shRNA stable line - need better reduction of target (Nov/05/2009 )

Hello all,
I recently made some shRNA stable lines in U2-OS cells. While the shRNA worked well in transient transfections (>60% reduction with ~60-70% transfection efficiency), the final clones I obtained after two rounds of selection exhibited <25% reduction compared to controls. The shRNA plasmid also has pCMV driven GFP expression, so I can see that all of my clones (~30) have clonal GFP expression, so integration occurred. I am assuming that the copy number is too low compared to the transient transfection, since GFP fluorescence is so sensitive. Any ideas on how to improve the copy number? Should I use higher concentrations of antibiotic for selection?

Since you have GFP+ cells indicating integration, I wouldn't think increasing antibiotic would do much (btw - what are you selecting with? puro or hygro or something else?). My first suggestion would be to see if your cells would tolerate more plasmid.

Does your plasmid use and IRES for the GFP? That can sometimes decrease the expression of other parts of your construct. Usually a problem for translated proteins, but might also affect shRNA? Not sure about that. Do you have a version w/o the GFP?

If retroviral delivery is possible, that might be better than a plasmid for generating stable shRNA.

skeuos on Nov 5 2009, 10:11 AM said:

I am selecting with 2 ug/ml puro. The plasmid does not have an IRES for the GFP; the shRNA is generated from a human U6 promoter for pol III. The same vector can also be used for retroviral infection, but I have been avoiding this approach due to the stricter regulations required and need for packaging lines. While my lines are GFP+, they do show a lower intensity of fluorescence than the transiently transfected cells.

skeuos on Nov 5 2009, 10:11 AM said:

Also, I meant raising the concentration of the antibiotic in a repeat attempt to generate stable lines, (i.e. raising the selection stringency following transfection), not raising the concentration of the antibiotic on my existing lines. I am already transfecting as much plasmid as I want to.

Dr Teeth on Nov 5 2009, 01:11 PM said:

Yeah, I figured that's what you meant with the puro. Sometimes cells do funny things to escape puro selection (things not related to your target gene), so I'd personally be hesitant to add more. But that's something you could address once you get the knockdown you need. I usually use 2ug/ml for shRNA selection post retroviral delivery; some people do use 3 and 4 ug/ml so you could try it . . .

How well does the puro kill your control cells? Are you getting the standard 100% killing withing 3-5 days?

Transient transfection will usually give you better knockdown than you get from stable lines.

just to make sure.... these are lines derived from single clones and not from pools of selected cells (b/c you can always screen more clones).

What method are you using for the transfection into your cells? Calcium phosphate usually gives long tandem arrays of the integrated plasmid (the problem is usually too much integration) while you get may get less DNA from an electroporation. Trying a different method may help get you more integration events.

You could try tansfecting in more DNA

Please correct me if I misunderstand your experiment/result. Did you generate stable lines by selecting the transfected cells? If so, then my first question is: did you linearize your plasmid DNA before transfection? Non-homolog end joining the linear DNA increases the copy number of integration.

Also, you can try:
1. SORT the top 10% GFP+ cells should give you better chance
2. Use IRES-GFP or IRES-puro for selection/sorting. The expression level of gene via IRES is generally 10-fold lower, so (stringent) selection gives higher copies.

skeuos on Nov 5 2009, 03:44 PM said:

2 ug/ml Puro is killing 100% of my control cells within 2-3 days. More details: I transfected my shRNA plasmids into my target cells on 35 mm plates, and after 2 days of recovery, trypsinized the cells and transferred them to a 150 mm plate with selection medium. After ~1 week, I isolated 48 individual colonies via sterile cloning discs and transferred them to 24-well plates for a second round of selection. Surviving clones with clonal GFP expression (>30) were expanded and evaluated for target knockdown. However, the best knockdown in any clone was about 25% compared to ~50% in transient transfections. However, while the transient transfection showed better overall knockdown, the transfection efficiency was only 60-70%, suggesting the knockdown was efficient in cells expressing the shRNA, so what I want is a clonal population of cells expressing the shRNA--hoping for better overall knockdown and having homogeneous target cells. While my clones do have the shRNA integrated exhibiting clonal GFP expression, I believe that the copy number or expression of the shRNA is much lower than in the transient, as seen by the decreased fluorescence intensity relative to transient. So, again what would be best to improve the amount of integration events?

I haven't tried linearizing the plasmids or FACS, so those might be worth a try.
As for transfecting more plasmid, I am already transfecting 2 ug into cells on 35 mm plates, so I don't want to increase that much more. Since I have checked over 30 clones, I don't think screening more at this point will help.
As for the IRES, I have used plasmids with these before, and while I always got clones with integration, the expression was always lower than with plasmids without the IRES, as you stated. Since I am already having expression problems, I am avoiding IRES vectors too.
Thanks for the input!

Dr Teeth on Nov 6 2009, 09:13 AM said:

Hmm - upping the DNA always runs the risk of increasing toxicity but since you're looking for clones to expand, might it be worth a try? Since you started from clones, I'm not sure how much FACS would improve things. It's possible, but theoretically each cell in your pop is identical since it's a clonal expansion, so they should all have the same level of knockdown and GFP expression. If you really don't want to increase DNA, at this point I'd consider the calcium phosphate and/or linearizing the DNA. I don't have experience with either (I usually use lipo2k with intact plasmid). If it's not too much trouble, I'd give all 3 possible combos of calcium phosphate and linearizing a shot, that way hopefully one will work out.

Actually just checked the lipo2k protocol, and it suggests using 4ug DNA for a 6-well plate, which is approx the same area as your 35mm dish. DNA tolerance always depends a bit on the cell line, but figured I'd give you some food for thought.