This is a shot in the dark -- I'm ignorant about PDF -- but I noticed that one of the character encodings supported by the Perl 5.8 Encode module is called "AdobeStandardEncoding". You might try converting the HTML/Unicode data to this encoding and see what happens... (The man page for the Encode module will explain ways to do the conversion.)

(I'll be glad to update this as soon as someone more knowledgeable weighs in -- there's a good chance I'm way off.) Also, maybe you can figure out which unicode characters are showing up, and create acceptable substitutes for these.