Monocyte-derived macrophages exhibit distinct and more restricted HIV-1 integration site repertoire than CD4(+) T cells

Abstract

The host genetic landscape surrounding integrated HIV-1 has an impact on the fate of the provirus. Studies analysing HIV-1 integration sites in macrophages are scarce. We studied HIV-1 integration site patterns in monocyte-derived macrophages (MDMs) and activated CD4(+) T cells derived from seven antiretroviral therapy (ART)-treated HIV-1-infected individuals whose cells were infected ex vivo with autologous HIV-1 isolated during the acute phase of infection. A total of 1,484 unique HIV-1 integration sites were analysed. Their distribution in the human genome and genetic features, and the effects of HIV-1 integrase polymorphisms on the nucleotide selection specificity at these sites were indistinguishable between the two cell types, and among HIV-1 isolates. However, the repertoires of HIV-1-hosting gene clusters overlapped to a higher extent in MDMs than in CD4(+) T cells. The frequencies of HIV-1 integration events in genes encoding HIV-1-interacting proteins were also different between the two cell types. Lastly, HIV-1-hosting genes linked to clonal expansion of latently HIV-1-infected CD4(+) T cells were over-represented in gene hotspots identified in CD4(+) T cells but not in those identified in MDMs. Taken together, the repertoire of genes targeted by HIV-1 in MDMs is distinct from and more restricted than that of CD4(+) T cells.

Figure 1. Chromosomal distribution of HIV-1 integration sites in monocyte-derived macrophages and CD4⁺ T cells infected ex vivo.

Inner red lines indicate the position at which HIV-1 was found and outer blue peaks indicate gene densities. In each chromosome, red bands indicate centromeres, darker bands are AT-rich, and lighter bands are GC-rich.

Figure 2. Conservation and frequency plots of 20 host nucleotides upstream of integrated HIV-1.

(a) Left panel shows the nucleotide selection specificity of HIV-1 with a serine residue at position 119 of the HIV-1 integrase protein in a T122 background. Right panel shows the nucleotide selection specificity of HIV-1 with serine substituted by a proline residue at position 119 of the HIV-1 integrase protein in a T122 background. (b) Left panel shows the nucleotide selection specificity of HIV-1 with a threonine residue at position 122 of the HIV-1 integrase protein in an S119 background. Right panel shows the nucleotide selection specificity of HIV-1 with threonine substituted by an isoleucine residue at position 122 of the HIV-1 integrase protein in an S119 background. χ test: *p < 0.01 and **p < 0.0001 compared to the same nucleotide positions in the left panel.

Figure 3. Frequency of HIV-1 integration events in genes encoding HIV-1-interacting proteins.

A total of three MDM data sets (two derived from this study; closed circles) and eight CD4⁺ T cell data sets (two derived from this study; closed squares) were compared to three sets of 5,000 random genes generated in silico. Details of data sets derived from other studies (open symbols) are found in Table 2. HTLV-1 integration site data set served as an independent control. Two-tailed Mann-Whitney U test with 95% confidence interval was used to test for statistical significance.

Figure 4. Gene hotspots for HIV-1 integration identified in MDMs and CD4⁺ T cells.

Genes in which at least two independent HIV-1 integration events were identified in MDMs (n = 75) and CD4⁺ T cells (n = 33) are shown along with the HIV-1-infected individuals’ samples from which these HIV-1-hosting gene hotspots were identified. HIV-1-infected individuals’ samples are colour-coded and strand widths are directly proportional to the number of independent integration sites. Underlined in red are gene hotspots that were identified at least four times in both cell types.

Figure 5. Frequency of HIV-1 integration events in genes linked to clonal expansion of latently HIV-1-infected CD4⁺ T cells.

(a) Each HIV-1 integration site data set was first divided into two groups: one comprises genes in which two or more HIV-1 integration events were identified (gene hotspots) and the other comprises genes in which only single HIV-1 integration events were identified. The frequency of genes linked to clonal expansion of latently HIV-1-infected CD4⁺ T cells in the two groups of genes for each data set was determined and the ratio was computed. (b) Ratio of frequency of genes linked to clonal expansion of latently HIV-1-infected CD4⁺ T cells in gene hotspots to in genes with single HIV-1 integration events. Autologous and heterologous HIV-1 data sets were combined here (closed symbols). Details of data sets derived from other studies (open symbols) are found in Table 2. Each random control comprised 5,000 random genes generated in silico instead of genes linked to clonal expansion of latently HIV-1-infected CD4⁺ T cells. Two random controls were analysed for each data set. Two-tailed Mann-Whitney U test with 95% confidence interval was used to test for statistical significance.