MicroRNA regulation and its effects on cellular transcriptome in human immunodeficiency virus-1 (HIV-1) infected individuals with distinct viral load and CD4 cell counts

Abstract

Background: Disease progression in the absence of therapy varies significantly in HIV-1 infected individuals. Both viral and host cellular molecules are implicated; however, the exact role of these factors and/or the mechanism involved remains elusive. To understand how microRNAs (miRNAs), which are regulators of transcription and translation, influence host cellular gene expression (mRNA) during HIV-1 infection, we performed a comparative miRNA and mRNA microarray analysis using PBMCs obtained from infected individuals with distinct viral load and CD4 counts.

Methods: RNA isolated from PBMCs obtained from HIV-1 seronegative and HIV-1 positive individuals with distinct viral load and CD4 counts were assessed for miRNA and mRNA profile. Selected miRNA and mRNA transcripts were validated using in vivo and in vitro infection model.

Results: Our results indicate that HIV-1 positive individuals with high viral load (HVL) showed a dysregulation of 191 miRNAs and 309 mRNA transcripts compared to the uninfected age and sex matched controls. The miRNAs miR-19b, 146a, 615-3p, 382, 34a, 144 and 155, that are known to target innate and inflammatory factors, were significantly upregulated in PBMCs with high viral load, as were the inflammatory molecules CXCL5, CCL2, IL6 and IL8, whereas defensin, CD4, ALDH1, and Neurogranin (NRGN) were significantly downregulated. Using the transcriptome profile and predicted target genes, we constructed the regulatory networks of miRNA-mRNA pairs that were differentially expressed between control, LVL and HVL subjects. The regulatory network revealed an inverse correlation of several miRNA-mRNA pair expression patterns, suggesting HIV-1 mediated transcriptional regulation is in part likely through miRNA regulation.

Conclusions: Results from our studies indicate that gene expression is significantly altered in PBMCs in response to virus replication. It is interesting to note that the infected individuals with low or undetectable viral load exhibit a gene expression profile very similar to control or uninfected subjects. Importantly, we identified several new mRNA targets (Defensin, Neurogranin, AIF) as well as the miRNAs that could be involved in regulating their expression through the miRNA-mRNA interaction.

Figure 1

A: Expression profile of differentially regulated miRNAs in HIV-1 infected versus uninfected subjects. (A) The Venn diagram displays the number and overlap of significantly differentially expressed miRNA (Benjamini-Hochberg adjusted, p < 0.05) among the LVL and HVL groups relative to the CT and within the infected groups. Hierarchical clustering of miRNA between CT and LVL (B); and CT and HVL (C). MicroRNA in the clustergram are dysregulated at a significance cutoff of p < 0.05. The dendogram depicting the clustering of samples is calculated using Complete linkage with Euclidian distance measure values. Color ranging from green to red indicates minimum to maximum ∆CT. Numbers on X-axis represent subject group. CT, uninfected controls; LVL, low viral load subjects; HVL, high viral load subjects. The clustergrams were generated using StatMiner analyses.

Figure 2

Independent individual validations of randomly selected differentially regulated miRNAs from Taqman array platform. qRT-PCR was used to validate the expression of selected miRNAs from the high throughput results (derived by StatMiner) using a specific primer and probe for each miRNA. (A) miRNAs selected from Control versus LVL comparison; (B) miRNAs selected from control versus HVL comparison. Fold increase/decrease was calculated based on normalization to U6. Average fold change for each miRNA represents fold change obtained from independent donors (N = 5 per group).

Figure 3

Expression profile of differentially regulated mRNA transcripts in HIV-1 infected versus uninfected subjects. (A). Hierarchical clustering of differentially regulated mRNA in LVL compared to control; and (B), CT and HVL. The probes in this clustergram are significantly differentially regulated (p < 0.05). Red indicates high, blue indicates low, and gray stands for no change in level of expression. Numbers on X-axis represent subject group; Y-axis represents the gene symbol. CT, uninfected controls; LVL, low viral load subjects; HVL, high viral load subjects.

Figure 4

Predicted interaction networks of genes significantly dysregulated in HVL compared to Control. (A) The interactions between genes were identified using STRING software. Colored lines indicate different sources of evidence for each interaction. Circles highlight the predominant clusters within the network. (B & C) Gene Ontology Enrichment analysis for biological processes using significantly differentially regulated genes in LVL and HVL groups in comparison with CT group. Genes that are differentially regulated in LVL group relative to the CT yielded 4 significant terms for biological processes (B), while those that are differentially regulated in HVL with respect to CT yielded 34 (C). Bar graphs are generated using average –lop of p value for each term on y-axis and term name on x-axis. CT, uninfected control; LVL, low viral load group; HVL, high viral load group.

Figure 5

Network of predicted miRNA-mRNA interactions by GroupMiR using results from miRNA and mRNA array and visualized by Cytoscape. Regression-based method was used to predict the potential miRNA (circles) that actively regulate mRNA (squares). Selected miRNA (four) were chosen from a list of miRNAs and their targets. Red and green within the mRNA represents up and down regulation, respectively. Each slice within the circle represents HIV-1 infected subject.

Figure 6

Expression of selected transcripts among the CT, LVL, and HVL groups. Unpaired Student’s t-test was used to assess significance between the three groups. (* = p < 0.05, ** = p < 0.01, *** = p < 0.0001). CT, uninfected controls; LVL, low viral load group; HVL, high viral load group.

Figure 7

Validation of HIV-1 regulated host cellular factors at the RNA and protein level. (A) Independent validations of randomly selected differentially regulated mRNAs from transcriptome analyses. qRT-PCR was performed to validate the expression of selected mRNAs using specific primer and probe pairs. Fold increase/decrease was calculated based on normalization to RPLPO. Average fold change for each mRNA represents fold change obtained from independent donors (N = 5 per group). (B) Inflammatory factors released by HIV-1 infected PBMC compared to uninfected control cells. Expression of CCL2, CCL8, CXCL5, IL6 and IL8 was monitored by ELISA in supernatants obtained from PBMCs infected with CXCR4-coreceptor utilizing virus (NL43), CCR5-coreceptor utilizing virus (YU2) or mock infected PBMCs (n = 7). * = p < 0.05; NS, Not significant. (C) Expression level of NRGN transcript in PBMCs infected with NL43, YU2 or mock infected cells. Post infection, RNA was isolated and qRT-PCR was performed using NRGN specific primers and probe and results were normalized to RPLPO control. Fold change was calculated using uninfected/mock infected PBMC controls. (D) Immunoblot of NRGN to measure the protein level in PBMC infected with HIV-1 virus or mock. Gag-p24, represents infectivity; and Actin, represents loading control. (E) Densitometry was performed to quantitate the fold change in signal intensity compared to uninfected (NT) and normalized with Actin level. NT, uninfected control; NL4-3, CXCR4 coreceptor virus; and YU2, CCR5 coreceptor virus.