Spatio-temporal dynamics of neocortical presynaptic terminal development using multi-photon imaging of the corpus callosum in vivo

Within the developing central nervous system, the dynamics of synapse formation and elimination are insufficiently understood. It is ideal to study these processes in vivo, where neurons form synapses within appropriate behavioral and anatomical contexts. In vivo analysis is particularly important for long-range connections, since their development cannot be adequately studied in vitro. The corpus callosum (CC) represents a clinically-relevant long-range connection since several neurodevelopmental diseases involve CC defects. Here, we present a novel strategy for in vivo longitudinal and rapid time-lapse imaging of CC presynaptic terminal development. In postnatal mice, the time-course of CC presynaptic terminal formation and elimination was highly variable between axons or groups of axons. Young presynaptic terminals were remarkably dynamic – moving, dividing to generate more boutons, and merging to consolidate small terminals into large boutons. As synaptic networks matured, presynaptic mobility decreased. These rapid dynamics may be important for establishing initial synaptic contacts with postsynaptic partners, refining connectivity patterns or modifying synapse strength during development. Ultimately, this in vivo imaging approach will facilitate investigation of synapse development in other long-range connections and neurodevelopmental disease models.

: Synaptophysin puncta colocalize with synapsin in the contralateral cortex. Histogram of synapsin intensities at synaptophysin-tdTomato presynaptic puncta. Coronal slices were made from the brains of mice that had been imaged in vivo. Slices were immunolabeled with anti-synapsin antibodies, and individual optical sections were imaged and analyzed. Synaptophysin-tdTomato puncta were identified using an automated ImageJ macro, then intensities of synapsin immunolabeling within puncta were measured and compared to neighboring background fluorescence. Data represent the percentage of puncta at each synapsin intensity level. Red dashed line, background fluorescence intensity within the same images (859.4 +/-51.7). Green arrow, mean synapsin intensity at synaptophysin-tdTomato puncta (1350.0 +/-72.5). Intensities at synaptophysin-tdTomato were significantly higher than background (paired t-test: p = 1.187 X 10 -8 , t(19) = 9.5; n = 20 images).

P25 P27
Supplementary Figure S3: Daily longitudinal imaging of the same brain volume. The absence of background uorescence in blood vessels, combined with distance from the cortical surface, allowed recognition of the same brain volumes for imaging on consecutive days. Z-stacks of the same volume were collected and analyzed on each day. The images shown are maximum intensity Z-projections of the stacks. The same brain volume was imaged and analyzed from P13-27, and every other day of imaging is shown for 1 mouse. Dashed lines indicate the same blood vessels at each timepoint. Scale bar, 50µm. Probability density histogram of the apparent diameters of all puncta between 0.5-5µm for all age groups analyzed (P13-P21). The data were fitted with two Gaussians (blue) using an unbiased Gaussian mixture model (GMM). Fitting with two Gaussians maximized R 2 (0.9655) and adjusted R 2 (0.9559) and minimized SSE and RMSE, when compared to fits with one or three Gaussians (adjusted R 2 : 0.8754 and 0.9454 for one and three curves, respectively). (B) Histograms of puncta sizes for all ages combined, clustered by size. Small (grey) and large (white) size histograms were generated by subjecting the data to cluster analysis, based on the optimized GMM parameters from A. Based on the analyses in A and B, a diameter of 2.5µm (red dot) was chosen as the cut-off for segregation into large and small puncta. (C) Probability histograms of puncta sizes for three age groups: P13-15, P16-18 and P19-21. Data were clustered into small (grey) and large (white) groups based on the GMM parameters derived in A. Segregation of small and large puncta based on a 2.5µm diameter cut-off was reasonable for all ages analyzed. In addition, adjusted R 2 was high (P13-15: 0.9262; P16-18: 0.8719; P19-21: 0.8681) for fits of all age groups with two Gaussians. Figure S5. Three-dimensional representations of dividing and consolidating presynaptic puncta. X/Y, X/Z and Y/Z planes of the time lapse image sequences presented in Fig. 4A-B were derived from maximum intensity projections for the indicated times. X/Z and Y/Z projections were made for the areas marked by cross-hatches in the X/Y projections. (A) The Y/Z planes verify division of puncta (red arrows) observed in X/Y planes and in the time lapse sequence in Fig. 4A. (B) Similarly, X/Z planes verify the consolidation of puncta (red arrows) observed in the X/Y planes and the time lapse sequences in Fig. 4B. Time-lapse movie of synaptophysin-tdTomato (yellow) in a P14 mouse. Tracks of puncta movement are shown over a period of 7.5 minutes. Tracks are color-coded for time. Movie is 150x real time, with one frame obtained every 30 seconds.

Supplementary Movie 2: Example of division of synaptophysin puncta.
Time-lapse movie of synaptophysin-tdTomato (white) showing an example of one large punctum dividing into two puncta of unequal sizes in a P18 mouse. Total duration of imaging is 10 minutes.
Movie is 150x real time, with one frame obtained every 30 seconds.

Supplementary Movie 3: Example of consolidation of synaptophysin puncta.
Time-lapse movie of synaptophysin-tdTomato (white) showing an example of two small puncta consolidating into a single larger punctum that is stationary in a P18 mouse. Total duration of imaging is 10 minutes. Movie is 150x real time, with one frame obtained every 30 seconds.