Quantification of brain-derived neurotrophic factor (BDNF)

Methods similar to GR immunohistochemistry were used for BDNF visualization. The primary antibody consisted of mouse anti-BDNF antibody (1:300, Developmental Studies Hybridoma Bank, University of Iowa). The secondary antibody was DyLight-488 horse anti-mouse IgG (1:1,000, Abcam).

35–40 confocal slices were collected per region of interest at 1 µm interval. Three slices (#6, #12, and #18) were chosen for the subsequent analysis. The fluorescent signal derived from BDNF was converted to an 8-bit image such that the intensity of the signal varied between 0 (no signal) to 255 (maximum). The number of pixels at each intensity was then measured. Out of all pixels with non-zero intensities, 99.2 ± 0.2% exhibited intensities lower than 150, demonstrating that the confocal images were not saturated. The normative histogram for each experimental group is depicted in Supplementary Fig. ². The relationship between intensity and the corresponding cumulative number of pixels was modeled using an exponential plateau function, Y = Ymax (1 − e^(−k×X)). Ymax in this equation estimates asymptote for the number of pixels with positive staining values. On the other hand, k, or rise constant, estimates relative preponderance of pixels with lower intensity values in the histogram.