The CRISPR nuclease systems greatly facilitate targeted genome modifications in mammalian cells. The outcome of genome editing depends on the involved DNA double strand break (DSB) repair pathways. While the classical non-homologous end-joining and the poorly defined alternative end-joining (alt-EJ) DSB repair pathways can cause imprecise repair and thus gene inactivations, the homologous recombination (HR) pathway often introduces precise modifications. Although CRISPR is highly efficient at inactivating single genes, it is inefficient at introducing precise genome modifications. Moreover, its efficiency at inactivating multi-locus DNA sequences such as highly repetitive endogenous viral elements also remains limited.
This thesis addressed these limitations by better characterizing DSB repair pathways in Chinese hamster ovary (CHO) cells - the most widely used production cell host for therapeutic proteins. In this thesis, I first aimed at identifying rate-limiting factors to improve HR-mediated genome editing. Second, I strove for studying approaches to inactivate repetitive endogenous retroviruses (ERV) presumably releasing viral particles into the CHO supernatant.
To identify factors limiting HR, we established two chromosomal CHO assays that measure HR activity based on the correction of a GFP loss-of-function mutation. By using knockdown and overexpression studies, we found that efficient HR-mediated genome editing depended on certain alt-EJ activities. Furthermore, we observed that alt-EJ contribution to HR correlates with the nuclease type and the location of the DSB site relative to the GFP mutation. These observations suggest that alt-EJ and HR repair pathways tightly interact and challenges the common perception of alt-EJ opposing HR. Finally, among the tested repair factors, high Mre11 nuclease and Pari anti-recombinase as well as low Rad51 recombinase levels were the most rate-limiting factors for HR in CHO cells. Counteracting these bottlenecks improved HR efficiency by 75%.
To inactivate repetitive ERVs, we transiently expressed a CRISPR-Cas9 nuclease that targets the gag gene of a specific transcriptionally active ERV group. Clones bearing a loss-of-function mutation in one particular ERV locus and corresponding mRNA produced considerably fewer particles loaded with viral RNA genomes. These findings indicated that a single ERV locus is responsible for the release of most, if not all, viral particles from CHO cells. Notably, ERV mutagenesis did not compromise cell growth, cell size or therapeutic protein production. In sum, this work provided novel strategies to improve HR-mediated genome editing and to inhibit viral particle release from CHO cells.