Perl processes the text, but saves the non-breaking space character as "\_"

Sorry, but I think this is likely not the explanation. Instead:

I'm simply using the regular expression to search for particular sequence of letter and replacing them to something else.

Probably this regular expression is making this change, or it could maybe be an encoding issue, but you haven't shown us any code at all, so we currently can't help you.

However, if you can post a Short, Self-Contained, Correct Example that reproduces the issue then I'm sure we can help. In this node I describe what information you should include in your post when debugging unicode issues.