic bacterial DNA was extracted using the QIAamp DNA Mini Kit (Qiagen, Valencia, CA, USA) following the manufacturer's protocol. The V3–V4 regions of the 16S rRNA gene were amplified using region‐specific primers with overhang adapters. Amplicon size was verified by capillary electrophoresis and subsequently purified. Library preparation was performed using the Nextera XT Index Kit (Illumina), which incorporates dual indices and Illumina sequencing adapters. After library validation, equimolar amounts of each sample were pooled and sequenced on the Illumina MiSeq platform (Illumina, San Diego, CA, USA) according to the manufacturer's instructions. FASTQ files were generated for downstream bioinformatic analysis.

2.12. Bioinformatic Analysis

Bioinformatic processing was performed using QIIME 2 (Version 2022.2). Raw paired‐end sequences were demultiplexed, trimmed, and quality‐filtered. Amplicon sequence variants (ASVs) were inferred using the DADA2 plugin. Taxonomic classification was performed using a naïve Bayes classifier trained on the SILVA 16S rRNA gene database, specific to the V3–V4 region. Alpha‐ and beta‐diversity metrics were calculated using core diversity analyses, and microbial community structure differences were assessed using principal coordinate analysis (PCoA) based on Bray–Curtis dissimilarity. Statistical comparisons were performed using PERMANOVA. Differential abundance analysis was performed using Linear Discriminant Analysis Effect Size (LEfSe). Taxa with statistically significant differences among experimental groups were identified based on a Kruskal–Wallis test (p < 0.05), followed by Linear Discriminant Analysis (LDA) to estimate the effect size of each differentially abundant feature.

2.13. Human Study Design

A cross‐sectional study was conducted to evaluate the frequency of PPARα polymorphisms and their relationship with the gut microbiota in a population of the urban area of Mexico City, Mexico. This study was conducted in the Department of Nutrition Physiology at the Salvador Zubirán National Institute of Medical Sciences and Nutrition (INCMNSZ) in Mexico City from January, 2013 to June, 2018 in adults who met the inclusion criteria (n = 168). The inclusion criteria were: Mexican mestizos, over 18 years of age, with a body mass index (BMI) >18.5 kg/m² and without any chronic disease, except for type 2 diabetes. Patients with type 2 diabetes must have had the disease for a period of between 1 and 7 years and have been treated for glucose control with metformin.

Exclusion criteria were a history of cardiovascular events, weight loss >3 kg in the last three months, cancer, HIV, kidney or liver disease, pregnancy, smoking, substance abuse, alcohol consumption, or having taken any medication. Any drug or medication that activates intestinal motility, laxatives, or antispasmodics 4 weeks before the study, antibiotic treatment 2 months before the study, treatment with probiotic/prebiotic/symbiotic foods and foods rich in fiber (>15 g of fiber), patients with a functional digestive disorder (constipation, diarrhea, dyspepsia, or functional abdominal distension determined by a questionnaire based on the Rome III classification), inflammatory bowel disease, irritable bowel syndrome, or other chronic gastrointestinal diseases, and patients who had undergone major abdominal surgery. Exclusion criteria for participants with type 2 diabetes only included concentrations greater than HbA1c 9.9%, fasting glucose 12.2 mM (220 mg/dL), total cholesterol 62.4 mM (240 mg/dL), and triglycerides 3.99 mM (350 mg/dL).

This study was conducted in accordance with the guidelines established in the Declaration of Helsinki, and all procedures involving human subjects were approved by the INCMNSZ Ethics Committee, which determined the gut microbiota and the presence of polymorphism (approval numbers 346, 793, 1165, and 2968). All participants were informed, and written informed consent was formally obtained.

2.14. Anthropometric and Biochemical Variables

Patients attended the assessment only once, where their medical history was obtained, anthropometric variables were measured, and serum biochemical parameters were measured in a blood sample. Stool samples were collected to determine the gut microbiota in isolated DNA. The anthropometric assessment included measurements of body weight and height. Fat and lean mass percentages were obtained using bioelectrical impedance analysis (Inbody 720; Inbody Co., LTD, South Korea) in the morning, after a 12‐hour fast. Body composition was performed with patients wearing light clothing and barefoot on the electrodes of the analyzer platform and with their arms holding the upper electrodes at a 15‐degree angle to their sides. All measurements were performed by a trained nutritionist. Also, a fasting blood sample was also obtained, centrifuged at 3000 rpm for 10 min, and then the serum was separated and stored at –80°C until analysis. Serum concentrations of glucose, total cholesterol, LDL cholesterol, HDL cholesterol, TAG, glycated hemoglobin, and C‐reactive protein were analyzed using an enzymatic colorimetric method with the Cobas C111 analyzer (Roche Diagnostic, Indianapolis, IN). Insulin (80‐INSHU‐E01 ALPCO) was measured by ELISA.

2.15. Genotyping

During the visit, an additional 5 mL blood sample was also obtained, from which DNA was extracted from leukocytes using the QIAamp DNA Blood Mini Kit (QIAGEN), as previously described [12]. The presence of the PPARα gene SNP (rs6008259) was evaluated by real‐time PCR using predesigned TaqMan assays (Roche) on a LightCycler 480 instrument (Roche) (n = 168). Genotypes were in Hardy–Weinberg equilibrium according to the χ² test.

2.16. Statistical Analysis

Results obtained in this study are presented as mean ± SEM. The unpaired Student's t‐test analysis was used to compare between groups. The Shapiro‐Wilk test was performed in all experiments to assess normality, and the ROUT method was used to identify outliers. No data were excluded in this study, as no outliers were detected. Spearman's correlations were performed between serum triglycerides and different taxa. Statistical analyses were performed using GraphPad Prism V. 10.6.1, and the statistical significance was considered at p < 0.05.

3. Results

3.1. PPARα^−/− Mice Gained Body Fat Despite Gaining Lower Body Weight

In order to understand the role of PPARα as a modulator of the gut microbiota, PPARα^+/+ and PPARα^−/− mice were fed for 10 days with AIN‐93 diet (Figure 1A). The results showed that PPARα^−/− mice gained 66% less body weight compared to PPARα^+/+ (p = 0.002) (Figure 1B), despite mice in both groups consuming a similar amount of diet during this period (Figure 1C). Interestingly, even though PPARα^−/− mice gained less body weight, the percentage of body fat mass was approximately 1.6‐fold significantly higher, and this was accompanied by a reduced percentage of lean mass in comparison with WT mice (Figure 1D). It is noteworthy that no differences were observed in the RER analysis as well in VO₂, suggesting that after the consumption of AIN93 diet there were no differences in the type of energy substrate used in both groups, as well as in energy expenditure assessed by indirect calorimetry (Figures 1E,F).

FIGURE 1.

Physiological characterization of the PPARα^−/− murine model. (A) Experimental design comparing PPARα^+/+ and PPARα^−/− mice. (B) Body weight gain, (C) food intake, (D) body composition, expressed as percentages of fat and lean mass (n = 20). (E) Respiratory exchange ratio (RER) and (F) Oxygen consumption (VO₂) (n = 10). Data are presented as mean ± SEM. Statistical differences are indicated by asterisks: **p < 0.01, ***p < 0.001, ****p < 0.0001.

3.2. PPARα ^−/− Mice Increased Hepatic and Circulating Levels of Lipids

Furthermore, PPARα^−/− mice showed significantly less liver weight (0.70 ± 0.03 g) than PPARα WT mice (0.86 ± 0.02 g) (p< 0.001) (Figure 2A). However, livers of PPARα^−/− mice had 8.7‐fold significantly higher triglyceride content than WT mice (Figure 2B). In addition, the results showed several serum biochemical abnormalities in PPARα^−/− mice, including 61% reduction in blood glucose, but 45% and 47% increased circulating levels of total cholesterol and triglycerides, respectively, and an increase of 25% of urea, compared to WT mice (Figure 2C–F).

FIGURE 2.

Hepatic and serum lipid alterations in PPARα^−/− mice. (A) Liver weight. (B) Hepatic triglyceride content. Serum measurements of (C) glucose, (D) total cholesterol, (E) triglycerides, and (F) urea. Data are presented as mean ± SEM. Statistical differences between PPARα^+/+ and PPARα^−/− mice are indicated by asterisks: **p < 0.01, ****p < 0.0001. n = 20.

3.3. PPARα^−/− Mice Developed a Decrease in Oxygen Consumption and an Increase in Goblet and Inflammatory Cells in the Colon

To assess whether PPARα absence led to metabolic changes in the proximal colon, we determined mitochondrial function in isolated mitochondria from this tissue, since this transcription factor is involved in the regulation of the oxidative capacity. We found that PPARα^−/− mice had a significant decreased in basal respiration compared to the PPARα^+/+ group (Figure 3A,B), indicating less oxygen consumption. Accordingly, the hypoxic state of this segment of the intestine, clearly indicated that the absence of PPARα produced a significantly less hypoxic state (85%) determined with the hydroxyprobe pimonidazole hydrochloride, compared to PPARα^+/+ mice (Figure 3C,D). To determine whether deletion of PPARα could induce morphologic alterations in the colon, we conducted histological analyses. There were not significant differences in the crypt depth and width in this segment of the intestine, however, the inter‐crypt distance significantly increased in the PPARα^−/− mice (1.2‐fold) (Figure 3E,F). Particularly, there was a significant increase in the thickness of the intercriptal lamina propria (76.8%) (Figure 3G), as well as an increase in the number of goblet cells (101%) (Figure 3H) in PPARα^−/− mice. Furthermore, between crypts, PPARα^−/− mice showed a significant increase of 54% in the total number of inflammatory cells (Figure 3I). The alteration in the hypoxic state in the PPARα ^−/− mice was accompanied with significant changes in the expression of several genes involved in gut barrier function and fatty acid oxidation among others, compared to PPARα^+/+ (Figure 3J). Thus, WT mice had a higher abundance of HIF1α and CPT‐1 proteins compared to PPARα^−/− mice, while a significant decrease in tight junction proteins such as zonula occludens‐1 (ZO‐1), claudin 1, and occludin was observed in PPARα^−/− mice (Figure 3K,L). These results indicate that the absence of PPARα leads to a decrease in oxygen consumption in the colon, promotes and increase in the proportion of goblet and inflammatory cells per crypt and a clear reduction in tight junction proteins, suggesting a clear impairment in colon homeostasis, possibly related to changes in gut microbiota.

FIGURE 3.

PPARα^−/− mice display reduced colonic hypoxia. (A–B) Mitochondrial respiratory states of isolated colonic mitochondria, measured by oxygen consumption rate (OCR). (C) Colonic hypoxic microenvironment in PPARα^+/+ and PPARα^−/− mice following pimonidazole injection. (D) Relative fluorescence intensity of colon sections; immunofluorescence analyzed with ImageJ for hypoxia assessment (n = 5). (E) Hematoxylin–eosin staining of ascending colon sections. (F) Crypt depth, width, and intercrypt distance. (G) Thickness of intercryptal lamina propria. (H) Number of goblet cells per crypt. (I) Total number of inflammatory cells. (n = 8) (J) Differential gene expression in colon between PPARα^+/+ and PPARα^−/− mice (n = 4) (K) Representative western blots and (L) densitometric quantification of tight junction proteins (ZO‐1, claudin‐1 and occludin), CPT1, HIF1α, and mitochondrial complexes in colon (n = 6). Data are presented as mean ± SEM. Statistical differences are indicated by asterisks: *p < 0.05, **p < 0.01, ***p < 0.001.

3.4. PPARα^−/− Mice Modified the Gut Microbiota Compared to PPARα^+/+ Mice

Gut microbiota analysis was performed between PPARα^−/− and PPARα^+/+ mice. The results showed a significant change in the alpha and beta diversity of the gut microbiota. PPARα^−/− mice had less α−diversity (Figure 4A), and the β‐diversity showed high dissimilitude compared to PPARα^+/+ mice (Figure 4B). The taxonomy of the gut microbiota was changed even at the phylum level whether PPARα^−/− mice showed a reduction in Firmicutes and elevation in Proteobacteria. Furthermore, there was a decrease in Desulfobacterota and Deferribacterota, and interestingly showed an increase in Verrucomicrobiota (Figure 4C). At the genus level as expected, PPARα^−/− mice significantly increased Akkermasia, Parabacteroides, Blautia and Lachnoclostridium (Figure 4D–G), whereas there was a reduction in the genus Bacteroides, Mucispirillum, Alistipes and Oscillibacter (Figure 4H–K). The results showed that the species that increased significantly in the PPARα^−/− mice according to the LDA score were Akkermansia muciniphila, Massiliprevotella massiliensis and Acutibacter muris, whereas in the PPARα^+/+ mice Ilebacterium valens, Bacteroides caecimuris, Turicimonas muris, and Odoribacer splanchnicus (Figure 4L).

FIGURE 4.

Gut microbiota composition differs according to PPARα genotype. (A) Alpha diversity measured by Shannon index. (B) Principal coordinate analysis of beta diversity. (C) Microbiota taxonomy at the phylum level. Relative abundance at the genus level for (D) Akkermansia, (E) Parabacteroides, (F) Blautia, (G) Lachnoclostridium, (H) Bacteroides, (I) Mucispirillum, (J) Alistipes, and (K) Oscillibacter. (L) Linear discriminant analysis (LDA) highlighting significant differences in bacterial abundance between WT (PPARα^+/+) and KO (PPARα^−/−) mice. Statistical differences are indicated by asterisks: ****p < 0.0001, n = 20.

3.5. The PPARα RS6008259 Polymorphism Modified the Gut Microbiota and Altered the Serum Lipid Profile in Humans

We then studied the relationship between PPARα and the gut microbiota in humans. To this end, we studied the presence of a SNP of PPARα that is common in the Mexican population. We studied the PPARα SNP rs6008259, which is located in the locus 4248 of the 22 chromosomes, in the long arm, and this polymorphism involves a change from G (common allele) to A (non‐common allele), and is a UTR variant of the 3’UTR. A previous study in the Mexican population showed an allele frequency of the allele G of 91.4%, and the frequency for the allele A of 8.6 %, whereas in the world population the frequency of the G allele is 73% and 27% for the A allele. In the population studied of 177 individuals, the frequency of allele G was 91.53% and for the A allele of 8.47%, that is very similar to that observed for the Mexican population, and it was in Hardy–Weinberg equilibrium (p = 0.218). Interestingly, similar to what was found in the PPARα^−/− mice animal model, subjects with the non‐common allele showed a significant increase in circulating triglyceride levels (84%). In addition, we also observed that these subjects had a significantly higher concentration of LDL cholesterol. On the other hand, similar to the PPARα^−/− mice model, subjects with the non‐common polymorphism had higher body fat mass (13.7%) and lower lean body mass (9.4%) compared to subjects with the common allele (Table 1).

TABLE 1.

Anthropometric and biochemical variables in humans genotyped.

Parameter	Commun (Mean, CI)	Non Common (Mean, CI)	p‐value
Age	45.95 (43.76‐48.14)	43.77 (38.58‐48.96)	0.43
Weight	83.43 (80.35‐86.52)	86.02 (78.02‐94.03)	0.54
BMI	32.62 (31.59‐33.65)	31.83 (29.33‐34.32)	0.55
Lean Mass	59.37 (58.09‐60.65)	53.78 (50.77‐56.80)	0.001
Fat Mass	40.63 (39.35‐41.91)	46.22 (43.20‐49.23)	0.001
Insuline	11.55 (9.87‐13.22)	11.65 (8.08‐15.22)	0.95
HbA1c	6.28 (6.08‐6.47)	6.74 (6.02‐7.46)	0.21
PCR	4.02 (3.27‐4.76)	2.82 (1.66‐3.97)	0.07
Glucose	109.5 (104.5‐114.5)	106.5 (98.9‐113.9)	0.49
Total Cholesterol	191.7 (184.17‐199.23)	208.5 (186.9‐230.1)	0.14
HDL Cholesterol	40.1 (37.98‐42.21)	39.20 (34.93‐43.46)	0.71
LDL Cholesterol	117.2 (111.7‐122.7)	146.3 (132.1‐160.6)	0.0004
Triglycerides	163.2 (152.8‐173.5)	301.7 (250.9‐352.47)	0.000006

Our results showed that the presence of the non‐common allele in the human population study was associated with changes in the gut microbiota similar to those obtained in PPARα^−/− mice: subjects with the non‐common allele had lower alpha diversity than those with the common allele (Figure 5A). In addition, we observed a clear dissimilarity in the gut microbiota between individuals with the non‐common allele and those with the common allele, according to principal component analysis (Figure 5B). Taxonomic results showed that subjects with the non‐common allele had an increase in the abundance of Bacteroidetes and a decrease in Firmicutes compared to individuals with the common allele (Figure 5C). In addition, subjects with the non‐common allele had more Proteobacteria than those with the common allele. Interestingly, at the genus level, five genera, including Akkermansia, Blautia, Lashnoclostridium, Mucispirillum, and Oscillibacter, showed a similar trend in abundance in subjects with the non‐common allele, as observed in PPARα^−/− mice (Figure 5D–O). Linear discriminant analysis showed that the most abundant species in subjects with the non‐common allele were Akkermansia muciniphila, Blautia producta, Muribaculum intestinali, Parabacteroides goldsteinii, Azosphirillum lipoferum, Alistipes finegoldii, Escherichia coli, and Streptococcus salivarius (Figure 5P).

FIGURE 5.

The rs6008259 SNP is associated with altered gut microbiota profiles. (A) Alpha diversity (Shannon index). (B) Principal coordinate analysis of beta diversity. (C) Microbiota taxonomy at the phylum level. Relative abundance at the genus level for (D) Alloprevotella, (E) Anaerostipes, (F) Bifidobacterium, (G) Fusicatenibacter, (H) Lachnospiraceae ND3007 group, (I) Monoglobus, (J) Mucispirillum, (K) Oscillibacter, (L) Subdoligranulum, (M) Akkermansia, (N) Blautia, and (O) Lachnoclostridium. (P) LDA highlighting significant differences in bacterial abundance. Serum measurements of (Q) triglycerides and (R) total cholesterol in the human cohort stratified by rs6008259 genotype: C = common allele, NC = non‐common allele. Statistical differences are indicated by asterisks: **p < 0.01, ***p < 0.001, ****p < 0.0001, n = 168.

It is noteworthy that in the animal model as in the human study, there was a positive and significant correlation between serum triglyceride concentrations and the relative abundance of the genera Blautia and Lachnoclostridium, whereas there was a negative and significant correlation between serum triglyceride and the genera Oscillibacter (Figure 6A–F).

FIGURE 6.

PPARα‐associated taxa correlate with serum triglycerides. Correlation analysis between serum triglycerides and relative abundance of (A) Blautia, (B) Lachnoclostridium, and (C) Oscillibacter in the murine model (n = 20). Correlation analysis between serum triglycerides and relative abundance of (D) Blautia, (E) Lachnoclostridium, and (F) Oscillibacter in the human cohort (n = 168). Raw data and Spearman's correlation coefficients are presented; p‐values are indicated in individual panels.

4. Discussion

This study provides compelling evidence that PPARα plays a crucial role in maintaining intestinal metabolic homeostasis and modulating host microbiota interactions with systemic metabolic consequences. Using a combined approach in both murine models and human subjects, we demonstrate that the absence or functional alteration of PPARα leads to disrupted lipid metabolism, altered colonic physiology, compromised intestinal barrier integrity, and significant shifts in gut microbial composition.

Although PPARα^−/− mice exhibited significantly lower body weight gain, they interestingly accumulated greater adipose tissue and exhibited reduced lean mass. This phenotype was associated with markedly elevated hepatic triglyceride content and dyslipidemia, as evidenced by increased circulating levels of triglycerides and total cholesterol, despite a similar caloric intake and unchanged energy expenditure. These findings align with prior studies showing that PPARα regulates fatty acid oxidation and lipid clearance pathways, particularly in hepatocytes and enterocytes, thus preventing ectopic lipid deposition and dyslipidemia under metabolic stress [13, 14, 15].

At the colonic level, our data reveal that PPARα deficiency leads to a reduction in oxygen consumption and physiological hypoxia in the intestinal mucosa. These results were accompanied by diminished expression of genes related to fatty acid oxidation and decreased HIF‐1α protein levels. These changes were associated with impaired abundance of tight junction proteins (ZO‐1, claudin‐1, and occludin), increased lamina propria thickness, and greater inflammatory cell infiltration. It is well‐established that mitochondrial oxidation of fatty acids by intestinal epithelial cells consumes local oxygen, creating a physiological hypoxic niche that stabilizes HIF‐1α and promotes barrier function. Loss of this metabolic‐epithelial signaling axis, as observed in PPARα^−/− mice, compromises epithelial integrity and predisposes to a pro‐inflammatory intestinal environment [16, 17, 18, 19].

This disrupted epithelial‐microenvironmental balance translated into pronounced alterations in the gut microbiota [20, 21]. Alpha and beta diversity were significantly altered in PPARα^−/− mice, with enrichment of phyla and genera previously associated with intestinal inflammation and metabolic dysregulation. Notably, taxa such as Bacteroidetes, Proteobacteria, Blautia producta, and Escherichia coli were overrepresented, while Firmicutes were depleted, a pattern shift commonly observed in obesity and metabolic syndrome [22, 23, 24, 25]. Similar microbial patterns were observed in human subjects carrying the non‐comun allele (A) of the rs6008259 polymorphism in the 3’UTR of PPARα, that also exhibited slightly increased body fat, decreased lean mass, and a dyslipidemic profile.

These findings highlight a gene–microbiota–metabolism axis wherein host genetic variation at the PPARα locus influences both metabolic outcomes and gut microbial profile. The rs6008259 SNP may alter mRNA stability or microRNA binding, leading to functional downregulation of PPARα expression [7, 26, 27, 28]. Although functional studies are needed to confirm this mechanism, the convergence of phenotypes across species strengthens the plausibility of a causal relationship. Furthermore, the observed correlations between serum triglyceride levels and the relative abundance of Blautia and Lachnoclostridium, both in mice and humans, suggest a possible microbiota‐mediated contribution to the lipid phenotype, though causality remains to be established. It is important to point out that the results obtained in humans are based on a cross‐sectional and correlational design; therefore, they should be interpreted as associations without implying causality, which requires future studies for confirmation.

The translational significance of these findings lies in the dual potential of PPARα as both a metabolic regulator and a modulator of host–microbiota interactions. Future studies can investigate whether activation of PPARα can restore colonic hypoxia, enhance barrier function, and correct microbiota dysbiosis, ultimately improving metabolic outcomes. Moreover, stratifying patients by PPARα genotype may offer a precision medicine approach to identify individuals who might benefit from such interventions [29].

There are some limitations to our study. The murine model clearly demonstrated the mechanism that occurs, particularly in the intestine, between the interaction of the gut microbiota with metabolic and gene expression changes in the intestinal epithelium. However, this study did not obtain human intestinal biopsies, which would allow validation of whether the changes in the intestinal mucosa of subjects with the PPARα variant are similar to those observed in the animal model. Future studies will be important to validate whether the PPARα variant studied has the same metabolic impact, although interestingly, the changes in the gut microbiota taxonomy were similar to those observed in the murine model. A second variable to consider in future human studies is the impact of diet on the gut microbiota and metabolic phenotype in subjects with and without the PPARα variant. This variable was only available for a subpopulation of participants. Although the exploratory analysis did not reveal relevant differences in dietary intake between the groups evaluated, the potential influence of other dietary patterns or nutritional factors on gut microbiota composition cannot be ruled out. Therefore, future studies should incorporate a more comprehensive dietary and nutritional status assessment to better understand its interaction with the gut microbiota. Another limitation of the study is the potential effect of sex on the variables analyzed. The stratified analysis in the human cohort showed no differences between men and women. However, it has been reported in mice that PPARα exerts a sex‐dependent differential effect on immune and inflammatory responses, which depends in part on fluctuations in sex hormones [30]. For this reason, the present study was conducted in male mice to avoid this confounding effect; however, in future studies it will be important to determine whether sex, as a variable, can influence the association between PPARα and the gut microbiota. Lastly, it cannot exclude the possibility that other transcription factors or epigenetic regulators contribute to the observed intestinal and metabolic alterations and further studies in intestinal epithelial tissue could clarify if other mechanisms are involved in subjects with SNPs variant [31, 32].

In summary, we identify PPARα as a key modulator of intestinal oxygen metabolism, barrier integrity, and gut microbial composition. Loss of PPARα function leads to a pro‐inflammatory intestinal environment and dysbiosis in gut microbiota, contributing to an altered metabolic profile.

Author Contributions

Mónica Sánchez‐Tapia: conceptualization, methodology, formal analysis, investigation, writing – original draft and visualization. Sandra Tobón‐Cornejo: conceptualization, methodology, formal analysis, investigation, writing – original draft and visualization. Martha Guevara‐Cruz: methodology, formal analysis, investigation. laura a. Velázquez‐Villegas: investigation, writing – original draft. Rogelio Hernández‐Pando: investigation. Adriana López‐Barradas: investigation. Rocio Guizar‐Heredia: investigation. Otoniel Maya: formal analysis. Janette Furusawa‐Carballeda: investigation. Rosa Rebollar‐Vega: investigation. Ariana Vargas‐Castillo: conceptualization. Nimbe Torres: writing – review & editing. Armando R. Tovar: conceptualization, supervision and writing – original draft.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The 16S rRNA gene and RNAseq sequencing raw sequence reads (fastq) are available at the NCBI Sequence Read Archive with BioProject IDs PRJNA1360570 and PRJNA1360570. The database is available upon a reasonable request to the corresponding author.