One naive approach would be to look at what IP address(es) they're using when proxying and blocking them. (Or detecting them in code and redirecting them to your domain, perhaps.)