Acute and Chronic Changes in Gene Expression After CMV DNAemia in Kidney Transplant Recipients

Abstract

Cytomegalovirus (CMV) viremia continues to cause significant morbidity and mortality in kidney transplant patients with clinical complications including organ rejection and death. Whole blood gene expression dynamics in CMV viremic patients from onset of DNAemia through convalescence has not been well studied to date in humans. To evaluate how CMV infection impacts whole blood leukocyte gene expression over time, we evaluated a matched cohort of 62 kidney transplant recipients with and without CMV DNAemia using blood samples collected at multiple time points during the 12-month period after transplant. While transcriptomic differences were minimal at baseline between DNAemic and non-DNAemic patients, hundreds of genes were differentially expressed at the long-term timepoint, including genes enriching for pathways important for macrophages, interferon, and IL-8 signaling. Amongst patients with CMV DNAemia, the greatest amount of transcriptomic change occurred between baseline and 1-week post-DNAemia, with increase in pathways for interferon signaling and cytotoxic T cell function. Time-course gene set analysis of these differentially expressed genes revealed that most of the enriched pathways had a significant time-trend. While many pathways that were significantly down- or upregulated at 1 week returned to baseline-like levels, we noted that several pathways important in adaptive and innate cell function remained upregulated at the long-term timepoint after resolution of CMV DNAemia. Differential expression analysis and time-course gene set analysis revealed the dynamics of genes and pathways involved in the immune response to CMV DNAemia in kidney transplant patients. Understanding transcriptional changes caused by CMV DNAemia may identify the mechanism behind patient vulnerability to CMV reactivation and increased risk of rejection in transplant recipients and suggest protective strategies to counter the negative immunologic impact of CMV. These findings provide a framework to identify immune correlates for risk assessment and guiding need for extending antiviral prophylaxis.

Keywords: CMV DNAemia; RNA-seq; kidney transplant; transcriptomics; transplant immunology.

Figure 1

Volcano plots of genes that are differentially expressed between (A) baseline and week 1; (B) baseline and month 1; (C) baseline and long-term in patients that developed DNAemia. Gray points represent genes that are not differentially expressed, green points represent genes that have a log₂FC ≥ 2 but not differentially expressed, blue points represent genes that are differentially expressed (FDR ≤ 0.1) but have a log₂FC < 2, and red points represent genes that are both differentially expressed (FDR ≤ 0.1) and have a log₂FC ≥ 2.

Figure 2

Bar plot of IPA canonical pathway analysis results comparing baseline and week 1 (first column), baseline and month 1 (second column), and baseline and long-term (third column) for immunologic pathways. The significance of an enriched pathway is represented by the length of each bar, where length corresponds to the -log(p-value). The number of genes enriching for a pathway is indicated in the bar. The vertical dotted black line corresponds to a p-value of 0.05. Positive IPA Z-scores are indicated in red, and negative Z-scores are indicated in blue. Gray coloring is used for pathways in which direction of change cannot be confidently predicted.

Figure 3

Heatmap of clustered DEGs across all timepoints for patients with CMV viremia. DEGs with |logFC| ≥ 2 at baseline versus week 1 post-viremia are shown. Red indicates increased expression, while blue indicates decreased expression compared with baseline. CMV IgG status of transplant recipient is indicated, with high-risk patients (D+/R−) shown in pink and intermediate-risk (R+) in orange. Timepoints are indicated as baseline (brown), week 1 (turquoise), month 1 (blue), or long-term (green) in bars at the top of the figure. Each column represents an independent patient sample.

Figure 4

(A) Time-course dynamics of IPA canonical pathways significantly enriched for with genes that were differentially expressed between baseline and week 1 post-viremia. Red indicates that the median of standardized gene expression of genes in a pathway are >0, while blue indicates that the median of standardized gene expression of genes in a pathway are <0. Each column represents a different time point, namely, baseline, week 1, month 1, and long-term. Most pathways have two significant time-trend as identified by time-course gene set analysis because some genes in a given pathway have a positive median standardized gene expression, while the other genes in the same pathway have a negative median standardized gene expression. (B) Examples of significantly enriched IPA pathways with two time trends, CD27 Signaling in Lymphocytes Pathway and Stat3 Pathway. Interferon Signaling Pathway has only one time trend.

Figure 5

Differential expression analysis at the long-term timepoint between DNAemia and no DNAemia. (A) Volcano plot of differentially expressed genes. Gray points represent genes that are not differentially expressed, green points represent genes that have a log₂FC ≥ 2 but not differentially expressed, blue points represent genes that are differentially expressed (FDR ≤ 0.1) but have a log₂FC < 2, and red points represent genes that are both differentially expressed (FDR ≤ 0.1) and have a log₂FC ≥ 2. (B) Bar plot of IPA canonical pathways enriched for patients with history of CMV DNAemia compared with those without history of DNAemia at the long-term time point. The significance of an enriched pathway is represented by the length of each bar, where length corresponds to the -log(p-value). The number of genes enriching for a pathway is indicated in the bar. The vertical dotted black line corresponds to a p-value of 0.05. Positive IPA Z-scores are indicated in red.

Figure 6

Model for immunological changes early and late after CMV DNAemia. (A) Concept figure illustrating the early immune response to CMV DNAemia. Cells infected with CMV including macrophages expressing CCR5 secrete CCL5, recruiting additional T cells, and undergo apoptosis from activated CD28+ T cells, which secrete IFN-g and stimulate phagocytosis in antigen-presenting cells. (B) Concept figure illustrating the late immune response to CMV DNAemia. Persistently infected cells including macrophages continue to present CMV antigen to CD8+ T cells, which have now become senescent, expressing LAG3 but not CD28−, and activating, expressing granzyme and triggering apoptosis in infected cells. Antigen-presenting cells have function impaired by decreased expression of MHC class II and CXCL3.