Transduced cell death in puromycin, MOI problems - GeCKOv2 screen

[Stefanos Skylakakis]

Hi everyone,

I've been having an issue with transduced cells dying in puromycin during the screen/ moi test using GeCKOv2. 

Normally, this is my protocol:

I seed/infect on day 1 at high confluence in complete media (about 25 million wild type eHAP cells per 145mm dish in 20ml, or 1.67 million per 6-well in 1.32ml - no spinfection). 

The next day, I replace media (2ml total for 6-well) to include 1ug/ml puromycin. I refresh the media (+puro) on day 3. 

On day 4, all the control cells (no virus) are dead, so I split (trypsin) and count the cells to calculate the MOI, or re-seed in case of the screen (yes, the confluence of the no-puromycin control is too high at this point, but the "30%" should just about be getting confluent). The efficiency of transduction is usually good, and I'll need 2 to 7ml of non-concentrated lentivirus per 145mm to get 30% (or 300ul - 500ul for 6-well). 

After splitting on day 4, if re-seeded in puromycin (at lower density, 7.5 million per 145mm dish), most of the transduced cells will die (about 2% survival relative to no-puromycin on day 6). I have also tried waiting to add puromycin until the next day (when the cells have attached) but observed a similar effect then. 

My first thought was the confluence was too high conferring some kind of puromycin resistance to neighbouring cells, so on day 2, I split, re-seed at low (0.332 million for 6-well) or high (re-seed everything) density into puromycin. On day 4 I count for MOI, and find about 30% for the high density, but 5% for the low density (the no-split was about 50%). I re-seeded all the conditions at 0.332 million cells per well into puromycin and on day 6 I counted again, and found 2% survival for the high density re-seed, 1.3% for the low density re-seed, and 1% for the normal condition (no splitting until day 4). 

All the above were done with no polybrene, I have recently also added polybrene (up to 8ug/ml) and seen a mediocre increase in MOI. 

My main concern is cell death after day 4. I don't know whether the cells repress expression of puromycin resistance after day 4, whether the lentivirus never integrated in the first place, whether trypsin sensitises the cells to puromycin, or if there is some other effect. On the no-virus control, all cells completely die in puromycin by day 4, so some cells must have been transduced by then. I don't overly stress the cells during splitting either (quick wash with warm PBS, 5 minutes in trypsin at 37 degrees, resuspended, counted, re-seeded within 20 minutes). 

The reason I use such high confluency in the first place is because I would need fewer dishes/flasks, and my impression was that the transduction efficiency might be higher (so less virus used). Would you say cell density is what is causing these issues? I don't spininfect but could try it if you think this would make a difference. 

I could try and avoid splitting on day 4 all together by seeding at low density on day 2, and keeping them in puromycin until they get confluent for the screen, probably something like 4-5 days (but do the MOI on day 4 to avoid high confluence in the control wells) - I haven't tried this yet. I have also thought of using PBS-EDTA instead of trypsin to reduce stress on the cells but have never tried this before and don't know if this would affect anything else. Maybe waiting until day 3 to add puromycin might also help reduce stress on these cells, and give more time for integration? Finally, it just occurred to me that perhaps I have a mutation in the VSV or PAX2 vector that would prevent the lentivirus from efficiently integrating into the genome, but I have sequenced important regions of these plasmids and didn't see any mutations.   

To note, I have completed the screen a couple of times despite the death, PCR etc yielded the desired bands just fine, and sequencing went okay with the data yielding some candidates but with low fdr values. I was thinking that losing representation during day 4 would have reduced the statistical power (even though all the other steps were done at 500 representation). 

Do you have any insights that could help me figure this out?

Many thanks in advance,

Stef

[Paola Dama]

Hi Stef,

I had the same problem and I gave the same your explanation.

And I suppose to do the same. I'll keep you updated next week. 

I'm very interested in this issue.

Thank you

Paola 

[Chunyang Ni]

Hi Stef,

Seeding density is a very improtant facor for cells to tolerate puromycin. You may need to try set up a kill-curve with a determined seeding density(typically 10%-20%) and stay with the density/puro concentration all the way through the selection.

I had the same experience when I first tried Brunello library: I seeded 7x10^6 cells with puromycin the day after transduction and they live pretty happy until I split them to 3x10^6 and all died. The good way is to seed them at 3x10^6 every time you need to replate them in puromycin.

I hope this may help.

Ni

[L Tee]

Hi,

I haven't come across this problem. I always seed my cells on day 1 and then add puro the day after (its important to be consistent between MOI titration and actual expt so do the same for both). I never go above about 1 million cells per well of a 6 well plate, I found the numbers used in the paper were much too confluent for me,  for one cell line I have used I even had to reduce this to 0.5 million per well. 

I always use polybrene (8ug/ml) and I always perform the spinfection as this does make a difference. I try and avoid splitting cells during puro selection though if I have to I allow a day for them to settle down again rather than splitting straight into puro, maybe they are more sensitive before they are stuck down

[Brian Iaffaldano]

Hi Stef,

I have seen this second round of selection you described and have assumed that there are some density dependent effects, where high density culture with a good amount of transduced cells can "protect" untransduced cells for a time.  Including a passage into puro earlier might help you get a handle on the real titer sooner.

Also, when passaging the cells I would be careful how much trypsin you are carrying forward and maybe add a gentle spin to remove it.

[Michael Baughn]

Never seed adherent cells directly into puro selection media, instead re-plate into normal media and then replace the media with your puromycin-containing media after a few hours (once the cells have a chance to re-attach to the plate).

Even puro resistant cells will fail to attach (and eventually die) if split directly into puro.

[Stefanos Skylakakis]

Thank you for your reply. I see. I suspected this before so waited for cells to attach before adding puro - my impression was there was still significant cell death but I didn't count to compare back then. My thought process was if I did the initial selection before cells attach it might be more stringent, so when I split them again the cells that have survived must be properly resistant - I can see that was wrong now. I will try waiting for attachment in combination with Chunyang Ni's suggestion to titrate puromycin down for lower seeding densities (when I titrated I must admit it was for high density so this is very likely to be causing problems also). 

Many thanks,

Stef

[Stefanos Skylakakis]

Hi Brian,

Thanks for your reply. You are probably right, the density is likely to be the major factor. Im generally careful not to carry over much trypsin but will keep this in mind.

Best,

Stef

[Stefanos Skylakakis]

Hi L Tee,

Great points. Do you happen to know (ballpark) how much difference spinfection makes to transduction efficiency? If you were to compare spinfection vs no-spinfection (both with 8ug/ml polybrene). 

Best,

Stef

[Stefanos Skylakakis]

Hi Chunyang,

This is actually very helpful. I knew the seeding density affected death in puromycin but it never occurred to me to titrate down the puromycin depending on the density. I guess 1ug/ml is not that little after all. I'll try a kill curve for 332k per 6-well with a couple of concentrations below 1ug/ml. Thanks for your help!

Best,

Stef

[Stefanos Skylakakis]

Hi Paola,

Thank you for your reply - it's reassuring to know I'm not the only one having this problem - I was getting worried as I didn't see any posts about this issue on the forum. I believe Chunyang Ni's suggestion might be able to solve this issue: so infect on day 1 (spinfect or not I guess is up to you), split on day 2 at low density - something like 332k per 6-well,  wait until day 3 to attach and then add puromycin - but make sure the concentration of puromycin has been selected for that density of 332k cells per well. I guess one could also use a different concentration at the start for high density, and a lower one when seeded at lower density, but the first suggestion is probably cleaner.

Let me know how it goes!

Best,

Stef