Toward ‘n of one’ gene therapies: Scalable ex vivo adenine base editing via lipid nanoparticles in a sickle cell disease model

Abstract Sickle cell disease (SCD) affects millions globally and results from a β-globin gene mutation causing abnormal hemoglobin and rigid, sickle-shaped red blood cells. Current gene therapies focus on inducing fetal hemoglobin (HbF) or adding a functional β-globin gene, but approved commercial products face limitations including manufacturing constraints and extremely high cost. To address these issues, we developed an ex vivo hematopoietic stem cell (HSC) editing approach using lipid nanoparticles (LNPs) to deliver adenine base editor (ABE) mRNA and small guide RNA (sgRNA), with the aim of recreating beneficial naturally occurring mutations within the gamma-globin promoter that enhance HbF expression and reduce sickling. Ultimately, this workflow can be adapted to other genetic disorders that could benefit from a single-nucleotide correction in HSCs. Our long-term goal is to build a versatile and comprehensive semi-automated platform for delivery of clinical “n of one” to “n of 100” ex vivo gene editing in a GMP-compatible setting. We performed editing experiments in human healthy donor CD34+ HSCs using one of two candidate sgRNAs targeting known repressor binding sites in the gamma-globin promoter. We optimized lipid nanoparticle (LNP) composition and concentration alongside cell density to balance cost efficiency, reduce potential toxicity, and minimize off-target editing while maximizing editing efficiency. Using in silico tools (CRISPRoff, COSMID, IDT) and iGUIDE-sequencing, we assessed genome-wide off-target editing. One sgRNA showed minimal off-target activity, while the second sgRNA demonstrated broader off-target effects, including editing at exonic sites, though none were associated with known oncogenes. Within 48 hours of LNP addition, both tested guides edited the target nucleotide in a dose-dependent manner, producing greater than 70% A-to-G edits and disrupting the repressor binding sites of BCL11A and ZBTB7A (LRF). Although bystander edits within the editing window were observed - a known limitation of ABE8e - the affected nucleotides lie within non-coding regions and are distant from oncogenic loci, suggesting low risk. Nonetheless, we are actively evaluating alternative base editor variants with narrower editing windows to further optimize editing specificity. No toxicity or impairment of hematopoietic function was observed, as demonstrated by preserved frequency and diversity of colony-forming units (CFU-GM, CFU-GEMM, BFU-E, CFU-E) compared to untreated cells. Assays of hemoglobin production in erythroid-differentiated HSCs following editing showed substantial upregulation of fetal hemoglobin by HPLC and a corresponding increase in F-cell frequency. In conclusion, this study establishes a robust and scalable ex vivo hematopoietic stem cell editing platform utilizing LNP-delivered adenine base editors, demonstrating high editing efficiency and favorable safety profiles. Through optimization of LNP parameters and rigorous off-target assessments, we highlight a pathway towards a cost-effective and clinically viable point of care base-editing therapy for SCD. These findings lay essential groundwork for broader applications in a personalized, GMP-compliant “n of one” gene editing strategy aimed at addressing diverse rare hematopoietic disorders.