siRNA with low mRNA levels - (Oct/09/2009 )

Hi everyone!

We do a lot of siRNA experiments with primary synovial fibroblasts and normally have no problems with transfection efficency (using nucleofection) or silencing (using 1-2ug of siRNA). Now, I tried to silence the chemokine receptor CXCR4 and it worked nicely in HeLa cells with 0.8 ug siRNA. The problem is that in synovial fibroblasts, even though they make functional protein, mRNA levels are around the detection limit (short half-life?). I tried up to 2 ug of siRNA and different time points, but I cannot see a change in the protein concentration. Does anyone know whether there is a connection between mRNA copy numbers/stability and siRNA efficiency? I asked Qiagen for advice and they told me, if there is enough mRNA to be translated siRNA should work. Well, it doesn't for me...

Maybe you just have a bad siRNA - has it been validated by somebody? Try using another siRNA that targets a different region of your gene.

Edit: Too early in the morning and I didn't completely read your post. I see now that it works for you in HeLa cells.

Is there any splice variants (or SNP) of CXCR4 that one expressed in synovial fibroblasts has no effect by the same siRNA worked in Hela cells?

Functional Screens on Oct 10 2009, 05:39 PM said:

I actually had the same thought and designed primers to detect the 4 different splice variants described for CXCR4. Synovial fibroblasts seem to express the 2 most common transcripts that are targeted by the siRNA and are detected by my primers. I don't know about SNPs however. How could I find that out?

SUI on Oct 11 2009, 11:32 PM said:

NCBI SNP database
http://www.ncbi.nlm.nih.gov/SNP/
How about any CXCR4 homolog or superfamily?

I checked for the SNPs - there is none described for the sequence where the siRNA is binding. It is known that CXCR4 has various antigenically different conformations. So it might well be that I don't detect all of the CXCR4 present on the cell with my antibody. Is it possible that the mRNA disappears too fast for changes to be measured and the silenced protein I cannot measured because I don't detect all conformations? Let's say I only detect with my antibody 1 of 5 possible conformations that are present on synovial fibroblasts, shouldn't I see the silencing in that specific conformation anyway? Or can one conformation be "preferentially"silenced and the expression of others be unchanged?

Lets take one step back.
When you said "CXCR4 mRNA levels are around the detection limit" in Synovial fibroblasts, did you mean the mRNA level before knockdown? If so, I may think CXCR4 is not expressed in Synovial fibroblasts until some activation and the protein you were detecting (by Western blot?) may be due to some cross reaction (but not in HeLa cells). Could you transfect CXCR4 expression vector (plus and minus specific and control siRNA) to Synovial fibroblasts and/or try to use some ligands to trigger endogenous CXCR4 expression and see what happens?

Functional Screens on Oct 14 2009, 05:23 PM said:

It is really possible that CXCR4 is not expressed in synovial fibroblasts and I actually think that is the most logical explanation, BUT expression of CXCR4 is published in synovial fibroblasts (also with FACS, like I did). I also see a band in Western blot, but of course all this could be unspecific. So, how to show convincingly that something is not expressed if others claim it is?
I stimulated with the CXCR4 ligand CXCL12 and got a response which I can totally block with siRNA against CXCR7, the alternative CXCL12 receptor. Since I didn't find a lot of mRNA for CXCR4 (it only comes up around 35 cycles), I concluded that there is no CXCR4. But the reviewers of course saw the publication showing the expression of CXCR4 in synovial fibroblasts and asked me to silence it. So, this is how I ended up trying to silence something that is probably not even expressed.
So maybe I should really try to overexpress CXCR4 and then show that with silencing of CXCR7 I cannot block CXCL12 responses anymore. Would that be convincing?