Well, this is a big can of worms. :)
Running remote code is going to present a crap storm of things to think about. You will most likely want to run it in a restricted and isolated app domain. Now lets assume though for simplicity that we are inside the "circle of trust" - you would need to construct a method to 'serialize' .Net code on the client and send it over the wire. It would then have to be compiled into a dll to run as native LINQ calls. This functionality would not be a part of MT itself, but if you found the right serialization fwk to support it, you should be able to plug it into the MT infrastructure for serialization.
Another "clunkier" option would be to build up your own representation of a Query much like the NHibernate Criteria that can be serialized.
Also note that you will be binding yourself to your types across app domains. So if you upgrade one service you could break the other.
As you can probably tell, this is an area that is frought with dragons. So take care. :)
-d