Mouse Brain Nuclei Isolation for 10X GEM-X Flex Gene Expression

Dear All,

I am a research assistant working in the neurologe field. As part of a project we are currently performing nuclei isolation for single-nuclei RNA sequencing from frozen mouse brain tissues. For the nuclei isolation we are using a protocol that has been optimized and used with good results in human brain tissues. The protocol consists of a homegenization step followed by a sucrose gradient, washes and filtration of the samples.

We are using the new GEM X Flex gene expression protocol for the other steps (fixation, hybridization, probe hybridization, GEM preparation and library construction).

The issues we are having are:

  1. The nuclei are clumping a lot, mostly after fixation and resuspension does not help much.

  2. We lose about 90% of the nuclei from the first count (After nuclei isolation and cleaning of debris) to the last count (before pooling of the samples and gem preparati on).

  3. Most nuclei have a nice form but some of them have a unregular nuclear membrane.

I was wondering if some of you is doing or has done a similar experiment using this protocol?
Do you perhaps have some suggestion for the nuclei isolation that might fit better for mice brains?

Thank you very much for your help :slight_smile:

Andrea

Hi Andrea,

Thanks for reaching out! We have run several Flex experiments (both with the previous v1 kit and the new GEM-X kit) on mouse tissue, with good success. Our nuclei isolation protocol is also homogenization-based, but also includes an iodixanol gradient-based demyelination step. The protocol can be found on protocols.io here. The only modification I have made is performing two additional filtration steps through 40 um cell strainers after the initial pass through the 70 um strainer (step 8.1.9). I have not had any issues with nuclei clumping; these additional straining steps may assist with that! We have also not seen any issues with nuclear membranes, though we haven’t investigated too thoroughly.

Attrition has definitely been an issue for us as well. The new GEM-X kit is has altered reagent chemistry that supposedly decreases the loss following fixation and probe hybridization, although we have still seen about a 40-50% loss after both steps. The post-hybridization loss seems to be an ongoing issue with this workflow in general (others have experienced the same issue); it might just be a numbers game at this point. We have been using fairly large tissue samples, so numbers have not been an issue for us (yet). I know 10x recommends having at least 25k nuclei going into hybridization, so it may depend on how many nuclei you are able to isolate from your samples.

Hopefully you can have greater success with this protocol!

Best,
Emma

Dear Emma,

Thank you very much for your very detailed answer and your protocol!

We will try the protocol this week, I hope it will get us better results :slight_smile: I have one question:
Why do you do 2 Filtration with different filters instead of just one? Did you notice too much debris by performing the filtration only one time?

Thanks in advance,
Andrea

Hi Andrea,

We found that the additional 40 um straining helped filter out some smaller debris that may have passed through the larger strainer. But your tissue may not require that!

Best,
Emma

Hi Emma,
Thank you for your answer and thank you again for your protocol! We have indeed seen a big improvement in pur nuclei number and condizioni.
However we are still unable to remove the clumps after fixation (we are following the 10x fixation protocol, over night 4 degrees).
Are you also using this protocol for fixation of the nuclei prior to rna seq? Do you have any clue how we could reduce nuclei from clumping?

We see under the microsocope that the clumps are hold together by rna and/or dna filaments, probably coming from nuclei that were damaged during the procedure.

Thanks in advance, this is helping us a lot :slight_smile:

Andrea