Presynaptic dysfunction in CDKL5 deficiency disorder

Cyclin-dependent kinase-like 5 (CDKL5) deficiency disorder (CDD) is a monogenic developmental and epileptic encephalopathy with onset in early infancy that is caused by mutations in the CDKL5 gene. CDD patients often exhibit profound neurodevelopmental delay, visual and motor deficits, and autistic-like manifestations, whereas epileptic seizures typically appear as early as the third week after birth. CDKL5 is a neuron-specific serine/threonine kinase that has been implicated in different cellular processes including neurite outgrowth, microtubule remodelling, and synaptogenesis. Animal models of CDKL5 deficiency have revealed phenotypes associated with defective neurotransmission. However, the potential role of CDKL5 in presynaptic processes and synaptic vesicle (SV) membrane trafficking remains unknown. In this project, we used a novel CDKL5 KO rat model to detect potential phenotypes that are linked to loss of CDKL5 function. Using a genetically encoded fluorescent reporter, we revealed that absence of CDKL5 results in defective SV recycling in an activity-dependent manner in primary hippocampal neurons. Using a molecular replacement strategy, we showed that the kinase domain of CDKL5 was able to restore the speed of SV endocytosis indicating that the catalytic activity of CDKL5 is essential for its role in SV recycling. In agreement, we revealed that CDKL5 mutants either lacking the kinase domain or containing kinase-inactive mutations reported in CDD patients were unable to rescue this impairment suggesting that defective presynaptic processes may contribute to the CDD onset. Since the kinase activity is critical for CDKL5-mediated SV recycling, we also explored whether the phosphorylation levels of its in vitro presynaptic substrate, amphiphysin 1 (Amph1), were altered in CDKL5 KO neurons. We revealed that CDKL5 does not exert its presynaptic role by phosphorylating Amph1 at S293. At the same time, this work showed that Amph1-mediated complexes are important for SV endocytosis in presynaptic terminals. Furthermore, we mapped the Amph1 motif that interacts with a different endocytosis protein, endophilin A1, and we demonstrated that the Amph1-endophilin A1 complex is essential for SV regeneration. Finally, the phosphorylation dynamics at Amph1-S293 dictates both Amph1-mediated interactions with endophilin A1 and SV endocytosis, indicating that phosphorylation-dependent Amph1-endophilin A1 interaction is essential for optimal SV endocytosis. Overall, this study offers the first evidence of a presynaptic role of CDKL5 that is mediated through its kinase activity and creates the basis for future research on presynaptic CDKL5 that could lead to potential treatments for CDD patients.