Research ArticleHEALTH AND MEDICINE

CMPK2 and BCL-G are associated with type 1 interferon–induced HIV restriction in humans

See allHide authors and affiliations

Science Advances  01 Aug 2018:
Vol. 4, no. 8, eaat0843
DOI: 10.1126/sciadv.aat0843
  • Fig. 1 Administration of pegylated IFN-α2b induces a decline in plasma HIV RNA and stimulates gene expression in activated CD4+ T cells.

    (A) Observed HIV kinetics after IFN-α2b in 19 participants. Boxplots are shown with overlaid values of in vivo log10 copies/ml of changes in plasma HIV RNA from baseline among patients who received IFN-α2b. Lines indicate longitudinal measurements from the same individual. (B) Infections were first simulated for 1000 days using three standard differential equations dx/dt = λ − dx − βxv, dy/dt = βxvay, and dv/dt = kyuv (λ is the production rate of uninfected cells, x is the number of uninfected cells, d is the decay rate of uninfected cells, β is the first-order rate constant of infection, v is the plasma HIV RNA level, y is the number of infected cells, a is the decay rate of infected cells, k is the first-order rate constant of virus production in infected cells, and u is the first-order rate constant of virus decay; see fig. S2) using established starting values (x = 106, y = 1, v = 105, λ = 105, d = 0.1, a = 0.5, β = 2 × 10−7, k = 100, and u = 5). Resulting values were then used as input into the same equations, separately modifying the indicated variable by the indicated factor and run for 7 days to reflect observed data in (A). We assumed that the effect of IFN-α2b on plasma HIV RNA can be due to changes in the numbers of susceptible cell numbers (x), changes in susceptibility of those cells (β), changes in the numbers of infected cells (y), and changes in the release of virus from infected cells (k); therefore, we varied x, y, β, and k in the models. Whereas varying x and y had minimal effects on modeling HIV viremia, varying β and k had significant effects on HIV kinetics that mirrored the observed data in (A). (C) Representative flow cytometry data and sorting algorithm for isolation of activated CD4+ T cells from total peripheral blood mononuclear cells (PBMCs) at baseline (Pre–IFN-α2b; blue) and 24 hours after injection with IFN-α2b (Post–IFN-α2b; red). SSC, side scatter. (D) Boxplots with overlaid points showing the fold changes for the 99 genes that significantly changed with IFN-α2b administration across the cohort, after adjustment for multiple comparisons. For each ISG, an individual point represents fold change data from a single person, and the boxplot represents summary fold change data across all people. (E) RNA sequencing (RNA-seq) reads from each person’s activated CD4+ T cells [CD3+/CD4+/CD38+/HLA-DR+; shown in (C)] were mapped and aligned to 2154 HIV reference genomes before and 24 hours after IFN-α2b. Displayed are representative alignment maps from three persons before and after IFN-α2b, demonstrating that the selection of these cells is adequate to testing how IFN-α2b restricts HIV replication. (F) RNA-seq reads from activated CD4+ T cells that mapped to HIV correlated closely with plasma HIV RNA levels, confirming that infection of these cells is tightly associated with plasma viremia and viral kinetics in (A).

  • Fig. 2 Identification of ISGs that are putative interferon-induced HIV restriction factors in activated CD4+ T cells.

    (A) Individual scatterplots of 13 of 99 ISGs that were significantly correlated with plasma HIV RNA decline after IFN-α2b administration, adjusting for multiple comparisons. BST2, a known IFN-inducible HIV restriction factor, is also shown for reference. We assumed that the induction of an HIV restriction factor had to precede the change in plasma HIV RNA: Therefore, the fold change in mRNA expression of each gene in activated CD4+ T cells after 24 hours of IFN-α2b is shown in the x axis, and the change in plasma HIV RNA level 1 week after IFN-α2b is shown the y axis. (B) Putative HIV restriction factors are not bystanders of IFN-α2b induction but appear to be specific for HIV. Each of the 99 ISGs identified in Fig. 1 is depicted by individual dots. The strength of the association of the induction of each ISG with plasma HIV RNA decline is shown as the Spearman’s P value in the y axis and by the size of each dot (scaled to represent the Spearman’s correlation coefficient). Genes that met statistical significance after adjustment for multiple comparisons are shown as green dots. To validate against false detection, Spearman correlations were also performed for the induction of the same genes and plasma HCV RNA decline (P values along the x axis), revealing no genes that were up-regulated in activated CD4+ T cells were significantly associated with HCV, after the adjustment for multiple comparisons.

  • Fig. 3 Covariate analysis of ISG expression reveals distinct regulation of CMPK2 and RSAD2.

    (A) Genomic positions of CMPK2 and RSAD2, magnified on CMPK2, illustrate a shared upstream region for both genes. (B) Pairwise correlation plots of fold changes for the putative HIV restriction factors and BST2 as reference demonstrate different clusters of gene expression and distinct patterns of regulation. (C) K-means clustering of principal components analysis grouping all points into four clusters. The same color scheme used to demarcate clusters in (B) is used in (C). Despite their shared upstream region, CMPK2 and RSAD2 show distinct regulatory patterns. BCL-G appears to be regulated separately from other reported and novel HIV restriction factors.

  • Fig. 4 Interferon-induced HIV restriction in vitro is dependent on CMPK2 and BCL-G.

    (A) IFN induces CMPK2 and BCL-G in multiple cell lines. Real-time quantitative polymerase chain reaction (qPCR) data showing the up-regulation of CMPK2, BCL-G, and MX2 (as a positive control) in response to IFN in six cell lines at 0, 6, 12, and 24 hours. (B) CMPK2 and BCL-G are ISGs in primary cells. Real-time qPCR data of CMPK2, BCL-G, and MX2 in response to IFN in activated CD4+ T cells ex vivo from two healthy donors who were not infected with HIV or HCV. (C) CMPK2 knockdown diminishes HIV restriction of HIV. CMPK2 knockdown was performed using pooled siRNAs in THP-1 cells (red) and compared to a scrambled siRNA (black). Cells were then inoculated with a wide range of HIV inocula (x axis) and split and either treated with IFN or left untreated. HIV replication was quantified by the TZM-bl luciferase assay. The log-transformed difference in HIV replication between IFN-treated and IFN-untreated cells for each condition is shown in the y axis. *P < 0.05; **P < 0.01; ***P < 0.005. (D) BCL-G inhibits HIV production. myc-tagged BCL-G, myc-tagged MX2, and an empty pMyc vector were transfected into TZM-bl cells. Cells were inoculated with a wide range of HIV inocula (x axis). HIV replication was quantified by the TZM-bl luciferase assay. The raw luciferase values, after the adjustment for baseline differences, are shown. Statistical significance in differences in luciferase activity between myc–BCL-G transfected cells and myc control cells are denoted by *P < 0.05; **P < 0.01; ***P < 0.005. Statistical significance in differences in luciferase activity between myc-MX2–transfected cells and myc control cells are denoted by §P < 0.05; §§P < 0.01; and §§§P < 0.005.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/4/8/eaat0843/DC1

    Fig. S1. Antiviral kinetics for HIV and HCV vary independently by person.

    Fig. S2. Standard virologic models highlight different possible actions of ISGs on HIV replication.

    Fig. S3. IFN-α2b administration results in changes in the proportion of activated CD4+ and CD8+ T cells.

    Fig. S4. Approach to quantifying and identifying ISGs in RNAs from longitudinally sampled cells.

    Fig. S5. IFN-α2b induces multiple MX2 isoforms in vivo.

    Fig. S6. IFN-α2b induces multiple isoforms of CMPK2 and BCL-G.

    Fig. S7. Baseline gene expression is not associated with plasma HIV RNA levels.

    Fig. S8. Mean ISG induction and plasma HIV RNA decline.

    Fig. S9. Interferon-mediated inhibition of HIV in vitro in multiple cell lines.

    Fig. S10. CMPK2 and MX2 are induced by IFN in THP-1 cells.

    Fig. S11. Knockdown of CMPK2.

    Fig. S12. IFN restriction of HIV is dependent on CMPK2 in THP-1 cells.

    Fig. S13. Expression of myc-tagged BCL-G and MX2 in vitro.

    Table S1. Participant characteristics.

    Table S2. RNA-seq metadata.

    Table S3. qPCR primer sequences and amplicon location by exon.

  • Supplementary Materials

    This PDF file includes:

    • Fig. S1. Antiviral kinetics for HIV and HCV vary independently by person.
    • Fig. S2. Standard virologic models highlight different possible actions of ISGs on HIV replication.
    • Fig. S3. IFN-α2b administration results in changes in the proportion of activated CD4+ and CD8+ T cells.
    • Fig. S4. Approach to quantifying and identifying ISGs in RNAs from longitudinally sampled cells.
    • Fig. S5. IFN-α2b induces multiple MX2 isoforms in vivo.
    • Fig. S6. IFN-α2b induces multiple isoforms of CMPK2 and BCL-G.
    • Fig. S7. Baseline gene expression is not associated with plasma HIV RNA levels.
    • Fig. S8. Mean ISG induction and plasma HIV RNA decline.
    • Fig. S9. Interferon-mediated inhibition of HIV in vitro in multiple cell lines.
    • Fig. S10. CMPK2 and MX2 are induced by IFN in THP-1 cells.
    • Fig. S11. Knockdown of CMPK2.
    • Fig. S12. IFN restriction of HIV is dependent on CMPK2 in THP-1 cells.
    • Fig. S13. Expression of myc-tagged BCL-G and MX2 in vitro.
    • Table S1. Participant characteristics.
    • Table S2. RNA-seq metadata.
    • Table S3. qPCR primer sequences and amplicon location by exon.

    Download PDF

    Files in this Data Supplement:

Navigate This Article