Could you post some sample data together with the code that you've tried? What you need to do should in principle also work with Unicode/UTF-8.

For example, you can use \p{Lu} to match a character with the unicode property "Letter, uppercase" (for a detailed list see perlunicode, in particular section "Effects of Character Semantics").  Uppercasing and lowercasing should work as well...

Update: to avoid unnecessary confusion, it's maybe worth mentioning that for a number of \p{...} expressions, there's the alternative [[:...:]] form. E.g. [[:upper:]] is the same as \p{IsUpper}. The \p{...} style is the more generic form, i.e. not all \p{...} expressions do have a [[:...:]] form. — BTW, the "Is"-prefix is optional, and you can use short or long forms. For example \p{IsLu} is equivalent to \p{Lu} or \p{UpercaseLetter} or \p{IsUpercaseLetter}.