CD90-Targeted Cocal-Pseudotyped Lentivirus As a Robust Platform for Human HSC Gene Therapy

Ex vivo gene delivery and editing in hematopoietic stem and progenitor cells (HSPCs) provide a promising approach for the treatment of many chronic conditions such as hemoglobinopathies, lysosomal storage diseases, and HIV. However, such approaches require highly specialized biomedical infrastructure limiting its accessibility to most patients who could benefit. The development of a targeted in vivo approach to HSPC gene delivery/editing is therefore necessary for the broadened use of gene therapies. The CD34 ⁺CD90 ⁺ subset of HSPCs represents a highly enriched population for both short-term and long-term hematopoietic lineage repopulation capacity as seen in the non-human primate (NHP) transplant model. Furthermore, ex vivo gene editing of solely CD34 ⁺CD90 ⁺ Human and NHP hematopoietic stem cells (HSCs) was sufficient to produce therapeutic thresholds for sickle cell phenotypic amelioration after transplantation. Thus, the development of novel gene therapy vectors targeting the CD34 ⁺CD90 ⁺ HSC population is crucial for more accessible and robust gene therapy platforms. Here we report the development of a novel CD90-targeted Cocal pseudotyped lentivirus system capable of targeted transduction of HSPCs in vitro and in vivo. Additionally, we compare transduction efficiency and specificity of this new technology with the more routinely used VSVG targeting methodology.

To produce targeted lentivirus, native targeting of VSVG or Cocal envelope to the LDL-Receptor (LDLR) was knocked out by amino acid substitution. An anti-CD90 single-chain variable fragment (scFv) was introduced onto the mutated capsids using CD8 hinge and ICAM transmembrane domains. Mutagenesis of Cocal envelope LDLR binding motif was performed by Gibson cloning and confirmed via nanopore sequencing. Successful mutated viral production was assessed using hydrodynamic titration methods followed by functional analysis of transduction on CD90 ⁺ cell lines. In vitro transduction specificity and efficiency assays were performed on Human GCSF-mobilized CD34 ⁺enriched apheresis samples and analyzed by flow cytometry. Humanized NBSGW mice were treated with targeted-lentiviral particles via intravenous (IV) and intraosseous (IO) injection. Mice were necropsied 4-8 weeks post injection. Transgene expression was analyzed by flow cytometry or luminescence imaging and confirmed by PCR on human-CD34 ⁺ FACS-sorted cells from necropsied mice.

CD90-Targeted Cocal-pseudotyped lentivirus (Cocal-CD90) transduced CD34 ⁺CD90 ⁺ HSCs 2.12-fold (p=.017) more robustly than CD90-targeted VSVG-pseudotyped lentivirus (VSVG-CD90) in vitro. Cocal-CD90 virus also showed significantly greater on-target transduction specificity of CD34 ⁺CD90 ⁺ HSCs (p=.0008) compared to untargeted wild-type Cocal-pseudotyped lentivirus (Cocal-WT); but similar specificity to VSVG-CD90 virus (p=.118). Unlike VSVG-CD90 virus, we did not observe decreased transduction efficiency of CD34 ⁺CD90 ⁺ HSCs using Cocal-CD90 virus in comparison to Cocal-WT virus. IV and IO injection of both CD90-targeted viruses in humanized mice resulted in transduction of immobilized human HSPCs (CD34 ⁺CD38 ^-CD133 ⁺) in the bone marrow. Transduced cells maintained multilineage differentiation capacity as seen by transgene expression of T-cell, B-cell, and myeloid compartments in the bone marrow, spleen, and peripheral blood at necropsy. As observed in vitro, Cocal-CD90 treated mice consistently displayed more robust transgene expression than VSVG-CD90 treated mice. Importantly, transduction of off-target organs (e.g., liver and lungs) post-IV injection with targeted virus was not detected by luminescence imaging during 8 weeks of follow-up or at necropsy.

We provide the first evidence of targeted engineering of a knockout Cocal envelope using the attenuation of native LDLR tropism coupled with a CD90-scFv fusogen protein. Cocal-CD90 virus has increased specificity for CD90 ⁺ transduction compared to Cocal-WT lentivirus, without lowering transduction efficiency of on-target cells. Furthermore, in comparison to targeted VSVG engineering, Cocal engineered virus does not decrease transduction efficiency of HSCs in vitro or in vivo. Coupled with Cocal lentiviruses' resistance to human serum inactivation and low immunogenicity, targeted Cocal viral vectors are more suitable for future in vivo gene delivery platforms than their VSVG counterparts.