I have a similar setup at this end. I frontend an Exchange server with a postfix MTA sitting in a DMZ. All inbound e-mail is received by postfix, scanned by Spamassassin, then relayed to the Exchange server. Virus scanning is done on the Exchange server. Outbound e-mail is virus scanned by Exchange server, then relayed to the postfix server for final delivery. Note: Outbound e-mail is not scanned by Spamassassin.
With the above in mind, I can give you a couple of pointers to configuration issues that will need to be addressed.
1) To relay inbound e-mail to your Exhange server...
If your using sendmail, look at the "mailertable" feature.
If your using postfix, look at the "transport" feature.
2) Configure MTA in DMZ to verify mailbox/user exists on Exchange prior to relaying to Exchange server.
The way I implemented the above was to configure postfix to issue an LDAP query against the Exchange server to verify the user/mailbox is valid prior to relaying. Without doing so, bogus recipients for your domain would be relayed to Exchange server and then bounced back to what is usually a forged sender address. If interested, I can post examples of how I implemented LDAP using postfix, but I know sendmail can be configured to do the same thing.
3) E-mail filtering
E-mail filtering can be implemented in many ways with either MTA. Since my postfix MTA only calls Spamassassin, I chose to implement SA filtering using postfix's built-in hooks. But there are some good third party apps available for free that will handle both virus and spam filtering by your MTA. I can think of a couple that I have used in the past.
amavis-new and
mime-defang
As for changing the e-mail header or adding a disclaimer... can't help you there, but if I remember correctly, mime-defang may have the capability to remove headers.
Good Luck!