Characterization of Regulatory Elements in the Slc4a1 (Band 3) Locus To Create Safe and Effective Vectors for Globin Gene Therapy

Effective gene therapy for hemoglobin β-chain disorders, β-thalassemia and sickle cell disease, requires efficient, safe delivery of globin genes into hematopoietic stem cells (HSCs). Engraftment of ∼25% of HCSs expressing a globin gene at ∼20% the level of endogenous α-globin would be sufficient to improve both diseases. The β-globin promoter is inefficient, requiring sequences from the locus control region (LCR) to increase expression. LCR enhancers are included in globin gene therapy vectors, but unfortunately are prone to cryptic splicing and polyadenylation, resulting in low virus titer. In addition, the LCR enhancers are, in theory, capable of activating neighboring oncogenes. To improve safety and efficiency of globin vectors we have developed a novel strategy by fusing the γ-globin gene to LCR-independent, erythroid-specific promoters. Band 3/AE1 is an erythrocyte membrane protein expressed from the Slc4a1 gene. We have previously shown in transgenic mice that a 1750-bp Slc4a1 promoter linked to γ-globin gene (pSlc4a1/γ) and flanked by the chicken β-globin insulator 5′HS4 (ch5′HS4), which contains both barrier and enhancer-blocking elements, is capable of erythroid-specific, uniform γ-globin expression at therapeutic levels (∼19.8% α-globin/transgene copy). Without ch5′HS4, the pSlc4a1/γ gene was prone to silencing. Lentiviral vectors with two copies of ch5′HS4 either internal or in the Long Terminal Repeat cannot be produced at high titer. We hypothesized that flanking the pSlc4a1/γ-globin gene with distinct barrier elements would prevent recombination and gene silencing. Using this strategy, we developed first generation lentiviral vectors in which pSlc4a1/γ is flanked by combinations of the ch5′HS4 insulator and barrier elements we have identified in the ankyrin and α-spectrin loci. To test the effectiveness of these lentivirus vectors in mouse models, we pseudotyped each one with an ecotropic envelope. All three were produced at high titer (>1x106 infectious units/ml). We transduced primitive mouse hematopoietic progenitor cells and detected γ-globin mRNA in >20% of spleen foci at levels as high as 17% of endogenous α-globin. In mice repopulated with transduced stem and progenitor cells, 11–15% of peripheral blood erythrocytes were positive for γ-globin 8 to 21 weeks post-transplantation. To establish the safety of the Slc4a1 promoter we used a high throughput real-time PCR-based assay to identify DNaseI hypersensitive sites (HS) in a 119kb region including Slc4a1. We have identified 6 HS and tested each for enhancer and enhancer-blocking activity. One HS (−355 to −112) increases reporter gene expression in a position- and orientation-independent fashion consistent with the properties of an enhancer. A second HS (−112 to 0) is active in enhancer-blocking assays, and deletion analyses indicate that this region may also contain a transcriptional silencer. A third HS in intron 1 (+910 to +1581) displays enhancer-blocking activity. Three HS have no activity. We are testing a second generation of pSlc4a1/γ lentiviruses in which the Slc4a1 enhancer and silencer are deleted. A third generation of vectors flank pSlc4a1/γ upstream with ch5′HS4, and downstream with either the ankyrin or α-spectrin barrier elements plus the Slc4a1 intron 1 enhancer-blocker to prevent activation of neighboring genes. We hypothesize that these new vectors will allow safe expression of therapeutic levels of γ-globin.