I've always been against doing this sort of thing because I've seen too many LF drivers blown in tiny reflex cabs because too much was demanded of them -- i.e. EQ'ing the crap out of the low end to compensate for poor cab response down there. However, two points have already been mentioned and I think they're good ones. First, high-passing the signal chain using a steep filter is a great safety measure. Second,
any cab with too high a cutoff (or driver-unload) frequency will have the same problem of potential LF driver damage. There are lots of cabs like that out there masquerading as bass cabs. So I think it comes down to this:
best practices. Understand what your instrument, tone, playing style, music style, and audience size will demand of the cab at the bottom end, make plans for all that, and execute them. Especially crucial with drop tuning, five-strings, or a bass-heavy tone. And above all, pay attention to what the LF drivers are doing -- visually as well as audibly. Oftentimes you can spot distress visually long before it shows up audibly, when it's often too late.
Finally, I think if you're going to use PA kit for BG, you should consider your midrange requirements as well as your low-end requirements. Fifteens and horns -- even twelves and horns -- have always made me cringe. In many cases I think bass players are better served by cabs that cross over low to large-ish cone midranges than by cabs that cross over high to compression drivers+horns. Most players I've seen using LF+horn cabs have an epiphany the first time they play through a cab with LF+midrange cone(s). Things suddenly open way up, and clarity and definition replace grate and sizzle. Problem is, you don't find many monitors built this way, for various reasons.
I'm still struggling mightily with the idea of $700 for a tiny cab with a twelve and a horn. But then I'm also struggling with the idea of $5,000 for a bass head+18" FH. So... eh.
