TreeCache can be used as a distributed cache. You can use it for HA deployments, including replication of session data. It is not directly related with redirecting http requests. For load balancing, you need to configure a load balancer software (e.g. mod_jk) just like you are currently doing.
If you are interested, load balancing with sticky sessions can be configured along with Buddy replication in TreeCache. You can replicate some data to all your computers, and other data not. If you are using sticky sessions, session data may be not replicated because all the web requests from the same web browser are directed to the same JBoss server.
you can review Buddy Replication configuration for TreeCache