Highlights
- •Chromatin and transcriptional profiling of HIV-1 integration sites in microglia
- •Architectural protein CTCF plays a role in HIV-1 integration and latent infection
- •TAD boundaries are a 3D genomic feature of HIV-1 integration in microglia and T cells
Summary
HIV-1 encounters the hierarchically organized host chromatin to stably integrate and persist in anatomically distinct latent reservoirs. The contribution of genome organization in HIV-1 infection has been largely understudied across different HIV-1 targets. Here, we determine HIV-1 integration sites (ISs), associate them with chromatin and expression signatures at different genomic scales in a microglia cell model, and profile them together with the primary T cell reservoir. HIV-1 insertions into introns of actively transcribed genes with IS hotspots in genic and super-enhancers, characteristic of blood cells, are maintained in the microglia cell model. Genome organization analysis reveals dynamic CCCTC-binding factor (CTCF) clusters in cells with active and repressed HIV-1 transcription, whereas CTCF removal impairs viral integration. We identify CTCF-enriched topologically associated domain (TAD) boundaries with signatures of transcriptionally active chromatin as HIV-1 integration determinants in microglia and CD4+ T cells, highlighting the importance of host genome organization in HIV-1 infection.
Introduction
Human immunodeficiency virus 1 (HIV-1), as a retrovirus, is defined by the ability to integrate its reverse-transcribed genome into the host DNA prior to proviral transcription by the cellular machinery. Embedded into hierarchically organized host chromatin of immune cells in the blood and in other tissue reservoirs, replication-competent proviruses represent the major barrier to a cure.1
HIV-1 integration, an essential step of the viral replication cycle, does not occur randomly within the host genome because HIV-1 recurrently targets transcriptionally active genes.2, 3, 4, 5, 6, 7, 8, 9
The integration process is dependent on the interactions of two viral proteins, integrase (IN) and capsid (CA), with their respective cellular binding partners lens epithelium-derived growth factor (LEDGF/p75)10 ,11 and cleavage and polyadenylation specificity factor 6 (CPSF6), which largely determine HIV-1 integration site preferences to gene bodies and transcriptionally active chromatin.12 ,13, 14
Within genomic regions of high transcriptional output, HIV-1 targets super-enhancers (SEs) and speckle-associated domains (SPADs).8, 9, 15
HIV-1 insertion patterns proximal to SEs contribute to HIV-1 integration hotspots in the transcriptionally active A1 subcompartment, whereas they are mostly depleted from B subcompartments15 and lamina-associated domains (LADs).13 ,16
A and B compartments, identified by genome-wide contact frequency mapping using chromosome conformation capture (Hi-C) and chromatin profiles, segregate chromosomes based on the long-range interactions within large (>1 Mb) gene-active and gene-inactive regions, respectively.17
On a more local scale (<1 Mb), DNA is folded into TADs characterized by a high frequency of internal DNA interactions that facilitate enhancer-promoter contacts.18 ,19 ,20
Genome folding into TADs is largely driven by cohesin-dependent loop extrusion until encountering CTCF.21 ,22 ,23 ,24 CTCF-enriched topological domain margins, referred to as TAD boundaries, are largely conserved among different cell types.18, 25, 26
Interestingly, chromatin folding mediated by CTCF has been suggested to influence HIV-1 expression and persistence in a recent chromatin profiling study using the assay for transposase-accessible chromatin sequencing (ATAC-seq) on actively and latently infected T cells.27
In patient-derived blood cells, simultaneous determination of integration sites (IS) and viral transcription and their mapping to host genome interaction matrices showed that frequent three-dimensional (3D) contacts coincide in regions where integration of transcriptionally active proviruses occurs.28
Furthermore, genome organization assessment suggested that distinct chromatin accessibility adjacent to intact proviruses could contribute to a selective advantage for the virus during long-term antiretroviral therapy (ART).28 ,29
These studies, underlining the link between dynamic changes in chromatin organization, HIV-1 integration patterns, and viral transcriptional control, are mostly derived from CD4+ T cells. The question that remains is what features of the host chromatin determine HIV-1 integration patterns and latency in other reservoirs, including gut-associated lymphoid tissues (GALT), bone marrow, and the central nervous system (CNS), all of which contribute to viral persistence.30, 31, 32
In the CNS, patient-derived samples support the existence of replication-competent proviruses, likely in microglia, years after ART.32 ,33 ,34 ,35 ,36 ,37 ,38 ,39
ART toxicity, microglial activation, and neuroinflammation, together with low levels of HIV-1 RNA, contribute to chronic inflammation and could lead to HIV-associated neurological disorders (HANDs) in people living with HIV (PLWHs).40 ,41 ,42 ,43 ,44 ,45 ,46
While there is currently no precise understanding of the nature of HIV-1 integrations in the brain of PLWHs, recent studies of HIV-negative patients mapped regulatory elements in major cell populations of the brain, highlighting the role of enhancers and the 3D genome in neurological diseases.47,48
This lays ground for an in-depth genomic analysis of the HIV-1 integration profiles in microglia.
In this study, we profiled HIV-1 ISs, transcriptional output, and chromatin signatures of HIV-1 targeted genomic regions in an unexplored brain reservoir, using microglia cell models. We report that HIV-1 ISs are found to be enriched in proximity to TAD borders marked by H3K36me3 gene body-associated chromatin and depleted of H3K9me3 heterochromatin in microglia and primary T cells. We observed that depletion of CTCF, a major architectural protein that insulates adjacent TADs, reduces HIV-1 integration levels and alters IS profiles. Our findings point to the importance of exploring the 3D genome organization and its functional implications for HIV-1 infection in blood and brain reservoirs, which could contribute to future therapy development.
Results
Microglia HIV-1 ISs map to introns of highly transcribed genes recurrently targeted in T cells
One of the major impediments in studying HIV-1 microglia infection is the limited tissue availability and cell numbers of primary human microglia. This specifically applies to downstream chromatin analysis when only a portion of cells is HIV-1 infected. We therefore chose a model cell line that resembles adult microglia in morphology, expresses several microglia markers, and has been established recently as an HIV-1 infection model in the CNS.49,50,51 The immortalized C20 microglia clone is derived from cryopreserved human microglia49 and was used here to first identify HIV-1 integration patterns. Three days post infection (3 dpi) with vesicular stomatitis virus (VSV)-G-pseudotyped HIV-1, viral integration, mRNA expression, and protein production were assayed to confirm productive infection (Figures S1A–S1C). Virus-genome junctions were amplified and sequenced by linker-mediated (LM) PCR sequencing, and 4,590 unique ISs were obtained using a dedicated processing pipeline (STAR Methods; Table S1). We evaluated the genomic characteristics of the ISs retrieved by comparing microglia model integration profiles with published profiles in CD4+ T cells (n = 13,544) (including our previous study; GEO: GSE134382) and monocyte-derived macrophages (MDMs) (n = 987).15, 52
Overall, chromosomal distribution of the HIV-1 integrations was comparable between the three cell types, and the majority of ISs were enriched within introns (58%), with 14.5% being in distal intergenic regions (Figures 1A and 1B ; Table S1). Genic integrations displayed higher similarity among T cells and microglia-like cells (Jaccard index, 0.209) compared with MDMs (Jaccard index, 0.096) (Figure 1C), with similar Gene Ontology (GO) terms mostly associated with chromatin modification-related processes (Figures S1F and S1G). In addition, highly targeted genes (harboring ≥ 5 HIV-1 insertions) in the microglia model were also frequently targeted in CD4+ T cells (56.8%) (Figures S1D and S1E) and transcriptionally active, as assessed by RNA sequencing (RNA-seq) (gene stratification is described in STAR Methods). Further global transcriptome analysis of microglia-like C20 targets showed that most integrations (91%) were within highly and medium transcribed genes (Figure 1D). Of note, HIV-1 ISs sequenced from infected human induced pluripotent stem cell (iPSC)-derived microglia displayed comparable insertion and transcription profiles (Figures S1H–S1J).
