Western diet promotes endometriotic lesion growth in mice and induces depletion of Akkermansia muciniphila in intestinal microbiota

Eight-week-old female BALB/cJRj mice, weighing 15–17 g, were purchased from Janvier Laboratory® (Le Genest Saint Isle, France) to establish the murine model of endometriosis. All animals received humane care following institutional guidelines, including being maintained under a standard 12-h photoperiod with ad libitum access to food and water throughout the study. All experimental procedures and animal care were approved by the institutional review board. The study protocol underwent thorough review and received approval from the “Comité d’Ethique en matière d’Expérimentation Animale” at Paris Cité University (CEEA 34), Paris (agreement #B75-1405, APAFIS #30,327–2019090511442754-v3). Female BALB/c mice were randomly assigned to one of two diets: (1) a purified control diet (CD) ( n  = 32), consisting of 17% total kcal from lard fat (D12450K, Brogaarden ApS, 3540 Lynge, Denmark), and (2) a matched purified Western diet (WD) ( n  = 32), containing 45% of total kcal from lard fat (D12451, Brogaarden ApS, 3540 Lynge, Denmark) (Additional File 1: Supplemental Table 1). This WD is considered a medium-fat diet (45%) compared to high-fat diets (60% and higher). Over a period of 4 weeks, mice were fed their respective diets. Following this dietary intervention, endometriosis was surgically induced in the mice by syngeneic transplantation of horn tissue, following a previously described protocol [ 19 ]. Prior to implantation, a preoperative gavage was administered to synchronize the estrous cycle of all recipient mice, with each receiving 56 μg/kg/day of 17β-estradiol (Provames®, Sanofi-Aventis, France) for 5 days [ 20 ]. The donor mice (3 mice in each CD and WD group) were fed the same diet as the recipient mice. The mice continued on the same diet until they were sacrificed. Upon sacrifice by cervical dislocation, uterine horns were surgically extracted and transferred into a Petri dish containing Dulbecco’s modified Eagle’s medium (DMEM) warmed to 37 °C (DMEM supplemented with 10% fetal calf serum, 100 U/mL penicillin, and 0.1 mg/mL streptomycin; PAA, Paris, France). Under a stereo-microscope (M651; Leica Microsystems, Paris, France), the uterine horns were longitudinally opened with microscissors, and 5-mm-length samples were prepared for grafting into the peritoneal cavity of recipient mice. Recipient BALB/cJRj mice were divided into two new groups: control diet with endometriosis (CD + OSE) and Western diet with endometriosis (WD + OSE). These mice were then anesthetized with isoflurane, intubated, and mechanically ventilated. An incision was made on the ventral midline, and a 5-mm donor horn fragment was sutured onto the parietal peritoneum using two 7/0 polypropylene stitches (Prolen®, Ethicon, Somerville, NJ). In all endometriosis mice, two pieces of donor horns were symmetrically sutured on each side of the abdominal incision to ensure comparable vascularization at host tissue sites. The two control groups underwent the same surgical procedure, with the exception of the horn graft, which was replaced by sutures alone. The cutis was sutured using a 6/0 nylon thread. Throughout the experiment, mice were weighed weekly (Fig.  1 ). The experiment was repeated twice, resulting in a total of 58 mice divided into four groups: the first experimental set comprised n  = 29 mice, while the second experimental set also comprised n  = 29 mice). Fig. 1 Experimental design Experimental design The implanted horn pieces were weighed extemporaneously. Evaluation of implant size and volume was conducted on days 20, 39, and 47 post- surgery using serial ultrasonography, as previously described [ 21 ], with the Vevo 2100 high-frequency ultrasound imaging system (VisualSonics®, Toronto, Canada). The ultrasound probe utilized had a 40-MHz center frequency (MS550) and an adaptable focal depth, providing a spatial resolution at the focus of 40 × 80 × 80 μm 3 . During the examination, mice were anesthetized with 1.5% isoflurane and placed on a heated stage while being restrained. The abdominal area was shaved using a depilatory cream, and ultrasound contact gel was applied to the abdomen. An image sequence capturing two-dimensional axial views of the endometriotic implant was acquired as the probe was swept from the upper to the lower abdominal wall of the mouse. The implant volume was calculated using the formula: TV (mm 3 ) = (L × W 2 )/2 [ 22 ]. All ultrasound examinations were conducted at the “Plateforme Imageries du Vivant (PIV) of Paris Cité University, INSERM U1016” Paris, France. Image acquisitions were performed by the same blinded operator. At day 49 post-implantation, animals were sacrificed by cervical dislocation. The endometriotic implants were then surgically removed and weighed. The right-side implant of each mouse was prepared for flow cytometry analysis. The left-side implant was divided into two halves: one half was promptly frozen in liquid nitrogen for subsequent RNA extraction, followed by reverse transcription quantitative real-time PCR analysis. The other half was fixed with 10% formaldehyde for further histological analysis. Multiple 5-μm paraffin sections were prepared. These sections were stained with hematoxylin and eosin or Masson’s Trichrome according to the manufacturer’s protocol. Immunochemistry was performed on slides using the Leica Bond RX automated system. The slides were subjected to unmasking at pH 6, followed by a 30-min incubation with an anti-Ki67 antibody (ab15580, Abcam) at a 1/500 dilution and with an leptin (Ob) receptor antibody (af497, R&D Systems) at a 1/100 dilution. After washing, the secondary antibody HRP conjugate was applied, and the signal was visualized using diaminobenzidine (DAB) with the “Bond Polymere Refine” kit (DS9800, Leica). Analysis of the immunohistochemistry images was conducted using the QuPath software ( https://qupath.github.io/ ). The immunohistochemistry images were stained with 3,3′ DAB and hematoxylin. In brief, a mean filtering process was applied to all images to facilitate color segmentation. For analyses of Masson trichrome staining, a fixed threshold was selected to distinguish the color corresponding to the stained segment. The selected regions were uniformly expanded with respect to filter threshold, gain, and magnification. For nuclear Ki67 quantification, a color deconvolution process was employed, resulting in the production of three images, one of which depicted the DAB staining. A QuPath plugin was used to analyze the nuclear staining patterns of the deconvoluted DAB images, specifically focusing on nuclei of epithelial cells. The quantification was expressed as a percentage of the area occupied by the color of interest (brown nuclei for Ki67 and Ob receptor marking and blue marking for collagen). Total mRNA was extracted from crushed samples using TRIzol reagent (Invitrogen, Carlsbad, USA). qRT-PCR was performed with a QuantiTect SYBR® Green RT-PCR Kit on a LightCycler 480 II instrument (Roche Applied Science, Indianapolis, USA). A panel of six genes, including four target genes and two reference genes (internal controls), was analyzed by quantitative RT-PCR using cDNA synthesized from each sample. Primers for RT-PCR analysis were selected using PRIMER3 software or were based on published sequences from a previous study [ 23 ] (Additional File 1: Supplemental Table 2). All chosen primers were aligned using BLAST software to avoid non-specific annealing and cross-amplifications. Primers were procured from Eurofins Genomics France (Nantes, France) and were used at a concentration of 10 nM in the PCR reaction. Quantitative PCR was conducted on a Light Cycler® 480 96-well apparatus (Roche Diagnostics, Manheim, Germany), with 160 ng of cDNA as the template. The PCR protocol began with an initial denaturation of 5 min at 95 °C, followed by 40 cycles of 10 s at 95 °C, an annealing phase for 10 s, 72 °C for 10 s, and concluded with a final melting curve. Across a broad range of RNA concentrations (via serial dilution), all primer sets demonstrated good linear correlation (slope ≥  − 3.6) and consistent priming efficiency across the various dilutions, as indicated by their Ct values. RT-PCR efficiencies for all genes were estimated to be ≥ 90%. The relative abundance fold changes of each target gene, compared to a set of internal controls, were determined using the formula − 2ΔΔCt. The set of internal controls comprised the geometric mean of two different reference genes: succinate dehydrogenase complex subunit A (SDHA) and glyceraldehyde-3-phosphate dehydrogenase (GADPH). Analysis of the results was performed using the LightCycler software. At the time of sacrifice, lesions were placed in 12-well plates filled with 1 mL of Roswell Park Memorial Institute (RPMI, Sigma-Aldrich, Saint-Quentin-Fallavier, France) medium on ice. The lesions were then crushed and digested, and the contents of each well were transferred to falcon tubes, filtered (70 μm), and centrifuged multiple times. Following automatic cell counting, 2 × 10 6 cells were isolated for staining. The cells were subsequently incubated with the appropriate labeled antibodies at 4 °C for 30 min in the dark, in PBS with 2% normal fetal bovine serum (FBS). Flow cytometry analysis was performed using a fluorescence-activated cell sorting (FACS) Fortessa II flow cytometer (BD Biosciences, San José, USA), following standard techniques. The monoclonal antibodies used included Zombie UV for selecting viable cells, CD45 BV510 for selecting immune cells, and CD11b APC-Cy7, CD11c PE, F4/80 BV711, CD80 PE Cy5, CD206 AF647, for characterizing macrophages. Additionally, CD45R/B220 PE Texas-Red, CD3ε FITC, and CD69 PerCP-Cy5 antibodies from eBiosciences (Thermo Fisher Scientific, Villebon-Sur-Yvette, France) and BioLegend were utilized. Data analysis was performed using the FlowJo software (Tree Star, Ashland, USA). The gating strategy for identifying peritoneal macrophages by flow cytometry is shown in Additional File 1: Supplemental Fig. 1. Retro-orbital bleeding of mice was performed a few minutes before sacrifice. Serum was obtained through two rounds of centrifugation. Blood samples were taken more than four hours after the start of the diurnal phase, corresponding to a morning fast [ 24 ]. The necessary volume per mouse was 15 μL of serum, which was not available for all mice. Lactate and glucose levels in the sera were quantified using the Cobas 8000 modular analyzer series (Diagnostics Roche, Meylan, France). At the time of sacrifice, the distal part of the colon was excised to extract the feces, which were subsequently stored at – 80 °C without a preservation medium. 16S rRNA gene amplification and sequencing were conducted using Illumina MiSeq technology, following the protocol outlined by the Earth Microbiome Project, with modifications to the MOBIO PowerSoil DNA Isolation Kit procedure for DNA extraction (available at www.earthmicrobiome.org/emp-standard-protocols .). Bulk DNA was extracted from frozen extruded feces using a PowerSoil-htp kit from MoBio Laboratories (Carlsbad, California, USA), with mechanical disruption through bead-beating. The 16S rRNA genes, specifically targeting region V4, were PCR amplified from each sample using a composite forward primer and a reverse primer including a unique 12-base barcode, designed using the Golay error-correcting scheme, which served to tag PCR products from respective samples [ 25 ]. We utilized the forward primer 515F with the following sequence: 5′- AATGATACGGCGACCACCGAGATCTACACGCT XXXXXXXXXXXXTATGGTAATTGTGTGYCAGCMGCCGCGGTAA-3′. In this sequence, the italicized part represents the 5′ Illumina adapter, the 12 X sequence indicates the Golay barcode, the bold sequence signifies the primer pad, the italicized and bold sequence represents the primer linker, and the underlined sequence denotes the conserved bacterial primer 515F. The reverse primer 806R employed had the sequence: 5′- CAAGCAGAAGACGGCATACGAGATAGTCAGCCAGCCGGACTACNVGGGTWTCTAAT -3′. Here, the italicized part corresponds to the 3′ reverse complement sequence of the Illumina adapter, the bold sequence represents the primer pad, the italicized and bold sequence denotes the primer linker, and the underlined sequence signifies the conserved bacterial primer 806R. PCR reactions were conducted using Hot Master PCR mix (Quantabio, Beverly, MA, USA), with 0.2 μM of each primer, 10–100 ng of template, and reaction conditions were as follows: 3 min at 95 °C, followed by 30 cycles of 45 s at 95 °C, 60 s at 50 °C, and 90 s at 72 °C on a Biorad thermocycler. The resulting products were visualized by gel electrophoresis and quantified using the Quant-iT PicoGreen dsDNA assay (Clariostar Fluorescence Spectrophotometer). A master DNA pool was then generated in equimolar ratios, subsequently purified using Ampure magnetic purification beads (Agencourt, Brea, CA, USA), and sequenced on an Illumina MiSeq sequencer (paired-end reads, 2 × 250 bp) at the Genom’IC platform (INSERM U1016, Paris, France). 16S rRNA sequences were analyzed using QIIME2—version 2019 [ 26 ]. The sequences underwent demultiplexing and quality filtering using the Dada2 method [ 27 ], with QIIME2 default parameters to detect and correct Illumina amplicon sequence data. Subsequently, a QIIME2 artifact table was generated. A phylogenetic tree was then constructed using the align-to-tree-mafft-fasttree command for phylogenetic diversity analyses, and alpha and beta diversity analyses were computed using the core-metrics-phylogenetic command. Principal coordinate analysis (PCoA) plots were used to assess variation between experimental groups (beta diversity). For taxonomic analysis, features were assigned to operational taxonomic units (OTUs) with a 99% threshold of pairwise identity to the Greengenes reference database 13_8 [ 28 ]. Unprocessed sequencing data are deposited in the European Nucleotide Archive under accession number PRJEB81557.

In the WD + OSE group, endometriosis lesions were larger compared to the CD + OSE mice (referred to as the controls in this experiment), as observed at the macroscopic level (Fig.  2 A (a)). Evaluation of implant volume by ultrasonography (Fig.  2 A (b)) on days 0 and 49 revealed a significant increase in the WD + OSE group compared to the CD + OSE group (69.7 ± 18.7 mm 3 and 148.3 ± 38.8 mm 3 , respectively; p  < 0.01). This difference was characterized by an accelerated rate of lesion growth in the WD + OSE group compared to the control group (Fig.  2 A (b) and Fig.  2 C). The weight of the implants before endometriosis induction surgery was similar in both the WD + OSE and CD + OSE groups (27.1 ± 6.7 mg vs. 27.3 ± 6.4 mg, respectively; p  = 0.96) (Fig.  2 B). Mice weights gain was not different between our groups except between CD and CD + OSE groups (3.0 ± 0.3 g vs. 1.6 ± 0.3 g, respectively; p  < 0.01) (Fig.  2 D). These data demonstrate that mice fed the WD diet exhibited lesions twice as large as those fed the CD diet 40 days after endometriosis induction, despite having equal grafted implant weights and equal weight gain between groups. Fig. 2 Promotion of endometriosis lesions by Western diet in vivo. A (a) Macroscopic view of the implants on day 49. (b) Ultrasonography images of peritoneal implants in mice on day 40. (c) Staining with hematoxylin and eosin of the lesion on day 49 (G: glandular cells; L: lumen; Sc: stromal cells). Original magnification × 100. B Implants weight per mice before surgical implantation to induce endometriosis in CD group ( n  = 16) and WD group ( n  = 15) (mg). C Volume evolution of the implants between day 0 (the day of the surgery) and day 40 evaluated by ultrasound (mm 3 ) in the WD + OSE ( n  = 14) and CD + OSE groups ( n  = 16). D Weight gain of mice between day 0 and day 49 (g) in the CD ( n  = 13), CD + OSE ( n  = 16), WD ( n  = 15), and WD + OSE groups ( n  = 14). Unpaired t -test ( B , C ), one-way ANOVA ( D ). ns, non-significant. * p  < 0.05, ** p  < 0.01, *** p  < 0.001. Data represent the mean and SEM from at least n  = 7 mice per group Promotion of endometriosis lesions by Western diet in vivo. A (a) Macroscopic view of the implants on day 49. (b) Ultrasonography images of peritoneal implants in mice on day 40. (c) Staining with hematoxylin and eosin of the lesion on day 49 (G: glandular cells; L: lumen; Sc: stromal cells). Original magnification × 100. B Implants weight per mice before surgical implantation to induce endometriosis in CD group ( n  = 16) and WD group ( n  = 15) (mg). C Volume evolution of the implants between day 0 (the day of the surgery) and day 40 evaluated by ultrasound (mm 3 ) in the WD + OSE ( n  = 14) and CD + OSE groups ( n  = 16). D Weight gain of mice between day 0 and day 49 (g) in the CD ( n  = 13), CD + OSE ( n  = 16), WD ( n  = 15), and WD + OSE groups ( n  = 14). Unpaired t -test ( B , C ), one-way ANOVA ( D ). ns, non-significant. * p  < 0.05, ** p  < 0.01, *** p  < 0.001. Data represent the mean and SEM from at least n  = 7 mice per group According to current literature findings, endometriosis is characterized by significant immune disruption and a pro-inflammatory state, particularly in macrophages [ 9 , 29 ]. Hence, in our mouse model, we analyzed the expression of surface molecules on immune cells from lesions. The percentage of macrophages did not differ between lesions in the CD + OSE and WD + OSE groups (Fig.  3 A). Investigation of macrophage activation by examining the surface markers CD80, CD206, and MHC-II revealed an increased fluorescence intensity for the markers CD80 ( p  = 0.02) and CD206 ( p  = 0.03) in the WD + OSE group (Fig.  3 B, C, D). The percentage of lymphocytes in the lesions remained unchanged (Fig.  3 E), as did their activation, as indicated by the CD69 study (Fig.  3 F). Fig. 3 Macrophage activation by Western diet in lesions. A – F Flow cytometry analysis on macrophage within lesions (CD11b + F4/80 +) and lymphocytes (CD45 + CD3e +). Macrophage activation markers CD206, CD80, and CMH-II, as well as lymphocyte activation marker CD69, were assessed by mean fluorescent activity and SEM. CD + OSE group ( n  = 6), WD + OSE group ( n  = 6). T -test. ns, non-significant. * p % 0.05, ** p % 0.01, *** p % 0.001 Macrophage activation by Western diet in lesions. A – F Flow cytometry analysis on macrophage within lesions (CD11b + F4/80 +) and lymphocytes (CD45 + CD3e +). Macrophage activation markers CD206, CD80, and CMH-II, as well as lymphocyte activation marker CD69, were assessed by mean fluorescent activity and SEM. CD + OSE group ( n  = 6), WD + OSE group ( n  = 6). T -test. ns, non-significant. * p % 0.05, ** p % 0.01, *** p % 0.001 With the understanding that endometriosis lesions exhibit diverse levels of inflammation and fibrosis [ 30 , 31 ], we conducted a thorough examination of these characteristics in lesions induced within our mouse model. Our characterization revealed a significant increase in fibrosis markers (Col1a1) within the WD + OSE group, albeit without a significant increase in the inflammatory marker Cyclo-Oxygenase 2 pathway (Ptgs2) (Fig.  4 A). Histological assessment of lesion fibrosis was conducted utilizing Masson trichrome coloration (Fig.  4 B, C). The area of fibrosis was notably greater in the WD + OSE group compared to the CD + OSE group, with values of 16.9% ± 5.9 and 11.43% ± 2.3, respectively ( p  = 0.04) (Fig.  4 D). Cell proliferation was assessed by Ki-67 immunostaining (Fig.  4 F, G). Lesions within the WD + OSE group exhibited a higher proliferation index, as assessed by stromal cells positive for Ki-67, compared to lesions in the CD + OSE group, with values of 29.6% ± 5.3 and 13.3% ± 3.0, respectively ( p  < 0.01) (Fig.  4 E). Lesions in the WD + OSE group demonstrated increased fibrosis and proliferation yet did not activate the Ptgs2 pathway. Fig. 4 Impact of Western diet on lesion characteristics. A qRT-PCR assessment of Ptgs2 and Col1a1 mRNA expression in lesions. Representative sections of ectopic lesions stained with Masson’s trichrome in the CD + OSE ( B ) and WD + OSE groups ( C ). D The area of fibrosis was assessed based on the blue coloration in the CD + OSE and WD + OSE lesions. E Percentage of stromal cells positive for Ki-67 immunostaining. Representative sections of ectopic lesions showing Ki-67-positive stromal cells in CD + OSE ( F ) and WD + OSE ( G ) lesions. CD group ( n  = 7), WD group ( n  = 6), CD + OSE group ( n  = 8), WD + OSE group ( n  = 7). Unpaired t -test. ns, non-significant. * p  < 0.05, ** p  < 0.01, *** p  < 0.001 Impact of Western diet on lesion characteristics. A qRT-PCR assessment of Ptgs2 and Col1a1 mRNA expression in lesions. Representative sections of ectopic lesions stained with Masson’s trichrome in the CD + OSE ( B ) and WD + OSE groups ( C ). D The area of fibrosis was assessed based on the blue coloration in the CD + OSE and WD + OSE lesions. E Percentage of stromal cells positive for Ki-67 immunostaining. Representative sections of ectopic lesions showing Ki-67-positive stromal cells in CD + OSE ( F ) and WD + OSE ( G ) lesions. CD group ( n  = 7), WD group ( n  = 6), CD + OSE group ( n  = 8), WD + OSE group ( n  = 7). Unpaired t -test. ns, non-significant. * p  < 0.05, ** p  < 0.01, *** p  < 0.001 Recognizing the role of the leptin pathway in endometriosis lesion progression, we investigated the transcription of the leptin gene (Ob) and its receptor (Obr). The expressions Ob and Obr mRNA were found to be heightened in the WD + OSE group compared to the CD + OSE group (Fig.  5 A). The significant increase in the Ob receptor was also tested by staining with immunohistochemistry (Fig.  5 B, C, D). Glucose oxidation is believed to be elevated in endometriosis, potentially contributing to lesion growth [ 32 ]. Given that WD and the leptin pathway are also known to increase glucose metabolism [ 33 , 34 ], we examined whether glucose and lactate levels in serum were altered, serving as indicators of glucose oxidation levels. Glucose levels did not differ significantly between the CD and WD groups (6.8 ± 0.8 mmol/L vs. 7.6 ± 0.7 mmol/L, respectively; p  = 0.44). However, glucose levels were significantly lower in the CD + OSE group compared to the CD group (4.9 ± 0.9 vs. 6.8 ± 0.8 mmol/L, respectively; p  < 0.01) and similarly in the WD + OSE group versus the WD group (6.4 ± 1.4 mmol/L vs 7.6 ± 0.7 mmol/L, respectively; p  = 0.02) (Fig.  5 E). This reduction in glucose levels in mice with endometriosis could signify higher glucose consumption in these groups. Lactate production during glucose anaerobic metabolism was significantly elevated in the WD + OSE group compared to WD alone (11.7 ± 0.8 vs. 7.7 ± 1.0 mmol/L; p  < 0.0001). There was no significant difference between the CD + OSE and the CD groups (10.8 ± 1.9 vs. 9.3 ± 0.5 mmol/L, respectively; p  = 0.63) (Fig.  5 F). These findings suggest that mice in the WD + OSE group exhibited higher levels of glucose oxidation, aligning with leptin pathway activation and dietary influences. Fig. 5 Impact of Western diet on metabolic changes. A qRT-PCR assessment of Ob and Obr mRNA expression in lesions. B Percentage of stromal cells positive for Obr immunostaining. Representative sections of ectopic lesions stained with Obr antibody in the CD + OSE ( C ) and WD + OSE groups ( D ). Levels of glucose ( E ) and lactate ( F ) in mouse serum. CD group ( n  = 8), WD group ( n  = 7), CD + OSE group ( n  = 8), WD + OSE group ( n  = 5). ANOVA test with Bonferroni correction. * p  < 0.05, ** p  < 0.01, *** p  < 0.0001 Impact of Western diet on metabolic changes. A qRT-PCR assessment of Ob and Obr mRNA expression in lesions. B Percentage of stromal cells positive for Obr immunostaining. Representative sections of ectopic lesions stained with Obr antibody in the CD + OSE ( C ) and WD + OSE groups ( D ). Levels of glucose ( E ) and lactate ( F ) in mouse serum. CD group ( n  = 8), WD group ( n  = 7), CD + OSE group ( n  = 8), WD + OSE group ( n  = 5). ANOVA test with Bonferroni correction. * p  < 0.05, ** p  < 0.01, *** p  < 0.0001 We proceeded to investigate the composition of the intestinal microbiota using Illumina-based sequencing of the 16S rRNA gene. Principal coordinate analysis plots (PCoA) of the Bray–Curtis distance interestingly suggested an influence of endometriosis induction on the composition of the intestinal microbiota, both in mice fed a CD diet (Permanova p value = 0.013) (Fig.  6 A) and in mice fed a WD diet (Permanova p value = 0.004) (Fig.  6 B). The dietary changes distinctly altered the microbiota composition between the CD and WD groups, independently of the presence of endometriosis, as clearly illustrated in the third view of the PCoA (Fig.  6 C). Moreover, microbiota richness (alpha diversity), assessed through computation of the Shannon diversity index, was not affected by diet or endometriosis induction (Fig.  6 D). We then delved into the microbiota composition at various taxonomical levels (Fig.  6 G). We observed a major decrease or disappearance in A. muciniphila in the WD + OSE group (Fig.  6 E), whose relative abundance has been positively associated with a healthy state [ 35 , 36 ]. However, we did not find any differences among our groups concerning the Firmicutes/Bacteroidetes ratio, which is known to vary in cases of intestinal dysbiosis [ 37 , 38 ] (Fig.  6 F). Fig. 6 Impact of endometriosis on intestinal microbiota. A – C Principal coordinate analysis (PCoA) plots were used to assess the variation between the groups (beta diversity), analyzed by permutational multivariate analysis of variance. Representative visualizations of variations observed by the induction of endometriosis in both diet groups ( A and B ) or by the diet changes ( C ). D Shannon diversity index. E Relative abundance of A. muciniphila . F Firmicutes/Bacteroidetes ratio. G The 10 most abundant taxa (at genus level) among different study groups. CD group ( n  = 8), WD group ( n  = 7), CD + OSE group ( n  = 7), WD + OSE group ( n  = 5). ANOVA test with Bonferroni correction. NS, non-significant. * p % 0.05, ** p % 0.01, *** p % 0.001 Impact of endometriosis on intestinal microbiota. A – C Principal coordinate analysis (PCoA) plots were used to assess the variation between the groups (beta diversity), analyzed by permutational multivariate analysis of variance. Representative visualizations of variations observed by the induction of endometriosis in both diet groups ( A and B ) or by the diet changes ( C ). D Shannon diversity index. E Relative abundance of A. muciniphila . F Firmicutes/Bacteroidetes ratio. G The 10 most abundant taxa (at genus level) among different study groups. CD group ( n  = 8), WD group ( n  = 7), CD + OSE group ( n  = 7), WD + OSE group ( n  = 5). ANOVA test with Bonferroni correction. NS, non-significant. * p % 0.05, ** p % 0.01, *** p % 0.001

Endometriosis is a benign gynecological condition characterized by the presence of endometrial-like tissue outside the uterine cavity. It manifests with symptoms such as pelvic pain and infertility, significantly affecting the quality of life for affected women [ 1 ]. This condition is heterogenous, presenting with various phenotypes that correspond to a highly variable number and location of lesions [ 2 ]. The widely accepted implantation theory suggests that endometriosis stems from retrograde menstruation through the fallopian tubes into the abdominopelvic cavity [ 3 ]. While this phenomenon is common in cycling women, the prevalence of endometriosis stands at around 10%, indicating potential roles of individual [ 4 ] and environmental factors [ 5 ]. In humans, the development of a healthy gastrointestinal microbiota during early life is crucial for proper immune system maturation [ 6 ]. Newborns’ exposure to intestinal microbiota can modulate the risk of bacterial respiratory infections [ 7 ], indicating a strong cooperative relationship between the digestive microbiota and the immune system. This symbiosis is observed in many immuno-inflammatory diseases [ 8 ], notably influencing innate immunity and promoting inflammation. In endometriosis, the ectopic implantation of endometrial cells involves hormonal and immuno-inflammatory processes [ 9 ]. Previous studies have shown that immune-inflammatory changes may be associated with bacterial by-products or components that can modulate inflammation by either blocking or activating the memory response of macrophages, a phenomenon referred to as trained immunity [ 10 ]. The available data demonstrate alterations in the intestinal microbiota of women affected with stage III and IV endometriosis [ 11 , 12 ]. Notably, dietary modifications have shown potential in reducing pain associated with endometriosis [ 13 ], particularly through the adoption of anti-inflammatory diets [ 14 – 16 ]. These dietary changes may be linked to variations in the intestinal microbiota, potentially involving unidentified metabolic functions. Both diet and intestinal microbiota could be key factors in the pathophysiology of endometriosis [ 17 , 18 ]. The aim of this study was to investigate whether a Western diet, known for its proinflammatory properties, could affect the growth of endometriosis lesions and to explore the underlying mechanisms.

In this laboratory study, utilizing a mouse model of endometriosis, it was observed that endometriosis lesions exhibited twice the size in mice fed with a WD compared to those fed with a CD. Additionally, a significant diet-induced activation of the leptin pathway and macrophage activation within the lesions was detected. Concurrently, changes in the composition of the intestinal microbiota due to endometriosis induction were observed, characterized by a depletion of A. muciniphila in the intestinal microbiota of endometriosis-afflicted mice fed with a WD. The strength of our study lies in the novelty of the topic and methodological design: (i) To match the morphotype of endometriosis patients with a normal or low body mass index, we used a medium-fat die rather than a high-fat diet. In contrast, Heard et al. conducted their study using a WD (cited as “high fat”) [ 39 ]. We opted for a diet containing fiber and less fat from lard, favoring fat from soybean oil, as it seemed closer to a typical human diet [ 40 , 41 ]. Heard et al. employed an intraperitoneal endometrial cell injection model in mice, which does not allow for as precise a measurement as the longitudinal ultrasound method we used. However, our results are consistent with their findings, particularly regarding the higher number of intraperitoneal lesions in mice fed a WD. (ii) Analysis of the intestinal microbiota necessitates careful assessment of a wide range of inter-individual and temporal variabilities [ 42 , 43 ]. Studying it under strict dietary conditions, including continuation of the diet after induction surgery, together with the daily hormonal synchronization of the mice for 5 days prior to surgical implantation of the lesions, enhances the robustness and reproducibility of our analysis. Indeed, because mice do not spontaneously menstruate, donor mice must be hormonally treated to induce menstrual endometrium for transplantation [ 44 ]. Finally, to monitor changes in lesion size over time and to mitigate the impact of estrous cycles, we conducted longitudinal ultrasound monitoring. (iii) We have replicated the experiments, and our study groups include control groups for both the diet and the induction of endometriosis. The choice of the BALB/c strain over C57BL/6 facilitates the monitoring of single and cystic lesions by ultrasound. This immunocompetent strain, which tends to exhibit Th2/M2 dominance, demonstrates reduced sensitivity to the estrous cycle in terms of lesion development. Additionally, BALB/c mice have been shown to sustain lesion development over longer-term experiments, such as our dietary exposures [ 45 ]. This methodological approach strengthens the study and minimizes evaluation bias. To match the morphotype of endometriosis patients with a normal or low body mass index, we used a medium-fat die rather than a high-fat diet. In contrast, Heard et al. conducted their study using a WD (cited as “high fat”) [ 39 ]. We opted for a diet containing fiber and less fat from lard, favoring fat from soybean oil, as it seemed closer to a typical human diet [ 40 , 41 ]. Heard et al. employed an intraperitoneal endometrial cell injection model in mice, which does not allow for as precise a measurement as the longitudinal ultrasound method we used. However, our results are consistent with their findings, particularly regarding the higher number of intraperitoneal lesions in mice fed a WD. Analysis of the intestinal microbiota necessitates careful assessment of a wide range of inter-individual and temporal variabilities [ 42 , 43 ]. Studying it under strict dietary conditions, including continuation of the diet after induction surgery, together with the daily hormonal synchronization of the mice for 5 days prior to surgical implantation of the lesions, enhances the robustness and reproducibility of our analysis. Indeed, because mice do not spontaneously menstruate, donor mice must be hormonally treated to induce menstrual endometrium for transplantation [ 44 ]. Finally, to monitor changes in lesion size over time and to mitigate the impact of estrous cycles, we conducted longitudinal ultrasound monitoring. We have replicated the experiments, and our study groups include control groups for both the diet and the induction of endometriosis. The choice of the BALB/c strain over C57BL/6 facilitates the monitoring of single and cystic lesions by ultrasound. This immunocompetent strain, which tends to exhibit Th2/M2 dominance, demonstrates reduced sensitivity to the estrous cycle in terms of lesion development. Additionally, BALB/c mice have been shown to sustain lesion development over longer-term experiments, such as our dietary exposures [ 45 ]. This methodological approach strengthens the study and minimizes evaluation bias. However, the study does have minor limitations and/or biases: (i) Some of our experiments were conducted only once due to the challenge of obtaining large quantities of biological samples. However, modern techniques such as RT-qPCR and flow cytometry enable valid analyses to be performed even with small quantities. (ii) The amount of food ingested by the mice was not quantified, which could potentially lead to an overestimation of the effect of the WD. Nonetheless, the groups of mice were homogeneous in number and subjected to identical conditions, which should help stabilize variations in food intake. Regarding metabolic parameters, mice were not fasted prior to blood collection. Importantly, the BALB/c strain does not develop glucose intolerance when fed a high-fat diet, ensuring accurate assessment of metabolic effects. However, our study observed blood glucose levels akin to hypoglycemic conditions following fasting in this strain of mice [ 46 ]. Additionally, data suggests that mice consumed similar quantities of food without significant variations in their weights throughout the experiment. (iii) One limitation of our study is the lack of direct control over the estrous cycle, which could potentially influence lesion growth. A hypothesis could be that the WD could induce prolonged estrous or proestrus phases with estrogen dominance. However, based on the recent study by Skalski et al., which demonstrated no significant changes in the length of estrous phases when BALB/c mice were fed a WD, we believe the diet’s effect on hormonal cycling is minimal [ 47 ]. We used of a surgical induction model in non-ovariectomized mice, which closely mimics the human condition by preserving natural estrous cycles. This allows us to maintain the integrity of the hypothalamic-pituitary-ovarian axis and observe more physiologically relevant hormonal responses. However, we did not track the specific phase of the estrous cycle at the time of sacrifice, which could provide additional insights into the hormonal state during lesion analysis. (iv) We correlated the size of the lesions with the severity of endometriosis, although this may not necessarily be the sole indicator of severity. Endometriosis is a heterogenous disease, with various phenotypes described in humans, and there may not always be a clear correlation between symptoms and the anatomical distribution of lesions known in humans [ 48 ]. Some of our experiments were conducted only once due to the challenge of obtaining large quantities of biological samples. However, modern techniques such as RT-qPCR and flow cytometry enable valid analyses to be performed even with small quantities. The amount of food ingested by the mice was not quantified, which could potentially lead to an overestimation of the effect of the WD. Nonetheless, the groups of mice were homogeneous in number and subjected to identical conditions, which should help stabilize variations in food intake. Regarding metabolic parameters, mice were not fasted prior to blood collection. Importantly, the BALB/c strain does not develop glucose intolerance when fed a high-fat diet, ensuring accurate assessment of metabolic effects. However, our study observed blood glucose levels akin to hypoglycemic conditions following fasting in this strain of mice [ 46 ]. Additionally, data suggests that mice consumed similar quantities of food without significant variations in their weights throughout the experiment. One limitation of our study is the lack of direct control over the estrous cycle, which could potentially influence lesion growth. A hypothesis could be that the WD could induce prolonged estrous or proestrus phases with estrogen dominance. However, based on the recent study by Skalski et al., which demonstrated no significant changes in the length of estrous phases when BALB/c mice were fed a WD, we believe the diet’s effect on hormonal cycling is minimal [ 47 ]. We used of a surgical induction model in non-ovariectomized mice, which closely mimics the human condition by preserving natural estrous cycles. This allows us to maintain the integrity of the hypothalamic-pituitary-ovarian axis and observe more physiologically relevant hormonal responses. However, we did not track the specific phase of the estrous cycle at the time of sacrifice, which could provide additional insights into the hormonal state during lesion analysis. We correlated the size of the lesions with the severity of endometriosis, although this may not necessarily be the sole indicator of severity. Endometriosis is a heterogenous disease, with various phenotypes described in humans, and there may not always be a clear correlation between symptoms and the anatomical distribution of lesions known in humans [ 48 ]. The intricate relationship between diet and endometriosis presents promising avenues for further exploration. Many women with endometriosis suffer from functional bowel disorders [ 49 ], regardless of the presence of digestive endometriotic lesions [ 50 ]. Therefore, emerging research suggests that dietary modifications might alleviate these symptoms [ 51 , 52 ]. Previous studies have tested the administration of probiotics on a mouse model of endometriosis, resulting in a decreased size of the lesions [ 53 ]. However, data is still lacking on the metabolic shifts that could occur in endometriosis in response to dietary changes. WD is known to compromise gut barrier function, leading to increased intestinal permeability and the leakage of toxic bacterial metabolites into circulation, potentially contributing to the development of low-grade systemic inflammation [ 54 ]. This low-grade inflammation has already been demonstrated in endometriosis and could contribute to the more extensive development of lesions in our model [ 55 ]. A metabolic effect observed in the WD + OSE group is the stimulation of the leptin pathway, which is known to be involved in the migration and invasion of endometriotic cells [ 56 , 57 ]. Leptin is also recognized for its influence on glucose metabolism, exhibiting glucose-lowering effects [ 34 , 58 ]. Endometriosis cells rely on heightened glucose metabolism for rapid ectopic implantation and growth [ 59 ]. In vitro studies of endometriotic stromal and epithelial cells, validated by in vivo experiments on BALB/c mice, have shown increased glucose consumption and lactate production [ 60 ]. Our findings align with these observations, further illustrating an increase in glucose oxidation, characterized by heightened glucose consumption and lactate production in our WD + OSE group [ 61 ], supporting the role of leptins in lesion development. The Western diet significantly increased the size of endometriosis lesions. Drawing on well-established research regarding the impact of dietary changes on the digestive microbiota [ 62 ], our study reveals that endometriosis can independently alter intestinal microbiota regardless of diet. This underscores the role of intestinal microbiota in various inflammatory diseases, particularly in exacerbating immune responses and perpetuating chronic inflammation [ 63 , 64 ]. Using a mouse model of endometriosis where endometrial fragments are injected into the intraperitoneal cavity, Yuan et al. linked endometriosis with significant alterations in intestinal microbiota [ 65 ]. Chadchan et al. further expanded on this, demonstrating that lesion growth was attenuated by antibiotic treatment in a similar model [ 66 ]. They observed a reduction in lesion development following broad-spectrum antibiotic-induced microbiota depletion, which was reversible through fecal transplantation from mice with endometriosis [ 67 ]. Studies involving microbiota-depleted mice undergoing endometriosis induction and subsequent oral gavage of feces from mice without endometriosis or with endometriosis have been conducted. These studies demonstrated that mice receiving feces from mice with endometriosis developed as many and as large lesions as those with intact microbiota. These findings suggest a bidirectional relationship: induction of endometriosis modifies the microbiota, and conversely, specific microbiota compositions can accelerate lesion development. A hypothesis emerging from these findings revolves around the role of short-chain fatty acids produced by anaerobic fermentation of intestinal bacteria, potentially influencing the severity of lesions by promoting cellular proliferation and endometriotic lesion growth [ 67 ]. This theory finds support in a baboon model, which identified specific microbial shifts post-endometriosis induction, including a decrease in gram-negative and an increase in gram-positive bacilli [ 17 ]. Human studies have delved deeper, examining microbiota ratios linked to intestinal dysbiosis, reflected in an increased ratio of Firmicutes/Bacteroidetes, though such an observation was not evident in our model [ 68 ]. In our investigation of specific bacterial species variations among different groups, we noted a pronounced depletion of A. muciniphila in the intestinal microbiota of the WD + OSE group. This gram-negative bacteria, recognized for its reduced presence in metabolic syndrome [ 69 , 70 ], and inflammatory intestinal conditions like Crohn’s disease [ 35 ], has been shown to confer multiple health benefits when administered as a pasteurized probiotic, including mitigating low-grade intestinal inflammation and enhancing gut integrity [ 71 ]. Notably, the abundance of A. muciniphila remains stable in mice in the WD group but is completely depleted in those in the WD + OSE group. This suggests a potential relationship between its absence and increased severity of endometriosis, aligning with emerging evidence of A. muciniphila ’s protective role in health. Considering the recent study suggesting Fusobacterium as a causative agent in the endometrial cavity for endometriosis, the depletion of A. muciniphila in feces could indicate its role as one of the protective bacteria in the pathophysiology of endometriosis [ 72 ]. The exacerbation of endometriosis lesions, influenced by dietary factors and coupled with alterations in microbiota composition, is intricately linked to immune changes. These changes include the activation of macrophages within the lesions, a phenomenon extensively documented in endometriosis [ 73 ]. Interestingly, this macrophage activation appears to be more pronounced in the WD + OSE group compared to the CD + OSE group. We did not observe clear macrophage polarization towards a dominant M1 or M2 phenotype in our study. Several studies suggest a shift towards an M2 phenotype, although this remains contentious, especially across different anatomical sites. For instance, an M1 phenotype has been reported in eutopic endometrium [ 74 ], whereas a dominant M2 phenotype is observed in the peritoneal cavity [ 75 ] of mouse models of endometriosis. Surprisingly, it has also been documented that macrophage phenotype can vary with the stage of disease progression [ 75 ]. This disparity may originate from the concept of trained immunity, where macrophages are primed for a swifter and more robust inflammatory response. This phenomenon manifests when macrophages are exposed to bacillus Calmette-Guérin (BCG) in the context of endometriosis [ 10 ]. It is plausible that the WD, characterized by an increase in gram-positive bacilli, could mimic the effect of BCG, which is known to enhance lesion growth in endometriosis. Furthermore, the detrimental activation of trained immunity observed in mice fed a WD, particularly through the activation of the inflammasome, could also significantly contribute to this process [ 76 ]. Recently, A. muciniphila has been identified as playing a role in modulating local long-term innate immune responses, particularly in trained immunity, with potential anti-inflammatory effects [ 77 ]. Its absence in the microbiota of WD + OSE mice underscore its potential protective role in the context of endometriosis. This discovery not only deepens our understanding of how specific dietary choices can benefit patients but also opens up a promising new avenue for research. It would be pertinent to investigate whether the observed phenomena are reversible upon cessation of the medium-fat diet. Our experimental animal study should encourage human studies aimed at defining optimal diets for endometriosis patients and assessing the effects of transitioning to a healthier diet. Therefore, further comprehensive research is essential to unravel the intricate interplay between intestinal microbiota and endometriosis. In fact, the multifactorial nature of endometriosis links it to an elevated risk of irritable bowel syndrome, for which dietary interventions have shown efficacy [ 78 ]. Further investigation is needed into the disappearance of A. muciniphila observed in mice with the largest lesions, particularly regarding its potential protective effects, akin to those of a probiotic. Hence, it is crucial for studies to examine the intestinal microbiota as a potential biomarker for assessing the efficacy of probiotics or dietary modifications.

Our findings underscore a significant correlation between diet and endometriosis, elucidating substantial dimensions within the realms of immune response, metabolism, and bacterial composition. This study marks a pioneering endeavor in unraveling the impact of dietary patterns on endometriosis, providing valuable insights into the complex interplay between dietary habits and this condition.

Additional file 1. Supplemental Table 1. Composition of Western and control diets. Supplemental Table 2. Murine primers used for RT-qPCR analysis of lesions. Supplemental Figure 1. Flow cytometry gating strategy for macrophages and B lymphocytes in lesions. Additional file 1. Supplemental Table 1. Composition of Western and control diets. Supplemental Table 2. Murine primers used for RT-qPCR analysis of lesions. Supplemental Figure 1. Flow cytometry gating strategy for macrophages and B lymphocytes in lesions.