Molecular cloning and mRNA production
Table of Contents
ETRs were transiently delivered as either plasmid DNA (in cell lines) or mRNA (in cell lines, primary cells and in vivo). To this end, ETRs were cloned in an expression vector containing: (i) an upstream CMV promoter; (ii) an upstream T7 promoter for mRNA in vitro transcription (IVT); (iii) a downstream WPRE signal; (iv) a 3′-terminal stretch of 64 adenines (64A); and (v) a SpeI plasmid linearization site for mRNA IVT (Extended Data Fig. 1b). The coding sequences of the dCas9-based ETRs were previously described15. For plasmid-mediated gRNA expression, the crRNA sequences were cloned into a previously described expression vector containing a U6 promoter and the sequence of the Staphylococcus pyogenes Cas9 trRNA39. gRNAs targeting the CGI of the mouse Pcsk9 were designed using Chop Chop40 (https://chopchop.cbu.uib.no/) and selected according to high simulated activity and specificity. ZFPs and TALEs were designed and synthesized by Merck and Thermo Fisher Scientific, respectively, and subcloned into the mammalian expression plasmid in place of the dCas9 sequence. mRNAs were produced by IVT using the T7 Megascript Kit (Thermo Fisher Scientific, AMB1334-5) according to the manufacturer’s instructions. For the in vitro experiments, partially modified mRNAs were produced by IVT, including the following modifications to the standard protocol: (i) inclusion of the anti-reverse cap analogue 3´-O-Me-m7G(5′)ppp(5′)G (NEB, M0251) at a final concentration of 8 mM; and (ii) reduction of the GTP concentration from 7.5 to 2.5 mM. For the in vivo experiments, heavily modified mRNAs were produced by IVT, including the following modifications to the standard protocol: (i) inclusion of CleanCap-AG (Trilink BioTechnologies, N-7113) at a final concentration of 4 mM; and (ii) substitution of UTP with N1-Met-ψ-Uridine (Trilink BioTechnologies, N-1081) at a final concentration of 7.5 mM. mRNAs were then purified using the RNeasy Mini Kit (Qiagen, 74134). The quality and integrity of the mRNAs were assessed with a 4200 TapeStation System, and quantities were measured by a NanoDrop 8000. sgRNAs were synthetized by Axolab according to the a previously described nucleotide-modification scheme41. For in vitro or in vivo studies, the Cas9 mRNA was in vitro transcribed as described above or purchased from Trilink BioTechnologies (L-7606), respectively. Sequences of the ETRs and gRNAs used in this study are listed in Supplementary Table 6. The plasmids used in this study are available upon signing of a material transfer agreement.
Cell culture, treatment and engineering
Hepa 1-6 cells (CRL-1830, ATCC) were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Corning, 10-013-CV) supplemented with 10% fetal bovine serum (FBS; EuroClone), 1% l-glutamine (EuroClone, ECB3000D) and 1% penicillin–streptomycin (Euroclone, ECB3001D) at 37 °C in a 5% CO2 humidified incubator. The Hepa 1-6 Pcsk9tdTomato cell line was generated by nucleofecting 3 × 105 Hepa 1-6 cells with: (i) an HDR donor plasmid containing the 2A-tdTomato-polyA cassette within homology arms to exon 12 of Pcsk9; (ii) a Cas9-expression plasmid; and (iii) a plasmid expressing a gRNA targeting the last exon of Pcsk9 (ref. 42). tdTomato-positive cells were than sorted at single-cell level and amplified. The Hepa 1-6 Pcsk9tdTomato cell line is available upon signing of a material transfer agreement. Primary mouse hepatocytes from C57BL/6 male mice were purchased from Biopredic International as adherent monolayers on collagen-coated 96-well plates and maintained according to the manufacturer’s instructions. Supernatants of treated and control cells were collected at different time points and stored as one-time-use aliquots at −20 °C.
Gene-delivery procedures
For the in vitro experiments in the Hepa 1-6 Pcsk9tdTomato cells, 3 × 105 cells were transfected with either RNAs or plasmid DNAs using the 4D-Nucleofector X System (Lonza) in SF Cell Line solution (Lonza, V4XC-2032) and with the CM-137 pulse program. For the in vitro and in vivo experiments with LNPs A, B, C, D and E, these research-grade reagents were formulated by Precision NanoSystem (PNI) combining lipid mixes and RNA, the latter dissolved in a PNI proprietary formulation buffer. The lipid mixes are made of four different components dissolved in ethanol-based solution: an ionizable lipid, a helper lipid, cholesterol and 1,2-dimyristoyl-rac-glycero-3-methoxypolyethylene glycol (PEG-DMG). The chemical nature of the helper lipid differs among the formulations: (i) 1,2-di-(9Z-octadecenoyl)-sn-glycero-3-phosphoethanolamine (DOPE) for LNP A; (ii) 1,2-dioctadecanoyl-sn-glycero-3-phosphoethanolamine (DSPC) for LNP B; (iii) 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) for LNP C; and (iv) 1,2-di-(9Z-octadecenoyl)-sn-glycero-3-phosphocholin (DOPC) for LNPs D and E. The ionizable lipid as well as the molar ratio at which the four components were mixed are proprietary information of PNI. RNAs and lipids were mixed into the NanoAssemblr Ignite instrument (PNI) using microfluidic cartridges (Ignite NxGen Cartridge; PNI) with a total flow rate (TFR) of 12 ml min−1 and a flow rate ratio (FRR) of 3:1 (RNAs:lipids). Nitrogen-to-phosphate (NP) ratios of 6 and 9 were used in the initial LNP screening experiments, and this ratio was set to 6 for the remaining in vivo experiments. The lipid mix LNP D is available upon request from Precision NanoSystems (PNI) using the code iL00V77. LNP D formulated with the desired RNA can be directly purchased from PNI by signing an agreement. The estimated turnaround time is 1–2 months. Alternatively, the LNP D and the formulation device can be purchased from PNI. In this case, PNI will technically support the investigator in the setting of the formulation protocol. LNPs produced with the GenVoy-ILM reagent (NWW0042, 25 mM) were formulated following the manufacturer’s instructions with the NanoAssemblr Ignite instrument (PNI) and using microfluidic cartridges (Ignite NxGen Cartridge; PNI). Formulation parameters were set as follows: (i) TFR of 12 ml min−1; (ii) FRR of 3:1; and (iii) NP ratio of 4. All LNPs were concentrated using an Amicon Centrifugal Filter (MWCO 30 kDa), the ethanol was removed by a 3:1 dilution in 1× PBS (pH7-7.3, Mg2+/Ca2+– free) and LNPs were finally filtered manually through a 0.22-µm syringe. Particle sizes and their polydispersity index (PDI) were analysed using dynamic light scattering (DLS), and the RNA encapsulation efficiency and concentrations were determined using a RiboGreen plate-based assay. The results of the DLS analyses and RNA quantification of the LNPs used in this study are reported in Supplementary Table 7.
Flow cytometry and cell sorting
Flow cytometry was performed using CytoFLEX S (Beckman Coulter) and raw data were analysed using FCS Express v.7 (DeNovo Software) to extract the percentage of Pcsk9tdTomato–negative cells. When indicated, tdTomato-positive or -negative cells were sorted with a FACSAria Fusion Cell Sorter (BD Biosciences) as either bulk populations or at the single-cell level. The gating strategy for both the flow cytometry and the cell sorting procedures is reported in Supplementary Fig. 1.
RNA sequencing
Total RNA was extracted from 6 × 106 Pcsk9-silenced cells using the RNeasy Mini Kit (Qiagen, 74134) and quantified using the Qubit 2.0 Fluorimetric Assay (Thermo Fisher Scientific). Unstranded libraries were prepared with the NEBNext Ultra RNA Library Prep Kit for Illumina after rRNA depletion, and sequencing was performed using an Illumina NovaSeq 6000 platform (NovaSeq Control Software v.1.7) to obtain at least 30 million of 150bp-long paired-end reads per sample. Read quality was controlled with Fastqc v.0.11.9 (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/), and low-quality reads and the adapters were removed using Trim Galore v.0.6.6 (https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/) according to the following parameters: –quality 20, –length 25, –paired. High-quality remaining reads were aligned to the mouse reference genome GRCm38 using STAR v2.7.6a (ref. 43) with default parameters. Gene counts were quantified using the featureCounts function from the Subread package v.2.0.1 (ref. 44) and Gencode M25 as the gene model. Raw counts were corrected for biases due to different library preparations, if present, using the ComBat_seq function from the R Bioconductor package sva v.3.38.0 (ref. 45). Read distribution was estimated using the negative binomial generalized log-linear model implemented in the R Bioconductor package DESeq2 v.1.30.0 (ref. 46). Differential gene expression was tested using the nbinomWaldTest function and P values were corrected using the Benjamini–Hochberg approach.
WGMS
Genomic DNA was extracted from 6 × 106 Pcsk9tdTomato-silenced cells using the Maxwell RSC Cultured Cell extraction kit (AS1620) and quantified using a NanoDrop 8000. Libraries were prepared using an enzymatic approach for cytosine conversion with NEBNext DNA Ultra II Reagents and sequencing was performed using an Illumina HiSeq (HiSeq Control Software v.3.4) platform to produce at least 250 million 150-bp-long paired-end reads per sample. Read quality was controlled with Fastqc v.0.11.9, and low-quality reads and the adapters were removed using Trim Galore v0.6.6. with the following parameters: –quality 20, –length 25, –paired, –clip_R2 5. High-quality remaining reads were analysed using the Bismark read mapper methylation caller tool v.0.23.0. In detail, reads were aligned to both the converted and the unconverted genomes (GRCm38) using Bismark v.0.23.0 with default parameters. Duplicates were then removed using the deduplicate_bismark script and the methylation status was obtained using the bismark_methylation_extractor script. Then, the methylation call was loaded into the R environment and processed using the R Bioconductor package MethylKit v.1.16.1. Imported data were filtered using the filterByCoverage function (low count filter equal to 1 and high percentile equal to 99.9) and normalized using the normalizeCoverage function. Information from the different samples was merged using the unite function considering the positions covered in all replicates. The percentage of methylation was calculated with the percMethylation function and the correlation among the samples was determined applying the cor function (default Pearson method). Differential methylation analysis was performed using the R Bioconductor packages bsseq v.1.26.0 (ref. 47) and DSS v.2.44.0 (ref. 48). First, the object was created using the makeBSseqData function starting from the Bismark output. The DMLtest function was used for the normalization step and the differential analysis with the following parameters: smoothing = TRUE and smoothing.span = 500. Then, the callDML function was applied to determine the differential methylated loci (DML) setting as thresholds delta = 0.4 and p.threshold = 1 × 10−3. To exclude any confounding DMRs not associated with off-target methylation, the delta methylation threshold was set at 0.4; that is, the minimal value not calling any DMR in Cas9-treated cells, a negative control having no direct methylation activity. The DMRs were defined applying the callDRM function with the same thresholds. The DMRs identified were annotated using the annotatePeakInBatch function from the R Bioconductor package ChIPpeakAnno v.3.24.2 (ref. 49) using the Gencode M25 annotation and the following parameters: PeakLocForDistance = “middle”, FeatureLocForDistance = “TSS”, output = “both” and multiple = TRUE. DMRs of all treated samples were computed using as reference the same mock-treated controls.
Mouse handling and treatments
Eight-week-old C57BL/6N female mice were purchased from Charles River Laboratories. Procedures involving animal handling and care followed national and international law and policies and were approved by the Institutional Animal Care and Use Committee (authorization numbers 604/2020-PR and 233/2022-PR, provided by the Italian Ministry of Health). Housing temperature and relative humidity were 22 °C (±2 °C) and 55% (±5%), respectively. A 12-h light–12-h light cycle was used and all possible efforts were made to minimize the number of mice used and their suffering. For in vivo administration of either Cas9 or ETRs, mRNA-LNP solutions were diluted in PBS without calcium and magnesium (Corning, 21-031-CV). Subsequently, mice were randomly assigned to a treatment group and heated with an infrared lamp to obtain vasodilatation. Finally, 250 µl of LNP solution or PBS (herein defined as vehicle) were intravenously injected into the tail vein. For plasma analyses (see next section), around 200 µl of blood was collected from the retro-orbital plexus of each experimental mouse by using a non-heparinized micro-haematocrit capillary tube (Kimble Chase, CSX40A502), and then moved into an EDTA-sprayed blood collection tube (Sarstedt, 20.1288.100). Blood was then centrifuged for 10 min at 2,000g at room temperature. Purified plasma was finally collected from the supernatant and stored as one-time-use aliquots at −20 °C. For experiment termination and organ collection, mice were euthanized by CO2 inhalation and tissues (liver, spleen, lungs and kidney) were removed and snap-frozen for further molecular analyses. For partial hepatectomy, mice were anaesthetized by 2% isoflurane continuous inhalation. Before hepatectomy, mice were fasted for 4 h. Surgery was performed according to the Higgins protocol50. In brief, the abdominal skin was shaved, and a 2-cm upper midline incision was made beginning from the xyphoid. After opening the peritoneum, the liver was gently mobilized and exposed. The left lateral lobe was lifted, tied up and resected through 3.0 silk sutures (Ethicon, EH6823H) distal to the applied ligatures. Muscle and skin were closed in two layers with 4.0 Vicryl (Ethicon, V994H) and an autoclip wound-closing system, respectively. For postoperative analgesia, carprofene (5 mg per kg) was used by subcutaneous injection into the neck fat pad. Liver tissue and blood were collected during hepatectomy and at necropsy for molecular analysis and Pcsk9 plasma quantification. For isolation of hepatocytes, mice were first anaesthetized with isoflurane and the liver was exposed and perfused (32 ml per min) through the inferior vena cava with HBSS-HEPES 0.03% collagenase IV (Sigma). The digested mouse liver was collected, passed through a 100-μm cell strainer (BD Biosciences) and processed into a single-cell suspension. Cells were spun down and washed three times with successive centrifugations at different speeds (30g, 25g and 20g) for 3 min each at room temperature to obtain hepatocytes.
Plasma analysis
To quantify PCSK9, plasma from treated mice and supernatants from primary mouse hepatocytes were thawed and diluted 1:200 and 1:2, respectively. Dilutions were then loaded on a commercial pre-spotted ELISA kit according to the manufacturer’s instructions (R&D Systems, MPC900). Similarly, absorbance assays were used to quantify the levels of LDL-C (P/N 00018256040, Werfen), ALT (P/N 00018257440, Werfen), AST (P/N 00018257540, Werfen), LDH (P/N 00018258240, Werfen) and albumin (P/N 00 18250040), following the manufacturer’s instructions.
In vivo molecular analyses
Genomic DNA was extracted from snap-frozen tissues (around 30 mg) using the Maxwell 48 Promega RSC Tissue DNA Purification Kit (AS1610) according to the manufacturer’s instructions. Where indicated, editing and epi-editing efficiencies were quantified from purified hepatocytes (see ‘Mouse handling and treatments’). In these cases, genomic DNA was extracted using the Maxwell 48 RSC Tissue DNA Purification Kit (AS1610) from 1 × 106 cells according to the manufacturer’s instructions. Gene-editing efficiencies at the Pcsk9 locus were quantified using the T7 assay or targeted deep sequencing. For the T7 assay, a 765-bp genomic region encompassing the CRISPR–Cas9 binding site was PCR-amplified using the primers listed in Supplementary Table 8. PCRs were then processed using the Alt-R Genome Editing Detection Kit (IDT, 1075932), run on the Agilent ScreenTape System, and the percentage of editing was quantified according to the manufacturer’s instructions. For targeted deep sequencing, the promoter region or exon 1 of Pcsk9 were PCR-amplified using the primers listed in Supplementary Table 8. Libraries were then prepared using the NEBNext Ultra II DNA Library Prep Kit for Illumina for the Pcsk9 promoter or using the NEBNext Ultra DNA Library Prep Kit for Illumina for Pcsk9 exon 1, and sequenced using the Illumina MiSeq platform (MiSeq Control Software v.2.6). Sequencing data were analysed with CRISPResso2 v.2.8 (ref. 51) to detect nucleotide insertions and/or deletions. Reads were aligned to the boundary sequence around the putative cutting site (400 bp centred on the sgRNA complementary site for Cas9-treated samples or 300 bp centred on the ZFP-8 recognition site for EvoETR-8-treated samples) using bowtie2 v.2.2.5 (refs. 52,53) in paired mode and default parameters. After that, original fastq files were subset to retain only the reads mapping to the region of interest using the filterbyname module of the BBMap aligner v.39.01 contained in the BBTools suite (https://sourceforge.net/projects/bbmap/). The remaining reads were analysed with CRISPResso2 in paired-end mode setting the options for Trimmomatic software v.0.39 (ref. 54; http://www.usadellab.org/cms/?page=trimmomatic) to remove low-quality positions (score < 30) and Illumina adapters (--trim_sequences --trimmomatic_command trimmomatic --trimmomatic_options_string ‘ILLUMINACLIP:TruSeq3-PE-2.fa:2:30:10 MINLEN:100’). Then, each couple of paired-end reads was merged using FLASH v.1.2.11 (ref. 55) to produce a single contig, which was mapped to the input amplicon reference (promoter region or first exon, depending on the experiment). The sgRNA complementary site and the ZFP-8 target sequence were provided to focus the analysis on the target region, and the quantification window was set to 20 bp per side around the cut site (Cas9 samples) or the ZFP-8 middle point (EvoETR-8 samples). Identified alleles were quantified by measuring the number of reads and their relative abundance on the basis of total read counts considering only insertions and deletions. The percentage of CpG methylation at the Pcsk9 promoter or at the in-vitro-identified DMRs was quantified using targeted bisulfite deep sequencing (targeted BS-seq). Specifically, purified genomic DNA was converted with the EpiTect Fast Bisulfite Conversion Kit (Qiagen, 59104) according to the manufacturer’s instructions. Then, the promoter region of Pcsk9 and unintended DMRs were PCR-amplified using the primers listed in Supplementary Table 8. For the Pcsk9 promoter, libraries were prepared using the NEBNext Ultra DNA Library Prep Kit for Illumina. For the other DMRs, the NEBNext Ultra II DNA Library Prep Kit for Illumina was used. Libraries were sequenced using an Illumina MiSeq platform in paired-end mode (MiSeq Control Software v.2.6). Read quality was controlled with Fastqc v.0.11.9, and low-quality reads and adapters were removed using Trim_Galore v0.6.6. with the following parameters: –quality 20, –length 25, –paired, –rrbs. High-quality remaining reads were analysed using the Bismark read mapper methylation caller tool v.0.23.0 (ref. 56). In detail, reads were aligned to both unconverted and converted genomes (GRCm38) using Bismark with the –local parameter and the methylation status was obtained using bismark_methylation_extractor script. The methylation calls were loaded into the R environment and processed using the R Bioconductor package MethylKit v.1.16.1 (ref. 57). Imported data were filtered using the filterByCoverage function (low count filter equal to 10 and high percentile equal to 99.9) and normalized using the normalizeCoverage function. The data from the different samples were merged using the unite function considering the positions covered in at least one replicate per condition. The percentage of methylation at each CpG was calculated using the percMethylation function.
Statistics and reproducibility
Data were plotted and analysed using GraphPad Prism v.9 (GraphPad Software). When indicated in the figure legends, statistical significance was evaluated by using GraphPad Prism v.9 (GraphPad Software) and applying the described tests. All of the in vitro experiments were conducted with technical replicates (n ≥ 2) and the exact number of replicates is indicated in the respective legend. In vivo experiments were designed including multiple mice (n ≥ 3). The exact number of treated mice in any experimental group for any experiment is indicated in the figure legend.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.