I've experimented with heat forming uhmw around a threaded shaft. I made a two sided nut by simply clamping two softened pieces of uhmw around a threaded shaft. I used some wood to make 'clamping plates' that would allow the soft uhmw to press into the threads on the shaft. The wood pieces have a shallow groove cut into each one where the shaft will lie when it's pressed in a vice or a press. The end result is a two part nut that has 'wings' which can be used for mounting the nut, and also allow you to adjust the play by skimming material off the inside surfaces of these wings. In use, one side of this two part nut can be bolted solid to whatever it needs to propel, and the other wing can either be shimmed apart some to create some play, or clamped tight to eliminate play, or have some material skimmed off then bolted together to create a tight fitting nut. Or the second wing can also be used to hold the nut to it's home in the mechanism, and both wings can be shimmed or skimmed as needed to adjust the play. Whichever would work best in the room available for the nut.
I found the initial fit to be tight without any shimming or skimming, and if anything, there needed to be a thin shim between wings to create an easier turning shaft.
It's my feeling that a nut formed this way from uhmw is superior to one turned in the lathe, in the same way that a rolled thread in steel is superior to a turned thread. Smoother threads and more accurate.
During my experiment, I determined that an accurate two part press mold would be needed, as the uhmw faithfully takes on every little offset and wrinkle in the mold. You would be looking for accuracy without a next step machining operation, as this plastic is 'springy', and will take some skill to machine to an exact size. Best to mold it to finished size right away. Having said that, it wouldn't be hard to include markers in the mold to define where to drill mounting holes in the finished nut halves.
One thing about pressing the softened plastic around a threaded shaft this way is that the shaft tends to self-center. Of course your mold can always have alignment grooves to center the shaft automatically anyway. The two sides of the mold should have some means of aligning to each other as they are pressed around the shaft and softened uhmw pieces.
One drawback to a nut made this way is that the resulting threads in each nut half don't go full depth for a full half circle, but this could be offset by making the length of the nut longer.
Something which I thought of but didn't try is to coat the threaded shaft with a graphite lube or something, prior to the clamping. The idea was that some of that lube might end up locked into the plastic, giving a self-lubricating action.
Something else that helped propel me towards this experiment was the fact that the resulting nut made this way could be as long as you wanted it to be. The longer, the less stress per thread area, and the greater the accuracy in the leadscrew since more of it's threads are engaged in the nut at one time.