In my opinion, it depends on the structure you analyze. I mean, if you are interested in the finest processes, you definitely should not use GFAP. But you don't have to use Lucifer Yellow, which is tedious and time consuming. If you have a GFAP-GFP mouse it is enough. Or if you want, you can obtain a good labeling by electroporating (in utero) a GFAP-GFP plasmid (included in a transposon). If you are interested, I have these plasmids (García-Marqués and López-Mascaraque., 2013).

In my opinion, it depends on the structure you analyze. I mean, if you are interested in the finest processes, you definitely should not use GFAP. But you don't have to use Lucifer Yellow, which is tedious and time consuming. If you have a GFAP-GFP mouse it is enough. Or if you want, you can obtain a good labeling by electroporating (in utero) a GFAP-GFP plasmid (included in a transposon). If you are interested, I have these plasmids (García-Marqués and López-Mascaraque., 2013).

Well, if you want to see something striking as astrogliosis it is quite a good and reliable marker. However, as Jorge said, it depends a lot on what you really want to see. Also, keep in mind that not all astrocytes are gfap positive.

I would also suggest immunolabeling for glutamine synthase, as a more sensitive marker for astrogliosis. It won't tell you anything about structure, really, but GS immunoreactivity decreases prior to increases in GFAP during astrogliosis.

Dear Jorge,

That is a very kind offer! Thank you. Certainly, I will interested in looking at the GFAP-GFP and we are just in the preliminary phases of generating our own colony here in Australia, from breeders we have obtained from Jackson. However, quite a bit of the work that we have done looking at changes in Astrocyte morphology has been in the rat, and this unfortunately makes genetic solutions a bit more difficult. I also agree with Alison that GS may be a solution but the GS labelling that I've seen (I've not attempted myself) looks quite diffuse. However, it could be worth a try to get a sense of the overall dimensions that the cell occupies. Do you have any experience with GS labelling with fixed rat sections?

I would also suggest the study of GFP under specific promoter for your analysis, easier and not time consuming. GFAP-GFP mouse are widely used, however as Marco noticed, you can miss up to 30% of astrocytes population... So be really careful to misleading conclusions. You can use also GLT1-GFP mouse and Aldh1l1-GFP mouse. The last one (I worked for several years with this strain, ordered at Gensat) is described as a pan-astrocytous marker, however i would not recommend the use of this one if you are working in the spinal cord, as ALDH1L1 expression decreases a lot with aging in control condition. But it's reliable to study astrogliosis and you will track more astrocytes (see Molecular Comparison of GLT1 and ALDH1L1 Astrocytes In Vivo in Astroglial Reporter Mice, Yang et al. 2010, Glia).

I would suggest visualizing a protein called EZRIN, that is typically localizes to astrocyte end-feet. It is an actin binding protein specific to astrocytes. This could be a better way to visualize overall astrocyte structure, even better, when combined with GFAP. Antibodies are available from many vendors.

Just to clarify my motivation for asking.

My group has just finished a study looking at how exposure to chronic stress alters the astrocytes cytoskeleton, using GFAP labelling. Essentially, we found that stress caused them to undergo significant structural atrophy. To do this we obtained high resolution (100x) stacks (usually 150 sections with a 0.1nm step) and then passed them through deconvolution and then into a package that allowed us to create a digital reconstruction of the cell. From the models we obtained various morphological metrics (volume, process length etc). When we submitted this paper the reviewers did suggest that we replicate these findings with cell filling to determine the extent to which cytoskeletal remodelling correlated with the cells overall volume. Fortunately, however, they eventually let me off the hook on these experiments but I thought it was a good idea and would very much like to find the answer.

Hi Bence,

Can you recommend a good antibody for IHC for EZRIN? It looks interesting.

GFAP immunohsitochemistry is good to show location of astrocyte cell bodies and general process abundance, bit it is true that you may not be able to see all the finest prodesses. However, if you have a really good antibody you can see many astrocyte processes.

I have used glutamine synthase, GFAP, and S100. It is true that GFAP does not detect the finest details particularly of the protoplasmic astrocytes. On the other hand Glutamine Synthase does not detect the most fibrous of the astrocytes. The bottom line is that a one-size fits all stain is not going to reveal every subclass of astrocyte.

So a few other options to consider along with several great ones suggested above.

If you are interested in the volume area that an astrocyte is taking up, you may want to look into the brainbow mice. Having individual astrocytes express different colors would allow you to make more precise measures of total volume of the astrocytes, and also do some more complex 3D reconstructions using Z-stacks of images.

Alternatively, you could infect cells with a virus expressing GFP with a membrane tag (such as Lck) to anchor the GFP to the membrane. By doing so, you will get very nice resolution of the smaller processes that GFAP staining will not get you. Also, by titering the virus you can isolated expression of single astrocyte, making resolving the boundaries and structure of a single astrocyte easier to do.

I have worked with both the Glt-1:GFP and GFAP-GFP mice before, and I find the Glt-1:GFP mice are better and do an excellent job of labeling the majority of astrocytes in most brain regions. The GFAP-GFP mice are good, but the Glt-1 tend to be brighter and in regions like the cortex do a better job of labeling in my opinion.

Combining any of the above with well done GFAP staining should give you a strong measure of GFAP cytoskeleton versus process volume measurements with 3D reconstruction.

GFAP expression will provide a general overview of what is happening. THe study you describe sounds fascinating, time consuming and, presumably is now completed. Using retrospective samples, the Ezrin and nestin staining suggested above could be quite informative, as would the addition of vinculin. If you are going to do more studies going forward, there are modified Golgi-Cox stains that pick up astrocytes (Torres-Platas 2011, for example) and allow for more detail than GFAP. The downside, is it also stains a subset of neurons. I will look forward to reading your paper, Rohan

Hi Andrew,

I'd be really interested if you could recommend any good antibodies for Ezrin and Nestin. I also thought that Nestin was more a developmental marker?

The one I used gives very nice and specific staining in rat and mouse free floating brain sections. It used be sold by Zymed, but it is now acquired by Invitrogen.

http://products.invitrogen.com/ivgn/product/357300?ICID=search-product#specifications

It is a Mouse anti-ezrin monoclonal Ab, Clone 3C12, Cat #35-7300.

This is not a direct answer to your question but ff you are interested in visualizing the astrocyte endfeet that are next to endothelial cells (blood vessels) by immunofluorescence, I suggest using antibodies against aquaporin 4. I have not tried antibodies against the Kir's, but this may be good as well.

Dear Reno,

Any suggestions about a good antibody for AQ4?

Best,

Rohan.

I used the Chemicon polyclonal catalog number ab3594

rabbit polyclonal AQP4 (Millipore AB2218) - I my experience this one is very efficient, and works well in IHC, IF and WB

mouse monoclonal AQP4 - (Abcam ab9512) - alos good in IF and WB (don't know why, but in IF stains AQP4 in vascular units much better than this one above)

I have found that my GFAP (goat anti-GFAP from abCAM) antibody was sensitive to the length of time I post-fixed the tissue in 4% formaldehyde after perfusion. If I only postfixed the tissue for one hour, the fine astrocyte processes stained with GFAP and I could see the "bushy" cell morphology of cortical astrocytes. If I postfixed for 24 hr, I was not able to see the fine processes.

If you are interested in astrocyte morphology that includes primery/secondary processes you can use GFAP, however GFAP will not visualize terciary/quartery (very fine) processes. There is a new antibody from Abcam coupled to A488, which is very efficient and together with short fixacton of slices for 30 min max, gives quite impressive astrocyte staining. But still, when you compare it with LY labeled astrocyte (labeled via patch pipette) it is clear that GFAP is not suitable for morphological evaluation of astrocytes.

Hi Miroslava,

That sounds interesting is that a new GFAP antibody? Or something else?

Also, do you mean that to get good results you need to do a PFA transcardial fixation and then only post fix for 30 minutes?

Hi Rohan, we use either 200um slices for astrocyte identification after patch clamp recordings, these slices are fixed only for 20-30 minutes to get decent GFAP staining. For normal immunohistochemistry, we do transcardial perfusion with physiological solution + Heparin followed by PFA/phosphate buffer. Further, the isolated brain segment is left for 3 hours in PFA/phosphate buffer and after that it is transfered to 10-30% sucrose for cryoprotection. The catalog number of new antibody is Ebioscience 53-9892. Sorry for misinforming you with Abcam.

An important caveat in the above discussion has to include that many of these techniques are not suitable for using in human samples - transcardial perfusion for example. It is to be assumed that animal studies are done to model what is happening in human disease. Primates, including humans have very different astrocytes from rodents; not least of which is that they are very much larger and more complex (see Obberheim 2009, Journal of Neuroscience, volume 29, pp3276-87, and more recently, Han 2013, Cell Stem Cell, Volume 7, pp342-53). While Golgi-Cox staining may be the ideal stain to use, it may not be possible to use in samples that are already paraffin embedded. As such, GFAP remains the best practical stain to use, and to answer the original question, changes in GFAP can be used as a good approximation for overall changes in the morphology of astrocytes