Kumaraswamy Naidu Chitrala

Discrimination is a social adversity that is linked to several age‐related outcomes. However, the molecular drivers of these observations are poorly understood. Social adverse factors are associated with proinflammatory and interferon gene expression, but little is known about whether additional genes are associated with discrimination among both African American and White adults. In this study, we examined how perceived discrimination in African American and White adults was associated with genome‐wide transcriptome differences using RNA sequencing. Perceived discrimination was measured based on responses to self‐reported lifetime discrimination and racial discrimination. Differential gene expression and pathway analysis were conducted in a cohort ( N = 59) stratified by race, sex, and overall discrimination level. We found 28 significantly differentially expressed genes associated with race among those reporting high discrimination. Several of the upregulated genes for African American versus White adults reporting discrimination were related to immune function IGLV2‐11 , S100B , IGKV3‐20 , and IGKV4‐1 ; the most significantly downregulated genes were associated with immune modulation and cancer, LUCAT1 , THBS1 , and ARPIN . The most enriched gene ontology biological process between African American and White men reporting high discrimination was the regulation of cytokine biosynthetic processes. The immune response biological process was significantly lower for African American women compared to White women reporting high discrimination. Discrimination was associated with the expression of small nucleolar RNAs, long noncoding RNAs, and microRNAs associated with energy homeostasis, cancer, and actin. Understanding the pathways through which adverse social factors like discrimination are associated with gene expression is crucial in advancing knowledge of age‐related health disparities.

Differentially abundant microbial features. Taxonomic and functional profiles obtained with HUMAnN 3.0 were CLR transformed and differential abundance analysis was performed with MaAsLin2 setting FDR cutoff to 0.05. A, Bar plots of the top differentially abundant genera in the tumor versus HC and adjacent normal versus HC contrasts. B, A heatmap showing clustering of samples based on top differentially abundant enzyme classes. The plots were created with ggplot2 and pheatmap R packages.

Microbial loads and transcriptional profiles. DNA extracted from the same cells as for RNA was used to determine bacterial, fungal, and archaeal loads in the samples relative to human β-actin gene by qPCR (A). Nonhuman, nonribosomal RNA sequences were used as input for HUMAnN 3.0 for microbial taxonomic and functional profiling. Microbial phyla (B) and top 20 genera (C) identified in each of the study groups ranked by their average transcriptional relative abundances. D, A PCA plot based on CLR-transformed genus-level profiles (created with Phyoseq and microbiome R packages). E, Top 20 expressed enzyme classes ranked by their average transcriptional abundances (CLR-transformed counts) in each sample type.

Host pathways (GSEA genesets) correlating with microbial enzyme classes

HL expression by Cutibacterium acnes. Relative abundance of HL transcripts in individual samples. The transcripts were exclusively contributed by Cutibacterium acnes based on HUMAnN 3.0 results.

Host transcriptome. Sequences were mapped to human reference GRCh38 and quantified at the gene level using STAR aligner and HTSeq, respectively. DESeq2 was used to identify DEGs which are depicted in volcano plots for tumor versus adjacent normal (paired comparison; A), tumor versus HC (B), and adjacent normal versus HC (C)—the color code in the volcano plots corresponds to the fold change and FDR cutoffs (2 and 0.05, respectively). D, A PCA plot based on the most variable 1,500 DEGs, generated using “plotPCA" function from DESeq2 R/Bioconductor package. Lists of DEGs preranked by Wald statistic were used as input for GSEA to identify upregulated and downregulated pathways in each contrast (E–G) based on Hallmark gene sets (MSigDB 7.1).

Gene Set Enrichment Analysis results (Host pathways)

Microbial enzyme classes (EC) and metabolic pathway profiles.

HL expression by Cutibacterium acnes. Relative abundance of HL transcripts in individual samples. The transcripts were exclusively contributed by Cutibacterium acnes based on HUMAnN 3.0 results.