Since the first report of patients with ID harboring IQSEC2 pathogenic variants [6], approximately 120 causative variants in IQSEC2 have been reported thus far, in which over 90 variants are truncating [7, 8, 15, 16]. IQSEC2 protein contains 6 known functional domains, including (1) an N-terminal coiled-coil (CC) domain, (2) an IQ-like calmodulin-binding motif, (3) a Sec7 domain responsible for GDP–GTP exchange on Arf6, (4) a pleckstrin homology (PH) domain, (5) a proline-rich (PR) domain which is known to bind to the insulin receptor tyrosine kinase substrate of 53 kD (IRSp53), and (6) a PDZ domain which mediates binding to PSD-95 (Fig. 1, Supplementary Fig. 2). In this context, while pathogenic missense variants are clustered in the IQ, Sec7, and PH domains, truncating variants including nonsense, frameshift, and splice variants are distributed throughout the gene [6, 8, 17] (Supplementary Fig. 2). Generally, the truncating variants in the N-terminal side are predicted to result in a mutant transcript carrying a premature termination codon that is rapidly degraded by NMD, leading to severe phenotypes due to haploinsufficiency. In contrast, truncating variants affecting the C-terminal domain as well as missense variants in the functional domains may rather lead to a protein with residual function and are associated with milder phenotypes [10, 18, 19]. As for IQSEC2, however, truncating variants altering C-terminus that are presumed to escape NMD can also cause severe neurodevelopmental phenotypes [10], as observed in our patient carrying the nonsense variant p.Q1227* in the last exon of IQSEC2. We thus conducted functional assays to address this issue.

The mutant protein IQSEC2 Q1227* lacks the C-terminal region including PR and PDZ domains that bind to IRSp53 and PSD-95, respectively. Of particular interest, IQSEC2 has been shown to be localized at the PSD through the interaction with two major scaffolding proteins, IRSp53 and PSD-95, responsible for anchoring proteins to the cytoskeleton and mediating signal transduction following excitatory stimulation [1,2,3]. In this regard, we clearly demonstrated that IQSEC2 Q1227* lacked the specific localization ability to the PSD through the transfection assay of the mutant protein into cultured hippocampal neurons. This result was consistent with that of our previous report [3], and could account for the severe neurodevelopmental phenotype of our patient.

Also notably, Arf6 pull-down assay indicated reduced GEF activity of IQSEC2 Q1227*, although IQSEC2 Q1227* still preserves the Sec7 domain essentially involving GEF function for switching GDP and GTP on Arf6. In neurons, the interaction between IQSEC2 and IRSp53 is suggested to link the Arf6 pathway to the Rac1/Cdc42 pathway, which is implicated in spine morphogenesis [3, 20]. Indeed, Arf6 regulates dendrite and spine formation upstream of Rac1 [21, 22]. Hence, the interaction of IQSEC2 and IRSp53 may facilitate the formation and maintenance of dendrites and spines through activating Arf6 [3]. Consequently, the disrupted association between IQSEC2 Q1227* and IRSp53, due to the absence of the PR domain, can lead to diminished GEF function and impede the activation of Arf6. Thus, it is tempting to assume that compromised GEF activity of IQSEC2 at the PSD of excitatory synapses would also explain the observed neurological manifestations of our patient, together with the disrupted association between IQSEC2 Q1227* and PSD-95.

In light of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP) guidelines [23], it was unclear whether the c.3679C>T (p.Q1227*) variant in the IQSEC2 gene met “PVS1” (null variant in a gene where loss of function is a known mechanism of disease), because NMD was not expected for this variant. Nonetheless, our assays demonstrated that the variant satisfied the “PS3” (well-established in vivo and in vitro functional assays supportive of a damaging effect on the gene or gene product), in addition to “PS2” (de novo in a patient with the disease and no family history), “PM2” (absent or extremely low frequency in the general population databases), and “PP5” (reputable source recently reports variant as pathogenic, but the evidence is not available to the laboratory to perform an independent evaluation). Accordingly, this variant was classified as “pathogenic” according to the ACMG/AMP guidelines.

In conclusion, we identified a de novo nonsense variant in IQSEC2 leading to C-terminal truncation protein on a 5-year-old boy presenting with severe ID and epilepsy. We infer that the mislocalization of the C-terminus truncated IQSEC2 in neurons and impaired GEF activity could lead to severe neurological phenotypes of the patient, comparable with those caused by null variants in IQSEC2. This is the first study providing compelling evidence towards the understanding of how C-terminal truncations in IQSEC2 lead to severe neurological manifestations. Further studies and case series would enhance our comprehension of the molecular pathology underlying this neurodevelopmental disorder.