Plasma vesicular miR-155 as a biomarker of immune activation in antiretroviral treated people living with HIV

Abstract

People living with HIV (PLWH), despite suppression of viral replication with antiretroviral therapy (ART), have high morbidity and mortality due to immune activation and chronic inflammation. Discovering new biomarkers of immune activation status under ART will be pertinent to improve PLWH quality of life when the majority will be treated. We stipulate that plasma large and small extracellular vesicle (EVs) and their microRNA content could be easily measured biomarkers to monitor immune activation in PLWH. Venous blood samples from n = 128 ART-treated PLWH with suppressed viral load (≤ 20 copies/mL) and n = 60 HIV-uninfected participants were collected at five testing or treatment centers of PLWH in Burkina Faso. Large and small plasma EVs were purified, counted, and the mature miRNAs miR-29a, miR-146a, and miR-155 were quantified by RT-qPCR. Diagnostic performances of large and small EVs miRNAs level were evaluated by receiver operating characteristic (ROC) curve analysis and principal component analysis (PCA). Among the EVs microRNA measured, only large EVs miR-155 copies distinguished PLWH with immune activation, with AUC of 0.75 for CD4/CD8 < 1 (95% CI: 0.58-0.91, P = 0.0212), and 0.77 for CD8 T cells ≥ 500/µL (95% CI: 0.63-0.92, P = 0.0096). In addition, PCA results suggest that large EVs miR-155 copies may be a biomarker of immune activation. Since miR-155 may influence immune cell function, its enrichment in large EV subpopulations could be a functional biomarker of immune activation in PLWH on ART. This measure could help to monitor and diagnose the immune activation with more accuracy.

Keywords: HIV-1; MicroRNA; biomarker; extracellular vesicles; immune activation; miR-146a; miR-155; miR-29a.

Figure 1

EVs count and miRNAs expression in HIV negative and HIV infected with undetectable viral load study participants follow to CD4 T cells count. (A) Flow chart of study participants included in the analysis following CD4 T cells count. (B) Large and small EVs particles number quantification in flow cytometry analysis. (C, D), (E, F), and (G, H) present respectively EVs miR-29a, miR-146a, and miR-155 quantification and expression as copies per µg of RNA and copies per vesicle. An ordinary one-way ANOVA with Tukey’s multiple comparisons test was used for comparison between groups.

Figure 2

EVs count and miRNAs expression in study participants following to ratio CD4/CD8 T cells count. (A) Flow chart of study participants included in the analysis based on ratio CD4/CD8 T cells count. (B) Large and small EVs particles number quantification in flow cytometry analysis. (C, D), (E, F), and (G, H) present respectively EVs miR-29a, miR-146a, and miR-155 quantification and expression as copies per µg of RNA and copies per vesicle. An ordinary one-way ANOVA with Tukey’s multiple comparisons test was used for comparison between groups.

Figure 3

EV count and miRNAs expression in study participants following to CD8 T cells count. (A) Flow chart of study participants included in the analysis based on CD8 T cells count. (B) Large and small EVs particles number quantification in flow cytometry analysis. (C, D), (E, F), and (G, H) present respectively EVs miR-29a, miR-146a, and miR-155 quantification and expression as copies per µg of RNA and copies per vesicle. An ordinary one-way ANOVA with Tukey’s multiple comparisons test was used for comparison between groups.

Figure 4

EVs miRNAs content diagnosis performance in receiver operating characteristics curves analysis. Large and small EVs miRNA content was used to generate a Receiver Operator Characteristic (ROC) curve analysis to discriminate participants with different conditions. The participants with CD8 T cell < 500 cells/µL, CD4 T cell ≥ 500 cells/µL, and ratio CD4/CD8 ≥ 1 (group 1, n=10) were used as controls. Diagnosis performance of large EVs miR-155 copies for the discrimination of all participants (A, F, K), female sex workers (B, G, L), men who have sex with men (C, H, M), and female (D, I, N) and men (E, J, O) from general population with respectively CD4 T cell count ≤ 500 cells/µL, ratio CD4/CD8 < 1, CD8 T cell count ≥ 500 cells/µL. Wilson/Brown method was used to compute the area under a ROC curve.

Figure 5

Change in the EVs count and miRNAs expression in study participants Follow a combination of undetectable viral load, CD8 T cell count, and CD4/CD8 ratio. HIV infected participants on ART for more than six months and with undetectable viral load (< 20 copies/mL) were divided into three groups. The reference group is characterized by participants with CD8 T cell < 500 cells/µL, CD4 T cell ≥ 500 cells/µL, and ratio CD4/CD8 ≥ 1 (group 1, n=10). The second group is characterized by ratio CD4/CD8 ≥ 1 and CD8 T cell ≥ 500 cells/µL (group 2, n = 24), and the last group is characterized by ratio CD4/CD8 < 1 and CD8 T cell ≥ 500 cells/µL (group 3, n = 78). (A) Large and small EVs particles number quantification in flow cytometry analysis. (B–E), and (F, G) present respectively EVs miR-29a, miR-146a, and miR-155 quantification and expression as copies per µg of RNA and copies per vesicle. An ordinary one-way ANOVA with Tukey’s multiple comparisons test was used for comparison between groups.

Figure 6

Principal component analysis (PCA) plots of individuals large and small EVs microRNA content. Graphs of individual large and small EVs miR-29a (A, B), miR-146a (C, D) and miR-155 (E, F) content respectively expressed as copies per µg of RNA and copies per vesicle were performed in PCA for HIV negative and HIV+ ART-treated participants with undetectable viral load participants divided into four groups. Group 1 (participants with CD8 T cell < 500 cells/µL, CD4 T cell ≥ 500 cells/µL, and ratio CD4/CD8 ≥ 1); group 2 (participants with ratio CD4/CD8 ≥ 1 and CD8 T cell ≥ 500 cells/µL), group 3 (participants with CD4/CD8 < 1 and CD8 T cell ≥ 500 cells/µL), and group 4 (HIV negative independently to CD4 or CD8 T cells count and ratio CD4/CD8). Ellipses were drawn around the centroids of the clusters, representing 95% confidence intervals.