Quantitative Analysis of Lymphatic Development

To quantify the proportion of SOX7‐positive cells in Fig 1A, SOX7⁺ cells were calculated manually using ImageJ and subsequently divided by ERG⁺ endothelial cells in the entire section. Fig 1D–J were quantified within 2,100 μm from both sides of the midline. LEC front distance to the midline (Fig 1D) was an average of 10 random measurements taken from the midline to the nearest leading lymphatic sprouts. For vessel width (Fig 1F), three locations within 200 μm from the sprouting end of each lymphatic leading vessel were measured. Vessel width was taken from the average of seven representative lymphatic leading vessels in each embryo. PROX1⁺ nuclei in Fig 1H were an average of PROX1⁺ number quantified from the 7 representative lymphatic leading vessels in Fig 1F (200 μm from the sprouting end). For Fig 1J disconnected vessels (> 100 μm) or lymphatic endothelial cell (LEC) clusters (< 100 μm) are defined as PROX1⁺ NRP2⁺ cell population that is isolated from the lymphatic network. Only the PROX1⁺ nuclei in a single LEC cluster were quantified and shown in Fig 1I.
To quantify the % of PROX1⁺ cells within the indicated region in the cardinal veins (CVs) (Fig 5B), CVs were divided into a 12‐section pie chart. Region nearest to the dorsal aorta was designated as 0/360°. Within each region, the number of PROX1⁺ LEC progenitors and EMCN⁺DAPI⁺ endothelial cells were quantified. Each region from Fig 5A–C was the average of 20 and 17 vibratome sections from sibling controls (n = 5) and Sox7^iECKO mutants (n = 4), respectively. Only sections from middle to lower thoracic regions were included for analysis. Data were then plotted in a rose diagram using Visual Paradigm online Diagrams.
To quantify the number of PROX1⁺ LECs or LEC clusters associated with EMCN⁺ BV (Fig 5D and E), confocal images were taken randomly from each skin quadrants: left/ right cervical, thoraco‐cervical, and thoraco‐lumbar regions. To effectively quantify the PROX1⁺ LECs associated with LV, EMCN⁺ surface was created by the “surface” function in Imaris, which was then masked for PROX1⁺ cells. Then, PROX1⁺ cells were counted manually by carefully going through each image slices and scrutinized also in 3D projection. PROX1⁺ cells that belong to the greater lymphatic networks and isolated from BV were excluded. PROX1⁺ cell types were categorized according to Appendix Fig S10A and B. Where a PROX1⁺ cell or cluster appears to be associated with BV, a confocal image at a higher resolution (X40‐X63) would be taken for further confirmation. Total PROX1⁺ cells were finally normalized to the total BV surface area analyzed.
Vessel front density (Fig EV2B) was quantified from EMCN⁺ blood vessels 500 μm from the midline (700 × 3,072 μm) using the “surface” function in Imaris. The number of H3⁺ERG⁺ proliferative endothelial cells and total EGR⁺ endothelial cells from the same region were quantified with the “spot” function in Imaris.
Figure EV2D–H were quantified from a similar area of skins (3,500 × 2,000 μm), using the “surface” and “spots” function in Imaris. Green BV was the green area that co‐localized with EMCN⁺ surface. Total blood vessel (BV) surface density was a surface rendering of the EMCN⁺ vessel. Percentage of Cre recombination in BV was therefore green BV density/total BV density (Fig EV2D). Total lymphatic endothelial cells (LECs) were quantified by spots rendering PROX1⁺ nuclei. Total green LECs were determined by spots rendering of green cells, which were also PROX1⁺. The percentage of Cre recombination in the lymphatic vessel (LV) was therefore Green LECs/Total LECs (Fig EV2E). Branching points and lymphatic vessel width were quantified as described in Fig 1E and F (Fig EV2F and G). BV blood density was total EMCN⁺ BV surface (Fig EV2H).