Preclinical Evaluation of ALS20, a New and Improved Lentiviral Vector for Beta-Globinopathies

The ongoing clinical trials, based on the use of lenti-globin vectors for beta-globinopathies, indicate that current vectors require high number of integrations (~3-4 copies per genome) in a pancellular fashion to make a patient transfusion independent, regardless of the genotype. This increased VCN requirement could increase the risk of genome toxicity, limiting the application of these vectors and preventing their use in a reduced myeloablative regimen. To overcome this limitation we designed new vectors, using modification or inclusion of a variety of regulatory genomic elements aiming to increase expression of the beta-globin gene. We performed in vitro and in vivo studies to compare the ability of these constructs to express the therapeutic gene with a low number of integrations and reduced chimerism.

We screened new constructs using a CRISPR-Cas9 modified clonal cell line, HUDEPM#13, which derives from the previously described erythroid HUDEP-2 cell line (Kurita, 2013). Upon differentiation, HUDEPM#13 cells produce a hemoglobin variant (HbMut) that can be discriminated by liquid chromatography from the adult hemoglobin (HbA) produced by the transgene in the lentiviruses. Among our candidates we identified a construct, indicated as ALS20, that synthesizes HbA at high level at a single integration (VCN). Moreover, ALS20 produced significantly more HbA per copy than constructs currently utilized in clinical trials, which were reproduced based on the literature (Negre, 2015; Miccio, 2008; and Boulad, 2014) and designated CV-1, CV-2, and CV-3, respectively.

In erythroblasts differentiated in vitro from patients with SCD ALS20 produces, on average, 21% HbA at VCN=1(P<0.001). CV1, which proved to be the most powerful among the clinical vectors tested in HUDEP#M13, produces, on average, 14.9% HbA at VCN=1 (P<0.001). Overall, ALS20 expresses 40% more HbA per VCN in patients' cells, confirming results observed in the mutant HUDEP cell line.

In a parallel potency assay, we tested ALS20 using beta0/beta0 thalassemic specimens, which represent the most severe phenotype to correct, due to complete absence of HbA production. Statistical analyses indicate that ALS20 produces, on average, 32.5% HbA at VCN=1 (P<0.001). All results were obtained targeting a range between 0 and 3 integrations, using dilutions of the viral product.

Our ongoing bone marrow transplantation studies, using a semi-myeloablative conditioning based on busulfan administration on the beta thalassemia Hbb^th3/+mouse model, indicate that ALS20 is curative at VCN lower than 1. In fact, mice with an average of 0.8 copies per genome and 65% transgenic chimerism, present Hb levels of 13.6g/dL, reiterating the potential of this new lentivirus in vivo. In immunocompromised NSG mice, injection of human CD34⁺cells transduced with ALS20 did not trigger tumor formation or meaningful pathological changes. The safety of our vector has been further assessed using an in vitro immortalization assay on primary mouse BM cells, in which no clonal survival has been observed. Finally, genome integration analyses on human CD34⁺cells infected with ALS20 showed that none of the examined samples had clones that exceeded the 20% abundance threshold, in accordance with the expected level of safety for gene therapy-based approaches.

In summary, we have identified a powerful new lentiviral vector with an enhanced ability to synthesize hemoglobin with a low number of integrations. ALS20's performance has been demonstrated in specimens from patients with hemoglobinopathies as well as in mice affected by thalassemia. Treatment of thalassemia and SCD with ALS20 could reduce the risk of cytotoxic events due to high levels of integration, and also lessen the intensity of the myeloablative regimen to correct anemia in patients. This, in addition with its confirmed safety features, makes this construct an outstanding candidate for clinical trial.