AUTHORS
Qian Qi, Yichao Li, PhD2, Tiffany Yee3, Lance Palmer Jr, PhD4, Yan Ju5, Lei Han, PhD6, Sheng Zhou, PhD7, Yong Cheng, PhD8
AFFILIATIONS
1 St. Jude Children's Research Hospital, Memphis, Tennessee, United States
2 St. Jude Children's Research Hospital, Memphis, Tennessee, United States
3 St. Jude Children's Research Hospital, Memphis, Tennessee, United States
4 St. Jude Children's Research Hospital, Memphis, Tennessee, United States
5 St. Jude Children's Research Hospital, Memphis, Tennessee, United States
6 St. Jude Children's Research Hospital, Memphis, Tennessee, United States
7 St. Jude Children's Research Hospital, Memphis, Tennessee, United States
8 St. Jude Children's Research Hospital, Memphis, Tennessee, United States
ABSTRACT
Background: Understanding how integrated genetic elements engage host chromatin is essential for defining principles of genome regulation and improving gene-therapy safety. Lentiviral (LV) vectors are widely used to deliver ectopic DNA, yet how integrated cis-regulatory elements (CREs) interact with host chromatin across diverse three-dimensional (3D) genomic contexts remains poorly understood. To address these gaps, we performed a locus-resolved, multi-omic analysis of chromatin remodeling induced by therapeutic DNA integration. We transduced Jurkat cells with a LV that has been used to treat X-linked severe combined immunodeficiency (SCID-X1). This LV contains a codon-optimized human IL2RG gene driven by a EF1a promoter. We included two versions of this vector in our study. One flanked by a pair of 400bp condensed chicken beta-globin chromatin insulator elements (cHS4) and another identical vector with a 147 bp deletion in the 3′ cHS4, which removes the CTCF binding sequences. These two LVs are referred to as CTCF-LV and dCTCF-LV, respectively.
Results: We identified 172 integration sites for CTCF-LV and 157 for dCTCF-LV. Over 70% of the LV integration sites are within 1kb of those mapped in the primary cells derived from SCID-X1 patients treated with the same CTCF-LV, indicating consistency between our system and primary cells regarding LV integration profiles. Also, the majority of integration sites were located in promoters or within actively transcribed regions. Our HiC analyses revealed no significant differences on A/B compartment (50kb resolution) and topologically associating domains (TAD, 10kb resolution) boundaries for both CTCF-LV and dCTCF-LV. Interestingly, when examining local chromatin structure around the insertion sites, CTCF-LV integrations reduced local insulation scores (n = 20, p = 5.1 × 10⁻³, two-tailed t-test) and dCTCF-LV increased local insulation scores (n = 25, p = 3.4 × 10⁻²). To examine direct interactions between integrated LVs and the host genome, we performed Capture Hi-C using oligo baits targeting LV sequences. No significant differences were found in terms of number of loops per integration site, loop length, distribution of loop anchor chromatin state and target gene expression. However, loops with CTCF-LV showed stronger loop strength (ChiCMaxima peak score) and CTCF ChIP-seq signal. Chromatin accessibility analyses revealed a significant increase in ATAC-seq signals at peaks nearby dCTCF-LV, not CTCF-LV. Pol IIS2 ChIP-seq data showed that transcript elongation can extend beyond the dCTCF-LV provirus, up to 10 kb downstream of the integration sites, but not CTCF-LV. RNA-seq differential gene expression and splicing analyses revealed that the LV integration orientation is a key determinant of regulatory impact. Specifically, 87.5% CTCF-LV and 92% dCTCF-LV upregulated genes were oriented in the same direction as the LV integration site. For upregulated genes with same-orientation integration, exons upstream of the LV integration site showed markedly higher expression than downstream exons. For downregulated host genes, exon-level patterns also varied with integration orientation: same-orientation integrations showed preferentially depleted downstream exon expression while upstream exons were relatively preserved, whereas opposite-orientation integrations showed a more uniform reduction across both upstream and downstream exons. To comprehensively assess the various factors affecting the functional impact of LV integration on host cells, we divided 329 LV sites into two groups (perturbing and non-perturbing) based on whether the integration led to significant changes in gene expression (fold change >1.5, FDR< 0.01) and trained a random forest model. The model performed well (ROC = 0.7 and PRC=0.48) and also showed the strand orientation of LV integration is crucial.
Conclusions: A central finding is that integration of ectopic regulatory DNA induces pronounced local chromatin remodeling while largely preserving higher-order genome architecture. A second principle is that the relative orientation between integrated DNA and host genes is a dominant determinant of regulatory outcome, reflecting a balance between opposing LV-driven mechanisms. Together, these findings provide a gain-of-function perspective on chromatin regulation, demonstrating how ectopic CRE-containing DNA reshapes local genome function within a largely stable 3D chromatin landscape.
DISCLOSURES
Yichao Li, PhD: No relevant conflicts of interest to declare.