Transcriptomic Signatures of Zika Virus Infection in Patients and a Cell Culture Model

Abstract

Zika virus (ZIKV), a re-emerging flavivirus, is associated with devasting developmental and neurological disease outcomes particularly in infants infected in utero. Towards understanding the molecular underpinnings of the unique ZIKV disease pathologies, numerous transcriptome-wide studies have been undertaken. Notably, these studies have overlooked the assimilation of RNA-seq analysis from ZIKV-infected patients with cell culture model systems. In this study we find that ZIKV-infection of human lung adenocarcinoma A549 cells, mirrored both the transcriptional and alternative splicing profiles from previously published RNA-seq data of peripheral blood mononuclear cells collected from pediatric patients during early acute, late acute, and convalescent phases of ZIKV infection. Our analyses show that ZIKV infection in cultured cells correlates with transcriptional changes in patients, while the overlap in alternative splicing profiles was not as extensive. Overall, our data indicate that cell culture model systems support dissection of select molecular changes detected in patients and establishes the groundwork for future studies elucidating the biological implications of alternative splicing during ZIKV infection.

Keywords: Zika virus; alternative splicing; cell culture; patients; transcriptome.

Figure 1

Transcriptome analysis of ZIKV infection at early acute and late acute times in patients. (A) Principal component analysis (PCA) of normalized counts. The principal components are colored by infection time (gold—Early Acute, blue—Late Acute, and red—Convalescent) and the reported sex of the patients is represented as circles for female and triangles for male. Venn diagrams show differentially expressed genes at either early or late acute ZIKV infection times relative to the convalescent stage that were either (B) upregulated or (C) downregulated. The differentially expressed genes are defined as Log₂ FC > 1. (D) Normalized counts of CCL2 expression across each ZIKV infection stage. The average Log₂ FC between early acute ZIKV infection and the convalescent stage was 6.135 and the average Log₂ FC between late acute ZIKV infection and the convalescent phase was 1.378. Adjusted p-values for the upregulation of CCL2 in early acute and late acute phases of the infection were 1.28 × 10⁻⁶⁵ and 0.00112, respectively. (E) Normalized counts of HERC5 expression across each ZIKV infection time. The average Log₂ FC was 3.14 between early acute ZIKV infection and convalescent period with an adjusted p-value of 5.87 × 10⁻³⁷. (F) Normalized counts of CMPK2 expression across ZIKV infection times had an average Log₂ FC of 3.3 between early acute and convalescent timepoints and an adjusted p-value of 2.42 × 10⁻⁵⁰.

Figure 2

Alternative splicing analysis in ZIKV infected pediatric PMBC samples from early acute and late acute infection stages. (A) Pie charts show the percentage of each of the five alternative splicing events in the early acute and late acute patient samples when compared to patient samples from the convalescent phase. The alternative splicing events include skipped exons (SE), mutually exclusive exons (MXE), retained intron (RI), alternative to 3′ splice site (A3SS), and alternative to 5′ splice site (A5SS). (B) PCA plot PSI values from skipped exon events present between early acute vs. convalescent, late acute vs. convalescent, and early acute vs. late acute timepoints. Samples were also categorized by reported sex. (C) Venn diagram of overlapping significant skipped exon events (ΔPSI > |0.1| and False Discovery Rate [FDR] < 0.05) between early acute and late acute infection when compared to convalescent phase. (D–F) Skipped exon splicing graphs with PSI values plotted for each patient sample at each infection timepoint are shown. The three genes chosen were (D) KIF23 exon 18, (E) CYRIB exon 6, and (F) MARCHF8 exon 5 and were selected based on being significantly alternatively spliced in both infection times (Figure 2C). *** FDR < 0.05.

Figure 3

Analysis of differentially expressed genes following ZIKV infection in patients and an A549 cell model. (A) Venn diagram of shared genes upregulated following ZIKV infection in early acute and late acute patients (Log₂ FC > 1) compared to ZIKV infected A549 cells (Log₂ FC > 2). (B) Venn diagram of shared genes downregulated following ZIKV infection in early acute and late acute patients (Log₂ FC < 1) compared to ZIKV infected A549 cells (Log₂ FC < 2). (C) Gene Ontology (GO) biological process categories for differentially expressed genes from early acute, late acute and A549 ZIKV infections. Full category names and number of genes per category are found in Supplementary Tables S6–S8. The GO biological process in (C) that has a truncated description represents “biological process involved in interspecies interaction between organisms (GO:0044419)”. (D,E) Normalized counts of (D) ATF3 and (E) CXCL10 expression across ZIKV infection times are shown in the left panel and the middle panel shows the data in ZIKV and mock infected A549 cells. RT-qPCR validation of ATF3 and CXCL10 expression in A549 cells infected with ZIKV or uninfected (mock) is shown in the right shaded panel. Statistical significance of RT-qPCR was determined by student t-test. Error bars represent ± SD from three independent experiments. *** p < 0.001.

Figure 4

Comparison of alternative splicing changes between ZIKV infected patient samples and the A549 cell line infected with ZIKV. (A) Pie chart representing the percentage of each of the five alternative splicing events in the ZIKV infected A549 cells when compared to mock-infected A549 cells. (B) Bar chart showing the large number of significant skipped exon splicing events in the ZIKV-infected A549 cells and comparison to the two acute infection timepoints of the patient study relative to the convalescent samples. (C) GO pathways analysis of the significant skipped exon events for the ZIKV-infected A549 cells, the early acute, and the late acute infection times ranked by number of genes found. The truncated fifth pathway is biological process involved in interspecies interaction between organisms (GO:0044419). The number of genes within each GO term for the early and late acute genes are represented on the x-axis and the number of genes in ZIKV-infected A549 cells exceeding the x-axis are annotated on the chart. (D) Venn diagram showing the overlap of significant skipped exon events in the ZIKV-infected A549 cells and PBMCs from early and late acute infection timepoints in patients. (E,F) Splicing graphs for selected SE events namely Aprataxin (APTX) exon 3 and Erythrocyte Protein band 4.1 (EPB41) exon 14 with PSI values determined from the RNA-seq data reported by Michlmayr et al. (left panel) [45]. PSI values from RNA-seq analysis of ZIKV infected A549 cells compared to mock (middle panel), and RT-PCR validation of each event in the shaded portion of graph (right panel). * FDR < 0.05. Significance of RT-PCR validation of the SE events was determined by three independent experiments and student t-test. **** p < 0.0001.