A Low-Dose Prebiotic Fiber Supplement Reduces Lipopolysaccharide Binding Protein Concentrations in a Subgroup of Young, Healthy Adults Consuming Low-Fiber Diets

Abstract

Although the beneficial effects of fiber supplementation on overall health and the gut microbiome are well-known, it is not clear whether fiber supplementation can also alter the concentrations of lipopolysaccharide (LPS) binding protein (LBP), a marker of intestinal permeability. A secondary analysis of a previously conducted study was performed. In the randomized-order, placebo controlled, double-blinded, cross-over study 20 healthy, young participants consuming a low-fiber diet at baseline were administered a daily dose of 12g of prebiotic fiber compared with a placebo over a period of four weeks with a 4-week washout between arms. In this secondary analysis we hypothesized that the fiber supplement would reduce LBP concentration. We further hypothesized that lecithin cholesterol acyltransferase (LCAT) activity, a measure of high-density lipoprotein (HDL) functional capacity, would be altered. Fiber supplementation did not significantly alter LBP concentration or LCAT activity in the overall cohort. However, in a subgroup of individuals with elevated baseline LBP concentrations, fiber supplementation significantly reduced LBP from 9.27±3.52 to 7.02±2.32 ug/mL (p=0.003). Exploratory analyses found positive correlations between microbial genes involved in LPS synthesis and conversely negative correlations with genes involved in antibiotic synthesis and LBP. Positive correlations between LBP and multiple sulfated molecules including sulfated bile acids and perfluorooctanesulfate (PFOS), and ibuprofen metabolites were also found. These findings highlight multiple environmental and lifestyle factors such as exposure to industrial chemicals and medication intake, in addition to diet, which may influence the association between the gut microbiome and gut barrier function.

Introduction

Extensive evidence suggests that dietary fiber supplementation provides numerous health benefits. However, the majority of Americans consume less than half of the recommended amount of dietary fiber, with less than 5% meeting the recommended intake (1). Dietary fiber has been found to nourish and regulate microbial communities within the intestinal tract, significantly reducing the risk of all-cause mortality and disease (2–4). Beneficial gut microbes such as Bifidobacterium ferment fiber and play a vital role in promoting a healthy gut barrier by displacing pathogens and directly influencing mucosal homeostasis. They do this by synthesizing an array of metabolites including short-chain fatty acids (SCFAs) and indolepropionate (IPA) that stimulate mucus production by goblet cells and contribute to regulation of tight junction proteins, respectively (5–7). An impaired gut barrier is associated with the introduction of immunogenic endotoxins, including lipopolysaccharide (LPS), through transcellular and paracellular pathways, leading to disrupted mucosal homeostasis, inappropriate translocation of harmful particles, and dysregulation of cell adherence protein complexes (8,9). Such mechanisms have been observed in pathological conditions like obesity, type 2 diabetes, and Alzheimer’s disease (10–12), but they are also linked to Western diets low in dietary fibers (13).

Evolutionary pressures have promoted highly conserved host mechanisms of the innate and adaptive immune systems to modulate endotoxins, including LPS binding protein (LBP), high density lipoproteins (HDL), macrophages and monocytes (14). HDL regulate cholesterol content in immune cells by influencing lipid raft composition, thus modulating signaling functions involving raft proteins such as toll-like receptors (15,16). HDL can also bind, deactivate, and clear LPS together with its associated protein LBP (14). When LPS is introduced into the system through the gut, gums, or a wound it is bound by LBP and trafficked to pattern recognition receptors CD14 and toll-like receptor 4 (TLR4) on the surface of immune cells such as macrophages and monocytes. This in turn induces an immune response resulting in the release of pro-inflammatory mediators (17). HDL can bypass this pro-inflammatory action of LPS and deactivate immune cells by transferring LPS from immune cells to HDL-associated LBP (18). When LPS is bound in the HDL-LBP-LPS complex it can be deactivated via the enzymatic action of acyloxyacyl hydrolase (14,19). More recently, intestinally derived HDL were shown to sequester LPS before it enters the portal vein, preventing recognition by TLR4 on macrophages, and reducing overexpression of inflammatory molecules in the liver (14). Lecithin cholesterol acyltransferase (LCAT) plays an important role in the metabolism of HDL by esterifying and packing cholesterol into HDL particles. LCAT thus plays a critical role in the maturation and functional capacity of HDL particles (20,21).

LBP serves as a potent marker of endotoxin presence and has been shown to be an important alternative to the measurement of LPS. Whereas LPS is cleared quickly LBP concentrations can persist in plasma for more than 24 hours after the onset of infection (22). LBP has shown potential as a biomarker of disease risk in chronic diseases including cardiovascular disease and Alzheimer’s disease (23,24). Abnormally low LBP concentrations were identified in Parkinson’s disease related to gut permeability and intestinal microbial dysbiosis (25,26). On the other hand, higher concentrations of LBP have been observed in diseases such as leptospirosis, tuberculosis, and Alzheimer’s disease (27,28). LBP was recently shown to be a powerful independent predictor of Alzheimer’s disease risk (24,29).

This study aimed to uncover the effects of fiber supplementation on the function of HDL, as measured by LCAT activity, and on gut permeability, as measured by LBP concentrations, in otherwise healthy individuals consuming a low-fiber diet. We hypothesized that 4 weeks of fiber supplementation in individuals with habitually low fiber diets would reduce LBP concentrations and improve HDL function. We also performed exploratory correlation analyses to determine whether LBP concentrations or LCAT activity were associated with gut microbiome composition and function, as well as plasma metabolomic profiles.

Methods and Materials

The complete study design and participant characteristics of the multi-omic analysis clinical trial were described previously (5). Briefly, the trial was a double-blinded, placebo-controlled, randomized order crossover trial with 20 healthy adult participants (n=10 men, n=10 women) (Figure 1) aged between 18 and 45 years, all of whom had BMIs within the range of 23–32 kg/m2 and typically consumed less than 15 grams of fiber per day. Over a 4-week period, participants were first provided with either a daily dose of 12g of a prebiotic fiber mix or a visually similar placebo. This was followed by a 4-week washout phase and then a switch to the opposite treatment for another 4 weeks. The prebiotic mixture was composed of fructooligosaccharides, resistant starch, sugarcane fiber, inulin, gum arabic, xanthan gum, and berry fruit powders. Stool samples were collected within 24h prior to the blood draw, and blood samples were collected after a 12h overnight fast at the onset and conclusion of each phase. Stool samples were analyzed by shallow shotgun metagenomic sequencing on an Illumina NovaSeq platform by Diversigen (Houston, TX, USA) to analyze the gut microbiome composition and the abundance of microbial genes. Blood samples were analyzed through untargeted LC-MS metabolomics by Metabolon (Morrisville, NC, USA) to measure plasma metabolites, including microbially-derived metabolites. Additionally, anthropometric measurements, blood lipid and glucose concentrations, diet records utilizing standardized 24-hour dietary recalls, bowel movement patterns, and health questionnaire responses were collected.

Figure 1: Study Design of the Double-Blinded, Placebo-Controlled, Randomized-Order Crossover Trial.

The study included 20 participants who underwent a randomized-order, crossover study comparing the effects of a prebiotic fiber blend to a placebo. Each group, after random allocation, received either the prebiotic fiber blend first or placebo first for a four-week treatment phase, followed by a four-week washout period before switching to the alternative treatment for another four weeks. The p-values reported in Table 1 reflect the effect of treatment, as determined by a linear mixed model that accounted for both treatment and time. The interaction between treatment and time was not significant.

The clinical trial was registered at clinicaltrials.gov with the identifier NCT03785860. The study was approved by the Institutional Review Board of the University of California, Davis Protocol 1335956 and written informed consent was obtained from all participants, following ethical guidelines (5).

Measurement of LBP

Concentrations of LBP were measured in plasma using the human LBP ELISA kit (Abcam, ab213805), following the manufacturer’s protocol. Briefly, plasma samples were thawed, and aliquots were transferred to the provided microplate wells pre-coated with an LBP-specific antibody. After incubation and washing steps, a biotin-conjugated secondary antibody was added, followed by a streptavidin-HRP conjugate. Colorimetric detection was achieved with the addition of the substrate solution, and the reaction was stopped with the provided stop solution. Absorbance was measured at 450 nm using a SpectraMax M5 microplate reader (Molecular Devices., and LBP concentrations were calculated using a standard curve generated from the provided LBP standards.

Measurement of LCAT Activity

Plasma samples were also used to assess HDL function by measuring LCAT activity. An LCAT Activity Assay Kit (MAK107) was employed according to the manufacturer’s instructions. In brief, plasma samples were thawed and mixed with the provided LCAT assay buffer and substrate. The reaction mixture was incubated at 37°C for a specific duration, and the reaction was stopped by adding the provided stop solution. The generated product was quantified by measuring the absorbance at 340 nm using a SpectraMax M5 microplate reader (Molecular Devices). LCAT activity was calculated based on the change in absorbance and was expressed as units per mL of plasma.

Statistical Analysis

All statistical analyses were conducted using R software (version 4.2.2). For general analyses, a linear mixed effect model was employed with the lmerTest package (version 3.1.3) (30) to account for repeated measures within the data. Post-hoc tests were performed on these models using the emmeans package (version 1.10.4) (31). Exploratory analysis was conducted by dividing the participants into those with above-median vs. below-median LBP concentrations at baseline. Differences in counts of microbial genes for LPS synthesis and Bifidobacterium abundance at baseline, as well as differences in the change in Bifidobacterium abundance from pre to post fiber intervention were analyzed by unpaired t-tests comparing participants with above-median vs. below-median baseline LBP concentrations. We focused on Bifidobacterium changes for these analyses because this was the primary outcome of the original study and because statistically significant increases in bifidobacteria were detected in response to the prebiotic supplement.

We performed a repeated measures correlation analysis using the rmcorr package (version 0.7.0) (32) to explore relationships between metabolites, Operational Taxonomic Units (OTUs) from microbes and gene counts with LBP concentration and LCAT activity. Repeated measure correlation analysis enables us to evaluate the within-individual relationship across paired measures assessed on two or more occasions. Metabolites, OTUs, and gene counts were log-transformed, followed by correlation analysis with LBP concentration and LCAT activity. The results were then adjusted for multiple comparisons using the Benjamini-Hochberg method, with statistical significance set at a p-value of less than 0.05. For the repeated measure correlation analysis of microbial gene abundance data, the edgeR package (version 4.2.1) (33) was employed for gene preprocessing, including normalization using the `calcNormFactors` function. As part of the preprocessing, genes with low abundance were excluded by filtering out those with a maximum count per million (CPM) value below a cutoff of 1. In other words, only genes with a maximum CPM value equal to or above 1 were retained in the dataset. Subsequently, limma-voom, a component of the limma package (version 3.60.4) (34), was applied for data transformation. Limma-voom involves further filtering, where genes with low expression were removed based on their CPM values. The resulting dataset was then transformed using voom, a process that estimates the mean-variance relationship of log-counts and assigns precision weights to each observation. The transformed data, representative of normalized and filtered gene abundance, were then utilized in repeated measure correlation analyses with LBP ug/ml and LCAT activity.

Results

As reported previously, as part of inclusion criteria study participants had habitually low fiber intakes of <15 g fiber/d, or approximately half of the current recommended dietary allowance. 24-hour dietary recalls were conducted at baseline and regularly during the study to ensure that their low-fiber consumption was maintained (5). There were no significant changes in participant dietary fiber intake in the background diet over the course of the study (5). LBP concentrations did not differ significantly between the placebo and prebiotic arms (p > 0.05) and similarly, post-treatment LBP concentrations were not significantly different between the two groups (p > 0.05) (Table 1). Additionally, LCAT activity did not display significant differences between the placebo and fiber intervention arms before or after treatment (p > 0.05) (Table 1).

Table 1.

Pre- and Post-Treatment LBP Concentrations and LCAT Activity Levels in Placebo and Prebiotic Groups of Healthy Young Adults^*

	Placebo		Prebiotic		P-value
Variable	Pre	Post	Pre	Post
LBP (ug/mL)	5.99 ± 3.1	5.98 ± 3.21	6.62 ± 3.88	5.67 ± 2.65	0.446
LCAT Activity Index	1.053 ± 0.018	1.055 ± 0.024	1.047 ± 0.023	1.052 ± 0.02	0.718

^*Data are shown as means ± SDs. Measurements of pre- and post-treatment lipopolysaccharide binding protein LBP concentrations and LCAT activity on the placebo vs. prebiotic arm were compared with a linear mixed model (n = 20).

Abbreviations: LBP, Lipopolysaccharide Binding Protein; LCAT, Lecithin cholesteryl acyltransferase

However, in participants with above-median pre-treatment LBP concentrations, a significant decrease in post-treatment LBP concentrations was observed following fiber supplementation (p < 0.01) (Table 2). However, the sum of the LPS-producing related genes did not show a statistically significant difference in the above- vs. below-median LBP participants before and after the prebiotic supplementation, (Supplemental Fig. 1A). We explored whether there were other differences in the microbiomes of participants with above- vs. below-median LBP concentrations. We hypothesized that participants with above-median LBP concentrations would have less healthy microbiomes, with higher baseline counts of LPS producing genes and lower baseline Bifidobacterium abundances. We further hypothesized that since the participants with above-median baseline LBP experienced a reduction in LBP after the fiber arm, this may be due to higher increases in Bifidobacterium abundance in response to the fiber intervention. However, none of these comparisons were statistically significant. The baseline counts of LPS producing genes (Supplemental Fig. 1B), the baseline abundances of Bifidobacterium (Supplemental Fig. 1C), and the changes in Bifidobacterium abundance from pre to post fiber intervention (Supplemental Fig. 1D) were all not different between participants with above- vs. below-median baseline LBP concentrations.

Table 2.

Pre- and Post-Prebiotic Treatment LBP and LCAT Activity Levels in Healthy Young Adults with Participants Stratified by Above or Below Group LBP Median at Baseline^*

	Above Median			Below Median
Variable	Pre	Post	P-value	Pre	Post	P-value
LBP (ug/mL)	9.27±3.52	7.02±2.32	0.003	3.97±1.56	4.32±2.13	0.597
LCAT Activity Index	1.067 ± 0.01	1.057 ± 0.02	0.188	1.036 ± 0.01	1.036 ± 0.02	0.941

^*Data are shown as means ± SDs. P-values were calculated from a linear mixed effect model with a post-hoc analysis.

Abbreviations: LBP, Lipopolysaccharide Binding Protein; LCAT, Lecithin cholesteryl acyltransferase

A significant positive correlation was found between LCAT activity and high-density lipoprotein cholesterol (HDL-C) (r = 0.391, p-value = 0.018), low-density lipoprotein cholesterol (LDL-C) (r = 0.343, p-value = 0.034), and total cholesterol concentrations (r = 0.382, p-value = 0.018), which remained significant after adjusting for multiple comparisons (p < 0.05) (Table 3). A positive correlation between HDL-C and LBP concentration (measure r = 0.271, p = 0.035) was found, however it did not remain significant after correction for multiple testing, and no other statistically significant correlations were found between clinical characteristics and LBP concentration (Table 3).

Table 3.

Association of Clinical Characteristics of Healthy Young Adults with concentration of LBP and LCAT activity.

	LBP (ug/mL)				LCAT Activity Index
Variable	Rmcorr	95% CI	P-value	Padjust	Rmcorr	95% CI	P-value	Padjust
Age, y	0.02	−0.23, −0.23	0.864	0.930	0.12	−0.14, −0.14	0.370	0.576
Weight, kg	−0.05	−0.3, −0.3	0.708	0.930	−0.22	−0.45, −0.45	0.087	0.244
Height, cm	−0.08	−0.32, −0.32	0.563	0.930	−0.08	−0.33, −0.33	0.527	0.726
BMI, kg/m2	0	−0.25, −0.25	1.000	1.000	−0.16	−0.4, −0.4	0.222	0.444
Systolic Blood Pressure, mmHg	0.1	−0.15, −0.15	0.424	0.930	0.07	−0.18, −0.18	0.571	0.726
Diastolic Blood Pressure, mmHg	0.13	−0.13, −0.13	0.320	0.930	0.17	−0.09, −0.09	0.194	0.444
Fasting glucose, mg/dL	0.02	−0.23, −0.23	0.857	0.930	−0.05	−0.3, −0.3	0.730	0.851
Fasting insulin, μIU/mL	0.07	−0.19, −0.19	0.607	0.930	−0.01	−0.26, −0.26	0.932	0.932
Total Cholesterol, mg/dL	0.13	−0.13, −0.13	0.334	0.930	0.38	0.14, 0.14	0.003	0.018
HDL cholesterol, mg/dL	0.27	0.02, 0.02	0.035	0.489	0.39	0.15, 0.15	0.002	0.018
LDL cholesterol, mg/dL	0.02	−0.23, −0.23	0.855	0.930	0.34	0.1, 0.1	0.007	0.034
Cholesterol:HDL	−0.04	−0.29, −0.29	0.753	0.930	−0.03	−0.28, −0.28	0.835	0.899
Triglyceride, mg/dL	0.04	−0.22, −0.22	0.782	0.930	−0.13	−0.37, −0.37	0.327	0.572
non-HDL	0.04	−0.21, −0.21	0.757	0.930	0.28	0.03, 0.03	0.032	0.112

^*Rmcorr was used to assess the within-individual association between clinical variables and outcomes (LBP or LCAT activity) over time. Rmcorr accounts for the non-independence of repeated pre- and post-treatment measurements from each participant, and estimates the common within-person correlation. 95% CI = 95% confidence interval. Padj = p-values adjusted for multiple comparisons using the Benjamini-Hochberg method. Bold padj < 0.05 indicates statistically significant correlation after adjustment. HDL, high-density lipoproteins; LDL, low-density lipoproteins.

Abbreviations: Rmcorr, repeated measures correlation; LBP, Lipopolysaccharide Binding Protein; LCAT, Lecithin cholesteryl acyltransferase, HDL, High-density lipoproteins; LDL, low-denisty lipoproteins

In the assessment of relationships between 185 microbial genera and LBP, we found significant correlations between LBP concentrations and 3 microbes that were positively correlated and 2 that were negatively correlated (P<0.05), however these relationships were no longer significant after correction for multiple testing, and several were not present at detectable levels in the majority of participants (Supplemental Table S1). Similarly, 11 microbes showed positive correlations with LCAT activity, while 3 were negatively correlated (P<0.05). However, these findings also lost significance upon multiple testing correction (Supplemental Table S2).

Out of 890 plasma metabolites analyzed, 35 displayed positive correlations and 15 showed negative correlations with LBP concentrations (P<0.05), none of which remained statistically significant after multiple testing correction (Figure 2, Supplemental Table S3). LBP concentrations were predominantly correlated with a variety of lipids and xenobiotics. Notably, LBP was positively correlated with several sulfated bile acid species such as taurodeoxycholic acid 3-sulfate and taurocholic acid sulfate, the sulfated estrogen metabolite estrone 3-sulfate, as well as perfluorooctanesulfonate (PFOS), an industrial chemical pollutant that is also used in food packaging. Positively correlated metabolites also included those associated with xenobiotic metabolism pathways, including a positive correlation with multiple metabolites of ibuprofen. In contrast, negative correlations were observed between LBP and several palmitate-containing glycerophospholipids, exemplified by species such as 1-(1-enyl-palmitoyl)-GPC and 1,2-dipalmitoyl-GPC.

Figure 2:

Correlations of LBP Concentrations and LCAT Activity with Plasma Metabolites. A) The top 30 metabolites with statistically significant correlations with LBP concentrations are shown (p<0.05). B) The top 30 metabolites that showed a statistically significant correlation with LCAT activity are shown (p<0.05). None of these correlations remained statistically significant after correction for multiple testing. The metabolites are categorized by super pathways, showing the diverse biochemical domains they belong to, including amino acids, carbohydrates, lipids, cofactors and vitamins, nucleotides, and xenobiotics. LBP, lipopolysaccharide binding protein; LCAT, lecithin-cholesterol acyltransferase

In relation to LCAT activity, 15 metabolites were positively correlated and 18 were negatively correlated (P<0.05), which did not remain statistically significant after multiple testing correction (Figure 2, Supplemental Table S4). LCAT activity was similarly primarily associated with lipid metabolism, including positive correlations with lipids such as palmitate-, stearate-, oleate- and linoleate-containing glycerophospholipids and acylcholines, as well as with cholesterol and vitamin A (retinol).

Out of 2,067 microbial genes, 14 displayed a positive correlation with LBP concentrations, while 8 exhibited an inverse relationship (P<0.05) which did not remain significant after multiple testing correction (Figure 3, Supplemental Table S5). Among these, 3-deoxy-manno-octulosonate cytidylyltransferase (kdsB), which is involved in the synthesis of LPS by gram-negative bacteria, was positively correlated with LBP. Conversely, 2-deoxystreptamine N-acetyl-D-glucosaminyltransferase (btrM), responsible for the biosynthesis of butirosin, an aminoglycoside antibiotic complex that deactivates both gram-positive and gram-negative bacteria, was negatively correlated.

Figure 3:

Correlations of LBP Concentrations and LCAT Activity with Microbial Gene Abundance. A) The top 20 microbial genes correlated with LBP concentrations are shown (p<0.05), and B) the top 20 microbial genes correlated with LCAT activity are shown (p<0.05), measured by the repeated measures correlation coefficient (rmcorr). LBP, lipopolysaccharide binding protein; LCAT, lecithin-cholesterol acyltransferase; rmcorr, repeated measures correlation.

Two microbial genes were found to be positively correlated and 42 genes were negatively correlated with LCAT activity (P<0.05), none of which remained statistically significant after multiple testing correction (Figure 3, Supplementary Table S6). Notable genes inversely correlated with LCAT activity included UDP-N-acetyl-D-mannosaminuronic acid transferase (wecG), contributing to the synthesis of a lipid-linked intermediate in the formation of enterobacterial outer membranes, and sporulation kinase D (kinD), pivotal in activating a sporulation-regulatory protein. Furthermore, the vancomycin C-type resistance protein VanC (vanC), involved in conferring bacterial resistance to the antibiotic vancomycin, also displayed a significant negative correlation with LCAT activity.

Discussion

In this study, the effects of prebiotic fiber supplementation on plasma LBP concentrations and LCAT activity in generally healthy, young individuals consuming a low-fiber diet were investigated. There were no statistically significant differences in LBP concentration or LCAT activity between the placebo and prebiotic arms. The lack of effect of the fiber supplement on LBP concentrations in the overall cohort may be due to several potential reasons including that a higher dose of fiber or longer intervention may be necessary to induce measurable effects, and that these generally young, healthy participants may not yet be experiencing the deleterious effects of a low-fiber diet that is seen in older adults or those who have already developed disease (35,36).

We performed an exploratory analysis dividing the group into those with above-median vs. below-median LBP concentrations at baseline and found that in those participants with above-median baseline LBP, fiber supplementation significantly reduced LBP concentrations. Thus, even this low dose of fiber supplement, which was designed to achieve levels of fiber intake consistent with the recommended dietary allowance, benefitted young, healthy individuals who already had some degree of gut barrier impairment. Multiple cross-sectional studies have found plasma LBP concentrations to be negatively correlated with dietary fiber intake (36–38) as well as indices of healthy diet (e.g. the dietary inflammatory index) (39).

An interesting question is whether the changes in LBP concentration are evidence of changes in intestinal permeability or whether they are a reflection of the LPS load in the gut. Animal studies show that changes in plasma LBP are correlated with changes in paracellular flux directly measured in ex vivo intestinal sections using Ussing chambers (40), suggesting that LBP is a marker of intestinal permeability. We performed an additional analysis to determine whether the individuals whose LBP concentrations were above-median at baseline had increases in LPS-production genes after the fiber arm. Intriguingly, we found the opposite to be true. Whereas participants with higher baseline LBP concentrations (i.e. higher gut leakiness) had a decrease in LBP after the fiber arm, there were no changes in LPS production genes in their microbiota. In the participants whose baseline LBP concentrations were below-median (i.e. lower gut leakiness), there were no significant changes in LPS production genes in their microbiomes after the fiber arm (Supplemental Fig. 1A). The lack of significance suggests that the change of LPS-producing genes may not fully explain the LBP changes observed in the entire group. These data indicate that while LBP is likely a marker of intestinal permeability, it does not necessarily reflect LPS load in the gut. We further hypothesized that participants with above-median LBP may have had less healthy microbiomes at baseline (i.e. more LPS-producing genes and lower abundances of bifidobacteria) but that they had a stronger bifidogenic response to the fiber supplement compared to the participants with below-median LBP. However, none of these differences were statistically significant, suggesting that other differences in baseline microbiomes and microbiome changes in response to the fiber intervention accounted for these differences in changes in gut permeability. This study was likely inadequately powered to detect differences in these subgroups. Future studies with larger sample sizes are needed to better understand the fiber-mediated microbiome changes that confer the beneficial effects on gut permeability.

It is important to note that this study did not measure inflammatory markers such as C-reactive protein (CRP), TNF-alpha, or IL-6, which limits the ability to assess the anti-inflammatory potential of the fiber supplement. Future studies should include these measurements to better understand the possible anti-inflammatory effects of fiber supplementation. Additionally, the lack of statistically significant differences in LBP concentration between the prebiotic and placebo arms may be attributed to the relatively small sample size or to differences in the baseline LBP levels of participants, which may influence their response to the intervention. Future studies with larger sample sizes are needed to confirm these findings and to explore the relationship between baseline LBP concentrations and the effectiveness of fiber supplementation.

We performed exploratory correlation analysis to determine if there were any relationships between gut microbiome composition, function, or plasma metabolites and both LBP and LCAT activity, in order to elucidate some of the microbiome-associated factors that might influence gut barrier function and HDL function. Notably, we found positive correlations between LBP and multiple sulfated molecules. Certain sulfated bile acids like taurolithocholic acid 3-sulfate are used to induce pancreatitis in mouse models, which may suggest a pathological role for this molecule (41). Interestingly, we also observed a positive correlation between LBP and PFOS, an industrial chemical that is used in a variety of industrial and consumer products (42), as well as in food packaging (43,44), and which has been shown to be associated with an array of negative health effects including high cholesterol, increased liver enzymes, thyroid disorders, immunotoxicity and cancer (45,46). Elevated concentrations of sulfated bile acids and sulfated estrone could indicate an increase in the activity of sulfotransferase enzymes in the liver, which are regulated by a wide variety of nuclear transcription factors (47) and which may be linked with the gut microbiome (48,49). The correlations between LBP and ibuprofen metabolites point to a potential relationship between non-steroidal anti-inflammatory medication use and impairment of gut barrier function, as has been observed previously (50). Together these data point to an important potential link between xenobiotic exposure and detoxification, and gut permeability. These relationships need to be further explored in larger studies that are adequately powered to detect statistically significant differences.

Furthermore, several associations between LBP concentrations and abundance of microbial genes were found. For instance, the positive correlation between kdsB, which is involved in LPS synthesis by gram-negative bacteria (51), and plasma LBP concentrations, point to a potential relationship between LPS synthesis genes andLPS translocation across the gut barrier. On the other hand, the btrM gene, which is involved in the synthesis of an aminoglycoside that acts as an antibiotic (52,53) was negatively correlated with LBP, highlighting potential links between endogenous production of antibiotic molecules and lower translocation of LPS across the gut barrier. Together, these findings have generated interesting hypotheses for further exploration in future studies of gut permeability. The impacts of dietary interventions that can alter gut barrier permeability and the associated translocation of pro-inflammatory LPS, which has a well-established, pro-inflammatory role across multiple chronic disease states (54–57), are an important area for future research.

When examining the relationships with LCAT activity, inverse correlations were found with bacterial genes like wecG and kinD, which are involved in the synthesis of enterobacterial cell wall components and sporulation proteins respectively (58,59). Likewise, the vanC gene, which enables bacterial resistance to the antibiotic vancomycin (60), was also negatively correlated with LCAT activity. These intriguing exploratory findings suggest that there are potential as yet underexplored relationships between antibiotic exposure, and increased abundances of certain classes of bacteria (e.g. enterobacteria and spore-forming bacteria) that may be linked with decreased HDL functionality, which merit further study. Future research should incorporate hypothesis-driven comparisons to identify specific associations and provide a more comprehensive understanding of the mechanisms involved.

Conclusion

Collectively, this secondary analysis study showed that a low-dose fiber supplement taken for 4 weeks did not decrease gut permeability to LPS, as measured by plasma LBP, or improve HDL functionality, as measured by LCAT activity, in young, generally healthy individuals. However, those participants who had above-median concentrations of LBP at baseline did experience a reduction in LBP in response to the fiber supplement, indicating that even young, healthy adults can have some degree of gut permeability and that this can be ameliorated with a simple, low-dose, easy to use fiber supplement. Furthermore, exploratory correlation analyses uncovered potential links between gut microbiome function and both gut permeability and HDL functionality, highlighting the potential involvement of pathways in LPS synthesis, antibiotic synthesis, and antibiotic resistance. Exploratory analyses further revealed potential links between plasma metabolites related to xenobiotic detoxification and LBP concentrations, including industrial chemicals (PFOS), sulfated bile acids, and ibuprofen metabolites. These findings provide insights into connections between indicators of gut permeability, HDL function, and the gut microbiome—warranting future targeted studies to confirm findings and elucidate mechanisms. Our findings showcase the potential of affordable, easy to implement fiber supplementation as an approach to mitigate gut barrier dysfunction-associated inflammation in individuals who are low-fiber consumers, and point to potential environmental and lifestyle factors that could be contributing to impaired gut barrier function in addition to diet.

Abbreviations

CPM

Count per million

HDL

High-density lipoprotein

HDL-C

High-density lipoprotein cholesterol

IPA

Indolepropionate

LBP

Lipopolysaccharide binding protein

LCAT

Lecithin cholesterol acyltransferase

LDL

Low-density lipoprotein

LDL-C

Low-density lipoprotein cholesterol

LPS

Lipopolysaccharide

OTUs

Operational taxonomic units

PFOS

Perfluorooctanesulfate

TLR4

Toll-like receptor 4

VLDL

Very low-density lipoprotein