It's not something you can prevent. Concerning automated systems, registration behind nearly any verification check is easy to bypass, unless it requires verifiable private information, implementing a limit of messages per day is a poor idea (and would not have much effect here, as they rarely seem to get out more than ten messages at a time), monitoring private messages is controversial at best, so you are basically left with some algorithm monitoring all public messages for repeating patterns, and that is far from perfect as well.
If you see someone suspicious, report them and be done with it. It's easy to spot them once you know the patterns.
