Improved Lentiviral Vectors for the Cure of Hemoglobinopathies

Given that both Sickle Cell Disease (SCD) and beta-thalassemia (BT) are caused by mutations in the beta-globin gene, several lentivirus-based gene addition therapies have been developed. Results from recent trials indicate that the vectors used are safe; however, their efficacy inversely correlates with the severity of patients' hemoglobinopathy. The severity of the mutations (non-beta0 vs beta0) largely influences the outcome of the gene transfer. In fact, the data indicate that a relatively low number of integrations (in the range of 1-2 copies per genome) or vector copy number (VCN) is sufficient to cure patients whose mutations are categorized as non-beta0 and express relative high levels of endogenous hemoglobins (adult hemoglobin, HbA, and/or fetal hemoglobin, HbF). In contrast, the same level of VCN alleviates the transfusion regimen of patients with beta0 mutations, but it does not cure them. In addition, the lentiviruses currently used in clinical trials were engineered by different groups and to date no one has directly compared them side by side.

In light of these limitations, here we describe a study that supplies a platform for rapid screening of lentiviral vectors expressing curative hemoglobin, based on the correlation between VCN and the increase in HbA levels. We also compared newly generated lentiviral vectors to vectors currently used in clinical trials. Our ultimate goal is to generate a new vector that can increase the yield of beta globin expressed per VCN in patients' cells.

Using CRISPR-Cas9 we modified the erythroid Hudep-2 cell line (Kurita et al, 2013) to generate a clonal cell line, named Hudep #M13, which, upon differentiation, produces a hemoglobin variant (HbMut) that can be discriminated from that produced by the lentiviruses (HbA). In parallel, we immortalized erythroid progenitor cells isolated from a SCD donor (SCD #13), using the HPV16-E6/E7 expression system, which was introduced into the cells by lentiviral transduction.

Using Hudep #M13, we compared the correlation between gene transfer and the production of HbA for 5 novel lentiviral vectors, indicated as ALS16-20. Our new vectors include the Ankyrin insulator in the 3' LTR (Breda et al 2012), the full beta-globin gene (including the native introns), the full 3' enhancer region, a combination of different portions of the beta-globin promoter, as well as modifications and inclusion of novel genomic elements from the locus control region (LCR). Our ALS- constructs were then compared to lentiviral vectors currently utilized in clinical trials. These constructs were reproduced based on information available from the literature (Negre et al, 2015; Miccio et al, 2008; and Boulad et al, 2014) and indicated as CV-1, CV-2, and CV-3, respectively. All these vectors contain the beta-globin gene with deletions in intron 2, different portions of the beta-globin promoter and/or 3' enhancer region, and different elements and sizes of the hypersensitive sites (HS) of the LCR.

In Hudep #M13, linear regression analysis of the ratio of HbA to vector copy number (VCN) for each treatment, indicates that ALS17 and ALS20 yield roughly 40, 157 and 84% more HbA per copy than CV-1, CV-2 and CV-3, respectively. Similar increment in HbA% were confirmed on primary and immortalized (SCD #13) SCD erythroblasts derived CD34+ cells isolated from patients' blood. In these specimens, ALS20 maintained a 40% HbA increase compared to CV-1, when exploring a range of VCN from 0 to 3 with a linear mixed effects model. To assess the ability of these constructs to increase hemoglobin content in vivo, we are performing murine bone marrow transplants using thalassemic hematopoietic stem cells treated with CV1 and our two most powerful vectors.

Based on most recently reported data (Thompson et al, 2018), 1 copy of the vector we reproduced as CV-1, makes on average 6.8g/dL of HbA. Hence, 1 copy of our best vector has the potential to make up to 9.5g/dL HbA. This could lead to a much greater clinical impact for patient with hemoglobinopathies, especially those who require higher Hb production to become transfusion independent, like patients with the beta0 genotype. The completion of these studies will provide not only a comparative analysis of our new best vector to those already in clinical trial, but also a way to predict how much therapeutic hemoglobin per vector copy number will be produced in the clinical setting.