AbstractAlthough delivery of a wild-type copy of the mutated gene to cells represents the most effective approach for a monogenic disease, proof-of-concept studies highlight significant efficacy caveats for treatment of brain disorders. Herein, we develop a cross-correction-based strategy to enhance the efficiency of a gene therapy for CDKL5 deficiency disorder, a severe neurodevelopmental disorder caused by CDKL5 gene mutations. We created a gene therapy vector that produces an Igk-TATk-CDKL5 fusion protein that can be secreted via constitutive secretory pathways and, due to the cell-penetration property of the TATk peptide, internalized by cells. We found that, although AAVPHP.B_Igk-TATk-CDKL5 and AAVPHP.B_CDKL5 vectors had similar brain infection efficiency, the AAVPHP.B_Igk-TATk-CDKL5 vector led to higher CDKL5 protein replacement due to secretion and penetration of the TATk-CDKL5 protein into the neighboring cells. Importantly, Cdkl5 KO mice treated with the AAVPHP.B_Igk-TATk-CDKL5 vector showed a behavioral and neuroanatomical improvement in comparison with vehicle or AAVPHP.B_CDKL5 vector-treated Cdkl5 KO mice. In conclusion, we provide the first evidence that a gene therapy based on a cross-correction approach is more effective at compensating Cdkl5-null brain defects than gene therapy based on the expression of the native CDKL5, opening avenues for the development of this innovative approach for other monogenic diseases.