All experiments were sanctioned by the Home Office Animal Procedures Inspectorate. The Wistar rat colony was started from breeding pairs purchased from Charles River (UK). These were maintained as non-SPF on a 12:12 h light dark cycle beginning at 8.00 am. They were kept at a constant temperature with unrestricted access to food and water. The colony was maintained through random pair mating. Timed mating was carried out by placing a male and female together and checking for the presence of a vaginal plug every 4 h. The presence of a plug was taken to indicate successful mating and the time taken as gestational day zero, E0. Fetuses from timed pregnancies of Wistar rats were harvested after euthanasia of the mother, by intraperitoneal injection of an overdose of anaesthetic (sodium pentobarbitone), at gestation days, E17-E20 and brains were removed and processed as described below.

For histological analysis, brains were fixed in 4% paraformaldehyde for a minimum of 2 h, cryopreserved in 20% sucrose, and frozen in iso-pentane cooled with dry ice. Serial coronal sections, 20μm-thick, were obtained with a Leica CM cryostat and stained with haematoxylin and eosin.

To identify proliferating cells, 5-bromo-2'-deoxyuridine (BrdU, Sigma, Poole, UK) was administered to pregnant dams by intraperitoneal injection at a dosage of 60 mg/kg at E17. Fetuses were then recovered at E18-E21, after BrdU injection of their dams, and the heads fixed in ice cold 4% paraformaldehyde in phosphate buffered saline (PBS) at pH 7.3. Fixation was carried out for a minimum of 2 h and usually overnight. Coronal sections were incubated in 10% normal serum in PBS for 2 h prior to flooding with primary antibody, a mouse anti-BrdU antibody (Nova Castro Laboratories, Newcastle UK), diluted 1:1000 in 10% normal serum in PBS, overnight at 4°C. Sections were then washed in PBS (3 × 1 h) and flooded with secondary antibody, universal biotinylated horse anti-rabbit-anti-mouse secondary antibody (Vector Laboratories Ltd, Peterborough UK), at a dilution of 1:200 in 10% normal serum in PBS for 2–4 h. Visualisation was by avidin-biotin-peroxidase diaminobenzidine staining (Vector Elite ABC Kits). All stained sections were mounted in glycerin-albumin and images were captured using a Coolsnap digital camera (Princeton Instruments) fitted to a Leica DMLB microscope. Image-capture and analysis was carried out using Metaview software (Universal Imaging Corporation). Different regions of the coronal cortical section, the germinal matrix, the intermediate zone and the cortical plate, were identified and the number of positive cells in each zone was counted on a photomontages constructed from micrographs taken at x400 magnification across the cortex. These gave a strip of cortex 400μm wide from the ventricular ependymal layer to the surface of the cortex next to the pia.

In all experiments, a minimum of three measurements (BrdU positive cell counts in the region of the cortex under analysis) was taken for each individual to give an average for each zone for that individual. These averages from at least four separate fetuses from separate litters were pooled to give a mean ± SEM.

Cortical hemispheres from the brains of embryonic rats at E17-E20 were used to prepare primary cell cultures. 4–6 hemispheres were cleared of meninges and incubated in trypsin-EDTA solution (0.25%) at 37°C for 20 min. The solution was replaced with Neurobasal medium (Invitrogen, Paisley, UK) and the cells dissociated by repetitive pipetting through tips of decreasing bore size. The suspension was centrifuged at 1700 rpm for 5 min and the supernatant replaced with fresh media. Further pipetting was performed to break up clusters of cells before a sample was taken for viability staining with trypan blue (0.4%) and cell counting. The dissociated cells were plated in poly-D-lysine (0.05 mg/ml) coated 96 well plates at a density of 1 × 10⁵ cells/ml in Neurobasal medium which preferentially supports progenitor and neuronal cell types (not glial cells), containing B27 supplement (Invitrogen), 2 mM glutamine and penicillin-streptomycin (0.1 mg/ml). At 24 h the media was replaced with 50 μl fresh Neurobasal medium (media only) or, alternatively, cells were incubated in 50 μl 100% CSF from different age sources (as indicated in the figure legends) and the cultures were maintained at 37°C in a 5 % CO₂ atmosphere for a further 48 h.

Cell proliferation was determined using the luminescence-based Lumitech Vialight (LumiTech Ltd, Nottingham UK) high sensitivity cell proliferation and cytotoxicity assay kit (as per manufacturers instructions), which is based on the measurement of concentrations of ATP in the culture. Analysis of the luminescence reading obtained against known viable cell numbers for cortical cells from fetuses showed that the assay is linear for cell numbers up to 1 million cells/well. This assay can detect as few as 100 viable cells in the culture and thus is highly sensitive in comparison with other methods for determining the number of cells within a culture. It measures cells that have actually completed division, unlike alternative methods, which measure incorporation of nucleotides into DNA. Measurements were made on a Multilabel Counter (Wallac Victor2 1420, PerkinElmer, UK). Luminescence measurements were taken from control wells plated at the initiation of the culture and assayed immediately. These give control day zero luminescence readings. Results are expressed as fold of this day zero measurement, mean +/- SEM of at least three experiments involving at least three litters, each performed in triplicate.

In experiments to determine the effect of CSF on cortical cell cultures, CSF was removed by tapping the cisterna magna of Wistar rat fetuses immediately after extraction from the dams as described above. This CSF was added to the cultures as shown. The cisterna magna is an excellent site to collect uncontaminated CSF since the fluid space is large and access is through a thin membrane once the overlying musculature has been dissected. Occasional contamination with blood was caused by severing a blood vessel within the cisternal cavity and these samples were discarded. All samples were collected into sterile Eppendorf tubes and routinely centrifuged twice at 14,000 rpm to remove any cells or debris from the fluid, which was decanted into another sterile tube. Samples were frozen on dry ice and stored at -80°C until use. Between 5 and 50μl of CSF was collected from each fetus using this method and CSF from individual litters was pooled for the experiments.

The rat cortex (or more accurately the cortical plate) appears over days E17-E21. In order to examine this development in more detail, we analysed photomicrographs of sections through the cortex across this time scale. There was some increase in total thickness from E17 to E20 and a rapid appearance of cortical plate over the first 48 h (Fig. 1), demonstrating that this is a critical period in the generation and migration of the neurones of the cortex from the germinal epithelium through the intermediate zone into the cortical plate. By E19 the cortical plate was well established and the thickness of the germinal epithelium appeared to decline as the mass of migrating neurones left the intermediate zone (Fig. 1).

In order to assess patterns of germinal epithelium cell proliferation and migration in vivo, BrdU was injected into pregnant dams at E17 to analyse cell generation and location over the subsequent 2–4 days. Cells generated on E17 were present in the germinal matrix and intermediate zone on E18, began to move into the cortical plate by E19, and were more concentrated in the cortical plate by E20 and E21 (Figs. 2A and 2B).

Cortical germinal epithelial (GE) cells are usually maintained in vitro in a specialised Neurobasal medium containing B27 supplement (Invitrogen, Paisley UK), which preferentially supports the growth of progenitor and neuronal cell types by enhanced proliferation and self-renewal 2521. Histological analysis as shown in Fig. 1, suggests that it is important to consider the exact age of the cells at extraction, since the apparent rate of proliferation changes over a single day, implying that the response profile of the primary cells derived from these time points would be different. The growth of E17 – E20 cells (the days known to be critical for cortical development in the embryonic rat) in Neurobasal media is compared in Fig. 3. There was a time window over days E18 and E19 of gestation where the cells exhibited a large significant, proliferative response (P < 0.001 compared to time zero sample as analysed by Students t test). In contrast on either side of this time window, on days E17 and E20, the proliferation of the cells under these conditions was minimal and only significant at E17 (P < 0.05).

The above experiments show that there was an age-related variation in the ability of primary cortical cells to proliferate and that this has at least two important components, an age-related proliferation potential presumably dictated by developmental genes, and their in vivo environment which provides the support required for growth, proliferation, differentiation and migration.

Based on previous work pointing to a role for CSF in cortical development, we examined whether the cultured cortical cells from different fetal stages could be maintained in CSF alone. Primary cortical cells were taken from E17 and E19 fetuses whilst CSF was collected from fetal brains aged E18 through to E21. The medium was completely replaced by CSF such that cells were incubated in 100% CSF. In all cases CSF was able to maintain the E19 cells in a viable state (Fig. 4). In fact not only was cell viability maintained, but cells also proliferated in the CSF to an extent comparable to, or, in excess of, that observed with normal Neurobasal medium (Fig. 4). The response of the E17 cells was significantly reduced compared to the E19 cells but even with these cells, the CSF was able to maintain cell viability over the 72 h period. The only exception to this was E21 CSF, which was unable to support the survival of E17 cells, a perhaps not unexpected result given the temporal separation between these two components of the developing cortex in vivo.

Together these data illustrate that CSF alone is able to function as a medium for survival and proliferation of developing cortical cells in vitro and that CSF composition changes with time as reflected in its changing ability to support viability and proliferation.