I don't remember the technical details, but I've read that some (many/most/all?) email programs look at the first few bytes of a file to determine what KIND of file it is, then do the attaching and sending process based on that assumption

As I (not much) remember, the problem is that the email program "sees" a PDF as one kind of file, and therefore uses a coding type for that type of file... when a PDF is really a different kind of file... and the chosen coding corrupts the PDF in transmission

Solution I've seen is to ZIP the PDF and send that way