Figure 1.
Image acquisition and processing workflow.
(A) Image stacks of the muscle layer (of the mounted, submerged ureter) were acquired at increasing depth (z) from the adventitial to the proprial side, at proximal, middle and distal locations (i.e., three segments per ureter). (B) Given the curvature of the vessel, the region of interest (ROI) for quantitative analysis was adjusted to ensure reliable cell density estimation and to limit cross-talk of cells resident at other depths within the wall. (C) A stack of raw images, showing smooth muscle cell (SMC) nuclei stained with SYTO 13. (D) Raw images (panel C) were filtered using cellness filtering (step 1) to identify SMC nuclei. Subsequently, a ROI was applied and individual SMC angles (α) were determined. α was defined with reference to the longitudinal axis of the vessel (x-direction). (E) SMC angles were plotted as a function of depth to evaluate transmural changes in orientation, taking circularity of the data into account. (F) A kernel density estimation (KDE) plot was used to estimate the orientation distribution. On this KDE, octile lines were plotted to clarify changes in orientation dispersion with depth.
Figure 2.
Example image slices from a single image stack.
(A) Image slice towards the adventitial side, showing predominantly longitudinally oriented smooth muscle cells (SMCs). (B) Image slice towards the proprial side, showing a more disperse orientation. Scale bar: 50 µm. Arrows indicate relatively round cells which are excluded from further analysis.
Figure 3.
Smooth muscle cell (SMC) orientation in intact rat ureter is predominantly longitudinal.
Panel A shows the orientation distribution (3D plot) and the cell density (2D graph) as function of depth, averaged over 36 rat ureter segments. Normalized depth 0 corresponds to the adventitial side of the muscle layer and 1 to the proprial side. At the adventitial side (normalized depth 0 to 0.5) there is a high probability that the angle of the SMCs with respect to the longitudinal axis of the ureter is about 0°. Towards the proprial side the SMC orientation gradually disperses but remains centered around 0°, as further illustrated in panel B by the distributions at four distinct normalized depths (zn) as indicated. At these depths, standard deviations (σ) are given. In the distribution plot (A) and its top view (C) the curves delimit the octiles of the orientation distribution.
Figure 4.
Example orientation patterns of two ureter segments.
Panels A and B (view as defined in Figure 3C) show two examples of orientation patterns acquired in different ureter segments. The pattern in A remains longitudal from adventitial to proprial side, whereas the pattern in B disperses towards the proprial side.
Table 1.
Variability in dispersion among different stacks.
Figure 5.
Control sample: By orientation distinguishable smooth muscle layers in the small intestine.
In order to verify our image processing method, we applied the exact same preparation and staining method to a rat small intestine, an organ with clearly delineated longitudinal and circumferential smooth muscle layers [8], [10]. Panels A–C show orientation distributions analogous to the panels in Figure 4, at three different sites in the intestine. Panel D shows the average orientation distribution, calculated from panels A–C. All panels clearly show a transition from a superficial, longitudinal smooth muscle orientation to a circumferential orientation at the deep end. The patterns shown are all slightly shifted to the right by ∼10°, which is caused by the fact that the two pipettes used to mount the intestine were not perfectly aligned (i.e., microscopy images were rotated by ∼10°).