In Vivo Silencing of Regulatory Elements Using a Single AAV-CRISPRi Vector

Clustered regularly interspaced short palindromic repeats (CRISPR)-associated protein 9 nuclease (Cas9)-based transcriptional repressors (CRISPRi) have emerged as powerful tools for functional epigenetic silencing.¹ They rely on a nuclease-deficient Cas9 fused to a repressor domain (dCas9i) and a target-specific guide RNA. In contrast to gene editing approaches, dCas9-KRAB (Krüppel associated box domain)–mediated CRISPRi is generally reversible and titratable, and circumvents the risks associated with direct modification of the genomic sequence.

Adeno-associated viruses (AAVs) are promising nonintegrating vectors for gene therapy, but their limited packaging capacity makes them incompatible with large inserts such as the common S. pyogenes dCas9i. Here, we describe a CRISPR-based gene-silencing method in the heart relying on a single compact vector compatible with AAV delivery (AAVi).

We engineered an AAV vector containing both an S. aureus–derived nuclease-deficient Cas9 fused to a KRAB repressor domain² driven by a cytomegalovirus promoter and a guide RNA cassette under the control of the U6 promoter (Figure [A]). The cytomegalovirus promoter ensures robust expression and can be exchanged if cell type-specific expression is required. Organ and cell type selectivity can further be modulated by the selected capsid serotypes as shown for muscle tissue.³ The insert size of the engineered AAVi construct is 4.7 kb, reaching the upper limit insert size of the AAV genome. Nevertheless, the yields of AAV6 and AAV9 particles were comparable to those produced using an AAV vector with a smaller (3 kb) insert size (Figure [B]).

To show the functionality of AAVi for cardiac research, we silenced the NPPA (natriuretic peptide A) because NPPA is highly and specifically expressed in cardiomyocytes. We designed guide RNAs targeting the accessible Nppa proximal promoter region (AAVi^Nppa; Figure [A]) or without homology to mammalian genomes (AAVi^Ctrl). We packaged them into AAV6 particles to transduce cultured mouse HL-1 cardiomyocytes. AAVi^Nppa reduced Nppa gene expression in HL-1 cells in a dose-dependent manner after 1 week (Figure [C]). Due to cell proliferation–associated AAV dilution, the effect is then progressively lost (data not shown) as AAVi-mediated silencing is not heritable. Epigenome analysis using ATAC-seq (assay for transposase-accessible chromatin using sequencing) and ChIP-seq (chromatin Immunoprecipitation-sequencing) showed local loss of chromatin accessibility and a stretch of trimethylated histone H3 Lys-9 (H3K9me3)±2 kb around the target site, orchestrating loss of transcriptional activity (RNA sequencing) and confirming successful epigenetic silencing (Figure [C]). A genome-wide analysis of gene promoters revealed that de novo deposition of the heterochromatin mark was specific to the target promoter region of Nppa and did not spread to known interacting regions such as the Nppb promoter and the shared superenhancer (Figure [C])⁴. One H3K9me3 positive off-target site was identified at a nonregulatory site of the genome homologous with the 3′half of the spacer. We then confirmed the broad applicability of AAVi to different cardiomyocyte cell culture systems by silencing several genes in primary neonatal rat ventricular cardiomyocytes, human induced pluripotent stem cells (iPSC)-derived cardioids, and iPSC-cardiomyocytes (Figure [D]). For human iPSC–derived cells, we confirmed the effect of AAVi on NPPA RNA and protein levels, as representatively shown for cardioids by immunofluorescence and quantitatively for iPSC-cardiomyocytes by Western blot analysis. In total, 80% (17 of 21) of tested guide RNAs permitted efficient gene silencing documenting the high efficacy of AAVi.

Nppa transcription is induced during cardiac stress and heart disease. To demonstrate the robustness of AAVi in vivo, we aimed to antagonize this process in an experimental heart disease model. Therefore, we retro-orbitally injected 8-week-old male C57Bl/6 mice with a single dose of AAVi^Nppa or AAVi^Ctrl (n=6–7, 3x10E12 AAV9 VG/mouse). Seven days later, we implanted osmotic minipumps releasing phenylephrine (50 mg/kg per day) and angiotensin II (450 µg/kg per day) over 2 weeks to mimic neurohormonal stress signals and induce cardiac hypertrophy. AAVi^Nppa resulted in >7-fold downregulation of ventricular Nppa expression (Figure [E]). Single-nuclei RNA-seq analysis of cardiac nuclei further revealed that Nppa expression is detectable by snRNA-seq in 3.5% (41/1165) of cardiomyocyte nuclei (1151/1165 Tnnt2⁺) from mice treated with AAVi^Nppa compared with 95% (864/913) on AAVi^Ctrl (770/913 Tnnt2⁺), demonstrating homogenous downregulation of Nppa (Figure [E]). Notably, the efficacy of AAVi exceeded a previous study reporting a 32% reduction of Pcsk9 mRNA expression in mouse liver using an alternative single AAV-CRISPRi system.⁵ To show the effect on the epigenetic layer and the specificity of AAVi, we sorted cardiomyocyte nuclei from diseased hearts and analyzed changes in chromatin accessibility (ATAC-seq) and gene expression (nuclear RNA sequencing) on a genome-wide scale. AAVi^Nppa established closed chromatin at the Nppa promoter (Figure [F]), and Nppa was downregulated, while accessibility and expression of neighboring genes were not affected. Besides Nppa, we found only 25 genes mainly linked to antiviral response to be affected (n=3; q≤0.001).

In summary, we present an efficient AAV-based method for CRISPRi-mediated epigenetic silencing of gene expression in cardiac myocytes in vivo and in vitro. This functional epigenetic approach provides a novel and efficient way to modulate gene expression in the heart and could become a standard method for cardiovascular disease modeling and translational research.