Quantitative study of spatial and temporal variation in retinal capillary network perfusion in rat eye by in vivo confocal imaging

Figures

• 

  Figure 1. Experiment set-up. (a) Image shows the right eye of a rat undergoing in vivo imaging using the Nikon A1 system with a × 10 Plan Apo objective lens. The rat is anaesthetized and the trachea intubated for artificial ventilation and its right eye aligned with the objective lens. (b) shows a projected stack of confocal images from the superficial (SL) and deep (DL) vascular plexus within the rat retina at approximately 2 disc distance region. The rat retinal vasculature was perfuse labeled using FITC-Lectins. The image stack was depth color-coded from frame 1–76 (SL to DL) such that the most superficial vessels are in red and the deepest vessels in white as indicated in the color bar below. (c) is a semi-thin epoxy transverse section of a rat retina, stained with toluidine blue. Arrow heads point to retinal vessels seated within the retinal neuronal layers. The SL is embedded in the most anterior portion of the retina close to the inner limiting membrane (ILM) as indicated by the red arrow head. Two closely apposed layers are seated on the anterior and poster borders of the inner nuclear layer (INL) and together are considered as DL for in vivo imaging. ONL, IPL, RPE and CH refers to…

• 

  Figure 2. Kymograph analysis method. (a) Temporal projection over 292 frames of an in vivo time series (corresponding to the first 5 s of Movie S1). The pixel size is 4.93 µm and the frame rate is 58.25 fps. (b) Pixel-wise standard deviation of all frames in the Cy5 channel (red channel showing fRBCs). Centreline paths have been manually traced for four vessel segments from upstream to downstream according to fRBC flow direction. (c) Kymographs generated for each of the four vessel segments (colorised here for easy cross reference with (b)). Each kymograph is a 2D image with the first row containing the Cy5 channel intensities (ordered from left to right) along the centreline path in the direction of flow for the first frame, the second row containing those from the second frame and so on. The high intensity streaks within each kymograph reflect fRBC movement through time. Moving fRBCs appear as diagonal streaks. (d) Histograms indicating the amount of structure present at each angle within each kymograph. The peak in each histogram, marked with a red circle, corresponds to the dominant streak direction and is used to calculate fRBC speed for the corresponding vessel segment (speed is…

• 

  Figure 3. Spatio-temporal variation in fRBC flow speed—representative data and overall data. (a) and (c) show the standard deviation temporal projections of the fRBC signal for 30 s video sequences of the superficial (SL) and deep layers (DL) respectively. The overlaid color dotted lines show the vessel segments belonging to the SL or DL selected according to histology for measurement. (b) and (d) are plots of the kymograph-derived fRBC speeds computed for consecutive 3 s intervals (T1–T10) for each of the vessel segments (S1–S19) highlighted in (a) and (c) respectively. A large range of fRBC speeds is present in the vessel segments from both retinal microvascular layers as indicated by the variation in box plot medians between the vessel segments, especially in the SL vessels in (b). Less variation in speed was observed in the DL vessel segments as indicated in (d). The overall range of speeds found in the SL, (b), are higher than those found in DL, (d). Temporal variation in speed is evidenced from the spread of the speeds for each vessel segment (S) displayed as per time interval (T1 to T10). (e) and (f) show box plots of fRBC speed for the SL and DL vessels segments respectively for…

• 

  Figure 4. Spatio-temporal variation in capillary plexus fRBC count—representative data. (a) Single frame (~ 17 ms) from a 30 s fRBC time recording. Fluorescent labelled red blood cells may be observed as red spots. (b) Histology corresponding to the field of view in (a). (c) Pixel-wise standard deviation over all frames, superimposed with the centrelines of vessels selected using (b) as a guide. (d) is the image in (a) superimposed with white crosses showing the fRBC candidates selected by the spot detection algorithm that coincide with the selected vessels. (e) is a plot of fRBC count (light grey) frame-by-frame for all selected vessels, superimposed with a ~ 0.5 s moving average (black). Variation in the number of fRBCs can be observed throughout the time series. (f) is a fRBC count map created by assigning the value 1 to the pixels of each white cross (detected fRBC) in each frame, and then computing the pixel-wise sum over all frames. The fRBC count map has been color-coded to visualize the variation in cumulative counts in different vessel segments. Variation in cumulative fRBC counts can be observed at different locations along the same vessel segments. The highest cumulative fRBC…

• 

  Figure 5. Diameter measurements from in vivo and confocal images. (a) & (b) are box plots of vessel diameter for the SL and DL computed from the higher resolution in vivo images. Data was not collected for rat 6. (c) & (d) are box plots of vessel diameter of the same vessels measured from histology images. The colored symbols overlaid on the box plots are diameter measurements of each vessel.

• 

  Figure 6. Surgical procedures to prepare rat for in vivo imaging. (a)–(d) shows the steps involved in tracheotomy, (e)–(g) shows the steps involved in the jugular vein cannulation, and (h)–(k) shows the steps involved in femoral artery cannulation. (a) shows a ventral cervical incision in the throat region. The skin, subcutaneous fat has been cut to expose the sternohyoid muscle (yellow arrowhead). (b) shows retraction of the sternohyoid muscle pair to expose the trachea lying underneath. (c) shows the trachea with an incision made and a cannula inserted into the tracheal lumen. (d) shows the tracheal cannula securely tied to the tracheal using two 3.0 sutures. (e) shows the exposure of the left jugular vein by gross dissection. (f) shows the jugular vein being cannulated and two double looped 5.0 sutures for securing the inserted cannula to the vein. (g) shows the throat region with a 4.0 suture closing the skin incision after the jugular vein cannula and tracheal cannula have been securely inserted. (h) shows the exposure of the femoral artery in the left inguinal region of this rat after gross dissection to free it from the femoral neurovascular sheath. (i) shows the femoral artery…

• 

  Figure 7. Categorization of vessel segments. The retinal microvasculature was perfuse labeled for glycocalyx (green, using Lectin-FITC), f-actin (red, using phalloidin) and nuclei (blue, using Hoechst). (a–d) are histology projection images of A1, A2, A3, A4 and capillary vessel segments from the SL. (e and f) are capillary and venule segments from the DL. In all the projected images of the vessel segments, f-actin labeling (white arrows) a differing pattern of arrangement may be seen on the outside of vessel segments. F-actin filaments are densely populated around the A1 and A2 vessel segments in a circular pattern alongside the nuclei of vascular smooth muscle cells (open arrowhead). F-actin filaments are also present around A3, A4, capillaries and venule segments, with a reduced density and regular pattern of appearance. Scale bar in (f) applies to (a–f).