shRNAs work in vitro, not in vivo - (Jul/27/2010 )

I am facing an annoying problem, and I would be more than happy if somebody has some ideas regarding what might be happening here (sorry that my message will be a bit long).

I made shRNAs against two targets, and selected the two best ones for each target based on in vitro knock-down on luciferase reporters + endogenous mRNA.

Then I injected some mice...
- same cassette in the plasmid used in vitro and the viral vector used in vivo
- 100% transduction of the target cells (hepatocytes)
- reporter gene is expressed and doesn't get silenced
- siRNAs are expressed
- there's even a slight liver toxicity in a few mice
... and yet, there's no knock-down, with none of my four shRNAs.

To make it easier, some colleagues of mine did pretty much the same with one single shRNA after screening only with luciferase reporter - same virus, same dose, same protocol, same control, etc... - basically only the target is different, and it worked like a charm.

Oh, and of course it's unlikely that it's because of my targets, because at least one of them has been knocked-down previously in vivo by another group. And one of my four shRNAs is the one this other group published.

I didn't manage to make up an explanation that would make sense, so far... Anybody?? (pleeaase)

how did you check the expression of your target gene? WB, RT-PCR? Sometimes you cannot see the silencing at protein level, but it is clear at RNA level.

No, qPCR, using the same primers as the ones used for in vitro. And the cell line used is derived from the mouse genotype I used.

It seems that nobody had this problem before...?

1. One of your target gene has feedback regulation
2. The primer sets you used for genome-typing are not good for RT-qPCR or qRT-PCR especially when you have genomic DNA contamination.
3. Dosage issue or shRNA potency; the same dosage works for your colleague because his target gene can be silenced in a lower concentration of shRNA. You might want to screen more potent shRNA by playing around the ratio of shRNA vector/luc-target vector (e.g. 1:1 or 1:5). Then use the shRNA for your in vivo experiments.

Functional Screens on Sat Aug 7 20:24:07 2010 said:

Thanks for your reply!
Feedback regulation is I guess a possibility at least for one target... even if the same happens with the other target (already published).
About primers, I've tested a few sets already, all giving the same results, and my samples are all DNase-treated, and my RT- controls are negative.
For my luciferase screen, I have indeed used several ratios, 1:20, 1:4, 1:1, 1:0.2, with the idea to pick the more potent shRNAs, but they all showed the same trend... Would you expect that this has an influence?

I will definitely pick the shRNA always gave more than 70% knockdown in your 1:1 or 1:0.2 reporter assay. You might want to introduce your luciferase reporter in vivo for confirmation.

Functional Screens on Wed Aug 11 06:36:28 2010 said:

Well, most of my shRNAs give more or less 50% knockdown at these ratios, and at higher ratios the % of knockdown is increased (70-80%) but still pretty similar between different shRNAs. So luciferase didn't really help me in my choice. I then decided to pick the best ones based on endogenous knockdown in vitro (70-80%).
I thought that endogenous knockdown in vitro would be more predictive of the in vivo outcome compared to luciferase reporters... Do you disagree?

Have you tried synthetic siRNA instead of ShRNA, they are more potent and easier to design than shRNA/

I've been testing same for couple of month now and faced this problem twice. Solution is obvious - since I've tried to use hominidae species it improved significantly! It's cause reaction on syntheses produce more efficient ways of signals sent by hepatocytes cells...