One of the advantages you already stated that depending on protein of interest if it's already expressed in CHO cell line, researchers use that.
Apart from that, this question is more of a general type. I will try my best to logically put down the things.
You are right, it's more appropriate to study proteins using cultured neurons. But, there are many practical factors that may compel us to avoid cultured neurons.
Ease of Culturing cells
Its way easier to culture cell lines such as CHO/HEK293/HeLa compared to neurons. Neurons are very sensitive and culturing them needs serious expertise and still, things can get wrong. There are many labs across the world that have dedicated technicians to avoid individual differences between cultures which can topple up experimental results. In contrast, culturing common cell lines doesn't need that level of expertise (relative comparison). Furthermore, for neuronal culture, several serious ethical guidelines need to be followed and require a good amount of funding to set up the facility. In addition, to have neurons develop processes and ready for study generally 10-12 days are required which prolongs the length of study. With common cell lines, you can perform an unlimited number of experiments and get results in a relatively short amount of time.
Ease of expression
Sometimes, the endogenous level of expression is not sufficient enough to visualize the interactions. More importantly, you do not have a tool to visualize these endogenously expressed proteins in cultured neurons. At that point Transfection and overexpression comes in handy. Transfection and overexpression in cultured neurons are way more difficult compared to common cell lines such as CHO, HEK, HeLa, etc. The efficiency is very low. It needs serious optimization and still, everything may go in vain. If endogenous expression levels are sufficient enough, knock-in expression is an option. Still, it requires several months to have such knocked-in mice and the success of knock-in is very low with larger proteins such as fluorescent proteins. Indeed new techniques are getting developed such as viral methods etc. Still, they are not well established across the globe.
In the case of common cell lines, it's relatively easy to transfect/optimize the conditions to obtain abundant protein of interest for the study.
Thus weighing time, money, expertise, and whether or not cultured neurons absolutely necessary to draw meaningful conclusions generally determine which cells to choose.