First, be sure you are making your percoll gradients correctly. For 100% SIP (standard isotonic percoll) you mix 9 parts percoll with 1 part 10X HBSS (or PBS). Further dilutions are made with 1X HBSS or PBS.

Second, tissue homogenization is important since you are separating by density. You want a single cell suspension, as complete as possible. Cell clumps have more surface area, and are thus less dense. They will not partition correctly. I find that mechanical dissociation is adequate, but many protocols include enzymatic digestion. Depends on what you prefer.

Third, be careful when preparing the gradients. I use a syringe and needle to inject my layers, bottom up. This ensures that I don't disturb layers by pipetting on top. An easier method that requires less gradient setup is Frank 2006 http://www.ncbi.nlm.nih.gov/pubmed/16125247. I have had success with both 70/37/30/0 and 70/50/0 gradients. Maybe try both for yourself.

And be careful, you may not see cells, but they are often still there. Keep in mind the size of tissue you start with. I see a thick band when using a whole brain (rat), but nothing if using a piece smaller than a hippocampus. I almost never see a cell pellet either, but cells are present under the microscope. Microglia are very tiny. I suggest resuspending in a very small volume the first times you try - if you overdilute, a small aliquot could be missing your cells.

One thing to try is to suck up volume from the 70/37 interface and immediately plate on a petri dish. Just to see what you've got.

Anyways, just my 2 cents. Good luck. Once you get it, this is a very simple method to reproduce.

I will try to dissociate the tissue better (I will change the digestion cocktail a bit) and use both gradients to find out which works in my hands. Resuspending cells in a smaller volume may help - I haven't realized that it's not necessarily a problem if there is no pellet. Immediate testing of 70/37 interface content is a good idea.

Thank you so much for your advice!

Have you tried papain for digestion?

We've done it from adult human brain with mechanical dissociation, a trypsin digest and a percoll gradient. We had more tissue, however you can scale down to a smaller gradient tube than we used. http://www.ncbi.nlm.nih.gov/pubmed/16061966

We tried using this protocol on mouse brains (with papain digestion, but don't forget to add L-cysteine as well or the enzyme activity is rubbish) and got a pretty good yield, with one brain per 15ml falcon tube. It is better to make the gradients in 15ml falcons over 50ml ones, as you get better separations. Another tip is to make one of the middle layers up using SIP diluted with PBS + Phenol red, with the others diluted using standard PBS - in that way you can see the different layers much more clearly after centrifugation.

I agree with David B Kurland

I know this thread is old, but it seems like it is much simpler just to do a 50% isotonic percoll layer overlaid with HBSS.  The myelin and cell debris will be at the interphase and all of the microglia will be pelleted at the bottom.  With a more complex gradient (70/37/HBSS) you have a layer of microglia just floating in the tube, which can be hard to see and recover.  With a single 50% layer all of the supernatant can be safely aspirated and the microglial pellet simply resuspended and washed with HBSS before plating.  Better yields, simpler execution.  Why mess with complex gradients? What's the benefit?

Michael, although I've never seen this published, why don't you give it a try and see?

My guess is that you may also pellet denser things that you don't want, i.e., nuclei from broken cells. The complex gradient allows for better separation, I'd suppose. And could be "softer"  on intact cells because spins are usually for at least 30 mins.

As I noted above, I use a needle (filed down to remove the bevel) and syringe to "inject" my layers in reverse order. This way, it's really quite easy to set up the complex gradients.

Hi all,

I went back to these experiments in last weeks during my internship and I succeeded to obtain the cells!

I was looking for microglia and infiltrating peripheral leukocytes in mice infected by Trichobilharzia regenti. I used a pool of three mice to get enough cells because I isolated them from brains and spinal cords (separately). I used mechanical dissociation of the tissue and 30/70% Percoll gradient. I got many microglia and other cells.

Considering cell yields, the spinal cord was much worse (trash contamination), but still giving reasonable data.

Tomas

https://en.wikipedia.org/wiki/Trichobilharzia_regenti

Hi Mariana,

I don't know the maximum microglia yield because I stop acquisition after 10.000 cells for the brain and 5.000 for the spinal cord. From these cells, amount of microglia is as presented in the attached figure.

Considering the protocol, I use 5 ml Dounce homogenizer after cutting the tissue with blades. I use 30/70% Percoll gradient, resuspending the cells in 30%. Otherwise, it's mostly the same as in Lee and Tansey 2013.

If you search just for microglia, you can use 37% Percoll only, see attached paper.

Good luck.

Tomas

Article CD11c-Positive Cells from Brain, Spleen, Lung, and Liver Exh...

Hi guys. So the adult microglia exist in between 70 % -37% percoll, right? Are there any other cells in this area?

Hi Cuilian,

yes in this phase will be basically all immune cells (myeloid lineage, leukocytes, ...).

This recent publication out of Beth Stevens' lab describes something similar to what Michael David Hoos is describing (single percoll layer and pelleting to the bottom): https://www.biorxiv.org/content/early/2018/08/31/406140

I will say that although the yield may be higher, you should separate the microglia via MACS sorting (an extra 30-60 minutes) following collection to ensure a homogenous cell population for plating or other experiments.

I know this is an old question, but hopefully there is someone who can help me. I have been trying to isolate microglia using Percoll gradient (70/37/30) and despite I have seen the ring and it seemed as if I had plenty of cells, unfortunately all of them were dead... Do you think it is just the matter of time that cells "sit" in Percoll or something else? Any tips or tricks, please?

Julia Michalkiewicz - don't let them sit in the Percoll for any longer than you need, it's not really very nice to the cells at all. Perhaps check you're not spinning them too fast in the centrifuge as well, and use soft braking. I've never found them to be as tough as circulating immune cells. Keeping them cool once they're dissociated out might be helpful too. Good luck figuring it out!

Simon McArthur Thank you! I am trying to do everything as fast as I can an I am spinning them in Percoll gradient at 2000 rpm for 30 minutes without breaks... I am trying to do everything on ice and I still fail to obtain them alive.

Julia Michalkiewicz I can send you the protocol I use if you like, to see if there's anything obviously different? Could you DM me your e-mail address?

Julia Michalkiewicz are you mechanically or enzymatically dissociating? Using a dounce homogenizer without enzymatic processing, I improved my yields. It is quite easy to overdo the amount of strokes which can cause a lot of cell death. I would also echo the suggestion to keep the cells cool at all times.

A little more info about your protocol would be useful. Best of luck!

Elliot

You can find some information that may help you in this quite recent paper:

Article Dissociation of neonatal and adult mice brain for simultaneo...

Good luck !

Lucas

Hi Elliot J. Glotfelty !!

About the bounce homogenizer, I tried to mechanically dissociate my mice brains and I get a lot and I mean A LOT of cell death, how many strokes do you recommend? And how strong?

Julia Michalkiewicz I'm having the vey same issue, today I tried enzymatic digestion and then a 40/70% percoll gradient, I seeded them at 10k cells per well (96-well plate) with DMEM high-glucose + 10% FBS + 20 nM GM-CSF, are you trying to isolate adult mice microglia?

Thank you guys!

Marcela Cárdenas-Tueme sorry to hear that happened! What size dounce homogenizer are you using? I start by chopping the brain with a scalpel (in ice cold HBSS in a Petri dish). Transfer these pieces to the 15mL dounce in HBSS (on ice). Use the loose pestel first (5-10 strokes with rotation) and then the tight pestel for 5-10 strokes (with rotation). Over homogenization will cause death-- less is always better if the job is complete. The first time I did this, almost all of my cells died so it definitely takes a little practice.Transfer to conical tube and spin down for 8 mins at 700xg. Decant and add percoll to the pellet (37-40% with or without 70% cushion-- see papers below for alternative methods of doing this). Spin for 30mins at 500x g and complete isolation (if you need more details, below are some resources that are useful). Plate on PDL coated coverslips or plates =)

Here are a few good papers to look through:

Very helpful article that details centrifuge speeds etc. that I also use:

Article Isolation and Culture of Rodent Microglia to Promote a Dynam...

More in depth look at the protocol for rat brains (similar for mouse):

Article Isolation and Culture of Microglia

I also use the method outlined in this paper (all done at 4C or on ice for RNAseq/FACS sorting):

Article Single-Cell RNA Sequencing of Microglia throughout the Mouse...

Good luck!

Elliot

Thank you so much for your answer! Elliot J. Glotfelty

Marcela Cárdenas-Tueme no problem! good luck!

Hi everybody!

I followed Elliot J. Glotfelty recommendations and I successfully isolated microglia. I chopped my mice brains in ice cold HBSS which totally benefited the extraction.

PD. I added 20 nM GM-CSF in my cell culture and they're now proliferating.

Good luck everyone

Look at these😍

Hi everybody!

This thread is very useful! Thanks for your responses!

Marcela Cárdenas-Tueme and Elliot J. Glotfelty , do you mind sharing at what temperature do you perform percoll gradient separation?

The protocols outlined in the references shared by Elliot J. Glotfelty all suggest percoll centrifugation at 4C, whereas the sources elsewhere state using percoll at room temperature (18-23 C).

Thanks in advance!

Aleksandr Zakirov -- it depends on your experiment. For single cell sequencing or other RNA work I try my best to prevent ex vivo activation of the microglia (and drastic mRNA changes) and thus used the 4C (at this temp I also only use the 37% solution and pellet all the cells to the bottom of the tube). I've also used 18C when I am doing MACS sorting/plating of the microglia. If you are using the 37% solution and 70% Percoll cushion, the higher temp may allow for a better formation of the interphase layer between them. When the cells are plated, they respond appropriately to LPS and other stimuli.

Thank you very much for your response, Elliot J. Glotfelty !

As far as I understand, you are not performing red blood cell lysis step?

I presume that using single (e.g. 37%) percoll solution pellets microglia and red blood cells together, whereas having a gradient (30/70 or 30/37/70) separates microglia (at the interphase) from red blood cells (pellet at the bottom).

Aleksandr Zakirov I have never worried too much about the blood cells as I always perfuse the animals with ice cold HBSS prior to extraction (removing most of the blood from the brain). If plating the microglia, I use the 37%/70% percoll layers, which does separate the blood cells to the bottom. In addition, I perform a Cd11b MACS extraction of the cells in the interphase layer, which further purifies the population for plating. If I am moving the cells to FACS, the markers I use to analyze the cells should suffice in weeding out any cells that aren't of interest.

Best of luck!

I really appreciate your help, Elliot J. Glotfelty ! Thank you!

Aleksandr Zakirov no problem. Do let us know how your experiments turn out! Best of luck! Do take a look at this recent preprint that discusses prevention of ex vivo activation of microglia in the brain dissociation process: Preprint Single Cell Sequencing Reveals Glial Specific Responses to T...

Hiii! I'm so sorry for the late answer, my dad is very sick in the ICU and I've been out of my mind lately.

I usually do the percoll gradient centrifugation at room temperature, which would be around 18-23ºC, that's because just as Elliot J. Glotfelty said I'm not doing RNA sequencing/extraction nor anything to do with RNA. Also, I haven't been perfusing my animals lately because my pump is broken :(. However, with the 37%/70% percoll layers the majority of blood cells go to the bottom pellet and my cultures have been working just fine :)

Good luck and stay safe!

Marcela Cárdenas-Tueme , I am really sorry to hear about your dad's health issues. I hope he will get better soon!

To continue the topic, Elliot J. Glotfelty and Marcela Cárdenas-Tueme ,do you mind sharing how much microglia do you manage to extract per gram of tissue?

I tried 30/70% percoll gradient at 18C after dissociating adult rat brain tissue with octoMACS dissociator following the manufacturers protocol. I ended up with only ~1.5x10^5K cells, which seems to be much less than 2-3x10^6 cells reported in the literature.

I am now trying to figure out where am I losing my cells.

I really appreciate your advise! It's been extremely useful!

Aleksandr Zakirov you are likely losing cells during the homogenization stage. Less can be more when using the dounce. Marcela Cárdenas-Tueme also sorry to hear about you Dad. Sending good energy yours and his way! For the transcardial perfusion, you can also use a 10 or 20mL syringe as an alternative to using a perfusion pump (just takes a steady hand) . In terms of yield, II usually get abut 100,000 cells per mouse brain. Mouse/rat age will also affect your numbers (younger mice will have higher yields). I haven't tried this procedure with rats, but yields should be much higher. Check out this pub for more details on expected yields: Article Isolation and Culture of Microglia

Thank you both Aleksandr Zakirov and Elliot J. Glotfelty :)

I've never used octoMACS, I've used the dounce which can be a little tricky but I've been getting good results. I usually get around 80-100k of microglia per mouse brain (complete) or 10-15k per two hippocampi. I'm not sure about how many cells per gram, sorry, I haven't calculated that andddd I've never worked with rats.

Good luck

Stay safe guys