Geometric and network organization of visceral organ epithelium

Figures

• 

  FIGURE 1. Pectin biopolymer isolation of visceral epithelium (mesothelium). A similar isolation procedure was used to harvest epithelium from freshly euthanized mice. The visceral organ epithelia including lung (A), heart (B), cecum (C), and liver (D), were isolated and cell geometry assessed by machine learning (Liu et al., 2022). In polygonal cell geometries, the edges were defined as the junctions between 2 cells, and vertices marked points of contact among three or more cells. The number of sides between vertices determine the sidedness of the polygon. The polygon distribution in the epithelial tissues was color coded: 3-sided, orange; 4-sided, yellow; 5-sided, green; 6-sided, red; 7-sided blue; 8-sided, light blue; 9-sided, purple; and 10-sided, dark purple. Bar = 200 um.

• 

  FIGURE 2. Geometric features of non-proliferating visceral organ surface epithelium. (A) The stained and harvested epithelia were analyzed by a machine learning algorithm previously described (Liu et al., 2022). The polygon frequency distribution was compared in images obtained from lung, heart, liver and cecum samples. Similar polygon frequencies were observed with the heart demonstrating the greatest variability. (B) Morphometric analysis facilitated the calculation of cell surface area. The liver cell surface area was significantly greater than the pleural epithelial cells surface analyzed at residual volume (p < 0 .05). Note, the lung was uninflated during epithelial harvest. The box spans the interquartile range of the replicate samples with the median marked with an X and the whiskers defining the statistical data range.

• 

  FIGURE 3. Effect of lung expansion with inhalation. (A) The polygon frequency distribution was compared in images obtained from lung at RV and TLC. The polygon frequency distribution was not significantly different (p >0 .05) between inflated (TLC) and deflated (RV)lungs. (B) Inflation was associated with a significant increase in cell area (p <0 .05). The polygon distribution in the epithelial tissues was color coded: 4-sided, yellow; 5-sided, green; 6-sided, red; 7-sided blue; 8-sided, light blue. (C) Both inflated and deflated lungs were associated with wavy or interdigitated cell borders (colored dots). (D) The number of peaks (interdigitations) were more prevalent in the lung samples and increased linearly with increased side length: lung (TLC) R 2 = 0.687, lung (RV) R 2 = 0.671, heart R 2 = 0.542, cecum R 2 = 0.185 and liver R 2 = 0.052.

• 

  FIGURE 4. Characterization of cell boundary interdigitations using two methods. (A) The mean number of interdigitations per side was independently quantified by a minimum of two observers. (B) The redundancy index was defined as the length of the traced interdigitated boundary compared to the theoretical minimum distance between vertices. The redundancy index value was expressed as a ratio with 1.0 reflecting a straight border without interdigitation.In both graphs, the box spans the interquartile range with the median marked with an X in the whiskers defining the data range.

• 

  FIGURE 5. Characterization of network graphlets in epithelial tissues. An illustration of the analysis is shown for lung, cecum, heart, liver and liver. A mathematical tessellation in the form of hexagons was included for comparison. Borders around each region of interest reflects the requirement for a distance of four edges to quantify graphlet frequency. The visceral organs included (A) heart, (B) cecum, (C) lung, (D) heart, (E) hexagon, and (F) liver.

• 

  FIGURE 6. Scatter plot of the Graphlet degree Distribution agreement Distance (GDD) values for epithelia from the lung (A) and all four visceral organs (B). The plots demonstrate the GDD values (Epi-Hexagons, Epi-Random and Epi-Voronoi5) of individual biological tessellations: blue, lung RV; red, lung TLC; green, heart; purple, cecum and teal, liver. The plots displayed the CVTn scale from iteration 1 (black) to iteration 700 (light gray). The liver was the only statistically dissimilar epithelium (teal, p < 0 .05).