An integrated genomics approach towards deciphering human genome codes shaping HIV-1 proviral transcription and fate

Abstract

A large body of work has revealed fundamental principles of HIV-1 integration into the human genome.However, the effect of the integration site to proviral transcription activity has so far remained elusive.Here we combine open-source, large-scale datasets including epigenetics, transcriptome, and 3D genome architecture to interrogate the chromatin states, transcription activity landscape, and nuclear sub-compartments around HIV-1 integration sites in CD4 + T cells to decipher human genome codes shaping the transcription of proviral classes defined based on their position and orientation in the genome. Using a Hidden Markov Model, we describe the importance of specific chromatin states and genome architecture in the control of HIV-1 transcription activity. Additionally, implementation of a machine-learning logistic regression model reveals upstream chromatin accessibility, transcription activity, and categorical nuclear sub-compartments as optimal features predicting HIV-1 transcriptional outcomes. We finally demonstrate clinical relevance by interrogating the positions of intact proviruses persisting in patients under suppressive therapy and provide a compass compatible with clinical decision-making.

HIV-1 Proviral Transcription and Latency in the New Era

“…As such, this major biomedical challenge demands a comprehensive definition of the molecular rules modulating proviral fate before we can even leverage this knowledge in the clinical setting. Studies in the "new era" may require deep, integrated genomic approaches, combined with the interrogation of patient samples and the potential implementation of deep learning models to both predict and test molecular features contributing to proviral expression and persistence [75]. First, human genome codes shaping proviral transcription could be deciphered by combining open-source, large-scale datasets (including epigenetics, transcriptome, and 3D genome architecture) to interrogate the chromatin states, transcription activity landscape, nuclear sub-compartments, TADs, and chromatin loops containing the major regulator CTCF around HIV integration sites in CD4 + T cells (Figure 3A).…”

“…Therefore, new attempts to capture generalizable model(s) of latency, such as combining cell line data with ex vivo approaches [236], is often at the forefront of identifying and confirming new compounds. It is clear the varying transcription activity within a reservoir is influenced by the heterogeneous integrated nature within a proviral population [75] (Figure 2). Even so, we must also consider the added complexity stemming from the different origins, or genesis, of each unique proviral integrations within the same individual (Figure 1).…”

“…By surveying the genetic environmental landscape in where proviruses have embedded themselves, many were able to define peculiar preferential "behaviors" of this intriguing virus. RIGs, genes where the provirus was found repeatedly integrated in patient samples, began to reveal the integration site could largely influence what would become the position effect phenomena (Figure 3) [75], in where the integration site neighborhood shift proviral activity. Indeed, it is clear the partialities extend beyond the 1D view of simply a proviral insertion into a linear scale, as the importance of 3D nuclear architecture and temporal order of events in respect to cell state have become more apparent.…”

“…This extends to screening for new drug compounds. In the "new era", it will be imperative to integrate multiple approaches (such as multiple genetic datasets, epigenetics, transcriptomes, proteomics, 3D genome architecture, nuclear topology, chromatin states, transcriptional activity, proviral position, and intactness) followed by deep learning [75] to predict and experiment with emphasis at the single-cell level and across multiple models including patient samples, and interrogate "deep" (Figure 5B). Understanding how each molecular rule fits together (simultaneously or stepwise) and which rules govern individual proviruses (i.e., is it the same set of rules for all proviruses or are there different kits/sets of rules for different proviruses) will aid personalized therapeutic development and eventual eradication of HIV.…”

Dynamics of Viral and Host 3D Genome Structure upon Infection

“…Finally, a comprehensive analysis of epigenomic, transcriptomic, and 3D genome architecture (Hi-C) datasets in combination with machine learning was undertaken to discern patterns characterizing HIV-1 proviral transcription. Specific chromatin states, active nuclear sub-compartments, and unique positions as well as orientations with respect to human genes and regulatory elements were all found to correlate with proviral transcriptional activity in CD4 + T cells [ 82 ].…”