Development of an endothelial cell-restricted transgenic reporter rat: a resource for physiological studies of vascular biology

Abstract

Here, we report the generation of a Cre-recombinase (iCre) transgenic rat, where iCre is driven using a vascular endothelial-cadherin (CDH5) promoter. The CDH5 promoter was cloned from rat pulmonary microvascular endothelial cells and demonstrated ~60% similarity to the murine counterpart. The cloned rat promoter was 2,508 bp, it extended 79 bp beyond the transcription start site, and it was 22,923 bp upstream of the translation start site. The novel promoter was cloned upstream of codon-optimized iCre and subcloned into a Sleeping Beauty transposon vector for transpositional transgenesis in Sprague-Dawley rats. Transgenic founders were generated and selected for iCre expression. Crossing the CDH5-iCre rat with a tdTomato reporter rat resulted in progeny displaying endothelium-restricted fluorescence. tdTomato fluorescence was prominent in major arteries and veins, and it was similar in males and females. Quantitative analysis of the carotid artery and the jugular vein revealed that, on average, more than 50% of the vascular surface area exhibited strong fluorescence. tdTomato fluorescence was observed in the circulations of every tissue tested. The microcirculation in all tissues tested displayed homogenous fluorescence. Fluorescence was examined across young (6–7.5 mo), middle (14–16.5 mo), and old age (17–19.5 mo) groups. Although tdTomato fluorescence was seen in middle- and old-age animals, the intensity of the fluorescence was significantly reduced compared with that seen in the young rats. Thus, this endothelium-restricted transgenic rat offers a novel platform to test endothelial microheterogeneity within all vascular segments, and it provides exceptional resolution of endothelium within-organ microcirculation for application to translational disease models.

NEW & NOTEWORTHY The use of transgenic mice has been instrumental in advancing molecular insight of physiological processes, yet these models oftentimes do not faithfully recapitulate human physiology and pathophysiology. Rat models better replicate some human conditions, like Group 1 pulmonary arterial hypertension. Here, we report the development of an endothelial cell-restricted transgenic reporter rat that has broad application to vascular biology. This first-in-kind model offers exceptional endothelium-restricted tdTomato expression, in both conduit vessels and the microcirculations of organs.

INTRODUCTION

Endothelium forms a highly dynamic barrier that coordinates the passage of water, solutes, macromolecules, and cells between the blood and the underlying tissue. Endothelial cells are highly specialized, or differentiated, to achieve the functions of any given vascular location (2, 14). Such differentiation is classified on an anatomical basis, being either continuous, discontinuous, or fenestrated (2). Conduit vessels, like the carotid artery and jugular vein, and capillaries in organs like the skin, lung, central nervous system, and muscle, possess a continuous type of endothelium, where cells reside on an underlying basement membrane. Organs like the liver, spleen, and bone marrow possess a discontinuous, or sinusoidal-type, capillary endothelium, where the cells do not reside on a basement membrane. The kidney glomerulus, endocrine glands, and intestinal endothelium each exhibits fenestrae, which are contiguous transcellular openings. These anatomical classifications relate to the variable ability of endothelium to regulate permeability in accordance with site-specific environmental demands. Permeability of the continuous endothelium largely occurs through intercellular junctions, and to a lesser degree, through vesicular transcytosis, both of which are highly regulated processes (23, 30, 31). Physiological stimuli dynamically adjust the strength of cell-to-cell junctions and the rate of transcytosis to coordinate blood-to-tissue communication. In contrast to this process, discontinuous endothelium offers less resistance to permeability, and fenestrated endothelium serves to sieve molecules based on molecular size and charge, although in both cases, junctional integrity remains an important mechanism governing transcellular permeability. Despite these highly specialized phenotypes, continuous, discontinuous, and fenestrated cells retain an “endothelial” specification.

Adherens junction proteins contribute to the restrictive endothelial cell barrier in each of these cell phenotypes (10), and hence, are important determinants of endothelial specification. Vascular endothelial cell cadherin, vascular endothelial-cadherin (CDH5), is the principal transmembrane protein comprising an adherens junction (16). VE-cadherin displays a highly restricted expression pattern. It is expressed selectively in endothelium and not in the underlying smooth muscle cells. However, VE-cadherin may not be equally expressed in all endothelial cell phenotypes. For example, VE-cadherin was prominently expressed in pulmonary arteries, arterioles, and capillaries, but not in venules and veins, in post-mortem human lung (19). In contrast, VE-cadherin reporter mouse models typically reveal a more homogenous expression pattern. A 2.2-kb CDH5 promoter sequence was cloned upstream of Cre recombinase, and these animals were crossed with various Cre-responsive reporters (18, 24) to engineer constitutive or inducible expression of reporter genes. Approximately half of the endothelial cells isolated from these transgenic mice expressed the reporter gene early in development, and this percentage increased throughout development in blood and lymphatic vessels (18). Near-uniform Cre recombinase reporter gene expression was seen in adult endothelium, across multiple organs. The promoter sequence possesses regulatory elements sufficient to drive expression in endothelium, while reducing expression in other cell types (18). A similar promoter that also contains a 200.3-kb upstream region has been used to drive inducible expression of genes of interest (27, 29, 35, 37, 38) for physiological studies. These studies also confirm the VE-cadherin promoter sequence retains endothelial specificity.

While use of VE-cadherin driven reporter genes has been instrumental in advancing our understanding of endothelial cell fate and properties of angiogenesis and barrier function, mouse models are limited in some areas of vascular biology, especially for the study of chronic lung diseases like pulmonary arterial hypertension (7, 17). In this case, severe pulmonary arterial hypertension is characterized by evolving occlusive neointimal lesions thought to arise from endothelial apoptosis and exuberant hyperproliferation of apoptosis-resistant cells into the lumen. Mouse models of pulmonary arterial hypertension do not typically develop this form of the vasculopathy, with the exception of endothelial and hematopoietic prolyl-4-hydroxylase 2 null mice (7, 17). In contrast, rat models better recapitulate the pulmonary arterial hypertension phenotype characterized by neointimal occlusive lesions (1, 4, 36). Therefore, we sought to develop a transgenic rat model that would enable endothelial cell fate mapping during the evolution of chronic vascular disease states, like occlusive vasculopathies. To develop this model, we engineered the rat VE-cadherin promoter upstream of Cre recombinase and crossed this animal with a Cre-sensitive tdTomato reporter rat. Here, we report the resulting progeny possesses prominent reporter activity across multiple organs, in both males and females.

MATERIALS AND METHODS

Cloning of the CDH5 promoter from rat endothelium.

Generation of CDH5-cre recombinase rats was performed at the Genome Rat Resource Center at the Medical College of Wisconsin under protocols approved by the Institutional Animal Care and Use Committee. An approximate 2.5-kB fragment of the CDH5 promoter was cloned upstream of the codon-optimized HA-tagged Cre (Cre recombinase, iCre), and this expression cassette was subcloned into a Sleeping Beauty (SB) transposon vector (15). The SB method of transpositional transgenesis (TnT) was used to produce transgenic Sprague-Dawley (Crl:SD, Charles River Laboratories) rats by pronuclear microinjection, as we have previously described (21, 22). Three transgenic founders were produced, one of which demonstrated robust endothelial-specific Cre expression when crossed to the tdTomato reporter knock-in rat (Horizon). This founder was then backcrossed to the parental Crl:SD strain to establish a colony for further development and characterized (herein referred to as CDH5-iCre).

Animal care.

CDH5-iCre × tdTomato transgenic Sprague-Dawley rats were shipped from the Medical College of Wisconsin to the University of South Alabama for this study. Once received, they underwent a 30-day quarantine, after which time, pathogen testing revealed they were all pathogen-free. Until the time of the study, animals were group-housed in microisolation caging with enrichment, according to the established guidelines for care and use of laboratory animals. Rooms were on a 12-h:12-h light-dark cycle, and temperature was controlled. Sixteen animals were studied, including 10 females and 6 males. Procedures were reviewed and approved by the University of South Alabama Institutional Animal Care and Use Committee.

Animal surgery and tissue preparation.

Tissues and vessels of interest were harvested from transgenic rats to examine the expression and distribution of tdTomato fluorescent protein. For lung tissue, we used the agarose/gelatin-infused lung slice technique to evaluate tdTomato fluorescence, autofluorescence, and Hoechst nuclear dye (NucBlue, Invitrogen) fluorescence within a 300-μm-thick lung tissue. Rats were sedated with injection of pentobarbital sodium intraperitoneally, and an anesthetic plane was verified by lack of toe pinch response. Rats were then injected with 0.3-mL heparin intraperitoneally and following thoracotomy, the pulmonary artery was cannulated through the right ventricle. The left ventricle was nicked, and ~30 mL of Hanks buffered saline solution (HBSS; Gibco) was slowly perfused through the pulmonary circulation. Next, NucBlue (3 mL) was perfused through the pulmonary artery cannula, and a stopcock was used to infuse ~15 mL gelatin (6%; Thermo Fisher Scientific) solution into the pulmonary circulation. The right lung was then sutured at the hilum, and a tracheal tube was inserted, followed by infusion of ~20 mL agarose (2.7%; Thermo Fisher Scientific) solution to inflate the left lung. Ice was used in situ to solidify the gel en bloc. The left lung, heart, and trachea were excised en bloc and placed in cold HBSS on ice at 4°C for 1 h. The left lung was cut using a razor into an ~1-mm-thick slice at the hilum and then sliced into 300-μm-thick serial sections using a vibratome. Slices were added to a six-well, glass-bottom plate (cellvis) and placed in an incubator for 1 h. An Andor Revolution RS3 (Andor) microscope using IQ (Andor, v3.6.1) software was used to collect z-stacks (100 planes at ~0.4 µm/plane) of selected structures within the slices using the RSL red (emission: 625 nm), RSL green (emission: 525 nm), and DAPI (emission: 440 nm) filter sets.

Other tissues, including liver, spleen, mesentery, and brain, were harvested and placed in buffered saline solution. Tissues were then cut into thin (~0.5 mm) slices using a surgical knife. A representative section of each tissue was incubated with NucBlue (Thermo Fisher Scientific) for 60 min to visualize nuclei. Blood vessels, including the aorta, carotid artery, jugular vein, inferior vena cava (IVC), and basilar artery, were harvested, cut longitudinally, opened, and pinned onto SYLGARD 182 blocks (182 silicone elastomer, DOW Chemical), as previously described (13). Blocks were then immersed in buffered solution containing NucBlue for 10 min. Blocks were washed and placed in a 35-mm round glass culture dish (separated 100 μm from glass by two parallel supporting pins) containing PBS. All tissues and vessels harvested from rats in all three age groups (6–7.5 mo, 14–16.5 mo, and 17–19.5 mo) were prepared similarly. Prepared specimens were imaged using a Nikon A1R confocal microscope (Nikon Instruments).

Hyperspectral imaging.

Hyperspectral imaging analysis approaches were used to examine the distribution and expression of tdTomato fluorescent protein across a range of tissue types, while accounting for tissue type-specific autofluorescence (12). Hyperspectral z-stack images were acquired using a Nikon A1R confocal microscope and several objectives, including ×4 (Plan Apo λ ×4), ×20 (Plan Fluor ×20 multi immersion DIC N2), and ×60 (Plan Apo VC 60X WI DIC N2). Samples (both from transgenic rats and control rats) were excited using 405 nm for NucBlue, 488 nm for autofluorescence, and 561 nm for tdTomato and autofluorescence simultaneously, and the emission spectrum was collected using wavelengths ranging from 424 nm to 724 nm in 10-nm increments. All three laser lines were set to 20% illumination intensity. The z-step (axial) interval for z-stack acquisition was set at 30 µm, 2 µm, and 1 µm, for ×4, ×20, and ×60 objectives, respectively. Specimens were imaged first using the ×4 objective, and subsequently with ×20 and ×60 objectives. A confocal pinhole size of 38.3 µm and detector voltage gain of 130 were used. Similar imaging parameters and settings were used across all image data acquired. Imaging parameters were selected so as to maximize signal-to-noise ratio while minimizing photobleaching of tdTomato.

A spectral library containing pure spectra of each endmember (tdTomato, NucBlue, and tissue-specific autofluorescence) was generated to analyze spectral image data. To construct the spectral library, spectral images were acquired using HEK 293 cells, either transfected with tdTomato or labeled with NucBlue to obtain pure tdTomato and NucBlue spectral signatures, respectively. Tissues and vessel specimens from control Sprague-Dawley rats were imaged to obtain tissue-specific spectral signatures. Custom-developed MATLAB programs were used to linearly unmix raw spectral data into individual endmembers, including tdTomato, NucBlue, and tissue autofluorescence. Unmixed images of tdTomato were used to visualize and quantify the expression and distribution of tdTomato fluorescence signal in different conduit vessels and tissues.

Quantitative analysis.

The age-dependent distributions of tdTomato fluorescence signal in the carotid artery and the jugular vein were quantified using ImageJ analysis software (32). Images acquired using a ×4 objective were used for quantitative measurements. Maximum intensity projection images of tdTomato and NucBlue signals were generated from unmixed tdTomato and nuclei image z-stacks, respectively. A region of interest was selected to cover the maximum possible area occupied by the vessel while eliminating all areas outside of the vessel, as based on the NucBlue projected image. This region of interest was applied to the tdTomato maximum intensity projected image, and the tdTomato image was subsequently cropped to eliminate any effects from areas outside of the tissue, thus removing any potential bias from blank regions of the image. A lower-bound intensity threshold of 0.2% and 0.3% of maximum signal intensity (gray scale values of 126 and 200 out of 65535) was selected for carotid and jugular images, respectively, to define pixels that were tdTomato-positive. The tdTomato-positive fractional area of the thresholded image was then calculated. The same lower-bound intensity threshold was applied to all images (n ≥ 4 rats for each age group) in each group. The age-dependent tdTomato-positive fractional area in carotid artery and jugular vein was plotted for each age group. Statistical significance of variations in tdTomato-positive fractional area per age group were calculated using one-way ANOVA, with multiple comparisons and Tukey post hoc test. Significance was established at P < 0.05.

RESULTS

The rat CDH5 promoter shares limited homology to the equivalent mouse promoter.

The rat CDH5 start site was identified, and the promoter region resolved by alignment against the mouse sequence. The cloned mouse and rat promoters were 2,509 and 2,508 bp, respectively (Fig. 1, A and B). Whereas the previously reported mouse promoter fragment overlaid the first exon by 86 bp, the cloned rat promoter extended 79 bp past the transcription start site. These promoters shared sequence homology, equal to 60% using EMBOSS, 64.4% using Lalign, and 64.9% using AlignX (Vector NTI) (Fig. 1C). In both genes, translation initiation sites are located in the second exon separated by 11,285 bp and 22,923 bp from transcription initiation sites in murine and rat genes, respectively. Therefore, the first intron in the rat gene, while incompletely sequenced at this point, appears to be much larger (22,862 bp vs. 11,147 bp). Once cloned, the rat promoter was placed upstream of codon-optimized iCre with a HA tag and subcloned into a Sleeping Beauty transposon vector for generation of transgenic rats (Fig. 1D), as described in the materials and methods. Founders were bred and backcrossed into the parental strain to establish the colony. tdTomato fluorescence was tested among the progeny from the colony to assess for rat CDH5 promoter efficacy.

Fig. 1.

The rat vascular endothelial-cadherin (CDH5) promoter was cloned from pulmonary microvascular endothelial cells. Organization of the murine (A) and rat (B) CDH5 genes. The cloned murine and rat CDH5 promoters extend 86 and 79 bp, respectively, beyond corresponding transcription start sites. Translation initiation sites are 11,285 bases and 22,923 bp downstream from the transcription initiation sites in murine and rat primary transcripts, respectively. C: alignment of the mouse and rat promoter reveals moderate similarity (from Clustal Omega multiple sequence alignment). D: schematic of the piggyback transposon construct. ITR, piggyback inverted terminal repeats; PCDH5, rat CDH5 promoter; attP, PhiC31 recombinase sites; iCre, codon-optimized Cre recombinase; HA, influenza virus hemagglutinin (HA) tag; pA, bovine growth hormone polyadenylation signal.

tdTomato fluorescence is limited to the endothelium within the conduit vessel wall.

A spectral library containing the unique spectra of each endmember was compiled using image data from single-label control samples. Controls for tdTomato and NucBlue were prepared using HEK 293 cells that were either transfected with DNA encoding tdTomato or labeled with NucBlue. Controls for autofluorescence were prepared in an organ and tissue-specific manner. Tissues and vessels from control Sprague-Dawley wild-type rats were harvested. Tissues and vessels were prepared and mounted as described in materials and methods. Spectral images of each single-label control and tissue type were acquired, including tdTomato (Fig. 2A), NucBlue (Fig. 2B) and tissue-specific autofluorescence (Fig. 2C; example from jugular vein autofluorescence). Spectral libraries (Fig. 2D) containing the unique spectra of tdTomato, NucBlue, and the autofluorescence of the tissue of interest were constructed by extracting the pixel-averaged spectrum from regions of interest in images of control specimens (shown as red, blue and yellow regions in Fig. 2, A–C). Spectral libraries were then used for spectral analysis to unmix respective signals from each fluorescence signature.

Fig. 2.

Spectral library construction and linear spectral unmixing to distinguish tdTomato signal from tissue autofluorescence. False colored images of single label control samples of tdTomato (A), NucBlue (B), and autofluorescence of tissue specimens for each tissue type from Sprague-Dawley wild-type rats; a jugular vein is shown as a representative example (C). Spectral libraries (D) were constructed using pure spectra of each endmember that were obtained from selected regions of interest shown in A, B, and C. E: a representative image from a three-dimensional image stack (z-stack) of jugular vein acquired at ×20 magnification and false-colored by wavelength. Spectrally unmixed (nonnegatively constrained linear unmixing) images of tdTomato (F), nuclei (G), and tissue autofluorescence (H) for a given slice in the z-stack. Three-dimensional projections of the raw spectral image stack (I) and false colored (red denotes tdTomato, blue denotes NucBlue, and green denotes autofluorescence) unmixed images (J). Maximum intensity projections of unmixed images with (K) and without (L) autofluorescence signals. Three-dimensional project highlighting the intima (M, top) and lumen (M, bottom) of an open vessel. Scale bar = 50 µm.

A pixel-by-pixel analysis was performed using custom nonnegatively constrained linear spectral unmixing algorithms in the MATLAB programming environment (Fig. 2D). Raw spectral image data (Fig. 2E) were unmixed into individual spectral components, called endmembers: tdTomato (Fig. 2F), NucBlue (Fig. 2G), and tissue autofluorescence (for example, jugular vein autofluorescence shown in Fig. 2H) using a spectral library. To examine the distribution of tdTomato signal in three dimensions, a projection was created through the unmixed z-stack images (Fig. 2J). With this three-dimensional analysis, different nuclear orientations were apparent, where overlying endothelial cell nuclei are round, and underlying smooth muscle cell nuclei are oblong and oriented at 90-degree angles relative to endothelial nuclei. To assess the total tdTomato signal in the sample, a maximum-intensity projection was also created (Fig. 2K includes autofluorescence and Fig. 2L does not include autofluorescence). The maximum intensity projection was used to visualize the distribution and expression of tdTomato in subsequent figures. For mounted vascular specimens, a rotated three-dimensional view (Fig. 2M) was used to confirm that the distribution and expression of tdTomato was confined to the endothelium. As seen in the image, tdTomato fluorescence is limited to the apical side of the vessel wall, along with endothelial cell nuclei and is not associated with underlying smooth muscle cell nuclei, indicating the rat CDH5 promoter possesses the necessary gene regulatory elements to suppress expression in smooth muscle cells. To better resolve tdTomato fluorescence on a single-cell level, high-magnification images are shown in Fig. 3. Here, tdTomato-positive cells are seen adjoining negative cells. Endothelial nuclei are shown relative to the orientation of blood flow, and the perpendicular orientation of the underlying smooth muscle cell nuclei are visible.

Fig. 3.

Single cell tdTomato fluorescence. A: unmixed false-color image of the jugular vein vessel preparation (blue denotes nuclei, green denotes autofluorescence, and red denotes tdTomato). Double-side arrow indicates the axis of blood flow through the vessel. The arrowhead highlights the orientation of the oblong endothelial cell nuclei in the foreground that is perpendicular to the underlying and elongated smooth muscle cell nuclei. B and C: regions of interests shown on A (white boxes) are selected to visualize the expression pattern of the tdTomato fluorescence signal in single cells. Scale bar on A, B, and C = 10 µm.

tdTomato fluorescence in the conduit endothelium decreases as animals age.

We examined age-dependent distribution of endothelium-restricted tdTomato expression in carotid arteries (Fig. 4) and jugular veins (Fig. 5). Similar lookup tables were applied across all unmixed images so that tdTomato intensity was comparable across different image panels. When unmixed overlay images of vessel preparations obtained at young age (D, G, and J in Fig. 4 and Fig. 5) were compared with images obtained at middle and old age, it was readily observable that the intensity of tdTomato fluorescence in both the carotid artery and jugular vein decreased with age.

Fig. 4.

Age-dependent loss of tdTomato fluorescence in the carotid artery. A representative maximum intensity projection of a raw spectral z-stack image acquired using a ×4 objective at young (A), middle (B), and old (C) age groups. Raw spectral images were unmixed to identify three spectral endmembers—tdTomato in red, NucBlue in blue, and autofluorescence in green—and a false-colored merged image is created (D–F). This procedure was also applied for image data acquired with a ×20 (G–I) and a ×60 objective (J–L). Similar color look-up-tables were used such that the expression intensity is comparable across the images. It is readily observable that the tdTomato fluorescence signal is decreased with aging (compare D to E and F, G to H and I, and J to K and L). Double-side arrows in panels A, B, and C indicate the axis of blood flow through the vessel. n = 4, 6, and 6 animals for young, middle, and old age groups, respectively.

Fig. 5.

Age-dependent loss of tdTomato fluorescence in the jugular vein. Maximum intensity projections of the raw spectral image data revealed that the total fluorescence intensity of the tdTomato signal is decreased over age (A, B, and C). These results are easily visible by examining the unmixed merged images at × 4 magnification (D–F). Similar expression and distribution trends of tdTomato fluorescence were observed when visualized with ×20 (G, H, and I) and ×60 (J, K, and L) objectives. Double-sided arrows in A, B, and C indicate the axis of blood flow through the vessel. n = 4, 6, and 6 animals for young, middle, and old age groups, respectively.

tdTomato fluorescence reveals an endothelial cell microheterogeneity within the conduit vessel wall.

Images in Figs. 2–5 illustrate a heterogeneous tdTomato fluorescence pattern among endothelial cells within the vessel wall. Examples in Figs. 3, 4J, and 5J highlight this point, where immediately adjacent cells either did, or did not, display fluorescence. This differential fluorescence pattern signifies an endothelial microheterogeneity based upon the cell’s ability to drive iCre expression sufficient to activate the tdTomato reporter. Summary data from four to six rats across three ages, including young (6–7.5 mo), middle (14–16.5 mo), and old (17–19.5 mo) age groups, revealed that ≈50% of the vascular surface area was covered by red fluorescence in young animals, and further, that this relative expression pattern decreased significantly in middle- and old-age groups (Fig. 6). A similar expression pattern, and a similar age-related decrease in tdTomato fluorescence, was observed in both the carotid arteries (Fig. 6A) and jugular veins (Fig. 6B). We noted similar tdTomato fluorescence patterns in both male and female subjects. A similar expression of the tdTomato fluorescence signal was observed in other vessel preparations, including the basilar artery, inferior vena cava, and aorta (see Supplemental Fig. S1 at https://doi.org/10.6084/m9.figshare.12461894).

Fig. 6.

Quantitative assessment of the tdTomato fluorescence signal reveals a decrease with age in both males and females. Fraction area occupied by the tdTomato fluorescence signal is quantified in the carotid artery (A) and jugular vein (B) harvested from young (○, male, n = 2, and ●, female, n = 2), middle (□, male, n = 2, and ■, female, n = 4), and old (▽, male, n = 2, and ▼, female, n = 4) age groups. The percentage of the area occupied by the tdTomato fluorescence signal (expression) decreased with age in both carotid arteries and jugular veins. *P < 0.05 by one-way ANOVA with a Tukey post hoc test.

Organ microcirculations exhibit uniform tdTomato fluorescence.

We screened various organs for tdTomato fluorescence in the microcirculation, including the mesentery, liver, and spleen (Fig. 7). In all cases, red fluorescence was observed, signifying widespread expression, especially prominent in the microcirculation. This uniform capillary endothelial expression was also visualized in lung capillaries. Here, to better visualize capillaries, the circulation was filled with gelatin, the airways were filled with agarose, and lung was cut into ≈300-µm sections before imaging. Fluorescence intensity is shown in each channel (Fig. 8A), and composites reveal extensive tdTomato fluorescence in the alveolar capillaries (Fig. 8B). In Fig. 8B, a precapillary arteriole adjacent to the capillary network is negative for tdTomato fluorescence, highlighting endothelial heterogeneity within the microcirculation.

Fig. 7.

tdTomato fluorescence is observed among organs. tdTomato fluorescence was seen in the mesentery, hepatic, and spleen microcirculations. In each case, detection of the fluorescence was decreased with age.

Fig. 8.

The pulmonary circulation displays prominent tdTomato fluorescence, particularly in the microcirculation. A: fluorescence signals in each of the tdTomato, autofluorescence, and NucBlue channels is shown, with a composite image from young-, middle-, and old-rat age groups. Robust tdTomato fluorescence is seen in the alveolar-capillary endothelium. Fluorescence intensity decreases with age. B: a composite image reveals uniform tdTomato fluorescence in alveolar-capillary endothelium, although a precapillary arteriole is negative for fluorescence, illustrating heterogeneous cell phenotypes. Scale bar = 50 µm.

DISCUSSION

Here, we report development of a novel CDH5-iCre driven transgenic reporter rat that exhibits exceptional endothelial-selective tdTomato fluorescence. In conduit arteries and veins the reporter is restricted to the endothelium and is not seen in the underlying smooth muscle layer. Endothelium from more than 50% of the vascular surface area possesses reporter fluorescence in the conduit vessels of both males and females. The endothelial microheterogeneity visible within the intimal layer is notable; the reason for such heterogeneity has not been established. Prominent tdTomato fluorescence is seen in the parenchyma of various organs, especially in the lung’s microcirculation where occlusive neointimal lesions arise. Yet, in both the conduit vessels and the microcirculations, reporter intensity decreases significantly with age.

While a majority of endothelial cells in conduit arteries and veins express the reporter, tdTomato fluorescence was absent in cell clusters throughout the vessel wall. This differential expression pattern supports the notion of endothelial microheterogeneity, where immediately adjacent cells differ in the tdTomato fluorescence. Clonal niches have been previously reported in conduit blood vessels like the aorta. For example, Schwarz and Benditt (33, 34) identified endothelial cell clonal niches based on thymidine uptake as an estimate of replication rates. They found that not all endothelial cells in the vessel wall are replication competent. Rather, proliferation relies on the relatively high replication rates of few cells within clusters. A similar heterogeneity of endothelial cell proliferation was reported in pulmonary arteries and arterioles, especially during the development of pulmonary arterial hypertension. K_i-67 positive endothelial cells are seen in vascular locations that are remodeling to form occlusive lesions (9, 28). A hierarchy of single endothelial cell growth potentials exists among cells isolated from the vessel wall. Single cell cloning of endothelial cells isolated from conduit vessels reveals that few cells are highly replication competent (5, 20, 39). In contrast, endothelial cells isolated from the lung microcirculation exhibit a high proportion of replication-competent clones, suggesting it is an enriched endothelial progenitor niche (3); a molecular basis for these proliferative lung capillary endothelial cells has recently been identified (25). It is uncertain whether the presence or absence of tdTomato relates to endothelial cell proliferative potential, yet this reporter animal provides a way to track endothelial heterogeneity both in vivo and in vitro.

Endothelial cell dysfunction is a cardinal feature of aging. We report a significant loss of endothelial cell tdTomato fluorescence as the animal ages. Whether this decrease in fluorescence reflects a silencing of tdTomato transcription, an increase in tdTomato turnover, or some impairment in the fluorescent property of tdTomato remains to be determined. Nonetheless, it is noteworthy that this decrease in reporter efficacy tracks with the age-dependent loss of VE-cadherin within adherens junctions (8). The loss of VE-cadherin in the adherens junction contributes to both impairment of endothelium-dependent vasodilation (8) and increased permeability (6, 11, 26). Thus, tracking tdTomato fluorescence as a function of age may provide insight into vascular dysfunction that accompanies the aging process (6, 8).

In conclusion, we report a first-in-kind endothelium-restricted transgenic reporter rat. Endothelial expression of the tdTomato was ubiquitous in all circulations tested. Whereas a microheterogeneity in tdTomato fluorescence was seen in conduit vessels, the fluorescence pattern was uniform in the microcirculations of various organs, most prominently in the lungs. This transgenic reporter rat represents an ideal model for assessing endothelial remodeling in various chronic vascular disease states, like atherosclerosis and pulmonary arterial hypertension. The model will be useful for acute injury studies, such as ischemic and hemorrhagic stroke, where adherens and tight junctions are disrupted and where blood-brain permeability is compromised. It allows for assessment of endothelial injury and repair in the lung, as occurs during infection, where loss of capillary density and angiogenesis or vasculogenesis can be tracked fluorescently. tdTomato fluorescence can be measured alongside other fluorescent macromolecular tracers to assess sites of permeability within the vessel wall. However, tdTomato fluorescence decreases as the animals age, which represents a limitation for its use in aging models. Nonetheless, we anticipate this unique animal resource will have widespread application to physiological studies in vascular biology.

GRANTS

This work was supported by National Institutes of Health Grants HL-60024 (to T.S., M.A., S.L., T.R., M.T., R.B., and D.F.A.), HL-66299 (to T.S. and M.A.), HL-140182 (to M.L. and T.S.), HL-114474 (to A.M.G.), HL-116264 (to A.M.G.), OD026560 (to A.M.G.), HL-118334 (to D.F.A.), and HL-136869 (to C.M.F. and T.S.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.A., A.M.G., and T.S. conceived and designed research; M.A., A.M.G., N.S.A., C.M.F., S.J.L., M.S.T., M.T.L., R.B., D.F.A., and T.S. performed experiments; M.A., A.M.G., N.S.A., C.M.F., S.J.L., T.C.R., M.S.T., D.F.A., and T.S. analyzed data; M.A., A.M.G., N.S.A., C.M.F., S.J.L., T.C.R., M.S.T., M.T.L., R.B., D.F.A., and T.S. interpreted results of experiments; M.A., A.M.G., N.S.A., C.M.F., S.J.L., J.M.K., and T.S. prepared figures; N.S.A., S.J.L., and T.S. drafted manuscript; M.A., A.M.G., N.S.A., C.M.F., S.J.L., T.C.R., M.S.T., M.T.L., R.B., J.M.K., D.F.A., and T.S. edited and revised manuscript; M.A., A.M.G., N.S.A., C.M.F., S.J.L., T.C.R., M.S.T., M.T.L., R.B., J.M.K., D.F.A., and T.S. approved final version of manuscript.