Regional density of glial cells in the rat corpus callosum

Axons and glial cells are the main components of white matter. The corpus callosum (CC) is the largest white matter tract in mammals; in rodents, 99% of the cells correspond to glia after postnatal day 5 (P5). The area of the CC varies through life and regional diff erences related to the number of axons have been previously described. Whether glial cell density varies accordingly is unknown; thus the aim of this study was to estimate glial cell density for the genu, body and splenium -the three main regions of CC-, of P6 and P30 rats. Here we report that the density of CC glial cells reduced by ~10% from P6 to P30. Even so, the density of astrocytes showed a slight increase (+6%), probably due to diff erentiation of glioblasts. Interestingly, glial cell density decreased for the genu (-21%) and the body (-13%), while for the splenium a minor increase (+5%) was observed. The astrocyte/glia ratio increased (from P6 to P30) for the genu (+27%), body (+17%) and splenium (+4%). Together, our results showed regional diff erences in glial cell density of the CC. Whether this pattern is modifi ed in some neuropathologies remains to be explored.


INTRODUCTION
The corpus callosum (CC) is the main white matter tract in the brain and is involved in interhemispheric communication.This commissure is composed of axons and glial cells, while neuronal somata make up less than 1%.After postnatal day 5 (P5), glial cells predominate, accounting for 99% of all cells (Sturrock, 1976;Aboitiz & Montiel, 2003).The size, myelination, and density of fi bers in callosal subregions are related to the brain regions they connect (Doron & Gazzaniga, 2008).Thus thin unmyelinated fi bers are most dense in the genu and splenium (anterior and posterior CC), where they connect association and prefrontal areas, while large diameter myelinated fibers are concentrated in the midbody of the CC, where they connect primary and secondary sensory and motor areas (Aboitiz, 1992).Interestingly, electron microscopic analysis revealed that the number of callosal axons increases from 4.4 million at birth to 11.4 million at P5 and this number is maintained until the adult stage in rats (P60) (Gravel et al., 1990), although the number of axons of the visual cortex projecting through the splenium of the rat CC decreased by 15% from P15 to P60 (Kim & Juraska, 1997).On the other hand, myelination in the CC starts with non-compacted cytoplasmic sheaths of myelin at P12, and each oligodendrocyte sheathes an average of 13 axons.Later, when the compact myelin is formed, a single oligodendrocyte sheathes, on average, 15 axons (P17) (Bjartmar et al., 1994).Considering all this information, CC regions might diff er in glial cell density.Interestingly, it is not known if the glial cell density increases or remains constant through development, or if the distribution of glial cells and the astrocyte/glial cell ratios are diff erent among diff erent regions of the CC.Therefore, the aim of this study was to estimate glial cell densities in the CC at two ages (P6 and P30) and to determine if there are diff erences among callosal regions.

Histology
Six female rats per age group (P6 and P30) were anesthetized with an overdose of sodium pentobarbital (100 mg/Kg) and then decapitated.Whole brains were removed rapidly and dissected on ice.The hemispheres were separated and fi xed with 4% paraformaldehyde in phosphate buff ered saline (PBS, 0.1 M, pH 7.4) for 24 h and then transferred to 30% sucrose in 0.1 M PBS at 4 ºC for cryoprotection.Sagittal sections (30 μm; two consecutive from the right hemisphere and one the from left hemisphere = 3 slides per animal) were obtained using a freezing cryostat and collected on gelatinized slides for staining with propidium red to estimate glial cell number (Fig. 1 and 2).

Immunohistofl uorescence
Cryosections were blocked in PBS containing 1% bovine serum albumin, 1% goat serum, and 0.1% Triton X-100 for 1 h and labeled overnight at 4 ºC with a 1:500 dilution of antiglial fi brillary acidic protein (monoclonal mouse anti-GFAP, Milipore, USA).After incubation with the primary antibody, samples were rinsed 3 times in PBS and labeled for 1 h at room temperature with a 1:500 dilution of Alexa 488 goat antimouse antibody (Invitrogen, USA).Finally, slides with sagittal sections containing the CC were washed 3 times with PBS, stained with propidium red and mounted with Vectashield (Vector laboratories, Inc, USA).Slides were examined with an epifl uorescence microscope (Olympus BX60) coupled to a digital camera (Olympus DP70).Image-Pro MC for Windows was used for estimation of glial cell density.A Zeiss LSM 510 Meta confocal microscope (Zeiss, Göttingen, Germany) was used for confocal images, with wavelength 488 nm for excitation of Alexa 488 and propidium red.The emission wavelengths were 519 nm and 610 nm respectively.The z-stack images (4 to 5 consecutive confocal sections of 512 x 512) were obtained every 5 μm with stack size of X: 450 μm and Y: 450 μm and processed in Aim Image Examiner.

Glial cell counting
The population of cells immunostained by the GFAP antibody was contrasted with the number of propidium red-labeled nuclei for each slide and each CC region.Data analysis (counting) was performed using Image J software 1.41 (NIH, National Institute of Health, Bethesda, Maryland, USA).
A test square grid of 0.01 mm 2 in each section was used to estimate the number of propidium red-stained cell nuclei equivalent to the total number of glial cells including GFAP + cells corresponding to astrocytes (only process-bearing cells showing their soma in the plane of the section were counted).Six randomly chosen fi elds (as indicated in Figs.1D and 2D) were counted by progressive displacement of the test square grid.The astrocyte / glial cell ratio results from the estimation of the GFAP+ cells divided by the total number of glial cells labeled with propidium red.Data are presented as the mean ± standard error of the mean (S.E.M.) and nonparametric statistics were performed with the Mann-Whitnney-U test; p < 0.05 was considered signifi cant.

RESULTS
The pool of glial cells is reduced from P6 to P30 Our estimations of glial cell density began at P6 because the nuclear counts could be related to the estimation of glial cell number (99% after P5; Sturrock, 1976); P30 (considered as an adult stage) was selected for comparison.We found that the number of glial cells decreased from 1880 ± 45 / mm 2 (P6) to 1672 ± 27 / mm 2 (P30) (p = 0.00012, Fig. 3).On the other hand, the number of astrocytes (GFAP + cells) showed a + 7% increase, from 135 ± 1.4 / mm 2 at P6 to 144 ± 1.59 / mm 2 at P30 (p < 0.0001; Fig. 3).These results indicate that the pool of glial cells of CC is reduced and that a fraction of glioblasts diff erentiated into astrocytes between P6 and P30.

Glial cells are more abundant in the genu
The size, myelination, and density of fibers in callosal subregions are related to the brain regions they connect (Doron & Gazzaniga, 2008).We estimated glial cell density to be greater in the genu at P6 (2319 ± 32 / mm 2 ) than in the body or splenium (1616 ± 14 / mm 2 and 1704 ± 14 / mm 2 , respectively; p < 0.0003) (Fig. 4A).A similar pattern was observed at P30, but here glial cell density of the splenium was closer to that of the genu; glial cell density in the genu was 1824 ± 8 / mm 2 , whereas for the body and splenium there were 1405 ± 9 and 1788 ± 13 cells / mm 2 , respectively (p < 0.001) (Fig. 4B).Therefore, we found a higher density of glial cells in the genu than in the body or splenium, at both P6 and P30.
We then estimated the astrocyte / glia ratio at P6 and P30, based on GFAP immunoreactivity and nuclear staining.We consider that this ratio is a more accurate estimation because it considers the glial cell density, which varies among CC regions.Our estimations showed that the astrocyte / glia ratio increased by + 27% for the genu (from 0.0637 ± 0.0014 to 0.0858 ± 0.0008; p<0.0001), +17% for the body (from 0.0794 ± 0.0008 to 0.0951 ± 0.0014; p<0.0001) and +4% for the splenium (from 0.0766 ± 0.0008 to 0.0793 ± 0.0008; p=0.066) (Fig. 5).Overall, these results indicate that the density of astrocytes increases during postnatal development, because a fraction of glioblasts diff erentiate to astrocytes before P30.

DISCUSSION
The development of the CC is not homogenous through its three main regions; the genu, body, and splenium.Diff erent changes in callosal thickness occur due to axonal myelination, redirection, and pruning (LaMantia & Rakic, 1990;Luders et al., 2010).Nevertheless, glial cell density has not been explored among CC regions during postnatal development.Here, we report that the pool of glial cells decreased from P6 to P30.During this period of development, glioblasts (the dominant cell type in CC at P5) diff erentiate mainly to oligodendrocytes and astrocytes (Mori & Leblond, 1970;Sturrock, 1976).Surprisingly, when the glial cell density was estimated in the three regions of the rat CC, we found that glial cell density was greater in the genu at both P6 and P30 than in the body or the splenium.The functional signifi cance of the greater density of glial cells in the genu remains to be investigated.

Figure 1 .
Figure 1.The rat corpus callosum (CC) at P6. A, Sagittal sections of the rat brain were immunostained for GFAP (18 slides from 6 brains; 3 per brain) and propidium red.The CC was dissected in 3 regions: genu (G), body (B) and splenium (S) for estimation of glial cell density (n = 18 per region).Other brain regions are indicated for anatomical reference: hippocampal CA1 region (CA1), caudate-putamen (CPu), hippocampal dentate gyrus (DG), lateral ventricle (LV).Arrows showing anatomical orientation: dorsal (D), ventral (V), anterior (A) and posterior (P).B, Zoom of the square in A to show astrocytes from CC, notice the radial projections of astrocytes.C, Zoom of the square in B to highlight radial glia (arrow).Scale bars 200 µm (A), 80 µm (B), 20 µm (C).D, The representative sites ( * ) used for estimation of glial cell density are highlighted in a diagram of CC.Three slides containing sagittal sections of CC where used for each brain and each region of CC included 2 selected sites per slide for estimations (6 sites per region, per rat brain).

Figure 2 .
Figure 2. The rat corpus callosum (CC) at P30. A, Sagittal sections of the rat brain were immunostained for GFAP and propidium red (18 slides from 6 brains; 3 per brain).The CC was dissected in 3 regions: genu (G), body (B) and splenium (S) for estimation of glial cell density (n = 18 per region).Other brain regions are indicated for anatomical reference: hippocampal CA1 region (CA1), caudate-putamen (CPu), hippocampal dentate gyrus (DG), lateral ventricle (LV).Arrows showing anatomical orientation: dorsal (D), ventral (V), anterior (A) and posterior (P).B, Zoom of the square in A to show fi brous astrocytes from CC, notice that radial projections of astrocytes observed at P6 are absent at P30. C, Zoom of the square in B to highlight fi brous astrocytes (arrow).Scale bars 250 µm (A), 80 µm (B), 20 µm (C).D, The representative sites ( * ) used for estimation of glial cell density are highlighted in a diagram of CC.Three slides containing sagittal sections of CC where used for each brain and each region of CC included 2 selected sites per slide for estimations (6 sites per region, per rat brain).