The speakers are a home device, so designed to be on the Same IP Segment. I belive it uses something similar to Bonjour to detect speakers on the network using broadcast. If you can see the speakers between the two IP segments using broadcast then you should be fine.
I cant think of a way of making it work otherwise, as these were designed to be on the same network. Any traffic or packet shaping could cause audio delay between the speakers over the network and you might get speakers out of sync.