Genome-wide search for the genes accountable for the induced resistance to HIV-1 infection in activated CD4+ T cells: apparent transcriptional signatures, co-expression networks and possible cellular processes

Abstract

Background: Upon co-stimulation with CD3/CD28 antibodies, activated CD4 + T cells were found to lose their susceptibility to HIV-1 infection, exhibiting an induced resistant phenotype. This rather unexpected phenomenon has been repeatedly confirmed but the underlying cell and molecular mechanisms are still unknown.

Methods: We first replicated the reported system using the specified Dynal beads with PHA/IL-2-stimulated and un-stimulated cells as controls. Genome-wide expression and analysis were then performed by using Agilent whole genome microarrays and established bioinformatics tools.

Results: We showed that following CD3/CD28 co-stimulation, a homogeneous population emerged with uniform expression of activation markers CD25 and CD69 as well as a memory marker CD45RO at high levels. These cells differentially expressed 7,824 genes when compared with the controls on microarrays. Series-Cluster analysis identified 6 distinct expression profiles containing 1,345 genes as the representative signatures in the permissive and resistant cells. Of them, 245 (101 potentially permissive and 144 potentially resistant) were significant in gene ontology categories related to immune response, cell adhesion and metabolism. Co-expression networks analysis identified 137 "key regulatory" genes (84 potentially permissive and 53 potentially resistant), holding hub positions in the gene interactions. By mapping these genes on KEGG pathways, the predominance of actin cytoskeleton functions, proteasomes, and cell cycle arrest in induced resistance emerged. We also revealed an entire set of previously unreported novel genes for further mining and functional validation.

Conclusions: This initial microarray study will stimulate renewed interest in exploring this system and open new avenues for research into HIV-1 susceptibility and its reversal in target cells, serving as a foundation for the development of novel therapeutic and clinical treatments.

Figure 1

Morphology and surface phenotypes of stimulated and un-stimulated CD4+ T cells. A: CD4 + T cells stimulated with PHA/IL-2 (“P”) or CD3/CD28 coated beads (“B”) were observed on day 0 (un-stimulated (“R”)), day 3 and day 6. B: Flow cytometric analysis of cell surface phenotypes. “R”, “P” and “B” cells were stained with antibodies to CD25 (FITC), CD69 (PE-Cy5) and CD45RO (FITC). Numbers indicate the percentage of each subset. A majority population of cells emerged in “B” cells which expressed CD25, CD69 and CD45RO. We did the experiments in 19 biological replicates.

Figure 2

Overall patterns of 7,824 differentially expressed genes and the 6 representative profiles in the permissive “P” and resistant “B” cells. Patters were plotted on the heatmap using Treeview. Red represents up-regulated genes while green represents down-regulated genes. Hierarchical clustering is shown on the left. All 16 expression profiles identified by Series-Cluster analysis are shown in the middle and 6 summarized representative profiles are shown on the right.

Figure 3

Gene ontology (GO) analysis and significant functional genes. Left: Numbers of genes in the 6 representative profiles. Middle: significant functional categories and the number of genes in them. Each color represents one category and the size of each sector in a pie diagram is proportional to the number of genes in its category. Right: Merged functional categories in permissive “P_up”, “B_down” and “P_upB_down” cells and resistant “P_down”, “B_up” and “P_downB_up” cells.

Figure 4

Hierarchical GO categories of genes involved in induced permissive “P” and resistant “B” cells. GO categories trees were hierarchically built using the Gene Ontology Enrichment Analysis Software Toolkit (GOEAST): http://omicslab.genetics.ac.cn/GOEAST/). The category tree of response to stimulus related genes in “P_up”, “B_down” and “P_upB_down” cells is shown in A. Category trees of multicellular organismal development, cellular process and metabolic process related genes in “P_down”, “B_up” and “P_downB_up” cells are shown in B.

Figure 5

Interactions of significant functional genes with HIV-1 proteins. Filled circles represent genes involved in significant functional categories. Red represents genes in profiles “P_down”, “B_up” and “P_downB_up”, while green represents genes in profiles “P_up”, “B_down” and “P_upB_down”. Unfilled squares represent HIV-1 proteins. A: Overall interactions of significant functional genes with HIV-1 proteins; B: Functional categories of known gene interaction with HIV-1.

Figure 6

Co-expressed genes and their networks. Left: Number of genes in 6 representative profiles (Figure 3 middle left). Red nodes represent “key regulatory” genes while blue nodes represent other regulated genes. Node size represents the power of the interrelation among the nodes, and edges between two nodes represent interactions between genes (i.e. the more edges of a gene, the more genes connecting to it, the more central role it has within the network). Middle right: Name list of “key regulatory” genes with highest k-core. Green represents unknown genes; known functional genes were underlined. Right: significant function of these genes.

Figure 7

Schematic overview of the cellular genes altered in “B” cells. Up: Analysis strategy and results. Down: Schematic network was constructed by 245 genes involved in significant GOs (Figure 3) and 137 genes in co-expression networks (Figure 6) that can also be mapped on KEGG pathways. Processes and genes in profiles “B_up” and “P_down B_up” or “B_down” and “P_upB_down” are respectively represented by red or green nodes. Purple represents HIV proteins.

Figure 8

Actin cytoskeleton and endoplasmic reticulum associated protein degradation related genes in schematic network. A: Actin cytoskeleton related genes up-regulated in “B” cells. B: Endoplasmic reticulum associated protein degradation related genes up-regulated in “B” cells.

Figure 9

Verification of CCR5 and CXCR4 expression by flow cytometry. Cells of “R”, “P” and “B” were stained with antibodies to CCR5 (PE) and CXCR4 (PE) as described in Methods. Numbers indicate the percentage of gated subset that expressed CCR5 or CXCR4. These results have been replicated and confirmed in 10 individuals.