CRISPR/Cas9-mediated genome editing: evaluation of different experimental approaches to edit cell lines

CRISPR/Cas9 is one of the most successful technologies used for genome editing in eukaryotic models. Its versatile mechanism is based on the use of guide RNA (gRNA), designed to recognise a target sequence in the genome and guide the Cas9 enzyme to cut the targeted site. The induced double-strand break (DSB) can be repaired by cellular mechanisms such as non-homologous end joining (NHEJ), which can cause the insertion or deletion of nucleotides. This can induce frameshifts, resulting in the introduction of a premature stop codon and allowing gene knockout. Depending on the target gene, cell line, and biological purpose, the CRISPR/Cas9 experiment may require careful optimisation. In this study, we initially used CRISPR/Cas9 technology to induce the knockout of several genes in different cell lines, adopting different knockout approaches depending on the transfection efficiency of the target cell line or the final biological aim. The target genes were selected because they are involved in tumour pathways or play a key role in viral infections. We knocked out the ACOT8 gene in the HEK293T cell line by targeting exon 1 with a single gRNA to generate a model for studying HIV-1 infectivity in the absence of ACOT8, a thioesterase interacting with the viral HIV-1 Nef protein. Pseudoviruses produced in ACOT8-KO cells showed significantly reduced infectivity on target cells, an effect that was only observed for viruses with HIV envelopes. The production of HIV-1 QHO-pseudotyped viruses in ACOT8 KO cells re-transfected with an ACOT8 plasmid increased infectivity in target cells to levels comparable to those of the same pseudovirus produced in WT cells. This confirms the specific role of ACOT8 in mediating the production of more or less infectious viral particles. Secondly, the interaction between ACOT8 and NEF, and the resulting infectivity of the viral particles produced, were evaluated with the addition of NEF plasmids. This demonstrated that the ACOT8/Nef interaction in virus-producing cells leads to the production of more infectious viral particles, supporting a key role for ACOT8 and ACOT8/Nef in HIV-1 infectivity. We then used CRISPR/Cas9 technology to identify the role of certain genes in tumour pathways, starting with HADHA in pancreatic cancer cells (PANC-1). This gene encodes for a mitochondrial enzyme involved in lipid metabolism, which was found to be overexpressed in pancreatic cancer stem cells. We then used a two-guide approach to target exon 1 of HADHA, isolating a single KO clone for a line that proved difficult to transfect. Using the HADHA-KO PANC-1 model, we observed that loss of HADHA does not affect pancreatic cancer stemness, with the HADHA KO- pancreatic cancer stem cells (PCSCs) forming more compact and denser tumourspheres compared to wild-type pancreatic cancer stem cells. CRISPR/Cas9-mediated gene knockout was then applied to generate knockout cell models in order to evaluate the biological role of specific genes in two other cancer types, such as melanoma and osteosarcoma. In the first case, we targeted the RUNX2 gene, a transcription factor overexpressed in melanoma cells, in mouse melanoma cells (B16) with a single and dual gRNA approach. The two different experimental setups allowed us to obtain both RUNX2 KO cells and cells deleted for the Runt domain, essential for DNA binding. Both models will be used to better depict the RUNX2 role in melanoma progression. In the second case, we used CRISPR/Cas9 to target FBXW11 in osteosarcoma MG63 cells. The targeted gene encodes a ubiquitin ligase involved in the degradation of cellular products across different pathways. MG63 cells required a careful optimisation of the CRISPR/Cas9 protocol, since transfection of Cas9/gRNA plasmid did not work on these cells. Therefore, we used a smaller CRISPR plasmid to transfect MG63 cells previously transduced to stably express Cas9 protein, thus gaining a moderate increase in editing efficiency. Finally, we adopted transduction with virus-like particles (VLPs), a non-integrating and efficient system for delivering Cas9/gRNA ribonucleoprotein complex, thus allowing us to considerably increase editing efficiency in MG63 cells. After VLP transduction and confirmation of the proper editing efficiency on FBXW11, we isolated an FBXW11 KO MG63 clone that will be used as a model to evaluate changes in molecular pathways and tumour growth due to its absence. We further worked to optimise other applications of the CRISPR technology. In particular, we worked with the prime editing technology, an innovative CRISPR-based approach first developed in 2019. The technique allows to perform single nucleotide modifications by using a prime editor enzyme, composed of a Cas9 nickase and an MMLV reverse transcriptase, which copies the information carried by a specialised guide, called pegRNA, into the genome. We applied this technology in the context of cleidocranial dysplasia (CCD), a skeletal disease caused by mutations in RUNX2 gene. In our approach, we demonstrated its application in HEK293T by introducing the desired mutation (RUNX2, c.505 C>T) in these cell lines, which have been subsequently used as a model for prime editing experiments aimed at correcting the mutation. The procedure required the optimisation of different editing conditions to finally identify the best system for introducing or correcting this pathogenic mutation. We finally used CRISPR/Cas9 to knockout HLA-C gene in HEK293T cells. The generated cell lines would have been re-transfected with different HLA-C-expressing plasmids to create models for investigating the stability of the HLA-C/beta-2-microglobulin/peptide complex for various HLA-C allotypes. Our aim was to provide a more accurate classification of peptide binding stability of HLA-C allotypes and demonstrate its association with HIV-1 infection. Despite these initial experimental setups, another study conducted by Sarkizova et al. had already re-transfected HLA-C allotypes into HLA-deficient cells and experimentally determined the peptides binding to the different allotypes considered. We therefore decided to use these peptides for bioinformatic analyses on the binding association with HLA-C allotypes to further study the stability of the HLA-C complex. To achieve this, we used a tool (NetMHCpan4.2) to determine stability scores for the 21 most frequent human HLA-C alleles. This software allowed us to quantitatively determine the binding association between the considered HLA-C allotypes and specific pools of binding peptides. We then genotyped HIV-1 patients for HLA-C and assigned each patient an HLA-C stability score. Our results demonstrated that patients with rapid progression to AIDS and patients with HIV-1-associated neurological impairment had lower HLA-C stability values than patients with slower AIDS progression to the disease and patients without neurological deficits, respectively. Our study confirmed the role of HLA-C as a genetic factor that can determine different outcomes of HIV-1 infection.