AUTHORS
Jonathan Ross Lindgren, B.S., Giulia Pavani, PhD2, Tyler Papp3, Hamideh Parhiz, PhD4, Denise Sabatino, PhD5
AFFILIATIONS
1 The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
2 Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
3 Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States
4 Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States
5 Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, United States
ABSTRACT
Hemophilia A (HA) is an X-linked bleeding disorder due to a deficiency in coagulation factor VIII (FVIII). Current adeno-associated virus (AAV)-based gene therapy approaches target FVIII expression to hepatocytes; however, the liver sinusoidal endothelial cells (LSEC) are the primary endogenous source of FVIII synthesis. We hypothesized that targeted gene editing of LSECs at the endogenous F8 locus could provide therapeutic benefit by overcoming key barriers of current gene therapies, enabling durable FVIII expression. To achieve LSEC-targeted editing, we developed a dual-delivery platform consisting of antibody-conjugated LSEC-targeted lipid nanoparticles (LNPs) to deliver the CRISPR-Cas9 mRNA and guide RNAs (sgRNAs), together with a bioengineered LSEC-tropic AAV capsid for donor template delivery. LNPs were conjugated to LSEC-specific monoclonal antibodies (LYVE-1 or CD31) via maleimide-thiol chemistry. Our genome editing strategy uses promoter hijacking by targeting insertion of the human F8 cDNA into the 5′ untranslated region (UTR) or exon 1 of the endogenous F8 locus, allowing expression to be driven by the native F8 promoter. To evaluate LNP tropism, LYVE1-LNPs, CD31-LNPs, or unmodified LNPs containing ZsGreen mRNA were administered intravenously to hem A/CD4KO mice. Twenty-four hours later, hepatocyte and non-parenchymal liver cell fractions were isolated and analyzed by flow cytometry and immunofluorescence. LYVE1-LNPs demonstrated selective LSEC targeting, with 77% of LSECs (CD45−/CD31+/CD146+) expressing ZsGreen and minimal hepatocyte transduction (2%). CD31-LNPs achieved higher LSEC transduction (99%) but showed greater hepatocyte targeting (25%). Immunofluorescence staining of liver sections confirmed high LSEC-targeting (LYVE1+ZsGreen+) with both antibody- conjugated LNP formulations. To evaluate AAV and LNP co-targeting efficiency, mice received an engineered LSEC-targeted AAV (AAV-LSEC1) encoding eGFP (4 × 10^12 vg/kg), followed three weeks later by LNPs containing mCherry mRNA. Co-targeting was quantified as the percentage of GFP+mCherry+ LSECs. The highest co-targeting efficiency was achieved with LYVE1-LNPs (5mg), resulting in 23% dual-positive LSECs and minimal hepatocyte co-transduction (8%). These findings demonstrate the feasibility of simultaneous LSEC targeting using both AAV and LNP delivery platforms. Additional dosing regimens may further refine the ability to co-target a larger portion of LSECs. For F8 editing by promoter hijacking, sgRNAs targeting the 5′UTR and early exon 1 of the murine F8 locus were screened in vitro. Guide performance ranged from 6.0% to 64.3% on-target editing efficiency, with the top-performing guides generating 34.5%, 38.0%, 53.3%, and 64.3% editing. To evaluate the ability of the LNPs to target the endogenous sites of FVIII synthesis in vivo, sgRNAs (either guide 6 or guide 7 sgRNA) were packaged with Cas9 mRNA into LYVE1-LNPs, CD31-LNPs, or unmodified LNPs. Functional disruption of endogenous F8 expression in wild-type mice was assessed by FVIII activity (aPTT), while genome editing was quantified by Sanger sequencing of the F8 locus in isolated LSECs and hepatocytes as well as whole tissues (spleen, muscle (diaphragm), and lung). One month after LNP treatment, reductions in FVIII activity correlated with LSEC editing efficiency. CD31-LNP delivery of guides 6 or 7 produced the greatest reductions in FVIII activity (81.9% and 84.4% reduction in FVIII activity, respectively) and high genomic knockout scores in LSECs (79.0 and 68.0). LYVE1-LNP delivery of guides 6 or 7 resulted in 43.8% and 45.4% reduction in FVIII activity, with corresponding LSEC knockout scores of 34.7 and 32.7. These data show highly efficient, on-target editing of the F8 locus by LNP delivery of CRISPR/Cas9. Notably, LYVE1-LNPs showed no detectable editing in hepatocytes or extrahepatic tissues, whereas CD31-LNPs demonstrated low-level editing in the lung. Further studies to use these tools to deliver a donor template for targeted integration at the F8 locus are currently underway. Collectively, these studies establish novel LSEC-targeted LNP and AAV delivery systems capable of efficient and selective genome editing to the endogenous site of FVIII synthesis. This platform provides critical tools for targeted insertion of therapeutic F8 donor templates and may enable a durable, mutation-agnostic gene-editing strategy for the treatment of hemophilia A.
DISCLOSURES
Jonathan Ross Lindgren, B.S.: No relevant conflicts of interest to declare.