I doubt the "limitation" is so much in the technology as in the COST to produce lenses with higher resolving power.
By way of comparison, take a look at the DSLR world: Canon, for example, makes DSLRs up to 22.3MP in a full-frame 35mm sensor (EOS 5D mkIII). Naturally one wouldn't expect your average everyday lens, something that will look fine on a 12MP consumer DSLR, to take full advantage of that. So Canon produces their ultra-high-quality, professional-grade L-series lenses.
So you can get the rather basic EF 70-300mm f/4-5.6 IS USM lens for a comfortable retail price of about $500 (IS=Image Stabilizer, USM=Ultrasonic Motor autofocus drive).
Or there's the EF 70-300mm f/4-5.6L IS USM, clocking in at about $1500. Pretty much identical BASE specs, but for three times the price, you get better build quality, smoother operation, greater resale value... but mainly, much higher quality optics to take full advantage of those 20MP-and-up pro-grade cameras. (Most L lenses are also white, vs. the standard black finish, so other photographers will know from a distance that yer ballin' outta control.)
Translate that back to CCTV... and yes, it likely is POSSIBLE to build lenses that are *fully* suitable for higher MP counts... but at what cost? If a "3MP" lens was $100 and a 5MP lens was $300+, would you really want to spend the extra to eke out that little bit of extra clarity? In a few cases, maybe... in most, I'd expect not.
Keep in mind that a lower-res lens won't be inherently "worse" on a higher-MP camera... the higher pixel count will just show off its flaws and limitations more efficiently :)