sometimes the quantity of surfactant that you used is not sufficient to revest all the surface of your nanoparticles and this way it could aggregate; but it´s important not so high quantity because can form miscelles systems;
But I don´t know if you use this surfactant (Tween 80) to construct of your nanoparticle system?
I think it is not a good idea to leave the cells in ddH2O with no source of energy and improper isotonicitiy. Of course it depends on your ultimate aim. For example in my case I was using some peptides as a delivery agent but had the same problem of aggregation in DMEM. The precipitation is mainly due the change in pH and sometimes the serum protein used in DMEM might add to the problem. First try with OPTI-MEM with no FCS. If it does not work then use PBS with Glucose but never ddH2O. But in this case the final read out should be as early as possible and it might not be appropriate for many functional assay. In my case it was just uptake in a two hour window so had no issues.
I completely agree the last answer: cells in ddH2O won't survive for long...I'm not used to work with Zein nanoparticles, but I did with others... Do your particles have any coating? Do you use DMEM supplemented with FBS? On the other hand...I guess you did, but have you tried to reduce the nanoparticle concentration?
If I were you, I'd first try with PBS, then move to DMEM (without serums), and finally, with complete medium, Once you optimize the nanoparticle concentration, I'd go for the assay with cells.