Yes, it can. Treat the remote server like you would any other server sitting behind the CSS when you create the service. However, you will need to create a source group rule, and make use of the 'add destination service' statement to source traffic from the CSS -> server. Whatever IP address is specified within the source group will be the one the CSS uses when source NAT'ing the load balanced traffic on its way to the backend server. This ensures that the remote1 server will respond back to the CSS, which sends the response to the client (after NAT'ing it again) instead of the remote1 server responding directly to the client, which would likely reject the traffic.
Example:
service local1
ip address 192.168.100.1
active
service local2
ip address 192.168.100.2
active
service remote1
ip address 67.192.20.20
active
------
owner vip
content 192.168.100.200-80
ip address 192.168.100.200
port 80
add service local1
add service local2
add service remote1
active
------
group 192.168.100.200-80
ip address 192.168.100.200
add destination service local1
add destination service local2
add destination service remote1
active
