generalized solution to parse level headers:

$slurp.=$_ while <DATA>;
 
@data=split/^(=+)([\w\s]+)\1$/m, $slurp;
 
unshift @data,'','<Filename>';
 
while ( ($level,$header,$text,@data) = @data ) {
  print "  " x length($level),$header,"\n";
  #  print $text,"\n\n";        
}
 
__DATA__
[download]

OUTPUT

<Filename>
  Title of Page
  Literature
      Comments
  Microarray Data
      Comments
  Pathway Information
      Comments
  Aditional Info
[download]