There's no need to pull the first file into memory. This is a fairly standard "load the filtering file into a hash and check the other file against it line-by-line" problem, with the one extra twist that, if the first field from the filtering file is a match, another field needs to be checked against the second field.

#!/usr/bin/env perl
use Modern::Perl;

my %k;
open my $fd2, '<', 'file2.txt' or die $!;
while(<$fd2>){
    chomp;
    if( /([a-z]),(\d+)/ ){  # one lowercase character, a comma, and di
+gits
        $k{$1} = $2;
    }
}
close $fd2;

open my $fd1, '<', 'file1.txt' or die $!;
while(<$fd1>){
    my @w = split /, /;
    if( $k{$w[3]} and $w[5] =~ /$k{$w[3]}/ ){
        print;
    }
}
close $fd1;
[download]