Expression of the C99 720-670 biosensor in the living mouse brain

(A) A schematic presentation of the C99 720-670 biosensor. (B) A cranial window was implanted on the top of the brain for image acquisition. (C) Extensive expression of the C99 720-670 biosensor in the somatosensory cortex was verified by confocal microscopy in vivo. Scale bar: 50 μm. (D) A high magnification image corresponding to the square in Fig. 1C. Scale bar: 5 μm.

Z-section images of the brain expressing the C99 720-670 biosensor.

The C99 720-670 biosensor fluorescence signal was detected at approximately 100 μm depth from the brain surface. Scale bar: 50 μm.

Imaging processing workflow and data analysis

(A) Before measuring the acceptor over donor emission ratios (i.e., 720/670 ratios) on a cell-by-cell basis, which reports γ-secretase activity in individual neurons, four-step image processing steps were applied: 1) background subtraction, 2) median filtering, 3) 3D iterative thresholding, and 4) morphological filtering. Scale bar: 50 μm. (B) To elucidate the relationship between γ-secretase activity and those in neighboring neurons, the distance between neuron and neuron was first determined, then identified each neuron’s five closest neurons, and calculated the average 720/670 ratio of the five neighboring neurons. The Pearson correlation coefficient between each neuron’s 720/670 ratio and the average ratio of the five neighboring neurons was calculated.

Identification and removal of auto fluorescent objects

(A) Our four-step segmentation approach still could not perfectly remove wrongly assigned ROIs, such as shown in the right panel (red arrowheads). However, these ROIs displayed higher 660-680 nm emissions and thus significantly lower 720/670 ratios (below 1.5). (B) A scatter plot supported our observation: two populations of ROIs displaying 720/670 ratios above 1.5 and the ratios below, later of which were excluded from our data analysis.

A potential “cell non-autonomous” regulation of γ-secretase in mouse brains

(A) A representative image showing the expressions of the C99 720-670 biosensor (Field of view) and a pseudo-color image corresponding 720/670 ratios (Pseudo-color FRET). Scale bar: 50 μm. (B-D) Scatter plots showing the 720/670 ratio in individual neurons (X-axis) and the average Mean of the 720/670 ratio in five neighboring neurons in three independent mice. The number of neurons, correlation coefficient (r), and p-value are shown. Pearson correlation coefficient. *** p < 0.001

Validation #1 - A potential “cell non-autonomous” regulation of γ-secretase in mouse brains

(A) The two or ten closest neurons were identified, and the average 720/670 ratio of the two or ten neighboring neurons was calculated and plotted. (B) We verified a significant positive correlation between the 720/670 ratio and the average ratio of two and (C) ten closest neurons. The number of neurons, correlation coefficient (r), and p-value are shown. Pearson correlation coefficient. *** p < 0.001

Validation #2 - A potential “cell non-autonomous” regulation of γ-secretase in mouse brains

(A) The average 720/670 ratio of neurons within a 20 μm radius was calculated and plotted. (B-D) There was a significant positive correlation between the 720/670 ratio and the average ratio of neurons within a 20 μm radius in three independent mice. The number of neurons, correlation coefficient (r), and p-value are shown. Pearson correlation coefficient. *** p < 0.001

γ-Secretase inhibition cancels the “cell non-autonomous” regulation

(A) The dose-dependent accumulation of endogenous APP C-terminus fragments (APP-CTFs) by subcutaneous administration of DAPT evidences the inhibition of γ-secretase activity in mouse brains. (B and C) Neither significant correlation between the 720/670 ratio and the average ratio of the five closest neurons, nor (D and E) the average ratio of neurons within a 20 μm radius. The number of neurons, correlation coefficient (r), and p-value are shown. Pearson correlation coefficient. n.s. not significant