Check out this paper and the GEOdataset that comes with it (Series GSE56782; http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE56782) Using GEO2R you may find a set of microRNAs that is stable in this cell line. There are more such datasets on HUVECs that should allow you to find candidate normalizers (I would not call them controls). These candidates then can be tested in your context and hopefully are hardly affected by cytokines.
We use miR5701 and miR- -forgot the number, 400 something. Both were stable in a variety of very different types of skin cancers and normal tissues. These may work in HUVEcs as well, may be not. But screening exiting datasets should give you a very good idea what microRNAs might be good normalizers in your cell type or tissue of interest. Don't expect them to be perfect, but go with the ones that are the least interesting ones, the ones that never really change...
What internal control did you use, rather than RNU6_2. I think I am facing the same problem when my endothelial cells (HUVEC and HCAEC) are treated with inflammation cytokines. RNU6_B (from Qiagen) varies up to 2 cycles between groups even though I put the same amount of RNA in each reaction. Can you give me some suggestion?
we tried SNORD95 and SNORD44 that time with more or less success. You may have recognized that I asked this question somne time ago .... I think it is at least imortant to have more than one non-funtional miRNA as control. Good luck!