I've had some difficulties using antibodies to BrdU, and the protocol necessitates injecting animals with BrdU for 3-5 days before perfusion. In my hands, an antibody to Ki-67 protein reliably identifies newborn neurons without these difficulties (see my abstract from the last Neuroscience meetings in November in New Orleans)

John K. Young

Department of Anatomy

Howard University, Washington DC

[email protected]

Hi Jaun. BrdU is very useful as you can (for example) label which cells were proliferating at a specific time, and then leave the animals to age and fate map the cells and add various analogs such as EdU to see how proliferation changes over time. Ki67 is also useful, but can only tell you about the proliferation at the specific point in time you harvested the tissue (and it has been suggested it is also expressed by quiescent cells not just proliferating cells- J. Cell. Physiol. 206:6240635,2006). This paper has good materials and methods and the group uses BrdU a lot to characterise the effects of irradiation in their model, I used to work downstairs from them. What is the aim of your experiment and do you need advice on using and detecting BrdU, or quantification of the immuno?

HI Bobbi Fleiss! Thank ypou for your useful answer. let me tell you i already have the immunofluorescences and histochemistries for brdu and other proliferative markers as ki-67. but i want to show my results as "number for BrdU+ cells/mm2 in dentate gyrus" the problem is i dont have a standar protocol to quantify this result. do you know the general ideas to do it? thank you! :)

Hi Juan, I would refer to the paper that I attached above. It is critical to be able to identify the total area of the DG, which you can do by taking a low power image and then tracing around with a program such as Image J ( and calculating together with info on scale, taken by photographing a haemocytometer). Then, if you count every positive cell in the DG (from higher mag non-overlapping images) you will have the total BrDu+ in the DG. I do NOT suggest that you figure out the mm2 of a single image and try and do part of the DG- it is small, and irregular and this technique introduces error. Ideally, you would count the numbers in serial sections as per the paper, do you have serial sections, or consecutive ?

Hi Juan, I agree with Bobbi's first post. Calculating the number of cells (or density of cells) on a single image does not return a result that is necessarily representative of the entire region of the DG that stretches across many sections. Stereology (utilized in the paper Bobbi attached) employs systematic random sampling through the entire region of interest and easy to follow counting rules to ensure that the number you obtained is unbiased and precise. [I'm giving a webinar on Thursday, March 28th on stereology concepts and probe selection, if you'd like to learn more: http://www.mbfbioscience.com/upcoming-webinars ]

Counting cell profiles from 2D images (via ImageJ or manually) leads to bias because you are not differentiating between whole cells and parts of cells (cell profiles). Stereologic counting ensures that you have the opportunity to count each cell once and only once. Have you already prepared your tissue?

use Image J to count the BrdU positive cells by fixing threshold ,, another alternate is manually count number by using adobe photoshop CS6.

I have met no uniform standard protocol.

Use DAB or another chromogen to obtain a long-living preparation, then count all BrdU+ cells throughout your preparation within DG borders manually. Use serial sections (every 10th, e.g.). Since proliferating cells are pretty rare, you should analyse at least 6-8 sections for each animal (bilaterally) or even more (if you count only in one hemisphere, e.g. if you compare affected hemispere with an intact one). Count only the cells with visible nuclear border, it will allow you to avoid problems with counting the artifacts as positive cells. You can also try to count the cells at different stages of chromatin condensation differently (if you have such an interest): in most cases, at least two morphological classes can be clearly distinguished - cells with homogenous nuclear stain (in DAB method they really look like potatoes) and cells with lucid nucleus (though with visible nuclear border) with very intensively stained spots or punctae within it. Then take a low-magnification photo of the each DG section (where you have counted BrdU+ cells) and use ImageJ or any other available program to estimate the area of DG. If you use thick sections (40-50 um) you shoud also count through the whole depth of the section, but skip the first (or the last - as you wish) focal plane in order to avoid an over-estimation (in opposite case, you can twice count the same cell which was simply cut by the microtome and got to two adjacent sections).

Anyway, stereology is a "gold standard", but, IMHO, the absence of effect can be sometimes demonstrated even without such an effort- and time-consuming approach)