It depends how your causality is specified. It could be an interactive term added, or using mediation analysis assuming A affects B, which further affect C.
The question is technique without theoretical specifications. Once you figure out how theoretically they should relate to each other, you would find a better specification