Gene expression in blood reflects smoking exposure among cancer-free women in the Norwegian Women and Cancer (NOWAC) postgenome cohort

Abstract

Active smoking has been linked to modulated gene expression in blood. However, there is a need for a more thorough understanding of how quantitative measures of smoking exposure relate to differentially expressed genes (DEGs) in whole-blood among ever smokers. This study analysed microarray-based gene expression profiles from whole-blood samples according to smoking status and quantitative measures of smoking exposure among cancer-free women (n = 1708) in the Norwegian Women and Cancer postgenome cohort. When compared with never smokers and former smokers, current smokers had 911 and 1082 DEGs, respectively and their biological functions could indicate systemic impacts of smoking. LRRN3 was associated with smoking status with the lowest FDR-adjusted p-value. When never smokers and all former smokers were compared, no DEGs were observed, but LRRN3 was differentially expressed when never smokers were compared with former smokers who quit smoking ≤ 10 years ago. Further, LRRN3 was positively associated with smoking intensity, pack-years, and comprehensive smoking index score among current smokers; and negatively associated with time since cessation among former smokers. Consequently, LRRN3 expression in whole-blood is a molecular signal of smoking exposure that could supplant self-reported smoking data in further research targeting blood-based markers related to the health effects of smoking.

Figure 1

Descriptive statistics of study participants by smoking status for (A) age at blood collection, (B) body mass index (BMI) at blood collection, (C) alcohol consumption at baseline, (D) smoking intensity, (E) smoking duration, (F) time since smoking cessation (TSC), (G) pack-years, and (H) comprehensive smoking index (CSI) scores. Yellow, blue, and red coloured violin plots represent kernel density estimates for never, former, and current smokers, respectively. White boxes extend from the 25th to the 75th percentile, vertical bars inside the box represent the median, whiskers extend 1.5 times the length of the interquartile range right and left side of the 75th and 25th percentiles respectively, and outliers are represented as black dots. The green diamond shaped dot represents the respective mean.

Figure 2

Volcano plots for the test statistics in fully-adjusted models from the tests of differentially expressed genes (DEGs) in comparisons of (A) current versus never smokers, (B) current versus former smokers, and (C) former versus never smokers; and forest plots for the 10 top-ranked DEGs in tests of DEGs in comparisons of (D) current versus never smokers and (E) current versus former smokers. In volcano plots (A–C), red dots display up-regulated genes, blue dots display down-regulated genes, while grey dots display genes with FDR > 0.05; the x-axis presents log₂ fold-changes and the y-axis presents − log10 of FDR adjusted p-values; and gene names displayed are the 20 top-ranked DEGs in the respective tests. In forest plots (D and E), dots in the x-axis represent log₂ fold-changes and the y-axis represents DEGS with the lowest FDR adjusted p-values ranked from the top; the horizontal line for each gene represents their confidence interval; and the vertical blue dotted line represents no difference.

Figure 3

Distributions of expression values for the top-ranked gene (LRRN3) (A) among never (yellow) and former (blue) smokers and (B) among current smokers according to comprehensive smoking index (CSI) scores. In figure (A), boxes extend from the 25th to the 75th percentile, horizontal bars represent the median, whiskers extend 1.5 times the length of the interquartile range above and below the 75th and 25th percentiles, respectively, and outliers are represented as points. In figure (B), the red line represents the linear regression fit and the shaded grey area its standard error.

Figure 4

Distributions of expression values for the top-ranked gene (LRRN3) (A) among never (yellow) and current (red) smokers and (B) among former smokers according to time since smoking cessation (TSC). In figure (A), boxes extend from the 25th to the 75th percentile, horizontal bars represent the median, whiskers extend 1.5 times the length of the interquartile range above and below the 75th and 25th percentiles, respectively, and outliers are represented as points. In figure (B), the blue line represents the linear regression fit and the shaded grey area its standard error.

Figure 5

Summary of functional enrichment analyses for up- and down-regulated genes for the (A) GO(BP) and (B) REACTOME pathway databases. The colour of the dots indicates the adjusted p-value, where red dots represent the most enriched categories; the ‘GeneRatio’ indicates the proportion of genes overlapping between lists of differentially expressed genes (DEGs) and the genes in gene ontology categories. GO: gene ontology; BP: biological processes; CS-vs-NS: comparison of current smokers versus never smokers; CS-vs-FS: comparison of current smokers versus former smokers; U: Up-regulated genes; D: Down-regulated genes.