Genome Editing in Erythroid Progenitor Cells Mediated By Crispr/Cas9

Investigation of the genetic underpinnings of erythrocyte development holds great value not only in the development of potential therapeutics for hematologic disorders, but also for elucidating basic biological principles. A robust model system for studying erythropoiesis is the mouse fetal liver system, as murine fetal liver is predominantly composed of erythroid progenitor cells at 2 weeks gestation. Upon isolation, these cells can be cultured in the presence of erythropoietic cytokines and follow distinct phases of development, from immature erythroid progenitors to terminally differentiated erythrocytes with robust enucleation and hemoglobinization.

To date, loss of function genetic studies of erythropoiesis using the mouse fetal liver system have relied on mouse strains deficient in a gene of interest, or RNA interference inhibiting translation of a gene product of interest. Both strategies have limitations in terms of either time-intensive generation of genetically deficient mice, or inability of RNA interference to faithfully model homozygous deficiency, or haploinsufficiency. The development of genome editing technology based on a RNA-guided system for inducing targeted DNA double strand breaks (DSBs) raises the possibility of faithfully modeling homozygous deficiency or haploinsufficiency in a significantly higher throughput manner. This system consists of RNA-based Clustered Regularly Interspersed Short Palindromic Repeat (CRISPR) elements complexed with the Cas9 nuclease. Upon expression of both components in eukaryotic cells, a CRISPR single guide RNA (sgRNA), base pairs with a genomic target, guiding Cas9 to induce a DSB at that site. Genome editing then occurs at the site of the break if repair occurs via the non-homologous end joining DNA repair pathway, which produces mutational insertions, deletions, and substitutions during the process of DSB repair.

In this study, we aimed to develop a system for CRISPR/Cas9-mediated genome editing in murine fetal liver cells. We constructed a retroviral vector co-expressing the Cas9 nuclease and an sgRNA. We initially designed sgRNAs targeting 2 genes non-essential for erythroid development, Gata3, which encodes a transcription factor required for T-cell development, and Lcp2, which encodes an adapter protein required for signal transduction during T-cell activation. These genes were chosen in order to assay genome editing efficiency without the occurrence of negative selection against disruption of genes required for erythroid development. Transduction of fetal liver cells isolated on embryonic day 14.5 (E14.5) with a retroviral vector expressing Cas9 and an sgRNA targeting Gata3 resulted in editing of 38% of Gata3 alleles. Transduction of E14.5 fetal liver cells with vector targeting Lcp2 resulted in editing of 15% of Lcp2 alleles. No editing was detected in control cells transduced with a retroviral vector expressing Cas9 and a scrambled sgRNA. Genome editing was detected using the Surveyor nuclease assay, which quantifies allelic frequency of gene mutations resulting from DSB repair by non-homologous end joining. We next designed an sgRNA targeting the Bcl11a gene, which encodes a protein shown to be instrumental in the embryonic to adult globin switch in mice. Transduction of E14.5 fetal liver cells with vector targeting Bcl11a resulted in editing of 49% of Bcl11a alleles. We then assessed if constitutive expression of Cas9 and an sgRNA affects the ability of fetal liver cells to undergo terminal erythroid differentiation. Compared to cells transduced with vector expressing only GFP, fetal liver cells transduced with retroviral vectors expressing Cas9 and scrambled sgRNAs had no significant difference in enucleation rate, a marker of terminal erythroid differentiation.

In this study we demonstrate the ability to induce robust levels of genome editing at various genomic sites in mouse fetal liver cells using CRISPR/Cas9. We also demonstrate constitutive expression of Cas9 does not have any detrimental effect on enucleation. These results open the possibility of high-throughput modeling of homozygous genetic deficiency and genetic haploinsufficiency in studies of erythropoiesis.