When it comes to WiNG Express this is designed for simple(r) deployments, hence missing some advanced features like manual cluster setup. Anyway, WiNG controllers are not recommended to be used in HA (active / active) in order to extend adoption capabilities. When one goes down, you'd exceed maximum number of APs adopted by single controller resulting in adoption loss.
Also, there is an article about cluster members on GTAC knowledge base Timo is right - limitation is 2 members for a cluster.
As described above, when you max out both controllers, on a fail over you will lose adoption on half of the APs. What can possibly play in your favor would be if you are running the WLANs in bridge mode local and not tunneling traffic back through the controller. If a controller goes down, the AP will still function handling all traffic as normal. Users will not know. In bridge mode tunnel, if a controller goes down, then half of the users will be ringing your phone off the hook complaining they lost network connectivity.
Either way, local or bridging, half of the APs will not adopt to the running controller. Always use the N+1 number when doing HA configurations.