Methylation studies using arrays - (Jan/08/2007 )

Hi,

I'm trying to find tumor supressor genes and have done some bisulfite experiments ruling out my candidate genes. Since none of them were methylated I'd like to look more globaly. The question is how?
*Tiling arrays with antibodies against metylated cytosines...
*Expression arrays after treatment with 5-AZA or perhaps zebularine...
or something else???

Anyone having some advises? I've been reading about the different ways and searched pubmed and now I'm really confused...

Thanks in advance!

molbio,

can we all have a little bit more information?

are you looking at a cancer cell line? Can it be oncogenes that are triggering your cancer and not the silencing of tumor suppressor genes as the assays you state would not be able to determine this. Have you looked at the expression levels of your tumour suppressor genes as they could be affected by histone modifications and not DNA methylation alone.

Good luck.

Nick

I have both cell lines and solid tumors available. The expression levels of the candidate genes are indeed down-regulated which is partly why they are our candidate genes together with a LOH of the other allele. I´ve tried to do a ChIP assay trying to find histone changes in one of the genes without success (the method didn't work). Could a gene be turned off without DNA methylation and just histone modifications? Perhaps I should work some more with the ChIP technique, although it would be nice to have a more global view and not just focusing on a couple of genes...

I'm quite interested in doing an array after treatment of cell lines with 5-AZA but are anxious in how to find the correct dose so you don't implicate a lot of errors. How do you know that you've chosen the correct dose in order to turn the methylated genes on? And what gene to use as a reference gene? I guess house-keeping genes are not normally methylated...

5-aza-C dosage is dependent on cell line, so you will need to perform a titration over 3-5 days and analyse cell death to determine the levels of 5-aza-C required.

amounts of 1uM to 10uM have been flown about in the litereature however, if your cell line has cytidine deaminase activity it will be resistant to 5-aza-C

Nick

Thanks for your reply. I guess its good not to kill the cells while doing AZA experiments... :-) Should I go as high as possibly without killing the cells and then trust that its enough to reactivate genes?

I repeat one of my questions, could a gene be turned off without DNA methylation and just histone modifications?

Thanks!

of course you must test the reactivation of your gene of interest at the highest possible drug concentration by q-PCR.

and it is entirely possible to have gene silencing that is independent of DNA methylation, a pubmed search would yield a number of papers I would say.

Nick

"My gene of interest" - but if I don't know which gene I'm interested in... I guess I can test a gene which is normally methylated. p16 perhaps...

Sorry for asking a lot of questions, but since my lab isn't into these techniques I don't have anyone else to turn to...

... so here's another one...

I've been thinking about it and I would like to do both arrays on cell lines treated with AZA (or is zebularine a better choise?) and ChIP-CHIPs with an antibody against methylated cytosines. But is this feasible to do when having no experience doing immunoprecipitations? It seems as a bit tricky technique... Any advices on what to think about?

Thanks again

molbio,

arrays can be rather expensive, and there are many checks in place to see your fractions are suitable to put on an array.

Have a look at affy's protocols with respect to the quality controls along the way of performing your IP.

I would go ahead with the untreated cells for array assays just to get to know the technique, and it would help to have a data set for your cell line of interest (this may not be possible). people normally perform q-PCR on their IP fractions before putting them on an array. if there is more than 8-fold enrichment with the gene you know is methylated then it can be put on an array.

as for aza-dc treatments, again I would optimise the amount and length of time required for an effect and once the array protocol has been optimised, then go ahead with it.

N