Elements of a Chromatin Insulator Necessary for Consistent, High-Level Globin Gene Expression

Vectors flanked by the chicken b-globin locus hypersensitive site-4 (cHS4) insulator provide consistent, improved expression (chromatin barrier activity) and reduced genotoxicity (enhancer-blocking activity); features important for randomly integrating viral vectors. However, full-length 1.2Kb cHS4 reduces vector titers significantly. Initial studies mapped the enhancer blocking activity of cHS4 to the first 250bp ‘core’ in a copy-number dependent manner (Felsenfeld and colleagues). However, elements required for the barrier effect are not well defined. We studied various regions of cHS4 for chromatin barrier effect. Beta-globin-LCR lentivirus vectors were constructed with the core (sBG^C), 2 copies of the core (sBG^2C), 400bp (sBG⁴⁰⁰) and 800bp (sBG⁸⁰⁰) fragments (both including the core) and compared to the un-insulated vector (sBG) and the vector with the full-length cHS4 (sBG-I). Barrier activity (improved probability and consistency of expression from individual integrants) in MEL cells and the clonal progeny of hematopoietic stem cells was analyzed. Single-copy MEL clones showed that sBG-I clones had a significantly higher proportion of hβ+ cells/clone. However, proportion of hβ+ cells from all other vectors carrying various fragments of cHS4 were not significantly different from the un-insulated vector (Table 1). Interestingly, all clones carrying a fragment containing the core showed reduced clonal variegation (coefficient of variation of hβ expression). We next analyzed hβ expression from the sBG, sBG^C, sBG^2C, sBG⁴⁰⁰ and sBG-I vectors in primary and secondary thalassemia (Hbb th3/+) mice. In the primary mice and secondary CFU-S analyzed thus far (Table 2), it appears that all vectors with cHS4 insulator fragments containing the core provide some barrier activity, but the full-length insulator is necessary for maximum barrier effect. To corroborate whether elements outside the core contribute to barrier activity, we also tested vectors with inert DNA spacers in addition to the core (sBG^400−S, sBG-^800−S and sBG^1200−S). Single copy MEL clones confirmed that the sBG^400−S, sBG^800−S and ^1200−S vectors had significantly lower proportion of hβ+ cells/clone [41±4.4%, 41±7% and 28±6% respectively, n=36]. These data suggest that regions of 950 bp region of cHS4 outside the core contains elements that cooperate with the core for optimal barrier activity. These studies have important implications for vector design for gene therapy for thalassemia, where variability in expression from uninsulated vectors compromizes efficacy.