RT-QPCR of cells treated with 5-aza - which housekeeping gene to use, different cell lines (Jul/11/2007 )

A major problem when analysing gene expression data is the normalization using a suitable housekeeping gene as a reference. I have tried several housekeeping genes in different cell lines but I can't seem to find a stable one for several cell lines.

Some examples when using 5-aza-2-deoxy vs control:

18S and B-actin are good housekeeping genes for RWPE-1 normalised prostate cells, PC3 and DU145 prostate cancer cells while they are not for LnCAP prostate cancer cells, neither is MLN51 (my personal observations). These observations are based on data from a lightcycler, not bands in a gel.

I have searched the literature and other researchers have used the same genes for the same cell lines as mentioned above, which concerns me.

What do people think of this? Have people used these cell lines in combination with 5-aza?
Bram

dear Bram,
i work directly with all the above mentioned cell lines (exception made for MLN51) and i do a lot of qrt-pcr taqman as well...I've always normalized my data to 18s, and with or without treatment (from VitD3, NR ligands, HDACs inhibitor and so on) I've always got my 18s' within less than half cycle or maximum one cycle. all the times i didn't get it was because of my RNA or cDNA. so, i think u could use 18s with 5-aza with no problems!
hope it helps
seb