Innate immune defects in HIV permissive cell lines

Abstract

Background: Primary CD4+ T cells and cell lines differ in their permissiveness to HIV infection. Impaired innate immunity may contribute to this different phenotype.

Findings: We used transcriptome profiling of 1503 innate immunity genes in primary CD4+ T cells and permissive cell lines. Two clusters of differentially expressed genes were identified: a set of 249 genes that were highly expressed in primary cells and minimally expressed in cell lines and a set of 110 genes with the opposite pattern. Specific to HIV, HEK293T, Jurkat, SupT1 and CEM cell lines displayed unique patterns of downregulation of genes involved in viral sensing and restriction. Activation of primary CD4+ T cells resulted in reversal of the pattern of expression of those sets of innate immunity genes. Functional analysis of prototypical innate immunity pathways of permissive cell lines confirmed impaired responses identified in transcriptome analyses.

Conclusion: Integrity of innate immunity genes and pathways needs to be considered in designing gain/loss functional genomic screens of viral infection.

Keywords: Cell lines; HIV; Innate immunity; RNA-Seq; TCR activation.

Fig. 1

Differences between resting primary CD4+ T cells and laboratory cell lines in innate immunity. a Laboratory cell lines revealed differential permissiveness to HIV infection. Cells were infected using a VSV-G pseudotyped HIV virus. Viral infection success was assessed by FACS analysis of the expression of the virally encoded GFP reporter gene, and ranged from ~7 to 53 % according to the cell line infected. b Heatmap of expression values of innate immunity genes in resting CD4+ T cells and laboratory cell lines. The figure shows the expression values of 1473 innate immunity genes in resting CD4+ T cells from two donors (CD4_J3 and CD4_J4), and four human laboratory cell lines HEK293T, Jurkat, SupT1 and CEM. Cell lines were evaluated in 3 conditions: uninfected mock (Mock), heat-inactivated HIV vector (hiLV) and HIV vector-infected (LV). Complete hierarchical clustering of genes and cell samples was based on Pearson correlation of variance-stabilized read counts (Methods). Color scale indicated in the legend corresponds to z-scores of RPKM distributions per gene, ranging from green (low) to red (high) expression. Two prominent clusters of genes are highlighted: 249 genes with a high expression in resting CD4+ T cells and a low relative expression in all laboratory cell lines (dark blue square) and 110 genes with a low relative expression in resting CD4+ T cells and a high relative expression in all laboratory cell lines (cyan square). The genes within each of these two clusters are listed in Additional file 3: Table S2

Fig. 2

Activation of primary CD4+ T cells produces an intermediate expression phenotype in clusters of innate immunity genes differentiating resting CD4+ T cells from permissive cell lines. Distribution of expression levels of the 249 (a) and 110 (b) gene clusters represented as violin plots. Seven distributions are shown summarizing the average expression values in resting CD4+ T cells (dark red), activated CD4+ T cells at 8 h (red) and 24 h (orange) after TCR activation, and four human laboratory cell lines HEK293T (pink), Jurkat (light violet), SupT1 (magenta) and CEM (dark violet). Lines within the plots represent the median of such distributions. All distributions are significantly different from the resting CD4+ T cells in both panels (Wilcoxon rank sum test, Bonferroni p.adjusted <1E−3). Expression values on the y-axis represent the log10 transformation of the number of library size-normalized reads per kilobase of exonic sequence averaged within cell type (Methods)

Fig. 3

Defects in 3 selected innate immunity pathways in cell lines. a The figure represents a simplified view of the TLR7/TLR8, IFN-gamma and TNF-alpha signaling pathways. Boxes representing genes display the transcriptional levels detected in RNA-seq libraries of resting CD4+ T cells, the four human laboratory cell lines HEK293T, Jurkat, SupT1 and CEM -mock (MO), heat-inactivated (HI) and HIV-infected (HIV)- and 4 samples corresponding to Activated CD4+ T cells at 8 and 24 h after TCR activation. Inset describes the order of the libraries as well as the color-code scale of expression levels (log10 transformation of the number of library size-normalized reads per kilobase of exonic sequence) ranging from 0 (green) to ≥2.8 (red; lower limit of the 9th-decile of expression values). The expression levels indicated for IFNG and TNF convey the basal expression level before adding the stimuli. b Experimental validation of the functional integrity of the selected innate immunity pathways. The table reports the stimuli applied and the functional read-out measured 24 h after stimulation. A positive sign indicates positive detection of functional read-outs (transcript levels by RT-qPCR or phosphorylation of STAT1 by Western blot analysis). NT not tested, nd not detected

Fig. 4

Heatmap of expression values of paradigmatic genes involved in antiretroviral defense and signaling relevant to HIV biology. The figure shows the expression values of antiretroviral genes (APOBEC3G, TRIM5, BST2, MX2, GBP5, and SAMHD1) and signaling genes (JAK, STAT1, IFI16 and STING/TMEM173) in RNA-seq libraries of resting CD4+ T cells, cell lines HEK293T, Jurkat, SupT1 and CEM -mock (Mock), heat-inactivated (hiLV) and HIV-infected (LV)- and activated CD4+ T cells at 8 and 24 h after TCR activation (Methods). The color-code scale in the inset represents the expression levels (log10 transformation of the number of library size-normalized reads per kilobase of exonic sequence), ranging from green (low) to red (high) expression. Complete hierarchical clustering of genes was based on Pearson correlation of the expression levels. Complete hierarchical clustering of samples was kept as assessed in Additional file 6: Figure S4. JAK1, JAK2, TMEM173 and GBP5 had a fold change higher than 2 between resting CD4+ T cells and permissive cell lines (Benjamini–Hochberg adjusted p value <0.01, Methods)