Investigating the response of the butyrate production potential to major fibers in dietary intervention studies

The gut microbiota plays a crucial role in host health, and dietary fiber interventions are frequently used to modulate its composition and activity. Dietary fibers, particularly prebiotics, selectively stimulate the growth of beneficial bacteria, leading to various health benefits. Inulin-type fructans (ITF) are well-known prebiotics with a documented bifidogenic effect. However, the mechanisms underlying the broader effects of different fibers on the gut microbiota and the production of short-chain fatty acids (SCFAs) remain poorly understood. Previous studies investigating the impact of various fibers on the gut microbiota have shown inconsistent results, with the exception of the bifidogenic effect of some fibers. This inconsistency highlights the need for larger scale, harmonized studies to investigate the specific effects of different fiber types on the gut microbiome and its functionalities. Short-chain fatty acids (SCFAs), such as butyrate, propionate, and acetate, are important metabolites produced by gut bacteria through fermentation of dietary fibers. Butyrate, in particular, is essential for gut health and homeostasis. While *in vitro* studies and animal models have provided valuable insights, human intervention studies are critical for understanding the effects of fibers in real-world settings. The current study aims to address the limitations of previous studies by performing a pooled analysis of raw 16S rRNA gene sequencing data from multiple placebo-controlled intervention studies using a uniform bioinformatics workflow. The focus is on interventions using ITF, resistant starch (RS), and arabinoxylan-oligosaccharides (AXOS), with a specific emphasis on the impact on SCFA-producing potential, particularly butyrate production, and the role of pre-intervention gut microbiota composition.

Numerous studies have highlighted the role of the gut microbiota in maintaining host health and disease. Dietary fibers, especially prebiotics, are frequently used to modulate the microbiota towards a beneficial state. Inulin-type fructans (ITF) are a well-known example, exhibiting a bifidogenic effect—selective stimulation of *Bifidobacteria* spp. While earlier studies relied on culturing and selective molecular methods, next-generation sequencing has enabled comprehensive profiling of gut microbiota composition. However, results across studies have been highly variable, often lacking assertive conclusions beyond the bifidogenic effect. There's a general consensus that fibers stimulate microbiota growth and activity, but the precise mechanisms underlying health benefits remain largely unclear. Individual variation in microbiota composition before intervention has been cited as a significant contributor to study heterogeneity. SCFAs, fermentation byproducts, are implicated in maintaining gut homeostasis and improving health. The main SCFAs are acetate, butyrate, and propionate, with only certain bacteria producing butyrate and propionate. The bifidogenic effect of fiber supplementation may not be directly beneficial; instead, *Bifidobacteria*-derived lactate might fuel butyrate production through cross-feeding, ultimately promoting positive health outcomes. While *in vitro* and animal models are valuable, human dietary intervention studies offer the most relevant data. However, these studies are often complex and varied in design, microbiota profiling, and bioinformatics, leading to inconsistencies. Harmonizing bioinformatics and data analysis is a key step towards reducing this variability.

This study conducted a pooled analysis of 16S rRNA gene sequencing data from 14 placebo-controlled intervention studies involving ITF, RS, and AXOS. The studies included a total of 441 participants. A standardized bioinformatics workflow using DADA2 (v1.20) for ASV inference and RDP3033 for taxonomic classification was applied to all datasets. The lowest possible taxonomic resolution for comparison across studies, was genus level. Predictions of SCFA-forming pathways, including the terminal genes *but* and *buk*, were performed. Data were harmonized to include pre- and post-intervention time points for each subject. Linear mixed-effect models (lme4; lmer) were used, including studies as a random effect, to analyze relative response (RR) data (post-intervention values relative to pre-intervention values, capped at ±100%) and relative abundance changes (RA). Analysis focused on major taxa associated with butyrate and propionate pathways, as well as abundant non-butyrate/non-propionate producing taxa. The pre-intervention microbiota composition's influence was assessed by stratifying analyses based on *Ruminococcus* and *Prevotella* abundance. Correlations between *Bifidobacterium* responses and butyrate producer responses were also evaluated. Ethical considerations: The institutional review board of Hannover Medical School approved the waiver of informed consent due to the secondary analysis of already published data.

The pooled analysis revealed distinct effects of the three fibers on the gut microbiota's SCFA production potential. ITF treatment significantly increased bacteria carrying the *but* enzyme (9.15 [1.56, 16.76]; p < 0.05), mainly driven by *Anaerostipes* and *Faecalibacterium*. *Buk*-carrying bacteria, conversely, declined. RS treatment trended towards increased *but*-containing bacteria (8.70 [-0.97, 18.37]; p < 0.1), primarily due to *Agathobacter*. AXOS treatment showed no overall effect on butyrate-producing potential. All three fibers exhibited a bifidogenic effect (significant increase in *Bifidobacterium*), although the magnitude differed significantly across treatments. Pre-intervention *Ruminococcus* abundance influenced the butyrate response; low *Ruminococcus* levels for ITF and high levels for RS correlated with increased *but*-containing taxa. Conversely, high pre-intervention *Prevotella* abundance negatively correlated with butyrate responses for both ITF and RS. Correlations between *Bifidobacterium* responses and butyrate producer responses were inconsistent across fiber types; The Suc pathway for propionate production was largely unaffected, while the Pdiol pathway decreased significantly with all interventions. Fecal butyrate concentrations increased after ITF and RS interventions in several individual studies but not consistently to statistically significant levels.

This study demonstrates that ITF and RS selectively stimulate specific gut microbiota taxa, enhancing butyrate production potential. The bifidogenic effect, while present, appears largely independent of butyrate production in this study. The observed selective stimulation of butyrate-producing taxa is multifaceted, involving direct substrate utilization (e.g., *Faecalibacterium* and ITF), cross-feeding mechanisms (e.g., *Bifidobacteria*-derived lactate potentially fueling *Anaerostipes* growth on ITF), and nutrient competition (e.g., *Prevotella* potentially competing with butyrate producers). The contrasting roles of *Ruminococcus* in response to ITF and RS suggest different mechanisms of action for these fibers. The lack of significant changes in the overall AcCoA pathway, but a marked effect on *but*-carrying taxa highlights the importance of considering functional redundancy and the different contributions of various butyrate-producing bacteria. The limited data on propionate synthesis highlight the need for future multi-omics studies to investigate in detail the factors affecting propionate synthesis. The study emphasizes that the pre-intervention microbiota composition strongly influences the response to fiber intervention, underscoring the importance of personalized approaches.

This pooled analysis demonstrates the selective stimulation of butyrate-producing bacteria by ITF and RS, highlighting the importance of considering both functional redundancy and pre-intervention microbiota composition. While a bifidogenic effect was observed across multiple treatments, butyrate production responses were fiber specific and seemingly distinct from the bifidogenic effect, suggesting that multiple mechanisms are involved and that these should be considered in future intervention design. Future research should focus on using in-depth metagenomic analyses and in-vitro models to investigate species- and strain-level mechanisms governing butyrate and propionate production in response to different fibers. These findings underscore the potential for individualized dietary interventions targeting specific gut microbiota composition to achieve optimal outcomes.

This study's analysis was limited to the genus level due to inter-study variability in 16S rRNA gene variable regions. While efforts were made to standardize bioinformatics, some inherent variability in wet-lab procedures may have influenced results. The relatively small number of studies included for AXOS may also influence the generalizability of findings related to this fiber type. Moreover, while fecal butyrate levels were examined, limitations associated with fecal sampling could confound the interpretation of the results related to butyrate production. Only two studies provided detailed data on SCFA proportions, restricting a more complete analysis.