Serial intravital imaging captures dynamic and functional endothelial remodeling with single-cell resolution

Abstract

Endothelial cells are important in the maintenance of healthy blood vessels and in the development of vascular diseases. However, the origin and dynamics of endothelial precursors and remodeling at the single-cell level have been difficult to study in vivo owing to technical limitations. Therefore, we aimed to develop a direct visual approach to track the fate and function of single endothelial cells over several days and weeks in the same vascular bed in vivo using multiphoton microscopy (MPM) of transgenic Cdh5-Confetti mice and the kidney glomerulus as a model. Individual cells of the vascular endothelial lineage were identified and tracked owing to their unique color combination, based on the random expression of cyan/green/yellow/red fluorescent proteins. Experimental hypertension, hyperglycemia, and laser-induced endothelial cell ablation rapidly increased the number of new glomerular endothelial cells that appeared in clusters of the same color, suggesting clonal cell remodeling by local precursors at the vascular pole. Furthermore, intravital MPM allowed the detection of distinct structural and functional alterations of proliferating endothelial cells. No circulating Cdh5-Confetti+ cells were found in the renal cortex. Moreover, the heart, lung, and kidneys showed more significant clonal endothelial cell expansion compared with the brain, pancreas, liver, and spleen. In summary, we have demonstrated that serial MPM of Cdh5-Confetti mice in vivo is a powerful technical advance to study endothelial remodeling and repair in the kidney and other organs under physiological and disease conditions.

Keywords: Endothelial cells; Hypertension; Mouse models; Nephrology.

Figure 1. Tracking of endothelial proliferation and the fate of single glomerular endothelial cells over several days in the same glomerulus in control and in response to hypertensive injury.

(A) Representative images of fixed frozen kidney tissue sections from Cdh5-Confetti mice demonstrating single Confetti⁺ glomerular endothelial cells (GEnCs), and the low number of Confetti⁺ cells in control, but their high density in response to VEGF (0.25 μg/injection, 1 time) or N(ω)-nitro-L-arginine methyl ester (L-NAME) treatment (1 g/L in drinking water) for 14 days. Scale bars: 100 μm (for all panels). G, glomerulus. (B) Summary of the number of Confetti⁺ cells per glomerulus in control, VEGF, or L-NAME–treated mice. The data are shown as the mean ± SEM, n = 54 (Control), n = 42 (VEGF), and n = 32 (L-NAME) glomeruli analyzed from n = 8–10 mice/group, using ANOVA followed by Tukey’s multiple comparison test. A P value of less than 0.05 was considered significant. (C) Single projection images of multiple optical sections (Z-stack) of the same glomerulus visualized by serial intravital multiphoton microscopy (MPM) over time (at baseline, days 4, 7, and 10) of a control Cdh5-Confetti mouse, and during L-NAME treatment. Plasma was labeled with i.v. injected Alexa Fluor 594–albumin converted to gray scale in the images. Scale bar: 50 μm (for all panels). (D) Summary of Confetti⁺ cell number per glomerulus in control versus L-NAME treatment. The data are shown as the mean ± SEM, n = 13 (control) and n = 25 (L-NAME) glomeruli tracked in n = 5 mice/group, using ANOVA followed by Tukey’s multiple comparison test. *P < 0.05 (considered significant versus control group); ^$P < 0.05 (considered significant versus day 0 [baseline]).

Figure 2. In vivo serial MPM imaging of the clonal expansion and function of local GEnC precursors.

(A–C) Z-stack projection images of the same glomerulus at baseline (A) and at 7 (B) and 14 days (C) of continuous L-NAME treatment. Plasma was labeled with i.v. injected Alexa Fluor 680–albumin (gray). Arrows show blue (from afferent arteriole [AA]) and yellow (from efferent arteriole [EA]) clonal cell clusters derived from local blue/yellow EPCs in AA/EA, respectively. (D) Clonal cell expansion during hypertensive injury (n = 13, control, and n = 14, L-NAME glomeruli, n = 4–5 mice/group, using ANOVA with Tukey’s test). (E) In vivo staining of the endothelial glycocalyx in the same glomerulus as in B (Alexa488-WGA, green). Note the weak glycocalyx staining of YFP⁺ EPCs (arrows) compared with nonexpanding GEnCs (arrowheads). (F) Alexa Fluor 488-WGA fluorescence of nonclonal versus clonal GEnC regions (n = 3 mice, unpaired Student’s t test). (G) Progressive changes in albumin glomerular sieving coefficient (albumin GSC) (n = 6, control, and n = 12, L-NAME glomeruli, n = 5–7 mice/group, using ANOVA with Tukey’s test). (H) Albumin leakage (Alexa680-albumin, gray) into Bowman’s space from clonal (arrows) versus nonclonal capillary regions. Inset shows Alexa Fluor 680–albumin fluorescence with green dots equaling no signal, dark and blue dots equaling high signal. Note the higher urinary space albumin signal adjacent to nonclonal (**) compared with clonal capillaries (*). (I) Albumin GSC of nonclonal versus clonal GEnC areas (n = 6 mice, unpaired Student’s t test). (J and K) Z-stack projection images of the same glomerulus at baseline (J) and 7 days (K) after targeted laser-induced GEnC ablation. Clonal clusters derived from local yellow/blue EPCs at the vascular pole (arrows). (L) Clonal cell expansion after laser injury (n = 9, control, and n = 7, laser injury glomeruli, n = 3–4 mice/group using ANOVA with Tukey’s test). (M) Z-stack projection image of a multicolor (nonclonal) proximal AA transitioning into a clonal terminal AA and glomerulus (all cells are blue/red combination, arrows). (N) Renin immunofluorescence (green) with Confetti overlay confirming the terminal AA localization of clonal GEnCs (red/yellow arrows). (O) Schematic of EPCs localized at the glomerular vascular pole (terminal AA/EA in red/yellow, respectively) and their clonal propagation (arrows) into the glomerulus. Scale bars: 25 μm. Data are shown as the mean ± SEM. P < 0.05.

Figure 3. Clonal endothelial cell remodeling in different organs in response to hypertension-induced endothelial injury.

(A–G) Representative images of fixed tissue sections from different organs (A, kidney; B, brain; C, heart; D, lung; E, pancreas; F, liver; and G, spleen) of Cdh5-Confetti mice after 2 months of continuous L-NAME treatment. Inset magnifications show multicellular endothelial cell tracing units appearing in the same Confetti color. Scale bars: 50 μm (for all main panels), 10 μm (for all insets). Scatter plots summarize the distribution of clonal Confetti⁺ multicell tracing units. The x axis shows the cell density categories (4 categories: 1, 2, 3, or more than 4 cells observed per unit) and the y axis shows the number of identical Confetti-colored tracing units observed per microscope field for each category. Data are shown as the mean ± SEM, n = 4 mice, 3–6 fields/mouse.

Figure 4. Lack of circulating Confetti+ EPCs in the intact living kidney.

(A) Representative line (xt) scans of comparable size glomerular capillaries in Cdh5-Confetti and Ren1d-Confetti mice. Arrows in enlarged insets show circulating Confetti⁺ cells detected on line scan. (B) Summary of the number of circulating Confetti⁺ cells per field during 10 minutes of time lapse imaging in Ren1d-Confetti versus Cdh5-Confetti mice. Data are shown as the mean ± SEM, n = 3 each, 3–5 glomeruli per mouse. A P value of less than 0.05 was considered significant using ANOVA. (C) Single projection image of multiple optical sections (Z-stack) of an almost entirely clonal glomerulus (yellow Confetti⁺) visualized by intravital MPM of a control Cdh5-Confetti mouse after 60 days of continuous L-NAME treatment. Plasma was labeled with i.v. injected Alexa Fluor 594–albumin converted to gray scale. Scale bar: 25 μm. G, glomerulus.