1059. The Influence of a Chromatin Insulator on Interactions between Globin Regulatory Elements and Cellular Promoters in Erythroid Cells

Abstract

Integrating retroviral vectors have been developed for gene therapy of blood disorders because they achieve long term expression of the transgene. However, interactions between regulatory elements within integrating vectors and cellular genes have resulted in pathogenic proto-oncogene activation in a clinical trial (Science 302: 415-419, 2003). We have used a promoter trap design to evaluate interactions between globin regulatory elements and cellular genes (Mol Ther. 11: supp 1 S4, 2005). The trapping cassette consisting of a splice acceptor followed by the GFP coding sequences and a polyadenylation site was inserted immediately upstream of the 3|[prime]| LTR in a reverse orientation in a self-inactivating (SIN) lentiviral vector (Trap only). It was also inserted into vectors containing either hypersensitive sites 2, 3 and 4 (HS234) from the human |[beta]|-globin locus control region (LCR), HS234 and a |[gamma]|-globin gene driven by the |[beta]|-globin gene promoter (LCR |[beta]|G-Globin), HS234 and the DsRed coding sequences driven by the |[beta]|-globin promoter (LCR |[beta]|G-Red), or DsRed under the control of the MSCV LTR (MSCV-Red). Trapping efficiency was calculated as the percentage of GFP expressing cells among total transduced cells as estimated on Southern blots. In HeLa cells (n=7), the Trap only and LCR-containing vectors gave similar trapping efficiencies ranging from 9.4 to 11.6% while MSCV-Red gave a higher efficiency of 21.2|[plusmn]|9% (P=0.002). Trapping efficiency of the Trap only vector in K562 erythroleukemia cells was 10.1|[plusmn]|1.4% whereas those of the LCR, LCR |[beta]|G-Globin and LCR |[beta]|G-Red vectors were 19.6|[plusmn]|2.4, 21.8|[plusmn]|3.8 and 22.9|[plusmn]|5.3%, respectively (P<0.0001), and MSCV-Red was 14.1|[plusmn]|3% (P=0.0029). To investigate the mechanism of the LCR-mediated increased trapping efficiency, we isolated 248 randomly selected K562 clones transduced with the LCR |[beta]|G-Globin vector and 288K562 clones transduced with the Trap only vector. Forty-one and forty-three of these clones contained a vector genome, respectively. Three of the 43 Trap only clones expressed GFP whereas 8 of the 41 LCR |[beta]|G-Globin clones were positive for GFP expression consistent with the trapping efficiency observed in the population experiments. The integration site was mapped in 28 of the LCR |[beta]|G-Globin clones and in 40 of the Trap only clones; 71 and 75%, respectively, contained intragenic insertions with an equal distribution between tandem versus reverse orientations. There was no apparent bias for integration near promoters by the LCR |[beta]|G-Globin vector compared to the Trap only vector. The 1.2 kb chicken HS4 insulator was added to the 3' SIN LTR of the Trap only, LCR |[beta]|G-Globin and the MSCV-Red plasmid vectors; the integrated vector genomes have the insulator in both LTRs. The trapping efficiency of insulator (n=3) versus control vectors in K562 cells were as follows: Trap only |[ndash]| 10.1 versus 10.1% (P=0.78); LCR |[beta]|G-Globin |[ndash]| 8.2 versus 21.8% (P=0.03); and MSCV-Red |[ndash]| 4.4 versus 14.1% (P=0.02). These results suggest that the higher trapping efficiency of the vectors containing globin regulatory elements or the MSCV LTR reflect interactions that enhance the expression of cellular promoters which can be blocked by a chromatin insulator.