MSP-M bands in healthy people - (Feb/05/2007 )

Hello, I am a beginner in MSP and I have no one to ask for advise. I am interested in methylation profile of lymphomas. After several months without any result I finally succeeded to convert DNA and to do MSP. To start somehow I chose six genes that should be often methylated in lymphomas. I found primer sequences in various articles and I am trying to optimize my pcr reactions using methylated positive control and dna from healthy volunteers. I get bands where they should be (healthy dna-unmethylated, methylated dna-methylated), but MGMT and estrogen receptor seem to be methylated at all of my healthy volunteers ( I have DNA from five young healthy friends) and they give no or very faint unmethylated signals. I do not know what to think about it and it makes me suspicious about the accuracy of my work with primer sets for other genes. Should i try another primers for these genes? or should i try rt-pcr to find whether these people express MGMT and estrogen receptor?
thanks for reply
maxa

Hi maxa,

gratulation for succesful MSPing. I don't know why, but many people think, that DNA methylation occurs only in disease states. This is not true! DNA methylation is absolutely physiological and we could not live if it wasn't working properly! It is also not true, that DNA taken from human blood is not methylated - there is a fair amount of methylation which may be different between healthy subjects, and for example alcoholics (which show elevated methylation). If your five friends that donated the DNA, are male it should not be too surprising to find methylation signals in the estrogen receptor gene...
If your bands make sense for methylated and unmethylated DNA your MSP seems to work. To judge the methylation signal from blood (or saliva, or squamos cells) of volunteers, it is likely to have mixtures of methylated and unmethylated DNA. You would need to do real-time PCR to quantify the amount of methylation. But your findings are nothing to worry about.
If you want to be completely sure you should BSP verify your findings.

Hope that helps,

Krümel

Hi maxa,

okay, I think I got your point. What DNA did you use from your healthy controls? Just extracted from blood or did you seperate specific cells? Maybe you have mixed populations, which contribute to the findings?

K.

Hello Krumel,
I simply extracted DNA from blood. Should I have done any separations?What do you mean with "mixed populations"? Leukocytes mixed with other cell types? But even in case of several cell types I would expect at least both bands on electrophoresis, either methylated, either unmethylated. Or am I wrong?

thanks
maxa




[

Hi maxa,

I worked on DNA directly extracted from blood, too. We always found methylation levels between 70 and 90% for several genes. IF you work on blood, you might still find differences between healthy and affected persons, but you will need larger sample sizes as you may have a difference of 10%. I think this may be due to the fact, that you have a lot of different cells containing DNA in the blood. It may make sense to look at certain cell populations (after FACS separation). There you might find more pronounced differences. Did I get your point?

K.

Hello Krumel,
I think I understand. I will do my experiments on more healthy people. But still, if I find any healthy people giving methylated signals among other healthy people having unmethylated bands I cannot study methylation of this gene in relation with a disease, right? Or is it wrong to use whole blood?
Thanks
maxa

Hi maxa,

hmm, still more questions than answers: You said above, that you tested the MSP on healthy and unhealthy DNA - where did this DNA come from - tissue or blood, also?
I think, if you want to compare the amount of methylation of your promoters in the whole blood of patients and controls, you will have to perform quantitative real-time PCR. From the differences between U and M curves, you can calculate the %amount of methylation. I could imagine, that you will find something like 70% in controls and 90% in patients (or someting like that; but it is not so probable that you'll get 100ù in patients and 0% in controls, I suppose).
If you want an assay based on the information derived from whole blood, then use whole blood! If you are looking for some kind of biomarker, then this is the most convenient source of DNA.

Krümel

Hi Krumel,
I used DNA from peripheral blood.

But I think it is clear now. No more questions )
Thank you very much, you helped me a lot.
maxa