Whole-mount BrdU staining of proliferating cells by DNase treatment: application to postnatal mammalian retina

Cell proliferation plays an important role in embryonic development, tissue repair, tissue homeostasis, and various pathological processes. Several markers have been introduced over the years that allow for direct or indirect analysis of proliferation (1–4). Tritiated thymidine ([³H]-T) was first used by Taylor (1) to analyze DNA replication and later became a popular marker for the identification of dividing cells. [³H]-T incorporates into newly synthesized DNA during the S-phase of the cell cycle and, therefore, allows direct analysis of proliferation. However, technical difficulties associated with the detection of this thymidine analog led to the development of new nonra-dioactive nitrogen base analogs such as 5-bromo-2′-deoxyuridine (BrdU), which also incorporates into newly synthesized DNA during S-phase, but, unlike [³H]-T, BrdU can be detected by immunohistochemical techniques (4–8). The development of highly specific antibodies to BrdU (5,6,9,10) made it the most popular marker currently used to detect cell proliferation.

For proper detection, the labeled DNA must first be denatured. Most commonly used protocols for visualization of BrdU-labeled cells on tissue sections utilize incubation in 2 M HCl for that purpose (4). Hydrochloric acid causes denaturation, dissolution, deamidation, and degradation of the majority of proteins, which has a profound damaging effect on the specimen, resulting in poor morphology and often precluding co-staining for other antigens due to disruption of the epitopes for commonly used antibodies. To avoid problems associated with HCl treatment, the use of microwave irradiation, which removes proteins and allows for better physical access of the antibody to the labeled DNA, was suggested (8). However, microwave irradiation alone, without DNA denaturation, was found to result in a lower percentage of BrdU-positive cells and had to be combined with HCl treatment for quantitative assessment of proliferation (11). Gonchoroff et al. (6) demonstrated that DNase I pretreatment can be used to quantitatively expose BrdU-labeled epitopes in cell suspensions. We have also successfully used DNase I treatment to stain BrdU-positive nuclei on tissue sections (12).

Initially, mitotic cells in whole-mount preparations were recognized by the characteristic shape formed by the mitotic chromosomes stained with cresyl violet (13). [³H]-T was also historically used to visualize proliferating cells in retinal whole mounts (14). The only protocol currently used to visualize BrdU-labeled cells in whole mounts is a replica of the procedure used to visualize BrdU-labeled cells on tissue sections and utilizes HCl treatment to expose BrdU-labeled epitopes (15–17). The method has all the disadvantages associated with tissue exposure to HCl and is characterized by poor penetration of antibodies into the deep cell layers of the specimen, which precludes quantitative analysis of proliferation.

Here I describe a reliable and highly efficient procedure for the visualization of BrdU-labeled nuclei in whole mounts using DNase I treatment for exposing BrdU-labeled epitomes. We used this protocol to study cell proliferation in the postnatal primate retina.

Five 4-month-old green monkeys (Chlorocebus aethiops) received a single intravenous infusion of BrdU (10 mg/mL in 0.9% NaCl, pH 7.4; 125 mg/kg; Sigma, St. Louis, MO, USA) over a period of 1.5 h. Following the BrdU infusion, the monkeys were sacrificed and their eyes were enucleated. After eye enucleation, the retinas were separated from the pigment epithelium, fixed with 2% formaldehyde in 0.15 M Sörensen phosphate buffer overnight, and cut into four quadrants, which were stained as whole mounts (18).

Protocol 1 describes specific concentrations and the timing of each step of the procedure. The quadrants of the neural retina were sealed between two sheets of nylon mesh to stabilize the tissue and provide equal access of the reagents to all portions of the specimen. This needs to be done in phosphate-buffered saline (PBS) to avoid drying the tissue. Specimens were bleached with 6% H₂O₂ and incubated with 5 µg/mL proteinase K. Limited digest with proteinase K results in partial disruption of the limiting membranes surrounding the retina and facilitates penetration of the anti-BrdU antibodies and avidin-horse-radish-peroxidase (HRP) complex into the deep layers of the retina. Proteinase K was then deactivated by incubation with glycine and postfixing with 0.2% glutaraldehyde/4% formaldehyde in PBS.

Protocol 1. Whole-Mount Retina Staining for BrdU

To expose the BrdU-labeled DNA epitopes, we used a limited digest of the nuclear DNA with DNase I. The timing of DNase I treatment needs to be carefully optimized for different tissues because overdigestion or under-digestion leads to a reduced staining intensity. We found that 2 h incubation at 37°C was optimal for the retina. DNase I was subsequently inactivated by 10 min incubation at 70°C in 10 mM Tris-HCl, pH 7.5. This short incubation at 70°C does not result in protein denaturation and does not have a detrimental effect on epitope integrity. After overnight incubation in blocking solution, the retinas were incubated overnight with biotin-conjugated anti-BrdU mouse monoclonal antibody (Molecular Probes, Eugene, OR, USA). This antibody bound to BrdU-labeled DNA was detected using an immuno-peroxidase detection system (19), the VECTASTAIN® Elite ABC Reagent (Vector Laboratories, Burlingame, CA, USA), which allows significant signal amplification while maintaining a very low background. A variety of chromogens can be used to localize peroxidase activity in the tissue. We used the traditional substrate 3,3′-diami-nobenzidine tetrahydrochloride (DAB), which produces a brown precipitate. To further enhance the staining, we used CoCl₂, which modifies the precipitate and produces an intense dark-brown color. The tissue has to be preincubated with CoCl₂ prior to exposure to the DAB solution because CoCl₂ forms an insoluble precipitate in cacodylate buffer. The resulting dark-brown precipitate is insoluble in alcohol and clearing agents, allowing specimens to be permanently mounted in Permount (Fisher Scientific, Hampton, NH, USA).

We successfully used this protocol to study cell proliferation in the postnatal primate retina (Figure 1). A uniform staining was routinely obtained throughout the surface (Figure 1, A and B) and the thickness (Figure 1, C and D) of the retina, allowing quantitative analysis of proliferation in different regions. The procedure utilizes a limited digest with DNase I to expose BrdU-labeled epitopes, which, unlike HCl treatment, does not have detrimental effects on the epitope integrity and makes it very easy to combine BrdU staining with immuno-histochemical detection of any other antigen (Figure 1E). The retinal whole mount is approximately 250 µm thick. Therefore, this protocol can be easily optimized for studies of proliferation during early embryonic development as well as for the analysis of proliferation in other adult tissues, given that thick vibratome slices are used as a starting material. The only parameter that would need to be optimized is the duration of a DNase I digest, which can be easily done by performing a time-course experiment. Our protocol represents a valuable tool and a platform that can be further optimized for studies of proliferation in various biological systems.