Drop Seq Optimization of Primary Cells

[Taylor Adams]

I never have issues running Drop Seq with cancer cells (e.g. the 3T3/HEK experiment) and getting quality cDNA library bio-analyzer results and ideal sequencing results. 

Every time we try running cell suspensions from tissue (lung),we get very low cDNA library levels. I know that we should expect lower gene expression in general with primary cells, but I'm worried that having to increase our PCR cycles from 13 to 20 in order to get comparable cDNA library sizes is too big of a difference to be based purely on cells having lower mRNA levels. When using the same protocol for tissue digestions and single cell suspension formation, we get decent results running on a C1 Fluidigm platform, just not via Drop Seq.

Does anyone have any experience running Drop Seq with either primary cell cultures or single cell suspensions from tissue? 

[Johnson, James]

I don't have any experience with drop seq, but I have lots of experience dispersing tissues into single cells. Protocols must be extremely gentle (much less trypsin etc than cell lines on a plate). Otherwise cells are very unhappy and will get sick and shut down. I think that's the simplest explanation. I have no expertise in lung, but I'd start by optimizing the protocol for cell health (ideally with a single cell functional assay)

James D. Johnson, PhD

Professor, 

University of British Columbia

--
If your email account uses an email alias, you may have problems with the automated unsubscribe function. In that case, please email safisher at upenn.edu for assistance unsubscribing.
---
You received this message because you are subscribed to the Google Groups "Single Cell Biology" group.
To unsubscribe from this group and stop receiving emails from it, send an email to singlecellbi...@googlegroups.com.
To post to this group, send email to single...@googlegroups.com.
Visit this group at https://groups.google.com/group/singlecellbio.
For more options, visit https://groups.google.com/d/optout.

[Linas Mazutis]

Dear Adams,

We are not using Drop-Seq but we are using an alternative method inDrops RNA-Seq :) 

We also use primary cancer samples obtained from lung tissue. I have to say we do get decent amounts of cDNA, but obviously the yield is much lower than from cancer cell lines. We noticed one thing working with primary cells, which I think everyone has experienced at one point. Handling cells properly is a key. If you are using harsh protocol to disperse cells that might affect the amount/quality of cDNA, lysis etc.. You should keep cells cool during encapsulation step, that helps. Also, I hope you are making sure that cells enter the droplets? Sometimes cells can be very sticky and because of that they do sediment in the syringe before they reach a chip. Do you know the fraction of productive droplets, i.e. having a cell and a bead? 

Cheers,

Linas Mazutis

[Taylor Adams]

Thanks James,

I'm going to try switching to a less aggressive liberase enzyme digestion instead of the collagenase protocol we've been using and see if this improves our viability.

Taylor

[Taylor Adams]

Thanks Linas,

I haven't tried counting productive droplets with human tissue because our cell counts usually end up so low after digestion that I'm reluctant to waste cells by conducting a short run without the lysing agent, but the next time I harvest murine lung I can compare this to a 3T3 line to see if there's a difference.

Another, concern we have with our cell suspensions is the amount of debris we often see, particularly because it suggests there is substantial ambient RNA throughout our suspension. We observe debris even after pelleting and re-suspending the cell several times. Besides using a less aggressive dissociation method, do you have any advice on how to to mitigate ambient RNA accumulation? Or outright remove it from our cell suspension? 

Taylor

P.S. Impressive work on the inDrop design.

[Linas Mazutis]

Hi Adams,

You don’t really have to harvest murine lung. 

You could look at the inlet of a chip and count how many cells are entering the chip per unit of time (e.g. per 1 min).

Then you should count how many beads are entering the chip per same unit of time. 

If you know the frequency (or volume) of your droplets you could calculate "productive droplets" using Poisson equation.

My suspicion is that you may have very very low amount of productive droplets, which will lead you to very low amounts of cDNA.

We observe debris even after pelleting and re-suspending the cell several times. 

We also see debris (depending on a sample). Sometimes we spin cells in Lympholyte or Ficol or Optiprep. Typically we use Optiprep, but if the amount of cells is very low (e.g. <20.000 cells in total), then we just encapsulate cells without gradient centrifugation. As you can expect, debris increases the background read noise quite a lot.

Do you have any advice on how to to mitigate ambient RNA accumulation?

It depends on your experimental set-up and lysis conditions. If you are not performing cDNA synthesis inside the droplets you might add 10-100 times diluted RNAse into your cell suspension. The idea would be that RNAse would degrade ambient RNA during encapsulation, but once cells will enter the droplet the lysis reagents will immediately destroy RNAse. I have never tried this approach though.

Let me know if you have other comments/thoughts,

Linas

[Taylor Adams]

Hi Linas,

Are you using a high speed camera to count cell and bead arrival?

I think your idea to test the effects of diluted RNAse in our cell suspension sounds interesting and worth exploring. From now on I think we will test RNA levels of our tissue derived cell pellet supernatants and if we continue having issues then we'll run some tests of different RNAse concentration's effects on Drop Seq runs with relatively cleaner cultured cell suspensions and an ambient RNA spike in. 

If/when we get any results of this experiment I'd be happy to share. Thanks again for your advice.

Sincerely,

Taylor