Keep in mind that these overrepresentation analysis are primitive. All it says is that with a given probability more genes than expected by chance overlap with the annotated pathway. Also, many genes are redundant and unspecific, so despite that the term might contain "neuronal" the genes could just be bystanders and exist in many pathways.

That having said, HEK is obviously not a neuronal cell. What you maybe get are genes that are under the same molecular similar regulation and if you knockout a key regulator, then many change, and this is similar to what happens in neuronal cells. Still, don't overinterpret it. It's a cell line, in culture for decades, malignantly transformed, not neuronal, so a lot of confounders.