what can be the problem ?

The problem is most likely that your data isn't what you think it is.

Try the following script:

use strict;
use warnings;
binmode STDOUT, ':encoding(UTF-8)';
use Encode qw/decode_utf8/;
while (<>) {
    $_ = decode_utf8 $_;
    s/\W//g;
    print;
}
[download]

Some in- and output:

möp spaß     # input
möpspaß      # output
АБВГ-ДЕЖ/ЗDD # input
АБВГДЕЖЗDD   # output

So it preserves both German umlauts and Cyrillic characters.

Please actually read the article I gave you a link to, it contains advice on how to debug such problems.