Statistics

Six to nine separate immunocytochemical experiments were run for each primary antibody. Individual experiments were composed of at least two or three tissue sections of each animal from each group. Ten to 14 fields were analyzed for each brain region. Interexperimental differences were not statistically significant. Values represent the mean ± SD of experiments performed for each marker and treatment. The statistical study was performed by applying a two-tailed Student's i-test to the results of the quantitative analysis and by assuming

Morphometric Study of Neuronal and Astroglial Cells 99 -<-

Fig. 1. Immunostaining for GFAP (upper row), Nf-200 (middle row), and MAP-2 (lower row) in the hippocampal CAi area of control (A, C, and E) and WIN-treated (B, D, and F) rats. In the WIN-treated animals, no changes in the morphology of GFAP-IR astrocytes or in the number or thickness of the cytoplasmic projections were observed. Note that in WIN-treated animals, Nf-200 expression was increased and cellular projections were thicker and presented an irregular morphology. Compared to control, WIN-treated animals showed changes in the morphology of MAP-2 immunoreactive dendrites. They appear as thicker, irregular, and waved neuronal processes. Primary magnification 400X.

Fig. 2, Immunostaining for Nf-200 in the corpus striatum of control (A) and WIN-treated (B) rats. Note that in WIN-treated animals, there was an increase in Nf-200 expression in the two striatal areas. In the striatal patches these appeared as thicker and more irregular axons, while in the matrix these were observed as a fine network of neuronal processes. Primary magnification 400X.

Fig. 2, Immunostaining for Nf-200 in the corpus striatum of control (A) and WIN-treated (B) rats. Note that in WIN-treated animals, there was an increase in Nf-200 expression in the two striatal areas. In the striatal patches these appeared as thicker and more irregular axons, while in the matrix these were observed as a fine network of neuronal processes. Primary magnification 400X.

that data were normally distributed. The equality of variance for control and treated values was analyzed by an F test. For the same group interanimal differences as well as interexperimental differences were not statistically significant. Statistical significance was set to p < 0.05.

4. Notes

1. Regarding the selection of the proper tissue-processing procedure, attention has to be focused on the preservation of immunoreactivity, as well as tissue structure. Therefore, compromise will be necessary in order to preserve both immunoreac-

Fig. 3. Relative area of Nf-200-IR neuronal processes in two striatal areas: the matrix (StrM) (left) and the patches (StrP) (middle), and in the hippocampal CAi area (CAi) (right). Data are expressed as area of Nf-200-IR neuronal processes per |im2 of tissue. Note the increased area of Nf-200-IR processes in the three studied areas of WIN-treated animals. **p < 0.001; *p < 0.05; after two-tailed Student's t-test. Bars represent mean ± SD.

Fig. 3. Relative area of Nf-200-IR neuronal processes in two striatal areas: the matrix (StrM) (left) and the patches (StrP) (middle), and in the hippocampal CAi area (CAi) (right). Data are expressed as area of Nf-200-IR neuronal processes per |im2 of tissue. Note the increased area of Nf-200-IR processes in the three studied areas of WIN-treated animals. **p < 0.001; *p < 0.05; after two-tailed Student's t-test. Bars represent mean ± SD.

Fig. 4. Relative area of MAP-2-IR neuronal processes in the hippocampal CAi area. Data are expressed as area of MAP-2-IR neuronal processes per |im2 of tissue. Note the increased area of MAP-2-IR processes in this area. ** p < 0.001; after two-tailed Student's i-test. Bars represent mean ± SD.

Fig. 5. Area of GFAP-IR astrocytes in striatum (left) and hippocampal CAi area (right). Data are expressed as area of GFAP-IR astrocytes in pm2. Note that there were no significant differences in the astrocytic cell area between control and WIN-treated rats after two-tailed Student's t-test. Bars represent mean ± SD.

Fig. 5. Area of GFAP-IR astrocytes in striatum (left) and hippocampal CAi area (right). Data are expressed as area of GFAP-IR astrocytes in pm2. Note that there were no significant differences in the astrocytic cell area between control and WIN-treated rats after two-tailed Student's t-test. Bars represent mean ± SD.

Fig. 6. Relative optical density (ROD) of GFAP-IR astrocytes in striatum (left) and in hippocampal CA1 area (right). Data are expressed as ROD units. Note that there were no significant differences in the ROD values of GFAP-IR astrocytes between control and WIN-treated rats after two-tailed Student's t-test. Bars represent mean ± SD.

Fig. 6. Relative optical density (ROD) of GFAP-IR astrocytes in striatum (left) and in hippocampal CA1 area (right). Data are expressed as ROD units. Note that there were no significant differences in the ROD values of GFAP-IR astrocytes between control and WIN-treated rats after two-tailed Student's t-test. Bars represent mean ± SD.

tivity and tissue structure. The use of buffered fixatives based on paraformaldehyde with or without picric acid or glutaraldehyde renders good results for that purpose. The duration of the fixation, however, is also an important variable, since some antigens cannot be demonstrated in overfixed tissues (23).

2. Factors affecting labeling: To quantify immunocytochemical data, it is important to understand the kinetics involved in the labeling reaction (24). Different variables can affect the labeling intensity in immunocytochemistry. These variables include the concentration and exposure time to the antibody and other reagents, including the chromogen. All these variables are maintained as constant as possible to ensure reproducibility. Some variables will affect the optical density measurements of an immunocytochemically labeled section. Section thickness should be uniform in order to allow comparison of antibody labeling of different sections. Antibody penetration should also be equivalent in those sections. Reagent incubation times and dilutions should be identical in all experiments where data will be compared. The pH and temperature of the solutions should also be kept constant.

3. Another important point is the antibody specificity. In fixed tissues antibodies may bind nonspecific antigens (i.e., other than the specific antigen they were developed to recognize). This unspecific reaction could be produced by the conformational alteration of antigens in the fixed tissue. An immunocytochemistry protocol should always include appropriate control for antibody staining (i.e., omitting primary antibodies, absorbing the antibody with the appropriate antigen, etc.) (see, for example, ref. 23).

4. In order to obtain representative and reproducible values, it is important to take into account some methodological procedures. All sections obtained from the different experimental and control groups must be processed in parallel at the same time, under the same experimental conditions. Theoretically, the measurement of the density of a single specimen should always be the same value. Unfortunately, density values are only replicable under well-controlled conditions of measurement. All measurements should be performed by at least two independent observers in blind conditions.

5. In this chapter three different ways to measure immunocytochemically labeled nervous tissue have been described: the relative area covered by neuronal processes (dendrites and fibers), cell body area, and relative optical density. The relative area of cellular processes is the most difficult to achieve, the major problem involves the segmentation step (see Subheading 3.5). However, it can be resolved if the segmentation threshold is maintained from one field to another. On the contrary, it is easier to obtain the cell body area since it is usually less difficult to segment. Regarding ROD measures, it is important to remember that the background of the immunolabeled tissue section can affect the optical density values. Thus it is better to obtain the background parameter from each section out of the labeled cells and subtract this value from each cell ROD before statistically processing the values.

6. In the work presented in this chapter, the VIDAS-KONTRON image analyzer, which was a predecessor of the Zeiss KS-300 and KS-400 models that became available for the newest Windows versions (www.zeiss.de), was used. Our laboratory has used a Windows-based version of the Optimas (www.optimas.com) with excellent results. An image processing and analysis program for Macintosh or PC has been developed by the Research Services Branch (RSB) of the National

Institute for Mental Health (NIMH), part of the NIH, that can be downloaded at www.rsb.info.nih.gov. Most image analyzers do basically the same operations described in this chapter, with obvious speed differences according to the software-hardware combinations used. In the web pages of each image analyzer are also the necessary software modules for the different frame grabbers present in the market as well as very useful macros (small routines that can be loaded into the image analyzer for automatic operations). The experience of many scientists in different fields is available in the web sites of each image analyzer.

The results presented here show some interesting image analysis parameters that can be used for the morphological study in the CNS after drugs or toxicants exposure (see, for example, refs. 3, 4, and 25). In our experience these parameters (area, relative area, ROD, etc.) were extremely useful for the study of the effects of canabimimetic drugs on the young rat's CNS.

Was this article helpful?

0 0

Post a comment