A Novel Combination of Chicken Hypersensitive Site-4 Insulator Elements Improves Titers and Restores Full Insulator Activity.

Chromatin insulators separate active transcriptional domains and block the spread of heterochromatin in the genome. A prototypic insulator, the 1.2Kb chicken hypersensitive site-4 (cHS4) element utilizes CTCF and USF-1/2 motifs in the proximal 250bp. This smaller “core” cHS4 element provides enhancer blocking activity and reduces position effects. However, the cHS4 core sequences alone do not insulate viral vectors effectively. Two copies of the core, while effective in plasmid based systems, are unstable in viral vectors. In contrast, the full-length cHS4 has excellent insulating properties, but its large size severely compromises vector titers (Urbinati et al, Mol Ther 2009). Therefore, we performed a structure-function analysis of the full-length cHS4 in the context of self inactivating lentivirus-vectors to identify minimal insulator elements required for optimal insulation. Specifically, we analyzed transgene expression in the clonal progeny of primary murine hematopoietic stem cells using the secondary bone marrow transplant assay, and analyzed epigenetic changes in cHS4 and the transgene promoter in vitro in clonal integrants. As expected, the full length cHS4 insulator reduced clonal variegation in transgene expression, reduced position effects and blocked silencing-associated epigenetic modifications over the insulator core and the transgene promoter. However, while either the 5′ cHS4 250bp core or the 3′ cHS4 400bp sequences similarly effected only lower clonal variegation in transgene expression, when they were combined these 650bp sequences recapitulated the activity of the full length 1.2kb insulator, with minimal impact on viral titer. The distal 3′ 400bp fragment contains no consensus sites for USF or CTCF. However, ChIP analysis on proviruses carrying only the 3′ 400bp showed that it binds CTCF. USF-1 binding to the 3′ 400bp, however, only occurred when both the 5′ 250bp core and the 3′ 400bp fragment were present in the proviruses. Indeed, the silencing associated epigenetic marks over the 3′ 400bp region were blocked only when the vector carried both these ends of cHS4 insulator sequences, suggesting that USF-1 bridges the 5′ core and the 3′ 400bp to confer full insulator activity. Furthermore, the 650bp sequences or the full length insulator had the maximal reduction in clonal dominance in the in vitro immortalization assay of lineage negative primary murine hematopoietic cells (Arumugam et al, Mol Ther. In press). Our studies confirm and extend earlier observations on the 5′ 250bp insulator core and identify a new “core-like” insulator activity in the 3′ end of cHS4. The specific elements in the 3′ 400bp sequences that promote interaction with the 5′ 250bp sequences would be important to determine, and may be present in other insulators in and across the genome/s. In the meanwhile, new vector systems flanked by this optimized ‘650bp’ cHS4 sequences, can provide excellent insulation of the transgene without significant loss in viral titers and have important safety and efficacy implications for gene therapy. Our data have important implications in understanding the molecular basis of insulator function and design of gene therapy vectors.

No relevant conflicts of interest to declare.