I have two stable cell lines (in mouse L cells), each of which overexpresses one of two isoforms of WNT5A. From these cell lines, I've generated conditioned medium. Before treating cells with these media, I need to demonstrate that those proteins are present and that they have equal activity. 

I have western blots that show that the cells are producing the protein. It has been demonstrated by another lab that these isoforms equally inhibit canonical wnt signaling. 

I am using a TOPFlash reporter system in 3T3 cells to assess canonical activity. When the canonical wnt pathways are stimulated, the TOPFlash reporter generates luciferase. A luciferase assay then gives the amount of canonical activity in RLU. 

To assay the activity of the WNT5A media, I've been treating the TOPFlash cells with WNT3A conditioned medium to activate the canonical pathway, and with the medium that I'm trying to characterize to inhibit the canonical pathway. Equal inhibition of canonical activity should mean equal amounts of active protein. 

Here's the problem: rather than inhibit canonical signaling, the WNT5A isoform A conditioned medium has stimulated canonical signaling when used in combination with WNT3A conditioned medium. The WNT5A isoform A conditioned medium does not stimulate canonical signaling when used on its own. The WNT5A isoform B conditioned medium shows the expected pattern of activity -  that is, it does not stimulate canonical signaling on its own, and inhibits it in a dose dependent fashion when used in conjunction with WNT3A medium. This conundrum has thus far prevented me from being able to proceed with my experiments that depend on this medium being characterized. 

I am looking for a way to characterize the WNT5A isoform A and WNT5A isoform B conditioned media. This could be a suggestion that mitigates the issues in using the TOPFlash system, or a new method for characterizing the conditioned medium. I've attached a graph of my last attempt at characterizing these media.