Bioinformatic analysis reveals the expression of unique transcriptomic signatures in Zika virus infected human neural stem cells

Abstract

Background: The single-stranded RNA Flavivirus, Zika virus (ZIKV), has recently re-emerged and spread rapidly across the western hemisphere's equatorial countries, primarily through Aedes mosquito transmission. While symptoms in adult infections appear to be self-limiting and mild, severe birth defects, such as microcephaly, have been linked to infection during early pregnancy. Recently, Tang et al. (Cell Stem Cell 2016, doi: 10.1016/j.stem.2016.02.016) demonstrated that ZIKV efficiently infects induced pluripotent stem cell (iPSC) derived human neural progenitor cells (hNPCs), resulting in cell cycle abnormalities and apoptosis. Consequently, hNPCs are a suggested ZIKV target.

Methods: We analyzed the transcriptomic sequencing (RNA-seq) data (GEO: GSE78711) of ZIKV (Strain: MR766) infected hNPCs. For comparison to the ZIKV-infected hNPCs, the expression data from hNPCs infected with human cytomegalovirus (CMV) (Strain: AD169) was used (GEO: GSE35295). Utilizing a combination of Gene Ontology, database of human diseases, and pathway analysis, we generated a putative systemic model of infection supported by known molecular pathways of other highly related viruses.

Results: We analyzed RNA-sequencing data for transcript expression alterations in ZIKV-infected hNPCs, and then compared them to expression patterns of iPSC-derived hNPCs infected with CMV, a virus that can also induce severe congenital neurological defects in developing fetuses. We demonstrate for the first time that many of cellular pathways correlate with clinical pathologies following ZIKV infection such as microcephaly, congenital nervous system disorders and epilepsy. Furthermore, ZIKV activates several inflammatory signals within infected hNPCs that are implicated in innate and acquired immune responses, while CMV-infected hNPCs showed limited representation of these pathways. Moreover, several genes related to pathogen responses are significantly upregulated upon ZIKV infection, but not perturbed in CMV-infected hNPCs.

Conclusion: The presented study is the first to report enrichment of numerous pro-inflammatory pathways in ZIKV-infected hNPCs, indicating that hNPCs are capable of signaling through canonical pro-inflammatory pathways following viral infection. By defining gene expression profiles, new factors in the pathogenesis of ZIKV were identified which could help develop new therapeutic strategies.

Fig. 1

Comprehensive gene set gene ontology analysis. Gene ontology analysis was performed on hNPC genes whose expression was significantly altered following infection with ZIKV for 56 h. Transcript enrichment significance was determined using the human protein coding genome as the reference background. a Regulation of nucleic acid metabolism was the only biological process sub-ontology enriched with ZIKV up-regulated genes. Down-regulated genes were enriched for pathways related to cell cycle regulation and general metabolism. All significant terms are presented. b Cellular component enrichment indicated down-regulated genes generally associated with chromosomal regions, the extracellular matrix, and centrosomes. Up-regulated genes showed the most enrichment for the Golgi apparatus. Both up- and down-regulated genes were significantly enriched in nuclear sites. Enrichment scores were sorted by the most significant terms, and the top 11 compartments are represented. (q < 0.05)

Fig. 2

ZIKV DEGs mapped to canonical human disease gene sets. Using the MalaCards database of human diseases, diseases with established clinical significance to ZIKV infection were queried. ZIKV DEGs were then compared to reference gene list for each disease to determine a percentage of representation. Diseases related to congenital CNS disorders were most highly represented with ZIKV DEGs. Although less, representation of adult inflammatory and demyelinating diseases were also observed. (Bars % of ZIKV induced genes present in the reference list) (Open circles total number of genes in the disease reference list)

Fig. 3

Gene ontology functional network of biological process and immune system terms. All significantly altered genes greater than 0.5 fold-change were imported into Cytoscape and Gene Ontology clustering was performed with the ClueGO plugin. The human genome was used as the background with the GO biological process and immune system terms queried for enrichment. Terms were generated such that GO Term Fusion was implemented on pathways with a less than 0.01 q-value. Groupings with less than 3 connections were excluded from the final list of networks. Terms related to cytokine and chemokine production as well as general pathogen response formed networks independent of nucleic acid and macromolecule metabolic processes

Fig. 4

Divergent gene expression patterns in CMV and ZIKV-infected hNPCs. Genes significantly different between mock and cells infected with either CMV or ZIKV were processed identically and resulting significant biological pathways were mapped. Clustering was performed on the resulting pathways such that similar functions were grouped for clarity. The Cytokine/Chemokine and Pathogen Response clusters were enlarged to demonstrate the divergent pathway enrichment between ZIKV and CMV infection

Fig. 5

Immune specific biological pathways enriched in up-regulated genes. a The biological pathways associated with ZIKV transcript changes were filtered to include top immune specific pathways. b Using a FDR q-value <0.05, several top immune function pathways were enriched for the up-regulated gene set, although none of the pathways were enriched for significant down-regulated transcripts. The q-values for the pathway enrichment are annotated along the left panel as the dot plot. c Using tissue specific expression databases, up- and down-regulated genes were probed for immune related cells and tissues. Innate and adaptive immune system components were represented in both lists. (q < 0.05)

Fig. 6

Activation of NF-κB and STATs signaling pathway in ZIKV-infected hNPCs. Cytokines (IL-1 and TNF-α), PAMPs, CD40 and CXCL12 interact with their receptors resulting in activation of the NF-κB pathway. IFN-γ, IL-6 and IL-12 binding to their receptors induces STAT dimerization and translocation to the nucleus. Activation of both the canonical NF-κB and STAT signaling pathways, results in the production of numerous inflammatory cytokines.