Colonic mucosal and exfoliome transcriptomic profiling and fecal microbiome response to a flaxseed lignan extract intervention in humans

Abstract

Background: Microbial metabolism of lignans from high-fiber plant foods produces bioactive enterolignans, such as enterolactone (ENL) and enterodiol (END). Enterolignan exposure influences cellular pathways important to cancer risk and is associated with reduced colon tumorigenesis in animal models and lower colorectal cancer risk in humans.

Objectives: The aim of this study was to test the effects of a flaxseed lignan supplement (50 mg secoisolariciresinol diglucoside/d) compared with placebo on host gene expression in colon biopsies and exfoliated colonocyte RNA in feces and fecal microbial community composition, and to compare responses in relation to ENL excretion.

Methods: We conducted a 2-period randomized, crossover intervention in 42 healthy men and women (20-45 y). We used RNA-seq to measure differentially expressed (DE) genes in colonic mucosa and fecal exfoliated cells through the use of edgeR and functional analysis with Ingenuity Pathway Analysis. We used 16S ribosomal RNA gene (V1-V3) analysis to characterize the fecal microbiome, and measured END and ENL in 24-h urine samples by gas chromatography-mass spectrometry.

Results: We detected 32 DE genes (false discovery rate <0.05) in the exfoliome, but none in the mucosal biopsies, in response to 60 d of lignan supplement compared with placebo. Statistically significant associations were detected between ENL excretion and fecal microbiome measured at baseline and at the end of the intervention periods. Further, we detected DE genes in colonic mucosa and exfoliome between low- and high-ENL excreters. Analysis of biopsy samples indicated that several anti-inflammatory upstream regulators, including transforming growth factor β and interleukin 10 receptor, were suppressed in low-ENL excreters. Complementary analyses in exfoliated cells also suggested that low-ENL excreters may be predisposed to proinflammatory cellular events due to upregulation of nuclear transcription factor κB and NOS2, and an inhibition of the peroxisome proliferator-activated receptor γ network.

Conclusions: These results suggest that ENL or other activities of the associated gut microbial consortia may modulate response to a dietary lignan intervention. This has important implications for dietary recommendations and chemoprevention strategies. This study was registered at clinicaltrials.gov as NCT01619020.

Keywords: colon; enterolactone; fecal microbiome; gene expression; human intervention; lignan; secoisolariciresinol.

Copyright © American Society for Nutrition 2019.

Figures

FIGURE 1

Recruitment and study activities of the FlaxFX study. FFQ, food-frequency questionnaire; Sig, sigmoidoscopy.

FIGURE 2

Urinary ENL excretion (24 h) in participants at the end of 60-d placebo (triangles) and lignan (squares) intervention periods. Participants (n = 42) were categorized as low (open symbols) and high (solid symbols) ENL excreters based on median ENL excretion after lignan supplementation (23.4 µmol/24 h). ENL, enterolactone.

FIGURE 3

Venn diagrams representing detectable transcripts in colonic biopsies and exfoliated cells. (A) A total of 21,011 genes (including noncoding genes, pseudogenes, and protein-coding gene transcripts) were detected in both gut mucosal biopsy samples and exfoliated cells, accounting for 92% of genes in exfoliated cells and 55% of genes in colonic biopsy samples (n = 219 biopsy samples and 54 exfoliated samples). (B) Data preprocessing and filtering resulted in the identification of 15,120 genes from the biopsy samples and 4577 genes from the exfoliated epithelial cells samples.

FIGURE 4

Comparison of gene expression in stroma and epithelial tissue from biopsies and exfoliome in stool. The exfoliome signature arises from cells sloughed from both the small intestine and colon and comprises reads from a diverse array of cell types expected to be found in the intestinal mucosa. (A) Heatmap showing counts of genes that are reported to be primarily expressed at specific anatomic locations (stomach, pancreas, small intestine, colon). All genes with counts >400 are dark blue. (B) Heatmap showing counts of biomarker genes from each cell type and each data source. N = 219 biopsy samples and 54 exfoliated samples. Abbreviations in Supplemental Table 7 and at https://www.ncbi.nlm.nih.gov/gene/.

FIGURE 5

Overlap of DE genes. (A) Depiction of the overlap between the DE (FDR n = 54 samples). LDA identified genes (bolstered resubstitution error <0.227), and the genes that successfully passed the low variance filter. Testing for intervention effects in exfoliated cell samples revealed 973 DE (FDR <0.05) genes. Among those genes, 32 (star) were also identified in 2-gene LDA with bolstered resubstitution error <0.227 and passed the low variance postprocessing step. (B) Depiction of the overlap between the DE (FDR <0.05) protein-coding genes with respect to low- compared with high-ENL excreter status in biopsies (n = 219 samples) and exfoliated cells (n = 54 samples). See Supplemental Tables 2 and 3 for additional details. DE, differentially expressed; ENL, enterolactone; FDR, false detection rate; LDA, linear discriminant analysis.

FIGURE 6

Differentially expressed genes and upstream regulators in colon biopsies and exfoliated cells. Colon biopsies: TGFβ1 (A), IL-10 receptor (B). Exfoliated cells: NF-κB (C) and NOS2 (D). z Scores >2.0, P values <0.05. See Supplemental Tables 4 and 5 for additional details. Data are derived from 219 biopsy samples and 54 exfoliated samples that had a fold change of >1.3 or <−1.3 and an FDR of <0.05. IL-10, interleukin 10; NF-κB, nuclear transcription factor κB; TGFβ1, transforming growth factor β. Abbreviations in Supplemental Table 7 and at https://www.ncbi.nlm.nih.gov/gene/.

FIGURE 7

Exfoliated cell network. PPARγ signaling activity network was the highest scoring network, with low-ENL excreters https://www.ncbi.nlm.nih.gov/gene/.

FIGURE 8

Plot based on the first two PCoA axes from unweighted Unifrac. A line connects within-person placebo (open circle) and lignan intervention (star) gut microbiome samples, with each participant colored by log(µmol/24-h) urinary ENL excretion at end of the lignan intervention period (n = 33). ENL, enterolactone; PCoA, principal coordinates analysis.