BrdU-NeuN immunofluorescence

Using combined BrdU and NeuN labeling, we observed new cells with neuronal characteristics in the neocortex of adult macaques and rats. Although some studies have corroborated these findings, others have been unable to replicate the results. In the case of studies that produced positive evidence, the number of BrdU-NeuN positive cells reported in the neocortex is proportionately small relative to the dentate gyrus. When searching for evidence of new neurons in brain regions where their occurrence is likely to be relatively rare, sensitive methods must be used to minimize the likelihood of producing false negative results. Protocols for labeling tissue with BrdU and NeuN vary widely, and different approaches to this method may explain some of these discrepant results. Below we compare the protocols used to label brain tissue with BrdU and NeuN and provide quantitative analysis of the signal to noise ratio for the two commonly used methods. We find strong evidence that the protocol used to stain tissue for BrdU and NeuN in the studies which report no neocortical neurogenesis in the adult brain is associated with significantly diminished NeuN signal in the neocortex of rats and macaques.

Available BrdU antibodies recognize BrdU in single stranded DNA. Thus, the tissue must be subjected to denaturation steps before antibody-antigen binding can occur. These treatments vary greatly from protocol to protocol (see Wojtowicz, Nature Protocols) – denaturation steps are damaging to tissue and must be optimized to maximize histological quality. These steps range from a single 30 min pretreatment at room temperature in HCl to multiple steps in addition to HCl, such as incubation in heated formamide (2 hr at 65ºC) and boric acid (pH 8.5). In addition, the latter protocol is often combined with tissue slicing techniques which involve freezing.

To examine the possibility that different protocols produce variations in histological quality, we processed tissue from 4 rat and 3 monkey brains using the following two protocols, described below in detail: A) the protocol we have used to identify BrdU-NeuN positive cells in neocortex. This method involves minimal pretreatment and no freezing of the tissue; and B) the protocol used in studies which report no BrdU-NeuN positive cells in the adult neocortex (Kornack and Rakic, 2001; Bhadwarj et al., 2006 – each cite Kuhn et al., 1996 or a related paper from the same group for methodological details). The latter protocol differs from our method in that it involves freezing the brain tissue, as well as treating it with additional chemicals, including 50% formamide, which has solvent-like properties, and alkaline borate buffer.

For quantitative analysis, rat tissue was stained using protocol A and protocol B, as well as protocol B without freezing the tissue (the effects of freezing tissue on staining using Protocol A were not tested).  After reactions were complete, scans from the anterior neocortex of the rat were made using a Zeiss Axiovert confocal microscope with an Argon laser and 510LSM software.  Data analysis proceeded as follows:  One µm-thick optical sections were scanned from the anterior neocortex; 10 frames (115.16µm²) were obtained per animal using each protocol, holding the scanning parameters constant. Outlines of all NeuN stained cell bodies in the scans were traced using the measurement tool in the LSM software and the mean intensity for each cell body was determined (0-255 grayscale). The outline was then moved to an adjacent region, which did not contain any NeuN positive cell bodies, and a background optical intensity measure of the same size area was recorded. Every labeled cell was analyzed in this manner. The signal was defined as the mean intensity value for the cell bodies. The background was defined as the mean intensity value for the comparably sized regions containing no stained cell bodies. Signal to noise ratio was calculated according to the following formula:

(Signal – Background) / SD Background

We found that tissue stained for NeuN using Protocol A had a substantially higher signal to noise ratio than tissue stained for NeuN using protocol B. Freezing the tissue did not appear to have any measurable effect on NeuN staining in tissue stained with Protocol B.   Data were analyzed using one-way analysis of variance (F(2, 590) = 352, p < 0.0001; Tukey post hoc, p < 0.001 for both Protocol A versus B comparisons).

Examples of images of NeuN staining taken from rat tissue used in this analysis can be previewed as jpegs under Images . Additional examples of monkey tissue stained with these protocols are also provided.

Methods

PROTOCOL A – for 40µm-thick sections from non-frozen brain tissue.

Day 1

1). Incubate in 2N HCl:TBS for 30min

2). Incubate in rat anti-BrdU:TBS (1:200; Accurate) with mouse anti-NeuN (1:500; Chemicon) and 25 µl/ml of 10% Tween-20 overnight at 4ºC

Note: all reactions take place at room temperature unless otherwise noted; each incubation is preceded by 3 rinses in TBS (3-5 min each).

Day 2

1). Incubate in biotinylated anti-rat:TBS (1:250; Vector) for 90 min

2). Incubate in Streptavidin Alexa 568:TBS (1:1000; Molecular Probes) for 1 hr in the dark

3). Incubate in secondary goat anti-mouse Alexa 488:TBS (1:500; Molecular Probes) for 30 min in the dark

4). Rinse, mount, dry in the dark, coverslip under 75% Glycerol:TBS

PROTOCOL B – for frozen brain tissue (first cryoprotected in 30% sucrose).

Day 1

1). Incubate in 50% formamide/2xSSC (0.3M NaCl, 0.03M sodium citrate) for 2 hr at 65ºC

2). Rinse in 2xSSC for 5 min

3). Incubate in 2N HCl at 37ºC for 30 min

4). Rinse in 0.1M boric acid (pH 8.5) for 10 min

5). Rinse 3x in TBS

6). Incubate in TBS/0.25% Triton X-100/3% normal horse serum (TBS-TS) for 30 min

7). Incubate in rat anti-BrdU:TBS-TS (1:200; Accurate) with mouse anti-NeuN (1:500; Chemicon) overnight at 4ºC

Day 2

Same as Protocol A