| Protocol used |
Materials
The rabbit anti-EphA4 antibody recognizing the 11 carboxy-terminal amino acids of chicken EphA4 was generously provided by Dr. Elena Pasquale (The Burnham Institute, La Jolla, CA). The production and specificity of the antibody were previously described (Soans et al., [1994]). The monoclonal anti-S100 antibody was purchased from Sigma (St. Louis, MO). The rat anti-N-CAM monoclonal antibody (isoclone 12F11) was obtained from BD PharMingen (San Diego, CA). Fluorescein isothiocyante (FITC)-conjugated donkey anti-rabbit, Cy5-conjugated donkey anti-mouse, and Cy5-conjugated goat anti-rabbit antibodies were purchased from Jackson ImmunoResearch (West Grove, PA). AlexaFluor 568 hydrazide and AlexaFluor 568- and 488-conjugated goat anti-rat and anti-mouse secondary antibodies (highly cross-absorbed) were obtained from Molecular Probes (Eugene, OR). Dilithium salt of Lucifer Yellow CH (LY) was purchased from Calbiochem (La Jolla, CA).
Intracellular labeling of astrocytes with fluorescent dyes
Intracellular injection of astrocytes in lightly fixed tissue slices was performed as previously described, with some modifications (Buhl et al., [1990]). Male Sprague-Dawley rats, 1 month old, were deeply anesthetized with Nembutal (10 mg/100 g body weight). The animals were transcardially perfused with oxygenated Ringer's solution at 37¿C (0.79% NaCl, 0.038% KCl, 0.02% MgCl2 ¿ 6H2O, 0.018% Na2HPO4, 0.125% NaHCO3, 0.03% CaCl2 ¿ 2H2O, 0.2% dextrose, 0.02% xylocaine), followed by 4% paraformaldehyde (PFA) in PBS (pH 7.4, 37¿C) for 8-10 minutes. The brain was placed in ice-cold PBS and cut into coronal slices with a vibratome at a thickness of 100 m. The slices were stored in PBS at 4¿C until used.
The slices were placed under a 60¿ water objective (NA 1.4) and observed with an Olympus BX50WI microscope using infrared-DIC optics (Olympus, Melville, NY). Astrocytes in the upper blade of the dentate gyrus were identified by the shape and size of their somata. Glass micropipettes (OD 1.00 mm, ID 0.58 mm; resistances 100-400 M) were pulled on a vertical puller (David Kopf Instruments, Tujunga, CA) and backfilled with either 5% aqueous LY or 10 mM AlexaFluor 568 in 200 mM KCl. Astrocytes were impaled and iontophoretically injected with dye using 1-second pulses of negative current (0.5 Hz) for 1-2 minutes. After several cells were filled, the slices were placed in ice-cold 4% PFA for at least 1 hour. The slices were then ready to be immunolabeled.
Immunohistochemistry
For S100 double-labeling with EphA4 or N-CAM, a 1-month-old male Sprague-Dawley rat was perfused as described above, except that the 4% PFA was perfused for 20 minutes. Vibratome slices were cut coronally at a thickness of 75 m. These slices and slices containing dye-filled astrocytes were immunolabeled as described below, with all steps performed at 4¿C.
Slices were washed three times for 10 minutes each in PBS. The slices were incubated for 1 hour in blocking solution (PBS containing 3% normal goat serum, 1% cold water fish gelatin, 0.25% Triton X-100). Slices containing LY-filled astrocytes were then placed in working buffer (WB; PBS containing 0.3% normal goat serum, 0.1% cold water fish gelatin, 0.125% Triton X-100) containing either 7 g/ml EphA4 or 1:200 N-CAM antibody for 48 hours. Slices used in the somata distribution experiment were placed in WB containing 7 g/ml EphA4 or 1:200 N-CAM and 1:200 anti-S100. The slices were washed three times in WB for 10 minutes each and then placed in WB containing secondary antibodies at a concentration of 1:100 for 24 hours. For studying astrocyte morphology near boundaries, astrocytes near the EphA4 boundary were filled with LY, and astrocytes near the N-CAM boundary were filled with AlexaFluor 568. EphA4 was subsequently detected using goat anti-rabbit Cy5, and N-CAM was detected using goat anti-rat AlexaFluor 488. In S100-labeled slices, N-CAM was detected with AlexaFluor488, EphA4 was detected with FITC, and S100 was labeled with either Alexa568 or Cy5, respectively. Slices were washed in PBS three times for 10 minutes each. Slices were coverslipped using Gelvatol (Harlow and Lane, [1988]) and allowed to set overnight at room temperature before they were examined. Donkey serum was used throughout the procedure for double-labeled specimens.
Image acquisition and analysis
Specimens were examined using a Radiance2000 laser scanning confocal system (Bio-Rad, Hercules, CA) attached to a Nikon E600FN microscope (Kanagawa, Japan). A 60¿ oil immersion (NA 1.4) objective was used to image LY-filled astrocytes, and a 40¿ oil immersion (NA 1.3) objective was used to image S100 double-labeled slices.
Image visualization and analysis were performed using the program Imaris 2.7 (Bitplane, Zurich, Switzerland). Baseline subtraction and linear contrast stretch functions were performed on volumes to enhance contrast. Final images were prepared using Adobe Photoshop 7.0 (San Jose, CA). Images of astrocytes near boundaries were constructed by combining an average intensity projection of either EphA4 or N-CAM labeling with a maximum intensity projection of the LY- or AlexaFluor 568-filled astrocyte.
Histograms and graphs were generated using KaleidoGraph (Synergy Software, Reading, PA). All results are provided as mean ¿ SEM. Linear weighted sum (LWS) equals P ¿ (count)dp. The degree of polarization (P) was calculated by measuring the maximum extent of processes from the center of the soma toward the pia and toward the stratum granulosum and then dividing the larger value by the smaller. Astrocytes with longer pia-directed processes were given a positive value, and astrocytes with longer stratum granulosum-directed processes were given a negative value. Astrocyte spatial arrangement was tested for nonuniform distribution by means of a bootstrap procedure (Romano, [1989]). Briefly, the calculated Cram¿r-von Mises (CvM) goodness-of-fit value for the empirical data was compared with the CvM value obtained from 10,000 pseudo-samples (each with a sample size equal to tested dataset), each randomly generated by drawing from the null (uniform) distribution. The resulting P value equals the number of pseudo-samples having CvM values greater than the CvM value of the actual sample, divided by 10,000.
|
