2.1.

Animals

Experiments and procedures were approved by the local ethics committee on the use of laboratory animals. Experimental groups consisted of C57BL6/J control mice [14–16 weeks (n=7), 20–22 weeks (n=7)] and atherosclerosis-prone ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice [14–16 weeks (n=7), 20–22 weeks (n=6)]. Mice were obtained from Charles River, Maastricht, the Netherlands. ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice were fed a western diet starting at the age of $8\text{weeks}$ (WD4021.06; Hope Farms, Woerden, the Netherlands), while C57BL6/J controls were fed a chow diet (SSNIFF; Bioservices, Uden, the Netherlands). Mice were euthanized by a mixture of $\mathrm{C}{\mathrm{O}}_2$ and ${\mathrm{O}}_2$ , after which carotid artery segments (common part and bifurcation) were excised.

2.2.

Tissue Preparation and Staining

Excised carotid arteries, including bifurcation, were mounted in a perfusion chamber.^{12, 13} Side branches were ligated. The system was filled with Hanks Balanced Salt Solution (HBSS, pH 7.4) containing in mmol/l: NaCl 144, HEPES 14.9, glucose 5.5, KCl 4.7, $\mathrm{Ca}{\mathrm{Cl}}_2$ 2.5, $\mathrm{K}{\mathrm{H}}_2\mathrm{P}{\mathrm{O}}_4$ 1.2, and $\mathrm{Mg}\mathrm{S}{\mathrm{O}}_4$ 1.2. On average, intraluminal pressure was $50\mathrm{mmHg}$ (depending on the maximum pressure ligations could endure). All fluorescent markers were dissolved in HBSS and administered both intra- and extraluminally. Incubation started $30\text{minutes}$ prior to the start of image acquisition. Probes remained present during image acquisition, since this did not alter their labeling characteristics or the viability of arteries, nor reduce the image quality.

DNA/RNA markers SYTO41 $({\lambda }_{\max \text{emission}}=480\mathrm{nm})$ and SYTO13 $({\lambda }_{\max \text{emission}}=520\mathrm{nm})$ were used for labeling cell nuclei (final concentration $2.0\mu \mathrm{mol}∕\mathrm{l}$ ; Invitrogen, Breda, the Netherlands). Phycoerythrin (PE, ${\lambda }_{\max \text{emission}}=570\mathrm{nm}$ ) conjugated rat antimouse CD11b (MAC- $1\alpha$ or CD18) was applied as marker for inflammatory cells such as mature macrophages,¹⁹ monocytes, and granulocytes (final concentration $8\mu \mathrm{g}∕\mathrm{ml}$ ; BD Biosciences, Alphen aan den Rijn, the Netherlands). Oregon Green 488 conjugated CNA35 (CNA35/OG488, ${\lambda }_{\max \text{emission}}=510\mathrm{nm}$ ) was used for staining of collagen^{13, 18} (final concentration $1.5\mu \mathrm{mol}∕\mathrm{l}$ ).

Two mounted and visualized arteries ( ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice, $21\text{weeks}$ ) were also studied for lipid content. They were incubated overnight at $7°\mathrm{C}$ with the lipid marker oil red O (ORO, ${\lambda }_{\max \text{emission}}=560\mathrm{nm}$ ; Fluka Chemie, Buchs, Switzerland) dissolved in 3.7% formaldehyde solution (Merck, Darmstadt, Germany) to a final concentration of $12\mu \mathrm{mol}∕\mathrm{l}$ .^{14, 20} ORO staining resulted in a significant reduction in fluorescence of previously applied labels and loss of cell viability. Furthermore, arterial rings (approximately $1\mathrm{mm}$ long) from 3 isolated and previously mounted arteries ( ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice, $21\text{weeks}$ ) were labeled with ORO and transversally imaged for visualization of lipid content of plaques. After overnight incubation, arterial rings were labeled with SYTO13 ( $2.0\mu \mathrm{mol}∕\mathrm{l}$ , $30\text{minutes}$ ) and rings were cast in agarose gel (1.5% dissolved in HBSS, Gibco BRL 15510-027; Invitrogen, Breda, the Netherlands) as previously described.¹⁴

2.3.

Imaging

The TPLSM^{12, 14} consisted of a Biorad 2100MP TPLSM (Biorad, Hemel Hempstead, UK) with a pulsed ( $100\mathrm{fs}$ at the entrance of the scan head) and mode-locked $\left(800\mathrm{nm}\right)$ Ti:sapphire laser as the excitation source (Spectra Physics Tsunami, Mountain View, CA) connected to a Nikon E600FN microscope (60X, $\mathrm{NA}=1.0$ water dipping objective; Nikon Corporation, Tokyo, Japan). Damage to the sample was avoided by keeping excitation laser power as low as possible.²¹ On average, laser power of the excitation light was set at $15\mathrm{mW}$ for imaging in the tunica adventitia, $22\mathrm{mW}$ for imaging in the tunica media, and $30\mathrm{mW}$ for imaging in the tunica intima. In atherosclerotic lesions, laser power was maximized to $60\mathrm{mW}$ for imaging of plaque cores. Imaging rate was either $0.1\mathrm{Hz}$ with a pixel dwell time of $39\mu \mathrm{s}$ , or $0.3\mathrm{Hz}$ with a pixel dwell time of $0.12\mu \mathrm{s}$ combined with Kalman filtering for noise reduction $(n=3\text{cycles})$ . Fluorescent signals were detected by three photomultiplier tubes (PMTs): SYTO41, $460–480\mathrm{nm}$ (PMT I); CNA35/OG488, $520–550\mathrm{nm}$ (PMT II); CD11b/PE, $570–610\mathrm{nm}$ (PMT III); ORO, $540–610\mathrm{nm}$ (PMT III). From each PMT, separate images of ${512}^{*}512\text{pixels}$ were obtained, saved, and combined into a single RGB image. The filter setting for each PMT was tuned for minimal bleedthrough of fluorescent signal of the markers. Furthermore, the PMT gain was set to 100% in order to limit the required laser powers, while obtaining maximal signal intensity in deeper layers and atherosclerotic lesions of the arterial vessel wall.

Stacks of optical sections were collected for 3D reconstructions to screen for plaques in the entire mounted vessel; staining patterns in plaques differed from those of healthy carotid arteries. Plaque detection was checked using brightfield microscopy, where plaques showed up as dark and less translucent regions on a bright background. Image analysis was performed using Image-Pro Plus 6.0 and 3D-reconstructor 5.1 software package (Media Cybernetics Inc., Silver Spring, MDA).

2.4.

Histology

Histological sections were prepared to validate the penetration depth of TPLSM in plaques in mounted arteries. Subsequent to imaging of mounted arteries ex vivo, arteries (two C57BL6/J mice; two ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ , all $21\text{weeks}$ ) were perfusion-fixated and processed to 5 $\mu \mathrm{m}$ -thick histological sections and stained for collagen with Picrosirius Red²³ (0.1%, Klinipath, Duiven, the Netherlands). Histological sections were imaged using a Leica DM 5000B microscope (60X oil objective; $\mathrm{NA}=1.4$ ) and a Leica DC300FX digital camera (Leica Microsystems GmbH, Wetzlar, Germany).

2.5.

Statistics

Data are presented as medians (M) with interquartile ranges (IQR): M [IQR]. Results were tested for significance using the Mann–Whitney test. A value of $p<0.05$ was considered statistically significant. SPSS 13.0 software package (SPSS Inc., Chicago, IL) was used for statistical analysis.

3.1.

Carotid Arteries of C57BL6/J Mice

Carotid arteries of C57BL6/J mice contained no lesions (Table 1 ); The vessel wall structure was normal.¹² The tunica adventitia consisted of cells (SYTO41) embedded in collagen (CNA35/OG488). The vSMCs in the tunica media were homogeneously distributed and arranged perpendicularly to the flow direction. Endothelial cell nuclei were homogeneously distributed covering the tunica intima [Fig. 1a ]. No inflammatory (CD11b/PE-positive) cells were present in or adhering to the arterial wall in any of the control arteries. Labeled collagen was only sparsely observed as local spots in the tunica intima (14% of the mice at $15\text{weeks}$ 71% of the mice at $21\text{weeks}$ ), and not at all in the tunica media (see also Ref. 13). Hence, the endothelial cells appear nonactivated and nonpermeable for small molecules. The barrier function of internal elastic layer (IEL) and external elastic layer (EEL) appear intact, as deduced from the nonpermeability for the collagen probe.¹³

Fig. 1

Optical sections of tunica intima in a healthy common carotid segment of (a) C57BL6/J and (b)–(d) ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mouse (each $21\text{weeks}$ ). Imaging depth of optical sections (a)–(d) in the vessel wall is around $70\mu \mathrm{m}$ ( $z=0\mu \mathrm{m}$ at outside of vessel wall). Arteries were stained for cell nuclei (SYTO41; blue), CNA35/OG488 (collagen, green), and anti-CD11b/PE (inflammatory cells, red). Section (c) was stained with SYTO41 and anti-CD11b/PE. Bars: $20\mu \mathrm{m}$ . Drawing at top indicates position of displayed sections. (a) Cell nuclei of vSMCs (pink arrow) and endothelial cells (yellow arrow) are clearly visible. Collagen (green) is only sporadically labeled (encircled by white dotted line) and inflammatory cells (red) are absent. At comparable sites in common carotid segments of ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice (b), the tunica intima contains a stronger collagen signal (green) between endothelial cells (yellow arrow) and vSMCs (pink arrow). (c) Inflammatory cells (green arrow) were detected adhering to endothelial cells (yellow arrow). (d) Typical example of a class 1 plaque present in carotid bifurcation of ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice of both ages. Several inflammatory cells (red) are clearly distinguishable in the distorted tunica intima underneath the endothelial cells (yellow arrow). Furthermore, a CD11b/PE-positive foam cell (I) and a CD11b/PE-negative foam cell (II) are visible. Local collagen spots (surrounded by white dotted lines) are noticeable in the lesion area. No collagen or inflammatory cells are visible between vSMCs (pink arrow) in tunica media. Distribution and orientation of vSMCs in the tunica media are unaltered.

Table 1

Mouse characteristics and presence of atherosclerotic lesions in C57BL6/J and ApoE−∕− mice of 15 and 21weeks .

3.2.

Common Carotid Arteries of ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ Mice

In the common carotid artery of both 15- and $21\text{-}\text{week}$ -old ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice, no atherosclerotic lesions were found; the tunica adventitia and media were undisturbed. However, a very thin sheet of collagen [Fig. 1b] was observed between the endothelium and IEL of the tunica intima at both ages in 70% of the mice at $15\text{weeks}$ and 100% of the mice at $21\text{weeks}$ . In addition, inflammatory cells adhering to the luminal side of the endothelium were found in 85% of the mice at 15 weeks and 100% of the mice at $21\text{weeks}$ [Fig. 1c]. The number of cells adhering to the endothelium increased with age from 0.2 [0–0.6] cells per $(100\mu {\mathrm{m})}^2$ at $15\text{weeks}$ to 0.6 [0.3–1.2] at $21\text{weeks}$ $(p<0.05)$ . No inflammatory cells were present inside the arterial wall. These data suggest that endothelial cells were activated and permeable for the collagen probe, while IEL and EEL were still intact.

3.3.

Morphology of Lesion Development in the Carotid Bifurcation of ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ Mice

In the carotid bifurcation of most ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice of both 15 and $21\text{weeks}$ , atherosclerotic lesions were detected exhibiting strong collagen and cellular staining (72% and 100%, respectively; see Table 1). In all phases of lesion development, a thin sheet of collagen was clearly present in the tunica intima adjacent to the plaque area between endothelial cells and IEL. This indicates increased CNA35/OG488 permeability of endothelial cells in these regions, and possibly reflects increased collagen content in this layer. In young mice, plaques were of initial (class 1) or mild (class 2) type (Table 1). No advanced (class 3) plaques were found. In contrast, in mice of $21\text{weeks}$ , the majority of plaques were class 3 (Table 1).

In the initial phase of plaque development [class 1; see Fig. 1d], inflammatory cells were accumulated in the tunica intima underneath the endothelial cells. Furthermore, collagen in the lesions formed a thin sheet or local spots. In the tunica media underneath these initial plaques, the orientation and distribution of vSMCs were unaltered. No collagen or inflammatory cells were observed in the tunica media.

During further lesion progression, mild plaques developed (class 2). The first signs of a fibrous collagen-rich cap structure could already be observed at $15\text{weeks}$ [Fig. 2a ; $n=5$ plaques; class 2]. In addition, infiltrated inflammatory cells were present in the tunica intima. In class 2 plaques at 21 weeks [Fig. 2b], the area of arterial wall affected by the atherosclerotic process appeared to be larger. Collagen-surrounded areas contained a combination of inflammatory cells (e.g., foam cells; $n=4$ plaques) or poorly fluorescent areas without visible structures (i.e., lesion cores; $n=2$ plaques) [Fig. 2b]. The tunica media underneath all class 2 plaques was free of inflammatory cells. Furthermore, vSMCs were homogeneously distributed and their orientation seemed unaffected. Collagen between vSMCs was not labeled.

Fig. 2

Optical sections of typical mild plaques (class 2) as observed in tunica intima of ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice of (a) 15 and (b) $21\text{weeks}$ . Both optical sections are acquired at $z\approx 70\mu \mathrm{m}$ in the vessel wall ( $z=0\mu \mathrm{m}$ at outside of vessel wall). Arteries were labeled for nuclei (SYTO41; blue), collagen (CNA35/OG488; green), and inflammatory cells (anti-CD11b/PE; red). Drawing at top indicates position of both sections; bars: $20\mu \mathrm{m}$ . (a) Class 2 plaque in the carotid bifurcation of a young ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mouse containing several inflammatory cells and large amounts of collagen. Collagen has fibrous caplike structure (blue arrow) with a small and weakly fluorescent area (I) inside that contains no collagen and only a few inflammatory cells, both typical properties of a small (necrotic) core. In surrounding arterial wall, strongly labeled collagen without a bulblike structure is observable. (b) Part of a typical class 2 plaque in carotid bifurcation of ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice $\left(21\text{weeks}\right)$ , containing CD11b/PE-positive foam cells (green arrow) and round cells. Collagen is abundant, forming a fibrous cap (blue arrow) at edge of plaque. Adjacent to plaque area, endothelial cells are observable (yellow arrow). In the tunica media, vSMCs (pink arrow) are homogeneously distributed; their orientation is unaffected. No inflammatory cells or collagen are present in the tunica media.

Progression of a substantial number of lesions toward an advanced phase of atherosclerosis was only observed in $21\text{-}\mathrm{week}$ -old ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice (class 3; Table 1). In total, 60% of the advanced plaques [Figs. 3a, 3b, 3c, 3d ] consisted of strongly labeled collagen with a network-like appearance. Inside the collagen network, large lesion cores that contained only a little fluorescent signal were present. Infiltrated inflammatory cells were predominantly positioned at the edges between lesion cores and collagen network [Fig. 3a]. The shoulders of these class 3 plaques contained strongly labeled collagen [Fig. 3b]. In some shoulder areas of these plaques $(n=3)$ , strongly labeled collagen protruded beyond the edges of the lesion toward the healthy arterial wall [Fig. 3b].

Fig. 3

Optical sections of advanced plaques (class 3) in (a)–(c) tunica intima and (d) tunica media of carotid bifurcations in ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice aged $21\text{weeks}$ . Optical sections are acquired at (a)–(c) $z\approx 70\mu \mathrm{m}$ or (d) $z\approx 55\mu \mathrm{m}$ in the vessel wall ( $z=0\mu \mathrm{m}$ at outside of vessel wall). Arterial segments (a)–(d) were stained for nuclei (SYTO41; blue), collagen (CNA35/OG488; green), and inflammatory cells (anti-CD11b/PE; red). Segment (b) was stained for collagen only (CNA35/OG488). Bars: $20\mu \mathrm{m}$ ; drawing at top indicates position of sections. (a) Advanced plaque with a network-like collagen structure, several lesion cores with weak fluorescent signal (I), and many inflammatory cells (green arrow) and non-inflammatory cells (white arrow). (b) Plaque shoulder (red dotted line) containing large amounts of collagen, which protrude into the area adjacent to the plaque (lower part of the section). Blue is autofluorescence of elastic laminae. Note that the fluorescence in (b) is optimized for the intensity of the adjacent area. Thus, the fluorescence in the shoulder is saturated. Class 3 plaques also appear as an interconnected group of smaller lesions with numerous inflammatory cells, and several lesion cores (c). Plaque borders and lesion cores (I) are less evident; collagen is abundant with less defined appearance. (d) The tunica media flanking advanced plaques contains areas with disrupted vSMCs layer (encircled by red dotted lines) and vSMC with altered orientation (green arrow).

Other class 3 plaques (40%) manifested as interconnected groups of smaller lesions with less evident plaque borders [Fig. 3c]. Such lesions consisted of many small lesion cores with weak fluorescent signal surrounded by large amounts of labeled collagen with a less defined network-like appearance. Many inflammatory cells of various sizes and shapes were observed throughout the lesion.

Application of a lipid stain (ORO) in mounted arteries with class 3 lesions revealed that the dark lesion cores inside plaques had a lipid-rich content [Fig. 4 ]. Imaging of the content of the cores in fresh (but not viable) arterial rings labeled with ORO [Figs. 4b and 4c] demonstrated that part of the lipids was situated inside (foam) cells as small droplets.

Fig. 4

(a) Mounted artery of $21\text{-}\text{week}$ -old ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice, containing a class 3 plaque in the bifurcation, labeled for nuclei (SYTO41; blue), collagen (CNA354/OG488; green), and lipids (ORO; red). Bars: $10\mu \mathrm{m}$ . Image is acquired at $z=74\mu \mathrm{m}$ (outside vessel wall $z=0\mu \mathrm{m}$ ). Note the lipid content in core of the lesion, collagen in the fibrous cap (white arrow), and vSMCs in tunica media (red arrow). (b), (c) Optical sections of fresh arterial rings stained with ORO and SYTO13 (nuclei, green) further demonstrate that lipids (red) are mainly situated as small droplets inside cells in the lesion cores. As in mounted atherosclerotic arteries, the maximal depth of imaging in fresh arterial rings inside the plaque core was limited to $z_{\max }\approx 100\mu \mathrm{m}$ .

In the larger class 3 lesions $(n=4)$ , the fluorescent signal intensity drastically and abruptly decreased toward the lumen [Fig. 5a ]. As a result, we were unable to visualize the complete lesion area in depth [Fig. 5b); no fibrous cap or endothelial cell lining was visible using TPLSM. However, histological longitudinal sections stained for collagen revealed that such lesions do contain fibrous caps with endothelial cells [Fig. 5c].

Fig. 5

(a) Optical section and (b) 3-dimensional reconstruction of a z-stack obtained in the tunica intima of class 3 plaque in a $21\text{-}\text{weeks}$ -old ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice. The z-stack in (b) runs over a region from $z=45\text{to}84\mu \mathrm{m}$ with a step size of $1.5\mu \mathrm{m}$ in the arterial wall ( $z=0\mu \mathrm{m}$ at the outside of the vessel wall). Drawing at the top indicates position of section; bars: $20\mu \mathrm{m}$ . Artery was stained for collagen (CNA35/OG488; green), nuclei (SYTO41; blue), and lipids (ORO; red). (a) The lesion contains several lipid-rich lesion cores (I); the collagen has a network-like appearance. (b) The plaque lacks a fibrous cap. Unlike TPLSM images, histological sections of equivalent class 3 plaques (c) stained with picrosirius red (collagen, red) do contain a fibrous cap (green arrow) that consists of collagen. As in TPLSM images, lesion cores (I) are abundant.

The tunica media was affected in all class 3 plaques. Although no inflammatory cells were detected in this layer, the vSMC layer was disrupted and the orientation of vSMCs was altered [Fig. 3d]. Sometimes, collagen between vSMCs in these plaques was labeled, indicative of local disruption of IEL $(n=4$ plaques). The tunica adventitia of carotid arteries in both non-plaque regions [in ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ and wild-type mice; Fig. 6a ] and plaque regions [in ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice; Fig. 6b] contained large amounts of collagen with a typical wavelike appearance. No apparent changes to the collagen structure were found between plaque regions and non-plaque regions. Furthermore, the tunica adventitia contained numerous inflammatory cells and non-inflammatory cells in both plaque and non-plaque regions.

Fig. 6

Optical sections in (a) normal tunica adventitia of a C57BL6/J and (b) tunica adventitia of a plaque in an ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice, both $21\text{weeks}$ old. Drawings on top indicate position of sections (depth $\approx 20\mu \mathrm{m}$ ); bars: $20\mu \mathrm{m}$ . Arteries were stained for collagen (CNA35/OG488, green), nuclei (SYTO41, blue), and inflammatory cells (anti-CD11b/PE, red). In both sections, thin, undulating collagen fibers rearrange into thicker wavelike bundles of collagen. In between collagen, nuclei of cells are visible. (a) In the healthy artery, only a few inflammatory cells (white arrow) are present; (b) in the plaque, the number of inflammatory cells (white arrow) is much higher.

3.4.

Inflammatory Cell Density

During lesion development, the inflammatory cell density in plaques increased significantly with age $(p<0.05)$ and with plaque progression ( $p<0.01$ ; Table 2 ). The tunica media in plaque regions did not contain any inflammatory cells.

Table 2

Intraplaque inflammatory cell density and association between inflammatory cells and collagen (medians with interquartile ranges).

In the tunica adventitia, the total cell density tended to be higher in plaque regions than in non-plaque regions in mice of both ages {23.6 [31.1–17.8] cells per $\left(100\mu {\mathrm{m}}^3\right)$ vs. 17.6 [23.6–13.3]; $(p=0.07)$ }. The relative number of inflammatory cells in the tunica adventitia in mice of both ages was increased in plaque regions (0.58 [0.63–0.51]) compared to non-plaque regions (0.24 [0.36–0.18] $p<0.001)$ .

3.5.

Relationship Between Collagen and Inflammatory Cells in Developing Plaques

In plaques as a whole, the percentage of inflammatory cells in direct contact with collagen did not change significantly with age or plaque progression (Table 2). In fibrous caps, the median score of the association between collagen and inflammatory cells was significantly higher at $21\text{weeks}$ (1.0 [2–1]) than at $15\text{weeks}$ (0.5 [1–0]; $p<0.05)$ . Note that fibrous caps were present only in class 2 and class 3 plaques. Moreover, class 3 plaques were only present in $21\text{-}\text{week}$ -old ${\mathrm{ApoE}}^{\text{-}∕\text{-}}$ mice.