Two-Photon Microscopy of Vital Murine Elastic and Muscular ArteriesCombined Structural and Functional Imaging with Subcellular Resolution Megens R.T.A.a · Reitsma S.a · Schiffers P.H.M.b · Hilgers R.H.P.b · De Mey J.G.R.b · Slaaf D.W.a, d · oude Egbrink M.G.A.c · van Zandvoort M.A.M.J.a
Departments of aBiophysics, bPharmacology, and cPhysiology, Cardiovascular Research Institute Maastricht, Maastricht University, Maastricht, and dDepartment of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, The Netherlands Corresponding Author
Understanding vascular pathologies requires insight in the structure and function, and, hence, an imaging technique combining subcellular resolution, large penetration depth, and optical sectioning. We evaluated the applicability of two-photon laser-scanning microscopy (TPLSM) in large elastic and small muscular arteries under physiological conditions. Elastic (carotid) and muscular (uterine, mesenteric) arteries of C57BL/6 mice were mounted in a perfusion chamber. TPLSM was used to assess the viability of arteries and to visualize the structural components elastin, collagen, nuclei, and endothelial glycocalyx (EG). Functionality was determined using diameter changes in response to noradrenaline and acetylcholine. Viability and functionality were maintained up to 4 h, enabling the assessment of structure-function relationships. Structural vessel wall components differed between elastic and muscular arteries: size (1.3 vs. 2.1 μm) and density (0.045 vs. 0.57 μm–2) of internal elastic lamina fenestrae, smooth muscle cell density (3.50 vs. 1.53 μm–3), number of elastic laminae (3 vs. 2), and adventitial collagen structure (tortuous vs. straight). EG in elastic arteries was 4.5 μm thick, covering 66% of the endothelial surface. TPLSM enables visualization and quantification of subcellular structures in vital and functional elastic and muscular murine arteries, allowing unraveling of structure-function relationships in healthy and diseased arteries.
Copyright © 2007 S. Karger AG, Basel
Alterations in structural and functional properties of the arterial wall may lead to vascular diseases such as atherosclerosis in the case of elastic arteries, and hypertension in the case of small resistance-sized arteries. The understanding of vascular diseases would benefit greatly when such structural and functional properties could be studied simultaneously. Functional studies (investigating diameter changes and force development during control and intervention) are routinely done in isolated vessels mounted on two pipettes in a perfusion chamber (perfusion myography) [1, 2] or in a wire myograph . To combine structural data with functional information, an imaging technique is required which enables imaging in intact and viable vessels ex vivo.
The combination of simultaneously performed structural and functional studies excludes (immuno)histology. Required is a microscopy method that allows for optical sectioning (to obtain 3D reconstructions of the tissue) and has a subcellular resolution and penetration depth (>100 μm). Confocal laser scanning microscopy (CLSM) offers optical sectioning and 3D reconstruction, but has a limited penetration depth (<50 μm in murine carotid artery) and its resolution deteriorates with depth . Furthermore, photobleaching, -toxicity, and -damage caused by the (single photon) excitation light are substantial, which also limits the application of CLSM for functional imaging.
In 1990, two-photon laser-scanning microscopy (TPLSM) was developed by the group of Webb et al. [5, 6]. TPLSM is based on the principle of two-photon excitation , where simultaneous absorption of two near-infrared photons (total energy is equivalent to that of one photon at half the wavelength) leads to the excited state of fluorescent molecules. The probability of two-photon absorption depends on the square of the intensity of the excitation light, and only takes place in a very narrow volume at the focal position of the microscope. Because out of focus absorption and fluorescence are absent, the detected emission light always originates from the focal position independent of scattering. Typical features of TPLSM are enhanced depth penetration (dependent on sample characteristics, up to 250 μm in rat aortas ), good optical sectioning, and good resolution (maximum resolution approximately 0.3 μm in xy- and 0.9 μm in z-direction, dependent on the wavelength of the excitation light ). In addition, photobleaching, -damage and -toxicity are drastically reduced , enabling imaging of delicate structures in viable tissue. Recently, TPLSM was established as a valuable tool for imaging of blood vessels [4,9,10,11], skeletal muscle arterioles , and microvasculature of the human uterus . However, the blood vessels were very small , still treated with a fixative, freeze-thawed and sliced, or casted in hot (>40°C) agarose gel, resulting in structural alterations and loss of functionality .
In the present study, our aim was to demonstrate that TPLSM is a useful tool for imaging of fresh intact arteries mounted in a perfusion chamber. Together with appropriate fluorescent markers, this should allow exploration of the actual vessel wall properties at a subcellular level in three dimensions. The utility of our method is illustrated with examples of both structural and functional imaging of cell nuclei, elastin layers and collagen. Elastin is a major component of the extracellular matrix. In the arterial wall, elastic fibers assemble to form elastic layers, important for the elasticity of arteries. Alterations in elastin structure and content are linked with vascular diseases, arterial remodeling, and vascular smooth muscle cell (vSMC) proliferation [15,16,17]. Another major substance of the extracellular matrix of arteries is collagen. Collagen is the fundamental component of basement membranes and is a major determinant for maintaining vascular integrity (tensile strength) . The endothelial glycocalyx (EG), a very delicate layer covering endothelial cells (ECs), was studied in elastic arteries. Recently, interest in the EG has increased because of its possible role in hemodynamics , blood cell-vessel wall interactions [20, 21], and atherosclerosis .
Structural and functional integrity of mounted arteries was tested by assessing cell membrane integrity on the one hand and vasomotor responses to a vasodilating and a vasoconstricting substance on the other hand. Experiments were performed in both elastic (carotid arteries) and muscular arteries (uterine arteries and first-order mesenteric-artery side branches) of mice.
Experiments were approved by the local ethics committee on the use of laboratory animals. Procedures were in accordance with institutional guidelines. Carotid and first-order mesenteric arteries were derived from 12- to 20-week-old C57BL/6J mice (Charles River, Maastricht, The Netherlands), that were anesthetized by subcutaneous administration of a mixture of xylazine (15 mg Xylazin®/kg body weight; Ceva Sante Animale, Naaldwijk, The Netherlands) and ketamine (75 mg Nimatek®/kg body weight; Eurovet, Cuijk, The Netherlands). After euthanasia, the carotid arteries and the mesentery were isolated. Uterine arteries of non-pregnant mice were isolated from uterine tissue obtained from female C57BL/6J mice (12–16 weeks of age, Charles River) after euthanasia by CO2.
Elastic Arteries: Carotid Arteries. Segments of common carotid arteries (length ∼6–8 mm) were freed of adipose and connective tissue and carefully handled, only at their outer ends without stretching them, to keep them viable. To avoid contact with air, they were kept moist during the whole preparation procedure. Until further processing, arteries were stored (maximum 30 min) at 4°C in Hanks’ balanced salt solution (HBSS, pH 7.4), containing in mM: NaCl 144, HEPES 14.9, glucose 5.5, KCl 4.7, CaCl2 2.5, KH2PO4 1.2, and MgSO4 1.2.
Muscular Arteries: First-Order Mesenteric Arteries and Uterine Arteries. After dissection, the mesentery or uterus was pinned out on a Petri dish (coated with silicon) and kept moist with HBSS. Adipose and connective tissue were carefully removed. One segment (∼3 mm) of a first-order mesenteric artery without side branches or one segment (∼3 mm) of the uterine artery at midpoint of the arterial arcade was isolated and stored in HBSS at 4°C.
Mounting Procedure. After storage, each vessel was mounted in a homebuilt perfusion chamber (fig. 1a; IDEE, Maastricht, The Netherlands) filled with 10 ml HBSS (37°C). The artery was mounted on two glass micropipettes (tip diameters 120–150 μm for elastic arteries, and 80–100 μm for muscular arteries) and residual luminal blood was carefully removed by gently flushing with HBSS. To correct for the shortening of the artery during isolation , a transmural pressure of 100 mm Hg was applied (using a modified Big Ben sphygmomanometer, Riester, Jungingen, Germany), and the distance between the two pipettes was adjusted until the mounted artery was straight. After this length adjustment, transmural pressure was set at 80 mm Hg in order to mimic physiological conditions . All experiments were performed at 37°C (MC60 warm stage controller; Linkam Scientific Instruments, Tadworth, UK) in the absence of luminal flow. Imaging was performed in vessels at a transmural pressure of 80 mm Hg and was restricted to the central portion of the vessel segment.
|Fig. 1.a Schematic representation of a vessel perfusion chamber used to mount isolated arteries. The distance between the two pipettes is adjustable for correction of the arterial length after dissection. Fluid capacity is 10 ml. Inset: mouse uterine artery (external diameter ∼250 μm, length ∼3,000 μm) mounted on two glass pipettes with sutures, length adjusted, and pressurized to 80 mm Hg. b 3D reconstruction (size of arterial section = 309×309×315 μm) of a uterine artery stained for elastin and nuclei with superimposed coordinate system and a z-stack of optical xy-slices.|
Nuclei of viable cells were labeled with the viable cell membrane-permeable DNA/RNA markers Syto 41 (λmax emission = 480 nm) or Syto 13 (λmax emission = 510 nm; Molecular Probes, Leiden, The Netherlands). Propidium iodide (PI, λmax emission = 620 nm; Molecular Probes) was used to assess cell membrane integrity. Although elastin displays an autofluorescence signal , the autofluorescence is weak compared to the fluorescence of added probes. Therefore, we have used eosin (λmax emission = 545 nm; Molecular Probes) as a specific, strongly fluorescent elastin marker. Eosin allows us to visualize elastin in more detail and in combination with other dyes . Collagen was visualized by second-harmonic generation, SHG, a nonlinear scattering process which results in the emission of photons at exactly half the wavelength of the excitation light (λexcitation = 840 nm) [4,25,26,27]. The EG was visualized with fluorescein isothiocyanate-labeled wheat germ agglutinin (WGA-FITC, λmax emission = 515 nm, Sigma-Aldrich, Zwijndrecht, The Netherlands) . Wheat germ agglutinin is a lectin from Triticum vulgare, which targets N-acetylated glucosamine residues in hyaluronic acid and heparan sulfates. All fluorescent probes were dissolved in HBSS except for WGA-FITC which was dissolved in phosphate-buffered saline (pH 7.4) in order to prevent any interactions of the probe with glucose residues present in HBSS. Labeling was performed in the perfusion chamber, and fluorescent probes were added 30 min prior to the start of image acquisition. Fluorescent labels were not washed from the preparation during measurements.
Final probe concentrations were: Syto 41 1.5 μM, Syto 13 1.5 μM, PI 1.5 μM, eosin 0.25 μM, and WGA-FITC 2.6 μM. All probes were applied both intra- and extraluminally, with the exception of WGA-FITC that was perfused at 1 ml/h (carotid arteries) or 0.3 ml/h (mesenteric artery) for 30 min. When WGA-FITC is applied intra- and extraluminally, it also labels glycosaminoglycans in the tunica media and adventitia [unpubl. data]. To verify the specificity of WGA-FITC for the EG, arteries were flushed with hyaluronidase (2.6 μM, type IV-S, Sigma-Aldrich) for 60 min (1 ml/h at 37°C) after WGA-FITC labeling. This enzyme is known to break down the glycocalyx in vivo .
Viability/Functionality of Mounted Arteries. To illustrate the viability of arteries in our experimental setup, mounted carotid arteries were stained with a mixture of Syto 41 and PI. Images were recorded 15 min and 4 h after the start of the incubation in various segments of the common carotid artery. In between imaging, the mounted artery was kept at 80 mm Hg transmural pressure and 37°C. Functionality of the ECs and SMCs of mounted muscular (mesenteric) arteries was evaluated by testing the acute vasomotor responses (change in luminal diameter) to the vasodilator acetylcholine (1 μM, Sigma-Aldrich) during noradrenaline (1 μM, Sigma-Aldrich)-induced vasoconstriction . Both were administered extraluminally. Luminal diameters were measured in yz-scans recorded prior to and during stimulation (after stabilization of the vessel diameter). All functionality experiments were performed in arteries that were stained with Syto 41 and PI.
The perfusion chamber was positioned on a Nikon E600FN microscope (Nikon, Tokyo, Japan), coupled to a standard Bio-Rad 2100 MP multiphoton system (Bio-Rad, Hemel Hempstead, GB). The excitation source was a 140-fs-pulsed Ti:sapphire laser (Spectra Physics Tsunami, Mountain View, Calif., USA) tuned and mode-locked at either 840 nm (to visualize collagen by SHG) or 800 nm (for visualization of fluorescent probes). Laser light reached the sample through the microscope objective (×60 water dipping, numerical aperture 1.0, working distance 2 mm; ×40 water dipping, numerical aperture 0.8, working distance 2 mm). Maximum field of view was 206×206 μm for the ×60 objective and 309×309 μm for the ×40 objective. When necessary, further magnification (and thus higher pixel resolution) was achieved by optical zoom in the scan head.
Three photo multiplier tubes (PMTs) were used to detect the emitted fluorescent signals. For imaging of each of the fluorescent markers, PMTs were tuned corresponding to parts of the emission spectra of the fluorescent markers used: Syto 41, 470–480 nm (PMT 1); Syto 13, 500–540 nm (PMT 2); WGA-FITC, 500–550 nm (PMT 2); eosin, 530–560 nm (PMT 2) and/or 560–590 nm (PMT 3), and PI, 560–600 nm (PMT 3). Second-harmonic generation signal was collected at 400–450 nm in PMT 1 for excitation at 840 nm. For simultaneous imaging of a combination of fluorescent markers, each PMT was tuned for minimal bleed-through of the fluorescent markers used. To prevent photochemical and thermal damage to the arteries, laser power was kept as low as possible. Frame rate was 0.1 Hz with a pixel dwell time of 39 μs, or 0.3 Hz with a pixel dwell time of 12 μs combined with Kalman filtering (n = 3 cycles) for noise reduction. From each PMT, separate images of 512×512 pixels were obtained and combined into a single image. Images were recorded in the xy-plane. Inaccurate alignment of the pipettes in the perfusion chamber usually caused imaging of the artery in a slightly oblique plane. Series of xy-images at successive depths (z-stack) were collected for reconstruction of 3D images (fig. 1b). Luminal diameters were obtained from yz-images. To obtain yz-images with square pixels, z-step distance was equal to the pixel dimensions in xy-direction.
Image processing was performed using Image-Pro Plus 6.0 (Media Cybernetics, Silver Spring, Md., USA). The 3D Constructor 5.1 software (Media Cybernetics) was used for 3D reconstructions. Image analysis was performed by two independent observers.
Six elastic (carotid) arteries and six muscular (uterine) arteries were used for quantitative analyses of cell nuclei and elastin. vSMC density was expressed as the mean number of cells per 100 μm3 of vessel wall tissue. The overall vessel wall volume was estimated in each z-stack by application of an iso-surface on the 3D reconstruction. Iso-surface rendering yields a smoothed surface around all voxels above a chosen threshold. The applied iso-surface was based on the signal of cell nuclei (Syto 41). This is a good measure for total vessel wall volume since it includes all structures from the ECs in the tunica intima to fibroblasts, which are widespread in the tunica adventitia. However, it excludes EG volume. The iso-surface threshold was based on the averaged full width at half maximum value of three randomly selected line profiles (in the xz-plane) of each z-stack. The EC density is expressed as the mean number of cells per 100 μm2 of luminal surface area.
Glycocalyx thickness and glycocalyx coverage (ratio of glycocalyx-covered surface area and total surface area) were assessed in five carotid arteries. Analyses were performed by quantification of the glycocalyx volume and luminal surface area (as described in the previous paragraph). Mean glycocalyx thickness was calculated by dividing the glycocalyx volume by the glycocalyx coverage.
Vasomotor responses, i.e. functionality of arteries, were determined in three mesenteric arteries and measured as changes in luminal diameter in yz-scans of the vessels.
Results are presented as means ± SD and were tested for significance using the Mann-Whitney test (non-parametric test for two independent groups). A value of p < 0.05 was considered to be statistically significant. Interobserver reliability for cell counting was estimated by a ‘two-way mixed consistency intraclass correlation coefficient’. All statistical analyses were performed using SPSS 13.0 software (SPSS, Cary, N.C., USA).
Geometric Deformations. Figure 2 displays images of a carotid artery stained for elastin (red) and nuclei (blue). Figure 2a, b shows geometric deformations caused by dissection and lack of transmural pressure. Dissection induced longitudinal shortening of the artery in the absence of transmural pressure and length adjustment. The three elastic laminae were difficult to identify, appeared to be inconsistent and partially collapsed. Vessel wall thickness was 57 ± 6 μm. In contrast, when the artery was straightened and a transmural pressure of 80 mm Hg was applied in order to mimic physiological conditions (fig. 2c, d), vessel wall thickness was reduced to 33 ± 4 μm. Furthermore, the elastic laminae appeared to be homogeneous and unfolded.
|Fig. 2. Carotid artery labeled for elastin (eosin, red) and nuclei (Syto 41, blue) imaged at approximately the same z-position near the transition of the tunica media and tunica intima (a, c) and y-position (b, d). a, b Images were obtained in the absence of transmural pressure and length adjustment. Elastin bands appear to be folded, and have a wave-like appearance. Furthermore, the IEL (green dashed line) and the intermediate elastic layer (pink dashed line in yz-reconstruction) and external elastic layer (yellow dashed lines) are difficult to distinguish from each other. vSMC nuclei (blue arrow) are visible in blue. The yz-reconstruction (b) clearly establishes the elastic laminae to be folded. c, d The same artery in the presence of transmural pressure (80 mm Hg) and length adjustment. All three elastic laminae (external elastic lamina, yellow dashed line; intermediate elastic lamina, pink dashed line, and IEL, green dashed line) are clearly recognizable and relatively straight. SMC nuclei (blue arrow) are homogeneously distributed and their (mean) orientation is perpendicular to the longitudinal axis of the artery. Furthermore, a few EC nuclei (pink arrow) are visible (due to a slightly oblique imaging plane). The yz-reconstruction (d), reconstructed from a stack of xy-sections, demonstrates that the vessel wall is tight and contains three elastic laminae.|
Viability/Vitality of Vessels. Mounted arteries (n = 3) were stained with Syto 41 and PI to assess the viability of arteries. Images were recorded 15 min and 4 h after the start of the incubation in various segments of the arteries (fig. 3). After 15 min, 31% of the cell nuclei in the adventitial layer were PI positive. Only 2% of the vSMCs were PI positive, and no PI-positive ECs were detectable. After 4 h (fig. 3a), the number of PI-positive nuclei in the three tunicae was unaltered, which indicates that vital cells in the artery remained vital. Intraluminal air bubbles (fig. 3b) caused a shift from PI-negative to PI-positive labeling of 95% of ECs within 10 min. The number of PI-positive vSMCs and nuclei located in the adventitia was unaltered by intraluminal air bubbles. In contrast, after a single pinch with a forceps, the number of PI-positive SMCs and adventitial cells in the visualized part of the vessel increased to 33.3 and 47.5%, respectively. Parts of the EC layer at the site of pinching were denuded. Residual ECs at the pinching site were all PI positive. The functionality of three mounted arteries was assessed by measuring the luminal diameter change in mesenteric arteries during noradrenaline constriction (fig. 4). The average luminal diameter prior to stimulation was 246 ± 21 μm. During noradrenaline administration, the average luminal diameter declined to 182 ± 3 μm. Subsequent administration of acetylcholine induced a vasodilatation; luminal diameter increased to 231 ± 6 μm.
|Fig. 3.a Syto 41 (vital, blue)- and PI (cells with a compromised cell membrane, red)-positive cells in 3D reconstruction (approximate size 206×206×40 μm) of the central portion of a mounted mesenteric artery imaged 4 h after the start of the incubation. Only few PI-positive SMCs are visible (green arrow). ECs are only faintly labeled with Syto 41 and, hence, hardly visible. After application of intraluminal air (b), most EC nuclei became PI positive (red). A single pinch with a forceps (c) caused a local shift from Syto 41- to PI-positive nuclei of SMCs (green arrow) and ECs (white arrow) within 10 min.|
|Fig. 4. Typical yz-scans of a mesenteric artery in the abscence (a) and presence of noradrenaline alone (b), and both noradrenaline and acetylcholine (c). Nuclei of cells were labeled with Syto 41 (pseudo-colored green), and collagen was visualized with SHG (blue). A Gaussian filter was applied to reduce noise. Note that signal (especially the SHG) is weakened at the lower parts of the yz-scans.|
Elastin. Fig. 5b–e shows xy-slices of an elastic (carotid) artery, recorded at four z-positions (fig. 5a) with z = 0 μm being the outside of the artery. Fig. 6b–e are xy-slices of a muscular (uterine) artery, recorded at comparable positions in the arterial wall (fig. 6a). The external elastic lamina was clearly visible in both types of arteries. The structure of the external elastic layer in elastic carotid arteries was more compact (fig. 5c), whereas in the muscular uterine arteries, a mesh-like structure was observed (fig. 6c). In all carotid arteries under investigation, the appearance of an intermediate elastic layer (red) was clearly visible in the tunica media (fig. 5d). This third elastic layer was not present in the muscular arteries investigated (fig. 6c, d). The internal elastic lamina (IEL) was clearly distinguishable in muscular (uterine and mesenteric) arteries (fig. 6d) but less apparent in carotid arteries (fig. 5d). Fenestrae (non-fluorescent cylindrical patches in the IEL ) were visible in the IEL. Their mean diameter was significantly larger in muscular arteries (2.1 ± 0.2 μm) compared to elastic arteries (1.3 ± 0.4 μm, p = 0.02). Furthermore, the density of fenestrae was significantly higher in muscular (57.4 ± 9.9 per 100 μm2) versus elastic arteries (4.5 ± 3.9 per 100 μm2, p = 0.02).
|Fig. 5.a Schematic view of the arterial wall representing the positions of the xy-sections shown in b–e. b–e Optical xy-sections (206×206 μm) of a common carotid artery, recorded at four z-positions in the vessel wall (z = 0 μm on the outside of the vessel). b z = 12 μm. c z = 27 μm. d z = 45 μm. e z = 57 μm. White bars represent 20 μm. b Tunica adventitia with collagen (SHG of collagen, blue), and several thin elastin fibers (red). Furthermore, fibroblast-like cell nuclei (white arrow) populate the tunica adventitia (b, c). c, d Tunica media where elastin forms the external elastic lamina (red), which is compact of structure and completely surrounds the ‘cigar-shaped’ vSMC nuclei (blue arrow). d The external elastic lamina (yellow dashed lines) and intermediate elastic lamina (more intense red bands, pink dashed lines) are visible. e Tunica intima in the center, showing the external elastic lamina (marked by yellow dashed lines) and the intermediate elastic lamina (marked by pink dashed lines) which separates two layers of vSMCs, the IEL (surrounded by blue dashed line), vSMCs (cigar-shaped nuclei) and a monolayer of (elliptical) EC nuclei (pink arrow). f 3D reconstruction (size: 206×206×59 μm) of the same stack of images representing the intimal side of the vessel. All three elastic laminae, IEL, and intermediate (IMEL) and external elastic lamina (EEL), are visible at the transversal edges of the object. EC nuclei (pink arrow) are positioned on top of the (uniformly shaped) IEL (red), and vSMCs (blue arrow) are visible at the edges of the reconstruction. A small part of the tunica adventitia visible at the bottom of the reconstruction shows collagen (blue) and fibroblasts (for further details, see supplemental video files 1 and 2).|
|Fig. 6.a Schematic view of the arterial wall representing the positions of the xy-sections shown in b–e. b–e Optical xy-sections (137×137 μm) of a mounted muscular (uterine) artery, recorded at four z-positions in the artery (z = 0 μm on the outside of the vessel). b z = 8 μm. c z = 15 μm. d z = 20 μm. e z = 30 μm. White bar represents 10 μm. b Tunica adventitia with collagen (SHG, blue), nuclei of fibroblast-like cells (white arrow), and thin elastin fibers (red). c The start of the tunica media. The external elastic lamina is clearly visible in red and has a mesh-like structure with a single layer of ‘cigar-shaped’ vSMC nuclei (blue arrow) positioned directly underneath. The outside of the artery shows collagen (blue) and fibroblast-like cells. d A section made deeper in the tunica media, which also contains part of the IEL (surrounded by blue dashed line). In the tunica media, vSMC nuclei are surrounded by elastin fibers (yellow arrow) which are connected to the external elastic lamina (marked by yellow dashed lines) and IEL, visible in the center. The IEL is uniformly shaped and contains several fenestrations (red arrow) underneath, and a monolayer of (elliptical) EC nuclei (pink arrow). e Part of the tunica intima and the lumen. EC nuclei (pink arrow) are visible on the luminal side and vSMC nuclei on the subluminal side of the IEL (marked by blue dashed lines). f 3D reconstruction (displayed artery size 137×137×23 μm) of the z-stack exposing the intimal side of the muscular artery. Both elastic layers (external elastic lamina, EEL, and IEL), vSMC nuclei (blue arrow), and EC nuclei (pink arrow) can be distinguished at the transversal edges of the reconstruction. EC nuclei are clearly visible on top of the IEL (red). Furthermore, vSMC nuclei weakly shine through the IEL (due to large difference in intensity between the IEL and vSMCs) and are visible when viewed from the luminal site (for further details, see supplemental video files 3 and 4).|
Nuclei. Cell nuclei in the arterial vessel wall were labeled with Syto 13 (fig. 5, 6). Based on their morphology, orientation and position relative to the elastin bands, nuclei can either be distinguished as EC or vSMC nuclei in the tunica media, or as fibroblast-like cell nuclei in the tunica adventitia (for details, see the supplemental video files 2 and 4) . EC density was not significantly different between elastic and muscular arteries (11.8 ± 2.2 in elastic vs. 12.4 ± 3.8 in muscular arteries, per 100 μm2 surface area). vSMC density appeared to be significantly higher in elastic arteries (350 ± 52 per 100 μm3) than in muscular arteries (153 ± 31 per 100 μm3, p = 0.004). Intraclass correlation for interobserver reliability was 0.997.
Collagen. Collagen was visualized in elastic and muscular arteries (fig. 7) with SHG using backward propagated light . Since SHG is dependent on peak laser intensity , the SHG signal in arteries weakened with increasing depth and was only visible in the tunica adventitia of both elastic and muscular arteries at relatively high laser power (>30 mW). The shape of the collagen fibers in mounted elastic arteries (which were corrected for geometrical deformations) appeared to be tortuous (fig. 7a), whereas in similarly mounted muscular arteries (fig. 7b) the collagen fibers appeared to be more stretched out. Furthermore, in muscular arteries the SHG signal was weaker and required higher laser power (>60 mW).
|Fig. 7. SHG (blue) of collagen in xy-sections at comparable z-position in the tunica adventitia of an elastic (carotid) artery (a) and a muscular (uterine) artery (b), both pressurized and length adjusted. In the carotid artery (a), thin undulating collagen fibers (white arrow) rearrange to form thicker clustered collagen fibers (green arrow). In the muscular artery (b), fibers are less undulated and more difficult to distinguish (SHG signal is weak compared with the SHG signal originating from the carotid artery). Bars = 20 μm.|
Endothelial Glycocalyx. The EG was visualized in both muscular (fig. 8a) and elastic (fig. 8b) arteries. Mean thickness of the glycocalyx in carotid arteries was 4.5 ± 1.0 μm; mean glycocalyx coverage of the endothelium was 66 ± 6%. To verify the specificity of the WGA-FITC labeling, the glycocalyx layer in carotid arteries was degraded enzymatically by hyaluronidase perfusion. Hyaluronidase treatment reduced the glycocalyx coverage to 25 ± 2% (fig. 8c) but did not affect the number of PI-positive (i.e. viable) ECs.
|Fig. 8. Examples of the glycocalyx in a mesenteric artery (a) and a carotid artery (b), labeled with WGA-FITC (green), Syto 41 (nuclei, blue) and PI (nuclei of cells with a compromised cell membrane, red). b, c Glycocalyx of a carotid artery labeled with WGA-FITC (green), Syto 41 (nuclei of vital cells, blue) and PI before (b) and after (c) treatment with hyaluronidase. The coverage of the EC layer by glycocalyx is drastically reduced after treatment with hyaluronidase.|
In this study, we demonstrated that the combination of TPLSM and arteries mounted in a perfusion chamber is a valuable tool for simultaneous imaging of structure and function of elastic and muscular arteries. We have shown that TPLSM is a very suitable microscopic technique for detailed imaging of intact and vital arteries in three dimensions, mainly because of its combination of a large penetration depth with good spatial resolution and optical sectioning. Furthermore, it allows quantification of delicate structures in the vessel wall, as illustrated by the visualization of the EG through the vessel wall. Until now, this thin layer, situated on the luminal site of the tunica intima, was difficult to visualize. Mounting of arteries on micropipettes is an established method that allows studying of functional properties of arteries [1, 2]. When combined with TPLSM, it enabled detailed imaging of arterial structures in physiological conditions. Furthermore, we have shown that the combination of both techniques allows functional imaging of large arteries. Therefore, this experimental setup can contribute to unraveling structure-function relationships in healthy and diseased arteries.
TPLSM imaging of vital and mounted arteries has a number of advantages over existing methods for arterial imaging. (1) The viability of mounted arteries is maintained. This was illustrated by a vasoconstriction-vasodilatation protocol resulting in changes in arterial diameter (visualized in xz-images). In addition, this demonstrates the feasibility of imaging structural-functional relationships of large arteries, similar to those already performed in skeletal muscle arterioles . Staining and frequent TPLSM image acquisition of mounted arteries did not affect viability (up to 4 h after mounting), which is essential for prolonged imaging of functional processes. (2) The ability to control transmural pressure and luminal flow in mounted arteries will also enable exploration of pressure or flow-mediated structure-function relationships of arteries with subcellular resolution. (3) Structural deformation due to isolation of arteries can be resolved. The wave-like appearance of the vessel wall, typical for unmounted arteries  and histological sections, disappeared when pressure was applied and arterial length was adjusted. Therefore, the data presented in our study are more representative for the in vivo situation. (4) It enables extra- and intraluminal labeling, the latter being virtually impossible in non-mounted arteries, which improves structure-specific labeling and simplifies labeling of structures in the tunica intima. For example, WGA-FITC is only specific for the EG when administered intraluminally.
Despite all these advantages, TPLSM imaging of mounted arteries also has its limitations. Even though it provides good theoretical spatial (xy ∼0.3 μm) and axial (z ∼0.9 μm) resolution, TPLSM suffers from image deterioration in the axial direction. Especially at larger depths, structures appear somewhat blurred due to a relatively large point spread function. Therefore, deconvolution of image stacks may enhance contrast, axial resolution and the signal-to-noise ratio. However, it is difficult to determine the (true) point spread function inside the vessel wall in large arteries (which is a prerequisite for appropriate deconvolution). Additionally, according to the Nyquist criterion, the necessary pixel resolution needed for appropriate deconvolution must approach the maximal resolution (1,024×1,024 pixels, step size ≤0.3 μm). However, this maximal pixel resolution in TPLSM is often limited by practical causes. Imaging of large arteries at maximum pixel resolution is time-consuming, results in large data stacks which are difficult to handle (especially during image analysis and restoration), and increases radiation load to the tissue. In practice, in large arteries, a resolution of xy ∼0.6 μm and z ∼1.5 μm is more realistic.
The present study shows that our method enables identification of biologically interesting variations between different vessel categories. We have visualized several structural differences between elastic (carotid) and muscular (uterine and mesenteric) arteries. In the large elastic arteries an extra elastic layer was present in the tunica media, the structure of the external elastic lamina was more compact and the IEL contained significantly fewer and smaller fenestrae. IEL fenestrae have been implicated in electrical coupling between the endothelium and the underlying vSMCs [32, 33]. One consequence of this coupling, endothelium-dependent vasodilatation through vSMC hyperpolarization, is more prominent in small muscular arteries than in large elastic arteries . The vSMC density was significantly higher in elastic arteries. Apparently, vSMCs are smaller in elastic arteries. EC density was comparable. Collagen fibers in the tunica adventitia of elastic arteries have an undulating pattern, even when corrected for geometric deformations, whereas in muscular arteries the fibers have a straight appearance. Collagen in the tunica adventitia was visualized using its SHG. However, we were unable to detect SHG of collagen in other parts of the vessel wall, in contrast to the results by Zoumi et al. [ 25]. They established the existence of SHG signal in the tunica media in transversal sections of porcine coronary arteries. In our setup, lack of SHG signal in the tunica media and intima limits the applicability of SHG as a contrast agent for collagen. A newly developed collagen probe (CNA35/OG488) may be advantageous for studying collagen structures in the tunica media and tunica intima .
Existing methods to visualize and quantify the glycocalyx in blood vessels often depend on electron microscopy, which implies extensive tissue processing and sectioning [22, 36, 37], and excludes combination with functional studies. In the microcirculation, the glycocalyx can be visualized in vivo using fluorescently labeled lectins, proteoglycan-binding proteins or specific antibodies [21, 38]. However, the intravital microscopic techniques used in these models cannot be used in larger arteries due to limited penetration of the excitation light. A combination of TPLSM imaging and mounting of arteries proves to be a good approach to visualize the EG in larger arteries in three dimensions. Since mounted arteries maintain their viability and functionality and can be kept under more physiological conditions, the observed glycocalyx is probably representative of the in vivo situation. The mean thickness of the EG in carotid arteries in this study (4.5 ± 1.0 μm) is comparable with observations described in the literature (2.6 ± 0.5 μm in mesenteric arteries with a luminal diameter of ∼150 μm) . Mean EG coverage of the EC layer in common carotid arteries is 66%. This incomplete coverage of the EC layer by glycocalyx may be caused by several factors. First, WGA-FITC may only label part of the heterogeneously distributed components [40, 41] of the glycocalyx, causing areas of the glycocalyx to be unlabeled. Second, weakly stained areas of glycocalyx did not contribute to the total area when the glycocalyx also contained areas with a much stronger labeling. This is a consequence of the applied image analysis procedure based on the full width at half maximum of the intensity signal of WGA-FITC, averaged over the total endothelial area within each image stack. Third, although mounted arteries were handled with care and were still viable, it cannot be excluded that some areas of the very delicate glycocalyx were damaged due to the isolation (the lumen of arteries occasionally collapsed during isolation) and mounting procedure.
In conclusion, the described method for imaging of viable arteries is a significant step forward in the field of vascular research, potentially enabling investigation of structure-function relationships in healthy and diseased blood vessels.
The Bio-Rad TPLSM was obtained via a grant (No. 902-16-276) from the Medical Section of the Netherlands Organization for Scientific Research (NWO).
Supplemental Video Files
To view supplemental videos 1–4, please refer to http://www.karger.com/doi/10.1159/000098259.
Dr. Remco T.A. Megens
Department of Biophysics, Cardiovascular Research Institute Maastricht
Maastricht University, Universiteitssingel 50, POB 616
NL–6200 MD Maastricht (The Netherlands)
Tel. +31 43 388 1661, Fax +31 43 367 0916, E-Mail R.Megens@BF.unimaas.nl
Received: July 23, 2006
Accepted after revision: October 25, 2006
Published online: December 28, 2006
Number of Print Pages : 12
Number of Figures : 8, Number of Tables : 0, Number of References : 41
Journal of Vascular Research (Incorporating 'International Journal of Microcirculation')
Vol. 44, No. 2, Year 2007 (Cover Date: February 2007)
Journal Editor: Pohl, U. (Munich)
ISSN: 1018–1172 (print), 1423–0135 (Online)
For additional information: http://www.karger.com/JVR