Viral and physical methods of hematopoietic stem and progenitor cell (HSPC) gene therapy have many drawbacks like low efficiency and carcinogenicity. Red blood cell extracellular vesicle (RBCEV) holds natural features such as blood cell homing and other superiorities, including safety and cargo capability, that may serve as an ideal delivery vehicle. RBCEV produced by adenosine Triphosphate (ATP) depletion and calcium loading is associated with red blood cell (RBC) membrane calcium pump activity decrease and external anionic phospholipid exposure. However, the application of RBCEV to gene therapy is not well studied, and a better understanding of the biogenesis mechanism could stimulate the manufacture of RBCEV. Here, we developed a method to deliver therapeutic nucleic acids for gene therapy of blood cells by RBCEV, and proved protein degradation leads to RBCEV biogenesis, which will help large-scale production.

In this study, we utilized RBCEV for delivering nucleic acids to 293T cells, CD34⁺ cells, mouse bone marrow cells (BMCs), and mouse HSPCs to drive treated cells to express proteins as green fluorescent protein (GFP) and Cas9. First, we found that CD34⁺ cells, human and mouse BMCs strongly ingest RBCEV respectively at 90%, 60%, and 70% in vitro. After single and multiple intravenous injections, mouse HSPCs uptake ratio was 20% and 40%, respectively. Next, GFP expression in 293T cells was 80%, while CD34⁺, mouse BMCs, and HSPCs were 45%, 30%, and 20%, respectively, after the delivery by GFP mRNA and plasmid DNA loaded RBCEV. The functional GFP expression suggested that RBCEV can encapsulate and deliver nucleic acids to different cells. Following this, we applied RBCEV for gene therapy delivery to 293T and CD34⁺ cells using DNA or RNA of Cas9 and sgRNA intended to disrupt the BCL11a binding site upstream of gamma-globin, which will facilitate the globin switch and supplement the beta-globin deficiency. Genetic mutations at the specific location were detected in 293T and CD34⁺ cells by T7 endonuclease I cleavage and sequencing. Additionally, we observed an increase in gamma/beta-globin ratio in treated CD34⁺ cells without disrupting erythroid differentiation. Subsequently, we investigated the applicability of RBCEV for the treatment of wild-type and blood disorder mice. In wild-type mice, 10% and 20% HSPCs expressed GFP after single and multiple injections with GFP mRNA loaded RBCEV, while only 7% of positive HSPCs in mice received multiple injections with Cas9-copGFP plasmid DNA loaded RBCEV. For the beta-thalassemia mouse model (Hbb-b1tm1Unc Hbb-b2tm1Unc), the in vitro uptake in BMCs was 70%, similar to the wild-type mouse. The positive ratio of HSPC uptake and GFP expression after multiple injections was about 10%, probably due to the abnormal clearance in the blood system.

After studying the application of RBCEV, scale production represents another barrier to using RBCEV for treatment. Here, we analyzed RBCEV proteome and attempted to discover the limitations during production. Compared to RBC, the ubiquitination processes clustered in up-regulated proteins while the proteasome-related processes clustered in down-regulated proteins of RBCEV, revealing that protein degradation plays an important role in RBCEV biogenesis. Meanwhile, high level of reactive oxygen species (ROS) was generated in RBC and RBCEV. The proteasome inhibitor, MG132, reduced the RBCEV production by 30%. Natural reductant ascorbic acid eliminated ROS in RBCEV, possibly reducing the side effects caused by ROS in the application.

Our results demonstrated that RBCEV delivers nucleic acids to HSPCs and achieves gene therapy without affecting erythropoiesis. Protein degradation triggers RBCEV biogenesis and generates a high level of ROS in RBC and RBCEV that can be reduced by ascorbic acid. However, the efficiency of systemic treatment needs further improvement. Our study illustrates the application of RBCEV as a non-viral gene editing vehicle for HSPC, and the mechanisms of RBCEV biogenesis also benefit the scaling up production.