In Vivo RNA Delivery to Hematopoietic Stem and Progenitor Cells via Targeted Lipid Nanoparticles

Ex vivo autologous hematopoietic stem cell (HSC) gene therapy has provided new therapies for the treatment of hematological disorders. However, these therapies have several limitations owing to the manufacturing complexities and toxicity resulting from required conditioning regimens. Here, we developed a c-kit (CD117) antibody-targeted lipid nanoparticle (LNP) that, following a single intravenous injection, can deliver RNA (both siRNA and mRNA) to HSCs in vivo in rodents. This targeted delivery system does not require stem cell harvest, culture, or mobilization of HSCs to facilitate delivery. We also show that delivery of Cre recombinase mRNA at a dose of 1 mg kg–1 can facilitate gene editing to almost all (∼90%) hematopoietic stem and progenitor cells (HSPCs) in vivo, and edited cells retain their stemness and functionality to generate high levels of edited mature immune cells.

Nanoparticle formulation. Lipid nanoparticles were prepared via microfluidic mixing as previously described 2 . Briefly, an ethanol phase was prepared consisting of ionizable lipid, DSPC, cholesterol, PEG-lipid, and functionalized maleimide PEG-lipid at a ratio of 50:10:38:1.5:0.5 and then rapidly mixed with an aqueous solution of RNA in 10 mM citrate buffer pH 3.0 at an aqueous to ethanol volume ratio of 3:1. For in vivo uptake experiments, a lipid dye DiR (ThermoFisher) was included in the ethanol phase at 0.5 mol% for direct labeling of the nanoparticles. After formulation, LNPs were dialyzed using Slide-A-Lyzer 10K MWCO dialysis devices (ThermoFisher) for at least 18 hours at 4°C in PBS containing 10 mM EDTA.
For characterization, nanoparticle size and PDI were determined with Malvern ZetaSizer and RNA encapsulation efficiency was determined via Quant-IT RiboGreen (ThermoFisher) assay. Antibody conjugation. LNPs containing DSPE-PEG2000-Maleimide were dialyzed at 4°C overnight in PBS with 10 mM EDTA. To prepare antibodies for conjugation, antibodies were first reduced with 5 eq. of TCEP (5 mM stock solution in PBS, Sigma-Aldrich) at 37°C for 1 hr with gentle shaking. Excess TCEP was removed by passing antibody solution through a Zeba 7K MWCO desalting column (ThermoFisher). Concentration of reduced antibody was quantified via absorbance at 280 nm via NanoDrop (ThermoFisher). Antibody was then incubated with functionalized LNPs for one hour at room temperature with end-over-end mixing and then stored at 4°C prior to purification. Ab-LNPs were purified with size exclusion chromatography using bench-top qEVoriginal columns (Izon Sciences) with PBS as the mobile phase. Fractions containing Ab-LNPs were combined and then concentrated using Amicon ultracentrifugation filters (Sigma-Aldrich) and then stored at 4°C prior to injection. Protein concentration was determined by BCA assay (ThermoFisher).  In vivo CD45 knockdown. To assess silencing of CD45 in vivo, C57BL/6 mice (F, 6-10 weeks) were injected with varying doses of siCD45 LNPs via tail vein. 72 hours after administration, mice were sacrificed, and the right hind leg was dissected and then kept on ice prior to processing and flow cytometry staining. Knockdown was assessed by comparing the median fluorescence intensity of CD45 of treated LSK cells to the MFI of untreated control.
In vivo luciferase biodistribution. To assess biodistribution of Ab-LNPs, C57BL/6 mice (F, 6-10 weeks) were injected with 0.3 mg/kg of luciferase mRNA LNPs via tail vein. 6 hours after administration, mice were injected i.p. with 130 µL of a 30 mg/mL D-Luciferin (PerkinElmer) solution in PBS. Ten minutes after luciferin administration, mice were sacrificed, and major organs (lung, heart, liver, spleen, intestines, and femur & tibia, skin) were collected and imaged on an IVIS Spectrum imaging system (PerkinElmer) and luminescence was quantified with LivingImage software (PerkinElmer).
In vivo Cre delivery. Ai14 mice (6-10 weeks) were injected via tail vein with Cre mRNA Ab-LNPs. For analysis of bone marrow HSPCs, mice were sacrificed after 48 hours and then femur and tibia were processed for flow cytometry staining. For immune cell lineage tracing, 100 µL of whole blood was collected every two to three weeks and then stained for flow cytometry according to the protocol above.

Statistics.
Means were compared with either one-way ANOVA or two-way ANOVA for comparison between multiple groups. Dunnett's multiple comparison test was used when comparing to a control group while Tukey's multiple comparison test was used when comparing every mean to every other mean. For comparison between two groups, unpaired two-tailed T-test was used. All statistics were performed with GraphPad Prism 9.      Table S1. Nanoparticle physicochemical characterization before and after conjugation.  . Figure S7. Representative gating strategy for analysis of TdTomato erythrocytes at 14 weeks (n = 3 mice). Erythrocytes were gated based on live TER-119 + cells. TdTomato gate was drawn based on FMO controls.