James,

Excellent question, and one I also have been struggling with as well. Primary mouse T-cells are tricky, and if you look at them wrong they like to die. I have attempted various types of liposome based transfections with minimal success and high toxicity, similar to your nucleofection assay. I recently have been experimenting with lentivirus-GFP to assess transduction efficiency under different conditions. While the toxicity is much lower than other methods, my gene transfer efficiency is also lower. I am attempting to optimize this as we speak, so if I have success I will let you know.

Another thought (while significantly more work) is to use a Cre/lox model. I found that indeed a cFLIP flox mouse exists (PMID: 16043517). You could take this mouse and cross it to an inducible ROSA-Cre (expressed in all tissues, and only activated under tamoxifen or doxycycline treatment). Once the animals are of age, extract the t-cells and grow them in vitro. At the right time, you could treat with tamoxifen or doxycycline and efficiently recombine out cFLIP. This would not be transient knock-down, but complete knock-out... but your efficiency should be very high with toxicity low.

I know its an undertaking (we are doing it now in our models), but gene delivery to these cells just has not been efficient enough... ESPECIALLY if 50% of the cells are not transduced and they repopulate the animal you are reconstituting (which confounds your results).

I will keep you posted on my lentiviral success/failure... good luck.

~Adam

James,

Excellent question, and one I also have been struggling with as well. Primary mouse T-cells are tricky, and if you look at them wrong they like to die. I have attempted various types of liposome based transfections with minimal success and high toxicity, similar to your nucleofection assay. I recently have been experimenting with lentivirus-GFP to assess transduction efficiency under different conditions. While the toxicity is much lower than other methods, my gene transfer efficiency is also lower. I am attempting to optimize this as we speak, so if I have success I will let you know.

Another thought (while significantly more work) is to use a Cre/lox model. I found that indeed a cFLIP flox mouse exists (PMID: 16043517). You could take this mouse and cross it to an inducible ROSA-Cre (expressed in all tissues, and only activated under tamoxifen or doxycycline treatment). Once the animals are of age, extract the t-cells and grow them in vitro. At the right time, you could treat with tamoxifen or doxycycline and efficiently recombine out cFLIP. This would not be transient knock-down, but complete knock-out... but your efficiency should be very high with toxicity low.

I know its an undertaking (we are doing it now in our models), but gene delivery to these cells just has not been efficient enough... ESPECIALLY if 50% of the cells are not transduced and they repopulate the animal you are reconstituting (which confounds your results).

I will keep you posted on my lentiviral success/failure... good luck.

~Adam

Hello James,

Nucleofection leads to lot of cell death. I faced a similar problem and it becomes particularly problematic when the goal of such experiments is to measure/ modulate apoptosis. I would suggest that you use Accell-delivery reagents from Dharmacon, That is what I used. It is a passive delivery method and so one needs to culture the cells with the K/D reagents for atleast 3 days for efficient K/D at protein level. However, it is very good for delivery and Knockdown in primary suspension cells (I used for mouse T cells while I was differentiating naive cells into Th1/Th2) and leads to less cell-death.

Use retrovirus or lentivirus transduction (retrovirus does not need S2 safety labs).

Make sure you do NOT use the CMV promoter, it becomes silenced easily - especially in primary immune cells.

I'd suggest you use EF1a (human, very long though), U1 (human) or ES (short EF1a, only ~ 240 nt, EF1a without upstream regions and intron but I see pretty much the same strength in expression). Add wpre to improve mRNA stability and export.

For Tcells make sure you grow them in rich media (I'd add amino acids extra and Hepes buffer too maybe) and have them dense enough, they need to be pretty dense to benefit from the cytokine micro-environment.

Try this......http://www.ncbi.nlm.nih.gov/pubmed/19924146

I used acell delivery reagent from dharmacon and obtained good results, but is too expensive to be used for troubleshooting.

Thanks very much everyone for your helpful responses.

Adam - that's a great idea about crossing the flox mouse.

Linda + Vandanna - interesting about the Dharmicon reagent. I hadn't heard about it - my only experience with reagents was hearing anecdotally that Lipofectamine wasn't so good for primary cells.

Ulrike + Max - I think lentiviral transduction will be my next approach. Interesting to know the CMV promoter doesn't work so well - the people I have spoken to at Sigma tell me their shRNA is driven by a U6 promoter, but that if you want GFP, it is driven by a CMV promoter. Interesting also that it works better with activated T cells - I would prefer to have my T cells resting, but all the literature I've seen recommends activated T cells also.

Does anyone have experience of the best multiplicity of infection (MOI) range using lentiviruses to transduce mouse T cells? Is polybrene helpful? If the CMV promoter which drives GFP expression doesn't work so well in mouse T cells, are there alternative means of determining best MOI other than transduction of T cells with control lentivirus with CMV-GFP? The limited publications I've seen seem to recommend 24 stimulation of the T cells prior to two cycles of viral spinoculation on fibronectin plates at 48 and 72h - does this seem a good approach? I'm planning on ordering ready made lentiviral shRNA particles from Sigma, though this is rather pricey ($4000 for 1ml of 10 par9 viral particles) - are there more economic ways of doing it?

Thanks again everyone for your advice.

Why don't you just package your own virus? I managed to infect activated T cell, but needs at least a period of 7days.

@ James - Additional tip if you don't mind one more. Add the TBK1 inhibitor BX795 (Invivogen). It works to nearly double transduction efficiency on many primary cell types we've tried as well as ones unpublished. Human and mouse. PMID: 22779406

Polybrene is used to mask membrane charge in case of VSV-G pseudotyping. VSV-G needs particle and cell membrane to be very close for the intercalation into the membrane (VSV-G does not act as ligand to a receptor but binds to certain membrane fatty acids it seems). The common concentration final is 4 ug/ml for Polybrene.

For the selection: you can use GFP, just replace the CMV with a better promoter. This publication should help you. http://www.ncbi.nlm.nih.gov/pubmed/20033928

Unfortunately most company constructs are plain very "simplified" and not up to date with recent knowledge of transgene expression, so they still use U6 (risk of perturbated miRNA regulation networkt due to extreme strength of promoter, thus artifacts) and CMV (developed on 293T but not good in immune cells in general).

Many thanks everyone once again - these answers will really be a big help with my experiments.

Adam - I've been in touch with the group who made the floxed mouse. They have agreed to send me some + I will cross it with a ROSA-Cre-ER mouse from Jackson labs. This will take a while with quarantine etc., but seems a good approach.

Ulrike - I'm in the process of ordering the lentivirus from Sigma - I share your concerns about the U6 and CMV promoters, but with my backrgound as a transplant surgeon, I don't think my molecular biology abilities are good enough for making virus myself!!

we have very good results with nucleofection using stabilised siRNA (see http://www.ncbi.nlm.nih.gov/pubmed/18792414). Transfection is always 100% and knockdown can be >90% but of course depends on the specific siRNA used.

The only problem might be in vivo proliferation, because then the siRNA is diluted quickly. We have shown that using CFSE labelled CD4 T cells in vivo. So probably longterm in vivo assays involving T cell activation will probably not work.

James,

A quick follow-up. I have recently attempted transducing my freshly isolated primary t-cells with lentivirus GFP and had wonderful success. I simply plated the cells in complete media containing 10 MOI of virus and 4ug/mL polybrene. My transduction efficiency was greater than 80% visually by microscope (independent of CD28 stimulation) with little to no toxicity, but I could quantify it by flow if necessary in the future. Additionally, if you use a siRNA that is GFP tagged you could flow sort out the green cells and plate those to have a 100% infected population.

Good luck in your experiments,

~Adam

Thanks very much Adam - well done, that's great news!

Was the GFP on a CMV promoter? Were your T cells resting or stimulated?

All the best,

James

It was in fact on the cmv promoter, so that did not seem to be an issue. I am attempting to express Cre-recombinase so transient expression will suffice, but as others have stated cmv may not be the best for longer expression times.

The tcells were stimulated with cd3 with or without cd28 (we sometimes perform experiments either way depending on the response we are looking at). Both showed robust GFP expression independent of the cd28.

Good post... Hope this was helpful, and good luck!

~Adam

Brilliant Discussion! I have been thinking alot about this too, The collective wisdom I've received seems to be against Lentivirus for primary mouse T cells...I can't tell if anyone here is using Retroviruses instead of Lentivirus for shRNA into primary mouse T cells? I don't mind if the T cells have to be activated, but I do want integration.

And does anyone know-is there a U6 (or EF1a, thanks Ulrike!) -shRNA--MSCV LTR-hCD4 (or other surface marker) construct out there? I've been searching for months...I want to be able to enrich the transduced cells without FACS sorting, can do minimal mol bio, but it's not my specialty, can package up virus no problem.

Thanks!!

@Alexander Scheffold - have you tried that on large animals T cells, particularly bovine T cells?

no sorry, but it works on murine and human T cells so why not on bovine?

I use TransIT-TKO (Cat #MIR2154) from Mirus to transfect siRNA in mouse CD4 T cells. I 've got a great down regulation of the protein target and expected effects on cytokine release. Briefly, I transfect cells after isolation of CD4 T cells in 12-well plate without any coating. 24H later, I induce T cell differenciation as usual.

To confirm my data I tried to transduce CD4 T cells with retroviral particules. The infection efficiency is about 1% (GFP detection by FACS). Is there anyone that succeed in this kind of experiment?

Thanks in advance.

Any one tried to transfect fresh naive CD8 T cells without stimulation? I tried to transient transfect fresh naive CD8 T cells with miRNA mimics. Naive CD8 T cells stick tightly to well and are supreme hardly off well while sitting in none serum media for 6hrs.