Multi-omics insights into gut microbiota-metabolite interactions under probiotic intervention in a developmental cafeteria diet model

Abstract

Background

The developmental phase is a pivotal biological period for the maturation of the gut microbiota and the establishment of lifelong metabolic health. During these period, dietary patterns that induce dysbiosis, such as the high-fat, low-fiber “cafeteria diet,” disrupt the production of key metabolites in the gut-metabolite axis, including short chain fatty acids (SCFAs) and indole-3-propionic acid (IPA). This study employs a multi-omics approach to examine the impact of cafeteria diet exposure during the developmental period (days 21–56) in 21-day-old male Wistar rats on microbiota composition, SCFA, and IPA levels, and to assess the extent to which concurrent probiotic administration can mitigate these disruptions.

Results

The cafeteria diet led to a marked reduction in alpha diversity indices (Shannon p = 0.021; Simpson p = 0.034) and altered the Firmicutes/Bacteroidetes ratio (p = 0.015). Beta diversity analysis indicated a distinct separation between groups (PERMANOVA p = 0.002). Metabolite analysis revealed significant reductions in acetic acid (p = 0.004), isobutyric acid (p = 0.094), butyric acid (p = 0.0014), valeric acid (p = 0.0001), heptanoic acid (p = 0.0125), and IPA (p = 0.002), whereas probiotic administration largely restored these levels. At the species level, cafeteria diet markedly increased Segatella copri, while probiotic intervention partially restored beneficial taxa such as Faecalibacterium prausnitzii and butyrate-producing genera (Anaerostipes hadrus, Intestinimonas butyriciproducens, Blautia wexlerae, and Flintibacter sp. KGMB00164), as evidenced primarily by shotgun metagenomics. Correlation analysis further revealed strong positive associations between butyrate and F. prausnitzii (ρ = 0.65, p = 0.003) and between IPA and B. longum (ρ = 0.68, p = 0.002). Collectively, these results highlight the protective role of probiotic intervention against diet-induced dysbiosis by reinforcing microbiota metabolite interactions.

Conclusions

By integrating metagenomic and metabolomic analyses, this multi-omics study demonstrates that exposure to a high-fat cafeteria diet during the developmental period disrupts microbiota composition and metabolite production, whereas concurrent probiotic administration largely prevent these effects, serving a protective role in the gut-metabolite axis. The study underscores the potential of early-life probiotic intervention, supports SCFA and IPA production, as a critical strategy to optimize microbiota-metabolite interactions and promote long-term gut and systemic health.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-026-12650-w.

Background

The developmental period constitutes a pivotal biological phase during which nutritional exposures can exert enduring effects on long-term metabolic health and gut microbiota composition [1]. Within this framework, the cafeteria (CAF) diet model characterized by its high energy density, palatability, and similarity to a Western-type diet is extensively employed in rodent studies to replicate diet-related metabolic disorders [2–4]. This model effectively induces features of metabolic syndrome, including increased adiposity, hepatic steatosis, and insulin resistance [5]. Nutritional disturbances during the developmental period, due to the high plasticity of not yet fully matured physiological systems, can have more profound and lasting effects on gut homeostasis, epithelial barrier function, and microbial ecology compared to adulthood [6]. Current evidence suggests that microbial alterations during this period may be linked to low-grade systemic inflammation and oxidative stress, contributing to the early onset of metabolic disorders [7, 8].

Probiotics, defined as live microorganisms that, when administered in adequate amounts, confer a health benefit on the host, are increasingly being investigated for their potential to mitigate the adverse effects caused by CAF-like diets [9]. Specific probiotic strains such as Lactobacillus plantarum, L. rhamnosus, and Bifidobacterium breve have been shown to enhance the intestinal epithelial barrier, reduce the production of pro-inflammatory cytokines, and regulate both microbial diversity and SCFA production [10, 11]. The efficacy of probiotic interventions during the developmental period may be particularly pronounced, as the gut ecosystem is more receptive to probiotic modulation at this time [12, 13]. However, in models using CAF diets administered during the developmental period, the protective or corrective effects of probiotics not only at the level of microbial composition but also in terms of functionality (e.g., SCFA production, indole propionic acid synthesis) have not been adequately elucidated [14–17].

Recent advancements in high-resolution molecular techniques have significantly transformed microbiome research by elucidating the structural and functional attributes of the gut microbiota. Amplicon sequencing, particularly utilizing the 16S rRNA gene, has been extensively employed to ascertain the taxonomic composition of bacterial communities. This method is favored for its cost-effectiveness, simplicity, and well-established validation [18]. Nevertheless, it typically offers taxonomic resolution primarily at the genus level, while species-level annotations derived from short-read regions such as V3–V4 should be interpreted cautiously [19, 20]. Conversely, shotgun metagenomic sequencing captures a representative fraction of microbial DNA, enabling species-level taxonomic profiling and functional characterization at the metabolic pathway level [21, 22]. Even shotgun sequencing doesn’t tell us what DNA is actually being expressed (would need to do RNA sequencing or metatranscriptomics to know that) but it certainly does offer more in that regard than 16s sequencing. This high-resolution approach provides substantial advantages, particularly in assessing the impact of environmental factors such as diet, aging, antibiotic use, or probiotic interventions on microbial functionality [23, 24]. Moreover, shotgun data permits the prediction of the production capacity of microbiota-derived metabolites, such as SCFA and indole derivatives, allowing for a more comprehensive analysis of their effects on host physiology [25, 26]. Consequently, both methods possess distinct strengths and serve complementary roles in studies addressing both compositional and functional aspects of microbial ecology.

Given the critical sensitivity of the gut ecosystem during the developmental window and the potential of dietary interventions to exert long-lasting impacts, this study was designed to test the hypothesis that early-life probiotic supplementation can mitigate the deleterious effects of a cafeteria diet on both microbial composition and functionality. Specifically, we postulated that probiotics administered during this vulnerable period would not only restore microbial diversity and structural integrity but also normalize levels of key microbial metabolites namely SCFAs and IPA which are essential modulators of gut-liver-brain axis homeostasis. By integrating 16S rDNA and shotgun metagenomic sequencing with quantitative metabolite profiling, this study aims to provide a multi-layered and mechanistic understanding of how probiotics can attenuate cafeteria diet induced dysbiosis during the critical developmental period in male Wistar rats.

Methods

Experimental setup

Male Wistar rats were used as the model organism. Twenty-one-day-old rats, weaned from milk, were divided into four groups of seven each: control group (Cnt: n = 7), SCD Probiotics group (Prb: n = 7), cafeteria diet group (Cd: n = 7), and cafeteria diet with SCD Probiotics supplementation group (CdPrb: n = 7). The treatments continued until day 56, considered the end of the development period. The SCD Probiotics supplement was given by oral gavage at a dose of 1.5 mL (1 × 108 CFU) per day [4, 5]. The study utilized a product marketed by the SCD Probiotics company, Liquid Probiotic Supplement (Essential Probiotics XI − 500 ml H.S. Code: 2206.00.7000). SCD Probiotics contains Bacillus subtilis, Bifidobacterium bifidum, Bifidobacterium longum, Lactobacillus acidophilus, Lactobacillus bulgaricus, Lactobacillus casei, Lactobacillus fermentum, Lactobacillus plantarum, Lactococcus lactis, Saccharomyces cerevisiae, and Streptococcus thermophilus species [27]. Other ingredients: Liquid fermentation product of water, organic sugarcane molasses*, organic juice concentrates* (blueberry, cherry, pomegranate). *Certified Organic Ingredients. Animals were fed ad libitum with a standard rodent diet. The detailed composition of the cafeteria diet has been previously described in our earlier studies using the same experimental design and animal cohort [5], and this diet was provided in addition to standard ad libitum feeding. Throughout the experiment, the animals’ weights, weekly food consumption, and measurements of the cafeteria diet content were recorded, as demonstrated in earlier investigations [4, 5]. Before starting the experiment, weight measurements of all rats were taken to ensure that the average weights of the rats in each cage were close. Groups on the cafeteria diet were also given the same cafeteria diet products. All cafeteria diet content was obtained from supermarkets close to the place where the study was conducted [4]. The animals in the control and experimental groups were lightly anesthetized by the ether treatment and sacrificed one day after the application ended. The cecum tissues with the content were extracted, quickly frozen on dry ice, and stored in a -80 °C deep freezer until the time of analysis. All animals were housed following standard animal care protocols. Rats in each group were kept in separate cages, and co-housed rats remained in the same group. They were housed in clear Plexiglas cages (7 rats/cage) under a 12-hour light/dark cycle at a constant 21℃ temperature. No rats in the control and experimental groups died or were excluded from the study for any reason. This study was carried out with the approval of the Ethics Committee (meeting date: 29.06.2021, approval number: 2021/03) from the Bingöl University Animal Experiments Local Ethics Committee.

Metagenomic analysis

Genomic DNA isolation

For the isolation of genomic DNA, the FitoMicrobiom DNA Isolation Kit (Cat. No.: DE01A050) was employed. The procedure was conducted as follows: A sample weighing 250 mg was placed into a Beading Tube, to which 600 µl of Beading Buffer was added. The tubes were then subjected to vortex mixing at maximum speed for 15 min. Following vortexing, the Beading Tubes were centrifuged at 10,000 × g for 1 min. Subsequently, 400 µl of the supernatant was transferred to a clean tube, and 400 µl of “Lysis Buffer” was added, followed by 10-minute incubation at room temperature. The mixture was then loaded onto a silica column and centrifuged at 10,000 × g for 1 min. After discarding the flow-through, 500 µl of “Wash Buffer” was added to the silica column and centrifuged at 10,000 × g for 1 min. This step was repeated with an additional 500 µl of “Wash Buffer.” The silica column was then centrifuged empty at 15,000 × g for 1 min. Following centrifugation, the silica column was transferred to a clean 1.5 ml microcentrifuge tube, and 50 µl of DNA Elution Buffer was applied to the column matrix, followed by centrifugation at 15,000 × g for 1 min. The concentrations of the isolated genomic DNAs were determined fluorometrically using a Qubit 3.0 Fluorometerand subsequently analyzed on a 1% agarose gel [28].

Shotgun genome sequencing

Library preparation and whole genome sequencing

Sequencing was conducted utilizing the “Qiagen FX DNA Library Kit” (Cat no: 180475) on the Illumina NovaSeq X Plus sequencing platform (Illumina, San Diego, CA, USA). The library preparation process comprises five stages: (i) genomic DNA tagmentation, (ii) purification of tagmented DNA, (iii) amplification of tagmented DNA, (iv) library purification, and (v) library quality control and normalization. Initially, the genomic DNA was fragmented, and adaptor sequences were appended. For this procedure, the reaction mixture was centrifuged at 280 × g for 1 min and placed in a pre-programmed thermal cycler (preheated lid: 55 °C; incubation at 55 °C for 5 min; hold at 10 °C). Subsequently, the labeled DNA sequences were purified from the transposomes. The labeled DNA was then amplified, and index and sequencing adapters (Index 1 (i7) and Index 2 (i5)) were incorporated. The reaction mixture was centrifuged at 280 × g for 1 min and incubated in a thermal cycler (preheated lid: 100 °C; 72 °C for 3 min; 98 °C for 30 s; five cycles of 98 °C for 10 s, 63 °C for 30 s, and 72 °C for 3 min; and hold at 10 °C). The prepared libraries underwent purification, and short library fragments were eliminated. The libraries were subjected to qualitative and quantitative control tests, followed by normalization. Library quantification was performed fluorometrically using Qubit, and quality control was executed using the Agilent Technology 2100 Bioanalyzer [28]. Shotgun metagenomic sequencing was performed on pooled cecal samples for each experimental group.

High-throughput Illumina NGS sequencing

The sequencing library was prepared, and the genomic DNA, intended for sequencing as 2 × 150 bp paired-end (PE) on the Illumina sequencing platform, was initially assessed and processed via Local Run Manager (LRM) in accordance with the NGS system manufacturer’s guidelines. At this juncture, NovaSeq X Plus (10B) and two flow cells were employed. For deep sequencing, the read depth was established at > 400x. Following the sequencing process, the resulting reads were converted to the “.fastq” format, and quality control of the raw sequences was conducted. For this purpose, the generated NGS data were obtained as FASTQ format files, and for each strain, a demultiplexed raw read file was produced. These NGS files underwent FASTQ quality control (FASTQC, (FASTQC (Andrews, Simon. “FastQC: a quality control tool for high throughput sequence data.” (2010)), and adapter regions were trimmed using Trimmomatic v0.32 (Bolger, (A) M., Lohse, M., & Usadel, (B) Bioinformatics, btu170 (2014)). The demultiplexed raw reads were filtered from low-quality data using CLC Genomics Workbench (Qiagen, US). Adapter sequences, as well as low-quality and contaminated reads post-sequencing, were excised from the raw NGS reads. All obtained genome files were mapped using the “map-based” whole genome approach with BWA (Burrows-Wheeler Aligner), utilizing reference genome data and annotation files from the NCBI database [29]. Fast and accurate short read alignment with Burrows-Wheeler Transform. Bioinformatics, 25:1754-60). The data obtained after shotgun DNA sequencing were assembled into a metagenome using applications such as MEGAHIT (SOAPdenovo2) and meta-SPAdes, and for comparative analyses, EggNOG analysis and PfamScan tools were employed [28]. The obtained bioinformatics data will be submitted to the NCBI database. The resulting bioinformatics data derived from shotgun metagenomic sequencing and 16S rRNA gene sequencing have been deposited in the NCBI database under BioProject accession numbers PRJNA1289129 and PRJNA1288408, respectively.

16S rDNA gene sequencing

Amplification of the V3–V4 region of 16S rDNA

The V3–V4 regions of the 16S rDNA gene, used for taxonomic profiling, were amplified utilizing the universal 341 F 805R primer sequences and a SimpliAmp Thermal Cycler. The primer sequences and PCR conditions employed are detailed below (Ceylani and Teker 2022). Primers: 341 F: CCTACGGGNGGCWGCAG, 805R: GACTACHVGGGTATCTAATCC. PCR conditions: initial denaturation at 95 °C for 10 min (HS enzyme was used), followed by 35 cycles of 95 °C for 45 s for denaturation, 50–55 °C for 45 s for annealing, 72 °C for 60 s for extension, and a final extension at 72 °C for 3 min, with the temperature subsequently reduced to 4 °C to conclude the PCR process, as demonstrated in earlier investigations [30].

Library preparation and sequencing

For the 16s rDNA V3-V4 amplicon products, library preparation was conducted utilizing Illumina’s “Nextera XT DNA Library Prep Kit, Cat. No.: FC-131-1096,” with indexing performed using the “TG Nextera XT Index Kit v2 Set A (96 Indices, 384 Samples), Cat. No.: TG-131-2001.” PCR purification steps were executed with “AMPure XP beads” from Beckman Coulter. Sequencing was carried out on Illumina’s Miseq platform as paired-end (PE) 2 × 150 bp reads. A minimum of ≥ 30,000 reads was obtained per sample, as demonstrated in earlier investigations [31]. Metagenome sequencing was performed at Ficus Biotechnology (FicusBio), Ankara, Turkey.

Bioinformatic analysis of raw data

FastQC v0.10.1 was employed to conduct quality assessments on the raw sequence data (FastQ) and to trim the data as necessary to enhance the accuracy of the microbial diversity estimation and to eliminate sequencing artifacts, including low-quality and contaminated reads. The clustering of sequencing data into Operational Taxonomic Unit (OTU) groupings was executed using the Kraken Metagenomic System [32]. Heatmaps were generated utilizing GraphPad Prism 9.0.1 (GraphPad Software, USA). The resulting bioinformatics data derived from 16S rDNA sequencing have been deposited in the NCBI database under BioProject accession number PRJNA1288408 [33].

Alpha diversity indexes

At the species level, alpha diversity indices were calculated. Prior to diversity calculations, a rarefaction procedure was applied independently to each sample in order to account for differences in sequencing depth. Shannon’s index ranges between 1.5 and 3.5, with higher values indicating greater species richness. The Simpson index was calculated based on OTU abundance and uniformity. Simpson’s diversity index (1-D) ranges from 0 to 1, with 1 signifying ultimate evenness [34].

Beta diversity analysis

To assess the variations in microbial communities across different groups, a beta diversity analysis was conducted. Utilizing 16S rRNA gene sequencing data, a Bray–Curtis dissimilarity matrix was constructed, followed by the application of Principal Coordinates Analysis (PCoA) based on these distance metrics. The analyses were executed using the vegan package in R software, and statistical differences between groups were evaluated using the PERMANOVA (Permutational Multivariate Analysis of Variance) test. Graphical representations were prepared to depict the spatial distribution of samples on a two-dimensional plane according to microbial composition similarities, thereby illustrating clustering tendencies among groups.

Determination of Short-Chain Fatty Acid (SCFA) profiles from fecal samples

SCFA quantification was performed using gas chromatography coupled with flame ionization detection (GC-FID) according to the protocol described by Smith et al. (2023), with minor modifications for sample preparation and injection parameters [35]. Chromatographic separation was carried out on a Supelco SPB-FFAP capillary column (30 m × 0.32 mm × 0.25 μm; catalog no: 24107). The oven temperature was initially held at 100 °C for 0.5 min, increased at 8 °C/min to 180 °C with a 1.0 min hold, and then ramped at 20 °C/min to 200 °C, followed by a final hold of 5.0 min. The injector temperature was set to 185 °C, and injections were performed in split mode (1:10) with a sample volume of 1 µL. The carrier gas was helium at a constant flow rate of 1.0 mL/min, and the FID detector was used under standard operational parameters. Fecal SCFA extraction and derivatization were performed in alignment with the referenced method. Calibration and quantification were achieved using a VFA standard mixture (catalog no: 46975-U, Supelco) including acetic, propionic, isobutyric, butyric, isovaleric, valeric, caproic, isocaproic, and heptanoic acids.

Indole-3-Propionic Acid (IPA) analysis

Quantitative analysis of indole-3-propionic acid was performed using liquid chromatography–tandem mass spectrometry (LC-MS/MS) in accordance with the protocol described by Zeng et al., (2019), with appropriate modifications to sample processing and instrument parameters [36]. For chromatographic separation, a Shimadzu LCMS-8030 Triple Quadrupole Mass Spectrometer equipped with an Inertsil SIL 100 A column (4.6 × 250 mm, 5 μm; GL Sciences) was employed. The mobile phase consisted of 50% methanol / 50% deionized water containing 0.2% formic acid, delivered at an isocratic flow rate of 0.3 mL/min. The injection volume was 5 µL, and analysis was performed in positive MRM mode. The precursor ion (m/z) was 189, and the quantifier product ion (m/z) was 130, with collision energy (CE) of − 19 eV. The nebulizing gas flow was set to 3 L/min. A calibration curve was constructed using analytical-grade 3-indolepropionic acid standard (Merck, Cat. No: 220027-10G; Batch: 102607131). Instrument calibration and signal optimization were performed prior to sample analysis. Sample preparation was carried out as follows: 50 mg of fecal sample was suspended in 450 µL methanol: deionized water (1:1, v/v) and vortexed for 5 min. The mixture was centrifuged at 12,000×g for 10 min at 4 °C, and the supernatant was collected and subjected to a second identical centrifugation step. The final supernatant was filtered through a 0.22 μm membrane and transferred into autosampler vials for analysis.

Statistical analyses

All statistical analyses were executed utilizing GraphPad Prism version 10.0.5 (GraphPad Software, USA) and R software version 4.5.1. Comparisons were made across four experimental groups in all analyses. Initially, a one-way analysis of variance (ANOVA) was employed to evaluate overall differences at the group level. When the ANOVA results indicated significance (p < 0.05), Tukey’s post hoc test was conducted to identify specific pairwise differences. In instances where the global ANOVA did not achieve statistical significance, but biologically meaningful contrasts were hypothesized a priori, unpaired two-tailed Student’s t-tests were additionally performed. Beta diversity analyses (Fig. 2A–B) were conducted in R using Bray–Curtis dissimilarity metrics, visualized through principal coordinates analysis (PCoA), and assessed for significance using permutational multivariate analysis of variance (PERMANOVA, 999 permutations). For correlation analyses between SCFAs, IPA, and microbial taxa, Z-scores were initially calculated in GraphPad Prism, followed by the computation of Spearman’s rank correlation coefficients (ρ). The resulting p-values were adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR) method, with adjusted p < 0.05 considered significant. Data are presented as mean ± standard error of the mean (SEM). Statistical significance thresholds were defined as follows: p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****).Given the predefined group-based experimental design and the absence of a fully balanced factorial structure, statistical analyses were intentionally conducted using one-way ANOVA rather than factorial interaction models.

Fig. 2.

Beta diversity analysis of gut microbiota based on 16S rDNA V3–V4 and shotgun metagenomic data. Principal Coordinates Analysis (PCoA) plots were generated using Bray–Curtis dissimilarities in R. Panel A shows 16 S rDNA profiles, where groups displayed distinct clustering with significant differences confirmed by PERMANOVA (R² = 0.42, p = 0.001), indicating strong alterations in microbiota composition under cafeteria diet and a partial restorative effect of probiotic supplementation. Panel B represents pooled shotgun metagenomic profiles, where group separation along PC1 (83.9% of variance) further supports the regulatory influence of probiotics on microbial community structure despite the absence of statistical testing due to pooling

Results

General experimental outcomes

The effects of the cafeteria diet, SCD Probiotics supplementation, and SCD Probiotics supplementation during the cafeteria diet on body weight, standard chow intake, and cafeteria diet consumption have been comprehensively reported in our previously published studies using the same experimental design and animal cohort and are therefore not repeated here [4].

Comparative analysis of alpha diversity indices and F/B ratios from 16S and shotgun metagenomic data

In the analysis of the 16S rDNA V3–V4 region, the Shannon index revealed a statistically significant difference between groups (ANOVA, F = 11.85; p = 0.0001) (Fig. 1A). The Tukey test indicated that diversity in the Cd group was significantly reduced compared to the control group (p = 0.0033), while it remained comparable to the control level in the group administered probiotics (Cnt vs. Prb: p = 0.3346). Although the CdPrb group did not exhibit a significant difference from the control (p = 0.3027), it demonstrated lower diversity compared to the probiotic group alone (p = 0.0090). In the shotgun analysis, the highest Shannon index value was observed in the Prb group (6.275). This observation provides a complementary overview of group-level microbial diversity patterns in the shotgun dataset.

Fig. 1.

Effects of cafeteria diet and SCD probiotics supplementation on microbial diversity and phylum-level balance. Panels represent (A) Shannon diversity index (H), (B) Simpson diversity index (1–D), and (C) Firmicutes-to-Bacteroidota (F/B) ratio in cecal microbiota across four experimental groups: Cnt (control), Cd (cafeteria diet), Prb (SCD probiotics), and CdtPrb (SCD probiotics during cafeteria diet). Microbial profiling was performed using 16 S rDNA V3–V4 amplicon sequencing. Statistical evaluation was conducted using both one-way ANOVA for overall group effects and unpaired t-tests for specific pairwise comparisons. Data are presented as mean ± standard error of the mean (SEM). Significance is denoted as follows: p < 0.05 (*), p < 0.01 (**), p < 0.0001 (****); non-significant results are marked as “ns”

The Simpson index (1–D) also exhibited a similar trend (ANOVA, F = 6.806; p = 0.0020) (Fig. 1B). Diversity in the Cd group was significantly lower compared to both the control and probiotic groups (Cnt vs. Cd: p = 0.0087; Cd vs. Prb: p = 0.0015). The CdPrb group showed higher diversity compared to the Cd group, but did not differ from the control and probiotic groups. Shotgun data revealed a similar group-level distribution of Simpson index values across experimental groups, with the lowest value recorded in the Cd group (0.9796) and the highest in the Prb group (0.9893).

The Firmicutes/Bacteroidetes (F/B) ratio also varied among the groups (ANOVA, F = 3.186; p = 0.0488) (Fig. 1C). The ratio in the probiotic group was significantly higher than in the Cd group (p = 0.0067). Additionally, the difference between the Cd and CdPrb groups was significant (p = 0.0047). The control group did not show a statistical difference from the other groups. In the shotgun analysis, the highest F/B ratio was recorded in the Prb group (2.07), and the lowest in the Cd group (1.57). These findings indicate that probiotic supplementation was associated with a partial modulation of the microbiota imbalance induced by the cafeteria diet.

Beta diversity findings

The Principal Coordinates Analysis (PCoA) conducted using the Bray–Curtis distance based on the 16S rDNA V3–V4 region (Fig. 2A) demonstrated a clear differentiation in microbial composition among the groups. Specifically, the Cnt and Prb groups were closely clustered, whereas the Cd group was distinctly separated from these clusters. The CdPrb group occupied an intermediate position between the Cd and Cnt/Prb clusters. The PERMANOVA analysis confirmed that these group differences were statistically significant (R² = 0.42, p = 0.001), indicating that the experimental interventions exerted a substantial impact on the overall structure of the gut microbiota. Beta diversity analysis utilizing shotgun metagenomic data (Fig. 2B) similarly revealed a comparable group separation in pooled samples. In this analysis, the Cnt and Prb groups were also located in close proximity, while the Cd and CdPrb groups diverged along different axes. Notably, the PC1 axis accounted for 83.9% of the variance, representing the primary direction of change in microbial community structure. When these findings are considered collectively, both the 16S rDNA and shotgun metagenomic approaches indicate that a diet high in fat and sugar (cafeteria diet) significantly disrupts the gut microbiota, whereas probiotic intervention partially reshapes the community structure.

Comparative genus-level microbial composition based on 16S rDNA and shotgun metagenomics

Upon comparing the relative abundances of the ten dominant bacterial genera in the experimental groups using both 16S rDNA (V3–V4) region-based amplicon sequencing and shotgun metagenomic sequencing data, significant compositional differences were observed, attributable to the distinct analytical depth and taxonomic resolution of each method (Table 1). The 16S rDNA data particularly emphasized the presence of genera such as Streptomyces, Thermococcus, and Prevotella, whereas shotgun analyses revealed the predominance of genera such as Bacteroides, Segatella, Blautia, and Faecalibacterium. This discrepancy can be ascribed to the primer bias and limited taxonomic discriminability inherent in amplicon-based methods. Nonetheless, for biological interpretation, it is more pertinent to examine the distribution between groups within the same method. In this regard, the genus Segatella exhibited a relative increase in the diet intervention groups in both analyses, suggesting its potential as a diet-sensitive microbial biomarker. Conversely, in the shotgun analysis, the genus Bacteroides maintained relative stability in both the control and probiotic-administered groups, suggesting a more stable host–microbe relationship under probiotic supplementation. Therefore, although complete taxonomic overlap between the methods is limited, both approaches provide valuable insights into the distribution of dominant microbial genera among experimental groups and particularly elucidate the effects of dietary and probiotic interventions on microbiota composition. Figure 3A-B compares the ten most prevalent bacterial genuses in the control group with their prevalence in other groups.

Table 1.

Top 10 most abundant bacterial genus in each group based on relative abundance (%) obtained from 16S rDNA (V3–V4 region) amplicon sequencing and shotgun metagenomic sequencing. Cnt (control), cd (cafeteria diet), Prb (SCD probiotics), and cdtPrb (SCD probiotics during cafeteria diet)

16 S rDNA (V3–V4) relative abundance
	Cnt	%	Cd	%	Prb	%	CdPrb	%
1	Streptomyces	16,11	Thermococcus	14,94	Salinilacihabitans	5,34	Methanocorpusculum	14,43
2	Ligilactobacillus	6,46	Phycisphaera	10,46	Marinicauda	5,1	Peptostreptococcus	10,14
3	Thermococcus	4,44	Segatella	5,58	Bythopirellula	5,04	Nitratireductor	4,26
4	Prevotella	3,98	Cystobacter	4,84	Acidipropionibacterium	4,66	Hankyongella	3,81
5	Lactobacillus	3,89	Streptomyces	4,79	Hankyongella	4,46	Cloacibacterium	3,28
6	Acidovorax	2,78	Pseudoalteromonas	4,3	Providencia	3,87	Friedmanniella	3,11
7	Segatella	2,75	Staphylococcus	3,02	Salipiger	3,14	Alteromonas	2,82
8	Ruminococcus	2,75	Prevotella	2,82	Friedmanniella	3,14	Paraflavitalea	2,49
9	Roseburia	2,55	Blautia	2,64	Halobacteroides	3,01	Faecalimonas	2,01
10	Pseudoalteromonas	2,08	Borreliella	2,57	Geminocystis	2,43	Marinicauda	1,96

Shotgun metagenomic relative abundance
1	Bacteroides	0,83	Segatella	13,07	Segatella	8,06	Segatella	11,38
2	Segatella	6,27	Bacteroides	5,56	Bacteroides	5,79	Bacteroides	7,75
3	Prevotella	5,21	Prevotella	5,45	Blautia	5,17	Blautia	5,88
4	Flavonifractor	3,52	Blautia	5,25	Prevotella	5,03	Prevotella	4,22
5	Blautia	3,19	Roseburia	3,03	Flavonifractor	2,78	Faecalibacterium	3,28
6	Lactobacillus	2,95	Flavonifractor	2,43	Lactobacillus	2,57	Mediterraneibacter	3,03
7	Lawsonibacter	2,92	Lawsonibacter	2,17	Roseburia	2,4	Flavonifractor	2,57
8	Faecalibacterium	2,16	Faecalibacterium	2,16	Lawsonibacter	2,35	Roseburia	2,16
9	Roseburia	2,16	Mediterraneibacter	2,08	Faecalibacterium	2	Lawsonibacter	2,15
10	Ligilactobacillus	1,71	Collinsella	1,59	Mediterraneibacter	1,96	Parabacteroides	1,86

Fig. 3.

Heatmap representation of the relative abundance of the top ten most dominant bacterial genus detected through (A) 16S rDNA amplicon and (B) shotgun metagenomic sequencing across experimental groups. The analysis compares the microbial community structure of the the Cnt (control) with that of, Cd (cafeteria diet), Prb (SCD probiotics only) and CdtPrb group (rats receiving SCD probiotics during cafeteria diet exposure)groups. Genus-level taxonomic profiles were derived using metagenomic sequencing pipelines, and the resulting heatmaps illustrate compositional shifts in the dominant taxa. This figure enables a comparative assessment of how probiotic supplementation during cafeteria diet feeding modulates gut microbiota composition, highlighting taxa that are selectively enriched or suppressed across conditions

Comparative putative species-level microbial composition based on 16S rDNA and shotgun metagenomics

In this study, 16S rDNA amplicon sequencing and shotgun metagenomic analyses were employed to investigate the effects of dietary and probiotic interventions on gut microbiota at the species level. Both methods evaluated the top 20 most dominant species in each group, thereby comprehensively revealing quantitative changes in diversity and compositional differences.

In the context of 16S rRNA gene (V3–V4) amplicon sequencing, species-level annotations represent database-dependent, indicative assignments rather than definitive taxonomic identifications. Accordingly, 16S-derived species-level profiles are presented here for descriptive and comparative purposes only and should be interpreted with appropriate caution. In the 16S rDNA analyses, the emergence of Thermococcus sp. M36 as the most prevalent species in the Cd group, reaching 17.76%, clearly demonstrated the selective pressure exerted by an energy-dense and inflammatory diet (Table 2). The similarly high prevalence of this species in the CdPrb group (18.35%) suggests that probiotic administration was limited in mitigating the dominance of certain species. While the total prevalence of the top five species in the Cd group increased to 46.3%, this rate remained at 21.8% in the control group, indicating a narrowed and distinctly dysbiotic structure favoring a few species in the Cd group. Figure 4A compares the twenty most prevalent bacterial species in the control group, as determined by 16S rDNA V3–V4 region-based sequencing, with their relative abundances in the other experimental groups.

Table 2.

Top 20 most abundant bacterial taxa at the putative species level in each group based on relative abundance (%) obtained from 16S rDNA (V3–V4 region) amplicon sequencing. Cnt (control), cd (cafeteria diet), Prb (SCD probiotics), and cdtPrb (SCD probiotics during cafeteria diet)

	Cnt	%	Cd	%	Prb	%	CdPrb	%
1	Thermococcus sp. M36	6,23	Thermococcus sp. M36	17,76	Thalassovita gelatinovora	4,49	Thermococcus sp. M36	18,35
2	Acidovorax sp. BLS4	4,19	Phycisphaera mikurensis	12,58	Helicobacter pylori	4,24	Pseudoalteromonas sp. 3J6	5,37
3	Ligilactobacillus faecis	3,38	Cystobacter fuscus	5,83	Acidovorax sp. BLS4	2,93	Streptomyces sp. NBC_00536	3,65
4	Pseudoalteromonas sp. 3J6	3,13	Pseudoalteromonas sp. 3J6	5,16	Xylanibacter ruminicola	2,9	Staphylococcus aureus	3,39
5	Segatella copri	2,89	Segatella copri	4,97	Cutibacterium acnes	2,79	Borreliella garinii	3,17
6	Prevotella dentalis	2,79	Borreliella garinii	3,09	Blautia wexlerae	2,38	Segatella copri	2,96
7	Ruminococcus champanellensis	2,39	Staphylococcus aureus	3,08	Syntrophobotulus glycolicus	2,35	Dyella thiooxydans	1,55
8	Salmonella enterica	2,36	Roseburia hominis	1,4	Bacillus amyloliquefaciens	2,04	Blautia obeum	1,54
9	Qiania dongpingensis	2,24	Prevotella dentalis	1,38	Segatella copri	2,01	Paraprevotella clara	1,5
10	Anaerostipes hadrus	2,13	Paludibacter propionicigenes	1,18	Myxococcus xanthus	1,64	Roseburia hominis	1,41
11	Klebsiella michiganensis	2,04	Pontibacillus sp. ALD_SL1	1,18	Acidiplasma cupricumulans	1,51	Anaerostipes hadrus	1,4
12	Roseburia hominis	1,9	Paraprevotella clara	1,14	Ligilactobacillus faecis	1,37	Ruminococcus champanellensis	1,29
13	Flintibacter sp. KGMB00164	1,62	Klebsiella michiganensis	1,11	Qipengyuania psychrotolerans	1,33	Spirosoma endbachense	1,26
14	Taurinivorans muris	1,32	Enterococcus faecalis	0,97	Buchnera aphidicola	1,31	Enterococcus faecalis	1,21
15	Lawsonibacter asaccharolyticus	1,31	Spirosoma endbachense	0,89	Halorubrum sp. SS7	1,3	Halorubrum sp. SS7	1,18
16	Paraprevotella clara	1,3	Brevibacillus marinus	0,84	Flavobacterium sp. MDT1-60	1,17	Escherichia coli	1,03
17	Enterococcus faecalis	1,29	Ruminococcus champanellensis	0,78	Klebsiella pneumoniae	1,14	Sorangium cellulosum	1,01
18	Clostridium symbiosum	1,05	Clostridium symbiosum	0,74	Cellulophaga sp. L1A9	1,14	Klebsiella michiganensis	0,94
19	Marvinbryantia formatexigens	0,92	Escherichia coli	0,74	Halostella sp. PRR32	1,1	Enterobacter hormaechei	0,92
20	Methanobrevibacter olleyae	0,92	Blautia obeum	0,68	Ruminococcus bovis	1,1	Prevotella dentalis	0,9

Fig. 4.

Heatmaps showing the relative abundances of the twenty most dominant bacterial species identified in the control group and their corresponding abundances in the other experimental groups. A Results based on 16 S rDNA V3–V4 amplicon sequencing. B Results derived from shotgun metagenomic sequencing. Groups include Cnt (control), Cd (cafeteria diet), Prb (standard chow + probiotics), and CdtPrb (cafeteria diet + probiotics). This comparison highlights how cafeteria diet feeding and probiotic supplementation modulate the abundance of the taxa that are most prevalent under control conditions

In contrast, species-level interpretations based on shotgun metagenomic sequencing provide higher taxonomic confidence and are therefore used as the primary reference for species-level biological interpretation. Shotgun analyses provided complementary species-level profiles that further characterized these group-specific compositional changes (Table 3). In the control group, Segatella copri was the most dominant species at a rate of 5.87%, whereas in the Cd group, this value increased to 12.47%. The total proportion of the top five species was 16.7% in the control group, rising to 22.19% in the Cd group. These results indicate a reduction in diversity and the dominance of certain species in the Cd group. Although the total proportion of the top five species in the CdPrb group decreased to 18.1%, structural homogeneity at the control level was not achieved. Nevertheless, with probiotic support, beneficial species such as Faecalibacterium prausnitzii, Roseburia hominis, and Bifidobacterium adolescentis partially reemerged. Additionally, increases were observed in butyrate-producing species such as Anaerostipes hadrus, Intestinimonas butyriciproducens, Blautia wexlerae, and Flintibacter sp., indicating that probiotic intervention, while not providing full restoration, is associated with the relative enrichment of functionally relevant microbial groups. Figure 4B presents a comparative overview of the twenty most abundant bacterial species in the control group and their corresponding relative abundances in the other experimental groups, based on shotgun metagenomic sequencing data.

Table 3.

Top 20 most abundant bacterial species in each group based on relative abundance (%) obtained from shotgun metagenomic sequencing. Cnt (control), cd (cafeteria diet), Prb (SCD probiotics), and cdtPrb (SCD probiotics during cafeteria diet)

	Cnt	%	Cd	%	Prb	%	CdPrb	%
1	Segatella copri	5,87	Segatella copri	12,47	Segatella copri	7,26	Segatella copri	10,29
2	Flavonifractor plautii	3,9	Flavonifractor plautii	2,67	Flavonifractor plautii	3,06	Flavonifractor plautii	2,82
3	Lawsonibacter asaccharolyticus	3,24	Lawsonibacter asaccharolyticus	2,38	Lawsonibacter asaccharolyticus	2,6	Lawsonibacter asaccharolyticus	2,35
4	Flintibacter sp. KGMB00164	1,87	Blautia wexlerae	1,69	Blautia wexlerae	1,75	Mediterraneibacter gnavus	1,82
5	Limosilactobacillus reuteri	1,82	Collinsella aerofaciens	1,54	Lactobacillus johnsonii	1,62	Blautia wexlerae	1,82
6	Lactobacillus johnsonii	1,6	Mediterraneibacter gnavus	1,41	Flintibacter sp. KGMB00164	1,56	Bacteroides uniformis	1,81
7	Pseudoflavonifractor gallinarum	1,49	Flintibacter sp. KGMB00164	1,39	Collinsella aerofaciens	1,54	Faecalibacterium prausnitzii	1,75
8	Oscillibacter hominis	1,48	Segatella hominis	1,32	Mediterraneibacter gnavus	1,28	Segatella hominis	1,75
9	Pusillibacter faecalis	1,35	Roseburia hominis	1,22	Pseudoflavonifractor gallinarum	1,21	Flintibacter sp. KGMB00164	1,33
10	Intestinimonas butyriciproducens	1,32	Oscillibacter hominis	1,11	Segatella hominis	1,18	Blautia obeum	1,13
11	Acutalibacter muris	1,22	Pusillibacter faecalis	1,1	Oscillibacter hominis	1,14	Oscillibacter hominis	1,09
12	Lactobacillus intestinalis	1,11	Pseudoflavonifractor gallinarum	1,03	Intestinimonas butyriciproducens	1,12	Collinsella aerofaciens	1,08
13	Mediterraneibacter gnavus	1,02	Ligilactobacillus murinus	1,03	Pusillibacter faecalis	1,09	Parabacteroides sp. AD58	1,07
14	Taurinivorans muris	0,96	Intestinimonas butyriciproducens	1,02	Acutalibacter muris	1,03	Intestinimonas butyriciproducens	1,06
15	Faecalibacterium prausnitzii	0,92	Faecalibacterium prausnitzii	1,01	Ligilactobacillus murinus	0,93	Helicobacter apodemus	1,06
16	Vescimonas coprocola	0,92	Blautia obeum	0,9	Clostridium scindens	0,92	Pseudoflavonifractor gallinarum	1,05
17	Ligilactobacillus murinus	0,91	Clostridium scindens	0,83	Faecalibacterium prausnitzii	0,84	Pusillibacter faecalis	1,02
18	Collinsella aerofaciens	0,85	Parabacteroides sp. AD58	0,77	Vescimonas coprocola	0,8	Ligilactobacillus murinus	1
19	Bacteroides xylanisolvens	0,84	Acutalibacter muris	0,76	Roseburia hominis	0,79	Clostridium scindens	0,84
20	Ligilactobacillus faecis	0,81	Clostridioides difficile	0,75	Limosilactobacillus reuteri	0,78	Lactobacillus intestinalis	0,83

In both analyses, sequencing reads were detected for Bacillus subtilis, Bifidobacterium bifidum, Bifidobacterium longum, Lactobacillus acidophilus, Lactococcus lactis, and Streptococcus thermophilus, which are included in the probiotic mixture. However, no reads were detected for Lactobacillus bulgaricus, L. casei, L. fermentum, L. plantarum, or Saccharomyces cerevisiae. This suggests that not all species in the administered probiotic mixture could colonize simultaneously. Considering biological barriers such as gastric acid resistance, mucosal adhesion ability, and microbial competition, it is understood that the limitations in the effectiveness of the product may result not only from its formulation but also from physiological conditions. In this context, it is considered that regular and long-term probiotic supplementation is necessary to maintain efficacy.

Short-Chain Fatty Acids (SCFA) profile

Short-chain fatty acids (SCFAs), the primary products of microbiota metabolism, serve as crucial indicators of the biochemical outcomes of host-microbe interactions. In our study, significant differences were observed between the groups, particularly in the levels of acetate, isobutyrate, butyrate, valerate, and heptanoic acid (Fig. 5).

Fig. 5.

Effects of cafeteria diet and SCD probiotics supplementation on cecal content short-chain fatty acid (SCFA) concentrations in male Wistar rats. Panels represent concentrations (µmo/g cecal content) of (A) acetic acid, (B) isobutyric acid, (C) butyric acid, (D) valeric acid, (E) heptanoic acid, (F) isovaleric acid, (G) isocaproic acid, (H) caproic acid, and (J) propionic acid. The experimental groups are: Cnt (control), Cd (cafeteria diet), Prb (SCD probiotics), and CdtPrb (SCD probiotics during cafeteria diet). Statistical evaluation was conducted using both one-way ANOVA for overall group effects and unpaired t-tests for specific pairwise comparisons. Data are presented as mean ± standard error of the mean (SEM). Significance is denoted as follows: p < 0.05 (*), p < 0.01 (**),p < 0.001 (***) p < 0.0001 (****); non-significant results are marked as “ns”

A significant group effect was observed in acetate levels (ANOVA, F = 9.489; p = 0.0004) (Fig. 5A). Acetate levels decreased in the Cd group (p = 0.0326), while a marked increase was seen in the group administered probiotics (p = 0.0004). Although partial improvement was observed in the CdPrb group, it did not differ significantly from the Cd group. A similar trend was noted in the analysis of isobutyrate (F = 5.269; p = 0.0094) (Fig. 5B). A marked decrease was found in the Cd group (p = 0.0116), while recovery was seen in the Prb group (p = 0.0182); no significant difference was detected in the CdPrb group. The highest variance explanation was obtained for butyrate (F = 9.204; p = 0.0006; R² = 0.5924) (Fig. 5C). Levels decreased in the Cd group (p = 0.0044) and increased in the Prb group (p = 0.0014). A partial improvement was also recorded in the CdPrb group (p = 0.0181). These findings underscore the protective role of probiotics in maintaining butyrate production capacity. The strongest group effect was determined for valerate levels (F = 14.85; p < 0.0001) (Fig. 5D). A decrease was observed in the Cd group (p = 0.0147), and a surprising decline was detected in the Prb group compared to the control group (p = 0.0276). However, the fact that the Prb group exhibited higher valerate levels than the Cd group (p < 0.0001) reveals a partial regulatory effect of probiotics. A significant group effect was also found in the analysis of heptanoic acid (F = 4.820; p = 0.0124) (Fig. 5E). Specifically, levels in the CdPrb group were decreased compared to both the control (p = 0.0262) and Cd groups (p = 0.0125), demonstrating a response profile distinct from other SCFAs.

Although no statistically significant difference was found in the levels of isovaleric, isocaproic, caproic, and propionic acids (Fig. 5F–J), a general tendency towards reduction in the Cd group and recovery in the CdPrb group was observed. Notably, the increase in isocaproic acid levels in the Cd group indicates a predominance of potentially pathobiotic species, while this increase appears to be suppressed in the CdPrb group. Overall, these findings demonstrate that the cafeteria diet markedly impairs SCFA production, and that probiotic intervention exerts a restorative effect particularly on acetate and butyrate levels. However, the effect varies by parameter across all SCFAs.

Indole-3-Propionic Acid (IPA) analysis

Indole-3-propionic acid serves as a functional biomarker, synthesized through the microbiota-mediated metabolism of tryptophan, and is recognized for its role in maintaining epithelial integrity and its potent antioxidant properties. In this study, IPA levels in the intestinal contents of the experimental groups were quantitatively assessed, with analysis of variance revealing significant differences among the groups (F = 11.91; p = 0.0002; R² = 0.6776) (Fig. 6). The group administered only probiotics (Prb) demonstrated significantly elevated IPA levels compared to the control group (p = 0.0245), the cafeteria diet group (Cd) (p < 0.0001), and the group receiving a probiotic-supplemented cafeteria diet (CdPrb) (p = 0.0490). The cafeteria diet was found to suppress IPA levels (Cnt vs. Cd; p = 0.0147), whereas probiotic administration partially ameliorated this reduction (Cd vs. CdPrb; p = 0.0392), although not completely restoring levels to those of the control group (Cnt vs. CdPrb; ns). These results suggest that probiotic intervention influences not only microbial composition but also the production of functional metabolites, and can particularly mitigate, at least partially, diet-induced reductions in IPA.

Fig. 6.

Effects of cafeteria diet and SCD probiotics supplementation on cecal indole-3-propionic acid (IPA) levels in male Wistar rats. Fecal IPA concentrations (ng/g cecal content) were determined using LC-MS/MS. Experimental groups include: Cnt (control), Cd (cafeteria diet), Prb (SCD probiotics), and CdtPrb (SCD probiotics during cafeteria diet). Statistical evaluation was conducted using both one-way ANOVA for overall group effects and unpaired t-tests for specific pairwise comparisons. Data are presented as mean ± standard error of the mean (SEM). Significance is denoted as follows: p < 0.05 (*), p < 0.01 (**), p < 0.0001 (****); non-significant results are marked as "ns"

Correlation of SCFAs and IPA with specific microbial taxa

Spearman correlation analysis identified significant associations between SCFAs, IPA, and specific bacterial species (Fig. 7A). Notably, a strong positive correlation was observed between butyrate levels and Faecalibacterium prausnitzii (r = 0.72, p = 0.003), as well as with Roseburia hominis (r = 0.68, p = 0.007). Acetate levels demonstrated a positive correlation with Subdoligranulum variabile (ρ = 0.65, p = 0.010) and Akkermansia muciniphila (ρ = 0.61, p = 0.014); although the absolute abundance of A. muciniphila remained low across all groups (≤ 0.12% relative abundance), this significant association suggests that even at low levels it may contribute functionally to acetate-linked barrier processes. Valerate levels were significantly associated with Anaerobutyricum hallii (r = 0.70, p = 0.005). IPA levels, in particular, exhibited a high positive correlation with Bacteroides thetaiotaomicron (r = 0.74, p = 0.002). Detailed information on all significant species-metabolite correlation pairs is provided in Table S1.

Fig. 7.

Spearman correlation matrix and standardized abundance heatmap of short-chain fatty acids (SCFAs), indole-3-propionic acid (IPA), and selected metabolically relevant bacterial taxa. A Pairwise Spearman’s rank correlation matrix illustrating associations between metabolite concentrations and the relative abundances of bacterial taxa. Numerical values represent Spearman correlation coefficients (r), ranging from − 1 (negative association) to + 1 (positive association). Color intensity reflects the magnitude and direction of the correlation. Statistical significance was determined using two-tailed Spearman’s correlation tests, with * indicating p < 0.05. B Heatmap showing z-score–standardized abundance profiles of metabolites and bacterial taxa across individual samples within experimental groups: Cnt (control), Cd (cafeteria diet), Prb (probiotic-treated), and CdPrb (cafeteria diet + probiotic). Values represent standardized deviations from the overall mean for each variable and are intended for visualization of relative enrichment or depletion patterns across groups. Importantly, values displayed in Panel B do not represent correlation coefficients. Together, Panels A and B provide complementary perspectives on metabolite–taxon associations and group-wise abundance patterns under different dietary and probiotic conditions

These relationships were further illustrated by the boxplot analyses. Specifically, in the Cd group, a significant reduction in acetate and butyrate levels was evident in the boxplots (p < 0.05), while the abundance of Subdoligranulum variabile and Roseburia hominis, associated with these metabolites, appeared at low intensity in the Cd group on the heatmap, though relatively more discernible in the control group. Similarly, in the Prb group, the increase in IPA was statistically significant in the boxplot (p < 0.01), and this increase coincided with a higher relative abundance of the IPA-associated species Bacteroides thetaiotaomicron in the heatmap. Furthermore, fluctuations in valerate levels in both the boxplot and the heatmap paralleled the abundance of Anaerobutyricum hallii, highlighting agreement between multiple analytical approaches.

The heatmap (Fig. 7B) visually reinforced the species–metabolite relationships by groups. Notably, in the Cd group, the prominence of Subdoligranulum variabile and Roseburia hominis in the control group, and in the Prb group, the dominance of Faecalibacterium prausnitzii and Akkermansia muciniphila, directly corresponded to SCFAs and IPA profiles. In the CdPrb group, increases in Clostridium sporogenes and Blautia wexlerae were also consistent with the metabolite profiles. Thus, when evaluating boxplot data, Spearman correlations, and heatmap results collectively, it becomes evident that significant changes in SCFA and IPA metabolites demonstrate consistent biological patterns with group-based abundance differences of specific bacterial species that produce or consume these metabolites.

Discussion

In this study, it was demonstrated that administering a high-fat cafeteria diet during the developmental period induces significant dysbiosis in the gut microbiota. This dysbiosis is characterized by notable reductions in SCFAs, particularly acetic acid, isobutyric acid, butyric acid, valeric acid, and heptanoic acid, as well as a suppression of IPA levels. Data obtained through a metagenomic approach revealed that these microbial and metabolic disturbances were partially, yet significantly, ameliorated by probiotic intervention. It should be explicitly noted that species-level profiles derived from 16S rRNA gene (V3–V4) amplicon sequencing in this study represent database-dependent, indicative taxonomic assignments rather than definitive species identification. Accordingly, 16S-derived species-level findings are discussed here within a descriptive and comparative framework, while higher-confidence species-level interpretation is primarily based on shotgun metagenomic data. In line with this framework, the recovery of microbial diversity and the production of functional metabolites were supported by strong correlations observed between the abundance of certain species and the profiles of SCFAs and IPA. These findings suggest that the effectiveness of probiotic application is derived more from the reshaping of gut ecology rather than direct colonization capacity, presenting translationally significant potential for mitigating diet-induced dysbiosis during developmental periods. Thus, the results support our hypothesis that probiotics can modulate microbiota–metabolite interactions to preservation of epithelial barrier integrity and metabolic balance.

Based primarily on shotgun metagenomic data the observed species-level alterations in microbiota composition underscore the significance of specific bacterial groups associated with intestinal homeostasis and metabolite production. Our findings indicate that the suppression of butyrate-producing commensals, such as Faecalibacterium prausnitzii and Roseburia hominis, aligns with both a reduction in alpha diversity and decreased SCFA levels particularly evident in the cafeteria died and probiotic groups compared to controls. While our data did not show a consistent increase of (A) muciniphila under probiotic supplementation, its presence and positive correlation with acetate levels align with literature highlighting its role in mucus layer integrity. Thus, the functional relevance of this species should be considered despite its low abundance in our dataset. The correlations among these three species have been identified not only in our experimental model but also in human studies as critical links concerning metabolic health, inflammation, and barrier functions [37, 38]. Furthermore, the strong correlation between Bacteroides thetaiotaomicron and IPA (r ≈ 0.74) is particularly significant, corroborating our independent histological findings of suppressed TNF-α and IL-1β expression in ileum and colon tissue post-probiotic treatment [6]. In this context, the literature highlights the importance of species such as Clostridium sporogenes, Bacteroides eggerthii, and B. uniformis in IPA production, supporting the notion that tryptophan metabolism bridges epithelial barrier function and immune response via microbial components. Recent reports have underscored the role of indole-propionic acid in modulating oxidative stress and maintaining epithelial barrier integrity, with reductions in IPA being consistently linked to accelerated aging and metabolic decline [39–41]. In complementary analyses using the same animal model, SCD probiotics were shown to restore microbiota composition, SCFA pools, and functional metabolites such as IPA, underscoring the consistency of their multi-omics impact [17].

The integrated assessment of shotgun metagenomic analyses and 16 S rDNA gene profiles suggests diet-associated predominance of species such as Segatella copri. This species has been documented in humans to be linked with both advantageous metabolic effects and inflammatory processes [42]. Its proliferation in response to a cafeteria diet underscores the sensitivity of microbial ecology to environmental stimuli. Similarly, the pronounced increase in Escherichia coli is indicative of the inflammatory effects of the diet, while its reduction following probiotic intervention aligns with histological findings, underscoring its translational significance. Conversely, the partial reduction in S.copri abundance observed with probiotic application suggests that the intervention may exert an indirect suppressive effect. This finding indicates that probiotics influence the ecological balance not only of the introduced strains but also of broader microbial networks. Cafeteria-style high-fat diets have been shown to disrupt SCFA profiles, including reduced butyrate production, in rodent models [43, 44]. In this context, our probiotic intervention may help selectively restore acetate and butyrate pools.

A significant observation at the taxonomic level is the selective detection of species such as Lactobacillus johnsonii, L. intestinalis, and Limosilactobacillus reuteri following probiotic administration. Although some of these species were not direct constituents of the administered probiotic mixture, they may have acquired a colonization advantage through niche expansion associated with probiotics or via cross-feeding mechanisms. Correlation analyses indeed revealed significant associations between metabolite profiles and Streptococcus thermophilus and Lactobacillus acidophilus, which are direct components of the probiotic formulation; this indicates that the effects of the introduced strains extend beyond themselves, creating expansive impacts within the ecosystem. Similarly, human studies have reported that probiotics can induce “ecological spillover effects” not confined to the administered strains [45, 46]. Thus, our findings support the notion that probiotic efficacy involves reshaping at the community level, beyond mere direct colonization. Emerging evidence further indicates that probiotic supplementation can mitigate cafeteria diet–induced reductions in microbial diversity and promote the stability of butyrate-producing taxa such as Intestinimonas and Blautia, both of which are functionally linked to SCFA generation and metabolic homeostasis [47, 48].

Our study further corroborates our previous findings. As previously documented, probiotic supplementation has been demonstrated to significantly protect hepatic antioxidant systems [5], maintain hepatic structural integrity [4], and preserve intestinal barrier functions [6] that were compromised by a cafeteria diet. Moreover, developmental probiotic intervention was also shown to ameliorate cafeteria diet-induced impairments in social behavior during adulthood [49], thereby extending the functional relevance of probiotic-mediated restoration beyond metabolic and structural domains. The congruence between the metagenomic and metabolite data from this study and the tissue-level results obtained from the same animal model further substantiates the multifaceted protective role of probiotics along the gut–liver axis. Specifically, the butyrate production and mucosal anti-inflammatory effects of Faecalibacterium prausnitzii, its direct association with SCFA output, the relatively higher presence of Roseburia hominis in the control group (as more clearly reflected in shotgun data than in 16S analysis) but its suppression in cafeteria and probiotic groups in the shotgun data, the role of Blautia species in acetate production and their potential correlation with IPA, the critical function of Anaerostipes hadrus in butyrate biosynthesis, and the strong correlation between Bacteroides thetaiotaomicron and IPA (r ≈ 0.74) collectively illustrate how the functional impacts of probiotic effectiveness may manifest at the species level. When these species-level associations are considered collectivelyboth in terms of restorative effects on SCFA production and IPA-mediated immunomodulatory mechanisms they closely align with our previous histological and biochemical findings.

Nevertheless, our study also contains certain limitations. Differences in taxonomic resolution between 16S rDNA and shotgun metagenomic methods can lead to some inconsistencies in reporting abundances at the species level; however, while 16S provides the advantage of higher sample size and broader coverage, shotgun sequencing offers superior functional resolution, thereby justifying their complementary use in this study. Moreover, given the diversity of human diets, lifestyle factors, and differences in microbial ecology, caution should be exercised when directly translating findings from animal models into the clinical context. In this framework, the observed reduction in butyrate-producing commensals such as Faecalibacterium prausnitzii and Roseburia hominis is consistent with findings reported in inflammatory bowel diseases [38] and aligns with our results as well. In contrast, although Akkermansia muciniphila has often been reported to decrease in metabolic disorders [37], this trend did not clearly emerge in our cafeteria diet model. Emerging evidence suggests that Intestinimonas butyriciproducens enhances host metabolic health by increasing butyrate production from dietary fructoselysine [48], while the genus Blautia has been shown to support colonic mucus function via SCFA secretion [50]. Taken together, these convergent and divergent findings demonstrate that diet-induced dysbiosis shows partial but not complete overlaps with human disease models.

Notably, these changes at the species level become even more significant from a translational perspective when considered alongside metabolite profiles. Our holistic multi-omic analysis, which integrates taxonomic composition (16S/shotgun) with functional outputs (SCFAs, IPA), reveals that probiotic supplementation not only restores ecological balance but also restructures metabolite-based pathways that are critical for host physiology. This overlap supports the notion that the protective effects of probiotics operate along both the gut–liver and gut–brain axes, thereby alleviating steatosis, neuroinflammation, and cognitive decline. Indeed, our previous findings demonstrated gut–liver axis modulation by probiotics [5], consistent with broader evidence emphasizing the integrative role of this axis in systemic metabolic regulation [51]. Similarly, accumulating studies highlight that probiotics exert beneficial effects on neural outcomes and cognitive performance through gut–brain interactions [52, 53]. In this context, the cafeteria diet model provides a robust experimental platform; meanwhile, probiotics stand out as system-level regulators capable of mitigating diet-induced disturbances through ecological, metabolic, and functional reprogramming.

Conclusion

This study illustrates that the dysregulation of gut microbiota and metabolite profiles induced by a high-fat cafeteria diet during the developmental period can be partially ameliorated through probiotic intervention. The integration of 16S rDNA and shotgun metagenomic data with SCFA and IPA measurements demonstrates that probiotics facilitate functional metabolite production by reshaping gut ecology. The findings suggest that critical species such as Faecalibacterium prausnitzii, Blautia wexlerae, and Anaerostipes hadrus play pivotal restorative roles in butyrate and acetate production, and the strong association of Bacteroides thetaiotaomicron with IPA is significant for metabolic balance. These multilayered data indicate that probiotic application supports structural and biochemical improvements not only in the liver but also in ileum and colon tissues, underscoring considerable translational potential for mitigating diet-induced dysbiosis.

Authors’ contributions

TC and HTT were involved in the conception, design, execution, data interpretation, and manuscript drafting phases of the study. HÖ, HA, and EŞ contributed to the completion of the experimental procedures, including animal handling, tissue sampling, and preliminary data analyses. TÜ was responsible for the bioinformatic processing and interpretation of metagenomic datasets. EA supervised all stages of the study, ensured methodological and analytical accuracy, and coordinated the final manuscript preparation and critical revisions.

Funding

This study was supported by the Inönü University Scientific Research Projects Coordination Unit (BAP) under project number FOA-2024-3587.

Data availability

The 16S rDNA amplicon and shotgun metagenomic sequencing datasets generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) under the BioProject accession numbers PRJNA1288408 and PRJNA1289129, respectively. Datasets related to short-chain fatty acids (SCFAs) and indole-3-propionic acid (IPA) analyses are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was carried out with the approval of the Ethics Committee (meeting date: 29.06.2021, approval number: 2021/03) from the Bingöl University Animal Experiments Local Ethics Committee.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.