Long-Term In Vivo Expression from a Self-Inactivating Lentiviral Vector Flanked by a Chromatin Insulator Element.

We have previously shown that self-inactivating lentiviral vectors infect quiescent hematopoietic stem cells (HSC), express long-term, resist proviral silencing in HSC and express in a lineage specific manner. However, their random integration into the host chromosome results in variable expression, dependent upon the flanking host chromatin (Mohamedali et al, Mol. Therapy 2004). Moreover, the recent occurrence of leukemogenesis from activation of a cellular oncogene by the viral enhancer elements calls for safer vector designs, with expression cassettes that can be ‘insulated’ from flanking cellular genes. We analyzed the role of the chicken β-globin locus hypersensitive site 4 insulator element (cHS4) in a self-inactivating (SIN) lentiviral vector in the RBC progeny of hematopoietic stem cells (HSC) in long term in vivo. We designed an erythroid-specific SIN-lentiviral vector I8HKGW, expressing GFP driven by the human ankyrin gene promoter and containing two erythroid-specific enhancer elements and compared it to an analogous vector I8HKGW-I, where the cHS4 insulator was inserted in the SIN deletion to flank the I8HKGW expression cassette at both ends upon integration. First, murine erythroleukemia (MEL) cells were transduced at <5% transduction efficiency and GFP+ cells were sorted to generate clones. Single copy MEL clones showed no difference in the mean GFP fluorescence intensity (MFI) between the I8HKGW⁺ and the I8HKGW-I⁺ MEL clones. However, there was a reduction in the chromatin position effect variegation (PEV), reflected by reduced coefficient of variation of GFP expression (CV) in I8HKGW-I clones (n=115; P<0.01), similar to in vitro results reported by Ramezani et al (Blood 2003). Next, we examined for expression and PEV in the RBC progeny of HSC, using the secondary murine bone marrow transplant model. Lethally irradiated C57Bl6 (CD45.2) mice were transplanted with I8HKGW and I8HKGW-I transduced B6SJL (CD45.1) Sca+Lin- HSC and 4–6 months later, secondary transplants were performed. Mice were analyzed 3–4 months following secondary transplants (n=43). While expression from both I8HKGW and I8HKGW-I vectors appeared similar in secondary mice (46±6.0% vs. 48±3.6% GFP+ RBC; MFI 31±2.6 vs. 29±1.4), there were 0.37 vs. 0.22 copies/cell in I8HKGW and I8HKGW-I secondary recipients, respectively (n=43), suggesting that the probability of GFP expression from I8HKGW-I vectors was superior when equalized for vector copy. The CV of GFP fluorescence in RBC was remarkably reduced to 55±1.7 in I8HKGW-I vs. 196±32 in I8HKGW RBC (P<0.001). We therefore, analyzed these data at a clonal level in secondary CFU-S and tertiary CFU-S. The I8HKGW-I secondary CFU-S had more GFP+ cells (32.4±4.4%) vs. I8HKGW CFU-S (8.1±1.2%, n=143, P<0.1x10E-11). Similarly, I8HKGW-I tertiary CFU-S also had more GFP+ cells (25±1.8%) vs. I8HKGW CFU-S (6.3±0.8%, n=166, P<0.3x10E-10). We also plated bone marrow from secondary mice in methylcellulose and analyzed GFP expression in individual BFU-E. The I8HKGW-I tertiary BFU-E had more GFP+ cells (28±3.9%) vs. I8HKGW BFU-E (11±5%, n=50, P<0.03) with significantly reduced CV (67 vs 125, n=50, P<6.6X10E-7). Taken together, the ‘insulated’ erythroid-specific SIN-lentiviral vector increased the probability of expression of proviral integrants and reduced PEV in vivo, resulting in higher, consistent transgene expression in the erythroid cell progeny of HSC. In addition, the enhancer blocking effect of the cHS4, although not tested here, would further improve bio-safety of these vectors for gene therapy for RBC disorders.