Optimization of lipid nanoparticle formulation for mRNA delivery to the heart

Introduction:: Although cardiovascular disease is the leading cause of death worldwide, innovation in heart failure (HF) therapeutics has been sparse¹. A primary reason behind the lack of therapeutic development is the inability to use phenotypic tissue-level approaches to discover novel therapies, such as complex in vitro models like microphysiological systems (MPS). In recent years, therapeutics that increase the expression of specific genes have been explored, but they have limited clinical translatability due to the lack of a safe and effective delivery system. A notable obstacle is that dense cardiac tissues suffer from low transfection efficiency of non-proliferative cardiomyocytes, diffusional barriers posed by the extracellular matrix of 3D cardiac muscle, and potential immunogenicity and carcinogenicity associated with viral vectors². Recent progress in the development of non-viral vectors like lipid nanoparticles (LNPs) holds great promise in overcoming these limitations and can make a breakthrough in cardiovascular medicine due to the transient nature of mRNA transfection. In this work, we developed a phenotypic tissue-level cardiac MPS, containing a heart micromuscle derived from human induced pluripotent stem cells³ (hiPSC) (Fig.1A,B), to identify LNP formulations that successfully diffuse into dense 3D cardiac tissue to transfect cells of the heart. For the LNP screens, we used a genetically modified hiPSC line (XR1 hiPSC) containing an eGFP gene downstream of a STOP cassette, which was removed upon CRE recombinase mRNA (CRE) delivery. We evaluated the effectiveness of an in vitro optimized LNP for mRNA delivery via intramyocardial injection in Ai6 mice.


Materials and Methods:: We screened different LNP formulations in 2D monolayers of hiPSC-derived cardiomyocytes (iCM) and the iCM-based cardiac MPS to assess LNP diffusion and gene editing efficiency. We developed a new class of LNPs containing an acid-degradable PEG-cholesterol designed to disrupt endosomes and improve delivery of their cargo intracellularly (Fig. 1C). We first quantified gene editing efficiency in the iCM monolayers by assessing the %eGFP+ cells using flow cytometry. We selected the best candidates to study their diffusion and gene editing in the 3D cardiac MPS. LNPs were delivered to microtissue via media channels overnight, where they diffused through the endothelial-like barrier that connects the media channel and tissue chamber (Fig. 1B). Volumetric nuclear segmentation based on the DAPI signal was used to measure %eGFP+ cells. We also quantified LNP diffusion from the tissue edges adjacent to the media channel to the center of the cardiac tissues. Next, we evaluated tissue physiology and gene editing efficiency in more complex micromuscles formed by isogenic 80% XR1 iCM and 20% hiPSC-derived cardiac fibroblasts (icFb). The efficacy of mRNA-LNP delivery in Ai6 murine hearts was assessed by intramyocardial injection and determining the organ distribution of two LNP formulations (candidate and standard LNP) containing firefly luciferase mRNA. Lastly, we injected LNPs containing CRE mRNA into the mouse heart. Like XR1-iCMs, the Ai6 line harbors a silent GFP variant cassette that becomes activated upon CRE-mediated deletion, allowing us to evaluate LNP efficiency and localization of expression within the heart.


Results, Conclusions, and Discussions::
We observed 50-55% eGFP+ iCMs for the LNP/CRE mRNA formulations containing a PEG-cholesterol ratio between 5% and 20%, regardless of the PEG molecular weight (1-2 kDa). Based on these results, we screened LNPs in the 3D cardiac MPS, identifying the formulations that exhibited the highest gene editing efficiency. We showed that when we increased the 2kDa acid-degradable PEG-cholesterol ratio in the formulations from 1% to 10%, the LNP diffusion to the center of the tissue increased significantly, resulting in ~25% eGFP+ iCM in the micromuscle (Fig. 1D,E). Complex isogenic micromuscles formed by 8:2 iCM-icFb were successfully edited in both cell types without specific preference, demonstrating the potential of using the 2kDa 10% PEG-LNP formulation as an mRNA delivery vehicle in tissues with a composition similar to the heart. Next, we measured the action potential duration and peak twitch amplitude of treated and control micromuscles before and over 3 days after LNP/CRE delivery. No negative effects on either the electrophysiology or contractile activity of the heart micromuscle were observed. Finally, our in vivo results in Ai6 mice illustrated that 2kDa PEG 10% LNPs containing CRE recombinase had high levels of accumulation in the heart and low off-target levels of liver uptake compared to a standard LNP formulation. Additionally, we observed significant eGFP expression in the left ventricle one week after injection (Fig. 1F). In conclusion, we demonstrated that by using the cardiac MPS for in vitro screening in dense 3D micromuscles, we identified a LNP formulation for significant in vivo mRNA delivery to the heart. We foresee using this effective LNP delivery system as a starting point to develop novel therapeutics to treat HF and genetic disorders in the heart.










Acknowledgements (Optional): :

References (Optional): : ¹Virani, et al. Heart Disease and Stroke Statistics—2021 Update. Circulation, (2021).
²Huebsch, et al. Metabolically driven maturation of human-induced-pluripotent-stem-cell-derived cardiac microtissues on microfluidic chips. Nat. Biomed. Eng, (2022).
³van Haasteren, et al. The delivery challenge: fulfilling the promise of therapeutic genome editing. Nat Biotechnology, (2020)