Why not? Three words: global magma ocean.
The energy of the final accretion of the Big Whack impact fragments melted the entire "crust" of the Moon. Any water that outgassed as vapor from the interior at that time would have been so highly energized by the high temperatures of the magma ocean that it would likely have had enough energy to escape out of the Moon's gravity well. Heck, some of that energy, combined with the chemical composition of the outgassing lunar crust and mantle, might well have dissociated some of the water molecules into other chemical forms.
By the time the surface cooled to the point where liquid water could actually exist (assuming a temporary atmosphere that would provide enough pressure for liquid water), the vast majority of the water that was "releasable" would already have been released, lost from the lunar gravity field, or altered into other chemical forms. So, the largest remnant of water would have been retained in the mantle. (Remember, the Moon's crust and mantle must have remained molten for long enough for a fairly high degree of differentiation, in which the relatively iron-rich magmas sank into the mantle and the relatively light aluminous feldspathic magmas floated to the top. It's not like the crust cooled to hardness in a matter of days or weeks after the final phases of that initial accretion.)
As for snowflakes, that's a different matter. At the very least, a fair amount of the water vapor that remained unaltered as it escaped the Moon's gravity had to have coalesced into tiny icy particles, which might have amalgamated into structures reminiscent of snowflakes. Some of that ice may well have settled back onto the Moon, and considering how long such particles might have hung around the Earth-Moon system, there's a chance some of it might well have accumulated at the lunar poles, in permanently shadowed locations. However, the time it took for the Moon and Earth to sweep up ice particles was probably less than the time from the Moon's crustal cooling to the end of the Late Heavy Bombardment, which probably scattered most of the accumulated ice back into space, or at least redistributed it into non-shadowed locations.
My feeling about the amount of water postulated for the lunar mantle is that, even during the most extensive periods of mare volcanism, the total amount of water available for liberation would have been so small (and the surface conditions now so near to a perfect vacuum) that deposition of liquid water would have been impossible, and only a tiny fraction of the liberated water would have ended up deposited in polar shadows.
Hopefully, all that made some degree of sense...
-the other Doug