The substitution below works for the sample provided, however this is the wrong way to do it, I've assumed that this is very simple HTML with nothing that would break a very simple minded substitution. (eg, what would happen when a button with the alternative text "Next >" ) There is a famous response to this on another site, but the Perl specific response is to use a HTML parsing module eg HTML::TokeParser::Simple which helpfully has extracting the content from a html file as the first example.

s/<[^>]+>//g;
[download]