Associations between dietary patterns and gene expression profiles of healthy men and women: a cross-sectional study

Abstract

Background: Diet regulates gene expression profiles by several mechanisms. The objective of this study was to examine gene expression in relation with dietary patterns.

Methods: Two hundred and fifty four participants from the greater Quebec City metropolitan area were recruited. Two hundred and ten participants completed the study protocol. Dietary patterns were derived from a food frequency questionnaire (FFQ) by factor analysis. For 30 participants (in fasting state), RNA was extracted from peripheral blood mononuclear cells (PBMCs) and expression levels of 47,231 mRNA transcripts were assessed using the Illumina Human-6 v3 Expression BeadChips®. Microarray data was pre-processed with Flexarray software and analysed with Ingenuity Pathway Analysis (IPA).

Results: Two dietary patterns were identified. The Prudent dietary pattern was characterised by high intakes of vegetables, fruits, whole grain products and low intakes of refined grain products and the Western dietary pattern, by high intakes of refined grain products, desserts, sweets and processed meats. When individuals with high scores for the Prudent dietary pattern where compared to individuals with low scores, 2,083 transcripts were differentially expressed in men, 1,136 transcripts in women and 59 transcripts were overlapping in men and women. For the Western dietary pattern, 1,021 transcripts were differentially expressed in men with high versus low scores, 1,163 transcripts in women and 23 transcripts were overlapping in men and women. IPA reveals that genes differentially expressed for both patterns were present in networks related to the immune and/or inflammatory response, cancer and cardiovascular diseases.

Conclusion: Gene expression profiles were different according to dietary patterns, which probably modulate the risk of chronic diseases.

Trial registration: NCT: NCT01343342.

Figure 1

Flowchart illustrating the significantly different transcripts according to scores for the Prudent dietary pattern.

Figure 2

Flowchart illustrating the significantly different transcripts according to scores for the Western dietary pattern.

Figure 3

The modified canonical pathways according to scores for the Prudent dietary pattern. Legend: Gene expression differences (≥ ± 1.2 fold change) in canonical pathways comparing high and low scores for the Prudent dietary pattern (A) For the men (B) For the women. P-values for Functional Analysis of dataset by IPA (Fisher’s Exact Test) are presented. % of genes significantly up-regulated and down-regulated in each canonical pathway are presented (number of genes differently expressed/number of genes in the pathways*100). (A) PDGF Signaling: P = 0.004, 14 genes significantly different; ERK/MAPK Signaling: P = 0.004, 29 genes significantly different; FcyRIIB Signaling in B Lymphocytes: P = 0.003, 10 genes significantly different; B Cell Receptor Signaling: P = 0.002, 24 genes significantly different; CTLA4 Signaling in Cytotoxic T Lymphocytes: P = 0.003, 17 genes significantly different; PI3K Signaling in B Lymphocytes: P = 0.0009, 23 genes significantly different; Leucocyte Extravasation Signaling: P = 0.0002, 32 genes significantly different. (B) Signaling by Rho Family GTPase: P = 0.02, 20 genes significantly different; B Cell Receptor Signaling: P = 0.02, 14 genes significantly different; Role of NFAT in Regulation of the Immune Response: P = 0.01, 16 genes significantly different; RhoGDI Signaling: P = 0.009, 17 genes significantly different; Cyclins and Cell Cycle Regulation: P = 0.007, 10 genes significantly different; Breast Cancer Regulation by Stathmin1: P = 0.0005, 22 genes significantly different; B Cell Development: P = 0.0004, 7 genes significantly different.

Figure 4

The modified canonical pathways according to scores for the Western dietary pattern. Legend: Gene expression differences (≥ ± 1.2 fold change) in canonical pathways comparing high and low scores for the Western dietary pattern (A) For the men (B) For the women. P-values for Functional Analysis of dataset by IPA (Fisher’s Exact Test) are presented. % of genes significantly up-regulated and down-regulated in each canonical pathway are presented (number of genes differently expressed/number of genes in the pathways*100). (A) Endothelin-1 Signaling: P = 0.01, 12 genes significantly different; PTEN Signaling: P = 0.01, 10 genes significantly different; iCOS-iCOSL Signaling in T Helper Cells: P = 0.008, 9 genes significantly different; Hypoxia Signaling in the Cardiovascular System: P = 0.006, 7 genes significantly different; Xenobiotic Metabolism Signaling: P = 0.005, 18 genes significantly different; NGF Signaling: P = 0.004, 10 genes significantly different; Neurotrophin/TRK Signaling: P = 0.00001, 11 genes significantly different. (B) IL-8 Signaling: P = 0.04, 15 genes significantly different; FXR/RXR Activation: P = 0.05, 9 genes significantly different; Atherosclerosis Signaling: P = 0.02, 12 genes significantly different; Role of Cytokines in Mediating Communication between Immune Cells: P = 0.02, 7 genes significantly different; Toll-like Receptor Signaling: P = 0.01, 7 genes significantly different; Role of NFAT in Regulation of the Immune Response: P = 0.01, 16 genes significantly different; Bladder Cancer Signaling: P = 0.002, 12 genes significantly different; Nf-κB Signaling: P = 0.002, 18 genes significantly different.