Background The CRISPR/Cas9 system has opened new perspectives to study the molecular basis of cerebral cavernous malformations (CCMs) in personalized disease models. However, precise genome editing in endothelial and other hard‐to‐transfect cells remains challenging. Methods In a proof‐of‐principle study, we first isolated blood outgrowth endothelial cells (BOECs) from a CCM1 mutation carrier with multiple CCMs. In a CRISPR/Cas9 gene correction approach, a high‐fidelity Cas9 variant was then transfected into patient‐derived BOECs using a ribonucleoprotein complex and a single‐strand DNA oligonucleotide. In addition, patient‐specific CCM1 knockout clones were expanded after CRISPR/Cas9 gene inactivation. Results Deep sequencing demonstrated correction of the mutant allele in nearly 33% of all cells whereas no CRISPR/Cas9‐induced mutations in predicted off‐target loci were identified. Corrected BOECs could be cultured in cell mixtures but demonstrated impaired clonal survival. In contrast, CCM1 ‐deficient BOECs displayed increased resistance to stress‐induced apoptotic cell death and could be clonally expanded to high passages. When cultured together, CCM1 ‐deficient BOECs largely replaced corrected as well as heterozygous BOECs. Conclusion We here demonstrate that a non‐viral CRISPR/Cas9 approach can not only be used for gene knockout but also for precise gene correction in hard‐to‐transfect endothelial cells (ECs). Comparing patient‐derived isogenic CCM1 +/+ , CCM1 +/− , and CCM1 −/− ECs, we show that the inactivation of the second allele results in clonal evolution of ECs lacking CCM1 which likely reflects the initiation phase of CCM genesis.

Cerebral cavernous malformations are slow-flow thrombi-containing vessels induced by two-step inactivation of the CCM1, CCM2 or CCM3 gene within endothelial cells. They predispose to intracerebral bleedings and focal neurological deficits. Our understanding of the cellular and molecular mechanisms that trigger endothelial dysfunction in cavernous malformations is still incomplete. To model both, hereditary and sporadic CCM disease, blood outgrowth endothelial cells (BOECs) with a heterozygous CCM1 germline mutation and immortalized wild-type human umbilical vein endothelial cells were subjected to CRISPR/Cas9-mediated CCM1 gene disruption. CCM1−/− BOECs demonstrated alterations in cell morphology, actin cytoskeleton dynamics, tube formation, and expression of the transcription factors KLF2 and KLF4. Furthermore, high VWF immunoreactivity was observed in CCM1−/− BOECs, in immortalized umbilical vein endothelial cells upon CRISPR/Cas9-induced inactivation of either CCM1, CCM2 or CCM3 as well as in CCM tissue samples of familial cases. Observer-independent high-content imaging revealed a striking reduction of perinuclear Weibel-Palade bodies in unstimulated CCM1−/− BOECs which was observed in CCM1+/− BOECs only after stimulation with PMA or histamine. Our results demonstrate that CRISPR/Cas9 genome editing is a powerful tool to model different aspects of CCM disease in vitro and that CCM1 inactivation induces high-level expression of VWF and redistribution of Weibel-Palade bodies within endothelial cells.