2.1.

Animals

All mice used in this study were kept under specific pathogen-free conditions and in accordance with Swiss federal regulations. To establish a murine aortic ex vivo model, apolipoprotein E (${\mathrm{apoE}}^{(-/-)}$)-deficient mice (The Jackson Laboratory, Bar Harbor, Maine) and C57BL/6 wild-type mice were used. All experiments were carried out at 20 to 30 weeks of age. Mice were fed a chow diet, and ${\mathrm{apoE}}^{(-/-)}$ mice were fed with a Western diet containing 21% fat supplemented with 1.25% cholesterol (2481, NAFAG, Switzerland) for five weeks from 10 weeks of age on to accelerate plaque formation, followed by a chow diet until experiments were performed.

2.2.

PDMS Embedding

A mouse was euthanized with $30\mathrm{mg}/\mathrm{kg}$ pentobarbital, and its vasculature was perfused with room temperature phosphate buffered saline (PBS) via cannulation of the left ventricle until blood was no longer visible in the perfusate. During section, the micro-catheters (5069, Medex, UK) were in situ inserted into the mouse aorta from both sides (into the aortic sinus through the left heart ventricle and into the abdominal aorta), glued (Histoacryl®, Braun Aesculap, Germany), and ligated together with surgical thread (Ethicon Prolene 3-0, Johnson & Johnson Medical GmbH, Germany). Then the aorta was extracted and dissected free of surrounding fat tissue under the binocular, as it would interfere with microscopical imaging due to additional light scattering and the increased distance of the atherosclerotic plaques from the microscope objective. The whole process was done in ice-cold DMEM medium (Invitrogen, Grand Island, NY). The branches of the aortic arch (innominate artery, left common carotid, and left subclavian artery) were occluded with tissue glue. Subsequently, the aorta-catheters ensemble was embedded in partially polymerized [polydimethylsiloxane (PDMS), Sylgard® 184, 1673921, Dow Corning, USA] and incubated at 37°C for 2 h until the PDMS polymerized sufficiently for further handling.

2.3.

Perfusion System

The circulation system that was employed to perfuse the aortas consisted of a pump (peristaltic pump: Control Company, USA; syringe pump: Infusion Pumps—Fresenius Kabi, 018020, Germany), standard medical care tubing (pyrogen free), and a warming plate to warm the perfusate before reaching the aorta in order to mimic in vivo conditions. Optionally, for long-term experiments (more than 24 h), the perfusate can be constantly refreshed with medium by a syringe pump and a drain for used medium that are both connected to the circulation (Fig. 1). The drain comprises a standard 0.22 μm syringe filter to avert bacterial contamination of the circulating medium. The speed level of the peristaltic pump was set accordingly to reproduce the physiological mean flow rate of the abdominal aorta in resting mice.¹⁶ Oxygen supply and ${\mathrm{CO}}_2$ clearance of the aortic tissue were allowed by passive diffusion through the surrounding PDMS.^14,15 For experiments, the embedded aorta was enclosed either in a conventional cell culture incubator (37°C, 5% ${\mathrm{CO}}_2$, humidified) or in the stage-top incubation chamber of a Zeiss LSM710 confocal microscope (37°C, 5% ${\mathrm{CO}}_2$, humidified).

2.4.

Two-Photon Microscopy

Two-photon microscopy was performed on a Leica SP5 microscope (Leica Mikrosysteme, Wetzlar, Germany) with TiSa-Laser excitation at 850 nm.

2.5.

Histology

Following time-lapse experiments, aortas were washed with PBS and fixed with 4% paraformaldehyde. The aortas were re-embedded in OCT (4583, Tissue-Tek, Medite, Germany), frozen down, and cryo-sectioned. Quantification of histological data was performed on 8 μm cryosections. In addition to standard hematoxylin-eosin staining, elastic fibers were stained using Elastin van Gieson (EvG) staining also named Verhoeff-van Gieson]. Lipid deposition was analyzed with Bodipy 493/503 (D-3922, Molecular Probes) staining. Leukocytes were detected using antibodies against CD45.2 (109804, BioLegend). A VCAM-1 (vascular cell adhesion molecule 1) antibody (109804, BioLegend) was used to visualize the endothelium. Apoptotic cells were detected using a caspase-3 antibody (9664s, Cell Signalling). Frozen tissue sections were fixed in 4% paraformaldehyde for 10 min, and nonspecific binding sites were blocked for 30 min with PBS containing 3% bovine serum albumin (BSA) and 0.1% Tween-20. Sections were incubated with the first and secondary antibodies in PBS containing 1% BSA. Between the steps, slides were washed with PBS. Images were acquired using an Olympus BX61 microscope (Olympus Europa Holding GmbH, Hamburg, Germany) equipped with ColorView and F-View cameras. Image acquisition settings were unchanged over the whole experiment. Digital images were acquired using the imaging software Cell^P (Olympus Imaging Europa GmbH, Hamburg, Germany) and processed using Adobe Photoshop.

2.6.

Necropsy and Organ Embedding

For comparison of the ex vivo results with in vivo conditions, ${\mathrm{apoE}}^{(-/-)}$ mice were intravenously injected with Cy5-conjugated aptamer. Mice were intraperitoneally euthanized with $30\mathrm{mg}/\mathrm{kg}$ of pentobarbital and then perfused with 15 mL of pre-warmed PBS followed by 30 mL of 2% paraformaldehyde. Subsequently, the heart and aorta were removed and incubated in 2% paraformaldehyde for 2 h, rinsed twice in PBS, and equilibrated in 30% sucrose-PBS ($\mathrm{wt}/\mathrm{vol}$) overnight. Tissues were embedded in OCT compound and snap frozen at $-70°\mathrm{C}$. Frozen samples were stored at $-70°\mathrm{C}$ until further use. Tissue sections of 8 μm were cut with a cryotome (HM560, Microm International GmbH, Germany) on to Superfrost® plus glass slides (J1800AMNZ, Thermo Scientific, Menzel GmbH & Co. KG, Braunschweig, Germany). Sections were air dried for 30 min and either stained immediately or stored at $-70°\mathrm{C}$ until further use.

3.1.

Aortic Tissue Viability

In a first step, we studied the viability of the aortic tissue after artificial perfusion. The PDMS-embedded aorta was enclosed in a cell culture incubator and perfused with cell culture medium using the peristaltic pump. After 24 h of perfusion, the aorta was fixed by perfusion of 4% paraformaldehyde, extracted from the PDMS block, re-embedded in OCT, cryo-sectioned, and stained. Eight cryo-sections are shown in Fig. 2. The first line corresponds to the aorta that was perfused for 24 h, while the second line shows an aorta that was fixed directly after harvesting. The four different stainings that were applied were hematoxylin and eosin (H&E), EvG, caspase-3, and VCAM-1 (vascular cellular adhesion molecule 1).

Fig. 1

System overview. (a) Schematic representation of the circulation system for the continuous perfusion of the embedded mouse aorta, consisting of the peristaltic pump, a warming plate, and the confocal microscope with stage-top incubator. (b) Schematic representation that shows the aorta connected to two catheters, embedded in PDMS, and perfused with cell culture medium. The basis of the heart is not depicted for clarity. Side branches of the aortic arch are ligated with tissue glue. Confocal imaging is possible with $m$ slices on $n$ positions with or without plaque. (c) Left, photograph of the connected aorta, including the basis of the heart, ligated with thread to two catheter tips. Abd. A.: abdominal aorta. Right: photograph of the aortic arch, with visible plaques at the lesser curvature of the aortic arch.

Fig. 2

Tissue viability assays. Aortic wall sections of aortas that were either perfused for 24 h with cell culture medium or directly processed after harvesting as a control (0 h). HE: Hematoxily-eosin staining shows the tissue structure. EvG: Elastin van Gieson shows the structure of elastin fibers (dark) and collagen fibers (pink). Caspase-3 staining is negative (i.e., no red dots), which precludes the presence of apoptosis. VCAM-1 staining (red fluorescent signal) stains activated endothelial surface (nuclei are stained blue). The continuous red lining proves the intactness of the endothelial layer.

The H&E staining is broadly used to visualize histo-morphology. It stains the cell nuclei blue and the (eosinophilic) cytoplasm and collagen pink. In the H&E staining, the perfused and nonperfused aorta sections are histo-morphologically similar. The collagen layers and the smooth muscle cells are intact (i.e., there is no sign of mechanical damage or necrosis). The EvG staining is generally used to stain extra-cellular matrix components, such as collagen and elastin fibers. Discontinuity of these fibers can be a sign of tissue necrosis. In the EvG staining, the perfused and nonperfused aorta sections show no discontinuities of collagen or elastin fibers. This precludes tissue necrosis. The caspase-3 staining is used as an indicator of apoptosis (programmed cell death). In the caspase-3 staining, neither the perfused nor the nonperfused aorta sections indicate signs of apoptosis. VCAM-1 is a mediator of endothelial leukocyte adhesion and is present on activated endothelium (i.e., in regions with disturbed flow), which includes endothelium on and around atherosclerotic lesions.¹⁷ The VCAM-1 staining can be used as an indicator of endothelial cell layer integrity. As shown in Fig. 2, the VCAM-1 staining produces a red fluorescent lining on the luminal margin, corresponding to VCAM-1 presence on the luminal endothelial cell surface (nuclei are stained blue). The spatial continuity of the VCAM-1 signal confirms the intactness of the endothelial cell layer. However, the VCAM-1 fluorescence signal from the perfused aorta is less intense than the signal from the nonperfused aorta. This could be an indication of endothelial stress that could be attributable to mechanical handling, 2 h of incubation without perfusion during PDMS polymerization, or perfusion of cell culture medium. Endothelial cells are very sensitive to mechanical damage (e.g., due to specimen handling) and to their chemical environment (e.g., composition of the perfusate).

These data demonstrate that the arterial tissue remains in a viable state after 24 h of perfusion, which is crucial to perform biological time-lapse experiments with the perfusion system.

3.2.

Confocal Scans

Figure 3 serves to visualize the confocal imaging and the receptor targeting concepts and shows two scans of an embedded aorta. Figure 3(c) shows a cross-section through the PDMS embedded aorta, containing an atherosclerotic plaque (yellow-pink freckles). The endothelium (orange) forms the border towards the lumen, which is filled with medium containing aptamer-Cy5 (coils with pink heads). Aptamer-Cy5 passively diffuses to the subendothelial space, where it is bound by intra-plaque macrophages via scavenger receptor A1. The sketch in Fig. 3(d) shows a macrophage with three surface bound aptamer-Cy5 molecules. The inset shows a macrophage that was stained with artificially perfused aptamer-Cy5 and imaged after tissue sectioning. The focal slice from confocal imaging [flat black rectangle in Fig. 3(c)] produces the fluorescent image, whereas the transmitted light reproduces the tissue contours, as illustrated by the sketches in Fig. 3(e) and 3(a), respectively. Figure 3(b) and 3(f) are transmitted light and confocal images from an aortic sample.

Fig. 3

Confocal imaging overview. (c) Schematic cross-section through the aorta containing a plaque. Coils with pink dots represent aptamer-Cy5. The black rectangle depicts the focal plane of the confocal microscope. In (e) and (a), fluorescence and transmitted signal, respectively, are shown schematically. In (b) and (b), real fluorescence and transmitted signal images are shown. (d) Schematic representation of a macrophage, with aptamer-Cy5 (coils with pink dots) bound to its scavenger A1 receptors (green). The inset shows the fluorescent signal of an intra-plaque macrophage, generated by the binding of circulating aptamer-Cy5. The image was acquired from a post-perfusion tissue section. The asterisk in (e) and (f) indicates the vessel lumen.

3.3.

Macrophage Receptor Targeting

As a model to demonstrate the ability of our setup to perform biological perfusate mediated receptor targeting and imaging time-lapse experiments, we chose an aptamer-Cy5 construct as a ligand and the scavenger receptor A1 (SRA1) as the target. The scavenger receptor A1 belongs to the family of endocytic pattern recognition receptors. It promotes intracellular signaling, as well as the attachment, engulfment, and digestion of pathogens.¹⁸ It is expressed by intra-plaque macrophages,¹⁹ leading to oxidized-LDL uptake and the formation of foam cells (lipid laden macrophages). A first experiment was designed to determine the saturation kinetics of aptamer-Cy5 in plaque tissue, as reported by its fluorescent signal. We chose positions on the aorta containing either plaque or no plaque (as negative control) to be repeatedly laser-scanned. The scanning interval was set to 10 min, and the experiment duration was set to 20 h. After the first scanning cycle, perfusion of the aorta with cell culture medium containing 0.5 μM aptamer-Cy5 was initiated. After the experiment, the mean fluorescence intensities of selected regions of interest inside and outside plaques were plotted over time (results not shown). The resulting curves showed a saturation behavior, reaching a plateau after $\sim 2\mathrm{h}$. Therefore, we devised the next experiment to initiate with a 2-h incubation step with 0.5 μM aptamer-Cy5 in cell culture medium, followed by a 22-h washout step with pure medium. Between the two steps, the entire circulation tubing, including the aorta, was rinsed with medium. The resulting time-plots of the mean fluorescence signal over plaque areas during the incubation step showed similar behavior as in the preceding experiment. In the washout step, after minor reduction, the signal reaches a second plateau. The nonplaque areas, serving as negative controls, only produce a weak signal after the incubation step. In the washout step, this signal decreases almost to background level (i.e., the signal intensity before perfusion with aptamer-Cy5). These results are depicted in Fig. 4. Confocal laser scans of two separate positions on the same aorta, each separated into the fluorescent signal from aptamer-Cy5, the transmitted light, and the merging of these two images are displayed. On each of these two scans, three regions of interest were chosen, two within the plaque (named 1 and 2) and one outside it (named n.c. for negative control), resulting in three mean fluorescence signal curves over time.

Fig. 4

Time-lapse confocal imaging. (a) and (b) Time-lapse experiment with aptamer-Cy5. Aptamer-Cy5 was circulated for 2 h, followed by a washout period for 22 h with cell culture medium. (a) and (b) represent two different positions on the aorta. Images from left to right are: aptamer-Cy5, transmitted light (the dotted lines indicate the plaque), merged. On each position, three regions of interest are defined: two within the plaque and one outside it. The mean fluorescent signal intensity of these regions is plotted over time. While negative control produces only a minor signal, on-plaque signal saturates rapidly and reaches a second plateau during the washout period. The scale bar in the images is 200 μm.

3.4.

Co-Localization of Receptor Targeting Signals

Figure 5(a) shows tissue sections of the mouse aorta that was used for the 24-h ex vivo perfusion experiment described in the paragraph above. In bright-field microscopy, the structure of the plaque can be distinguished from the physiological histological appearance of mouse aortas, i.e., there is an intra-luminal protrusion with increased granularity and disturbed elastin and collagen fiber continuity. In order to confirm the presence of a plaque, we used Bodipy lipid staining. Bodipy is a lipophilic fluorescent dye commonly used to stain lipid accumulations in histological specimens. In atherosclerotic plaques, intracellular lipid droplets (in foam cells) and extracellular lipid accumulations (necrotic lipid cores in advanced-stage atherosclerotic plaques) become stained. In our tissue section, the signal generated by the Bodipy staining spatially conforms to the protrusion observed in the bright-field, thus confirming the presence of atherosclerotic plaque tissue. Also, the fluorescent signal generated by aptamer-Cy5 is highly colocalized with the Bodipy signal. This proves that the enrichment of aptamer-Cy5 occurs in plaque tissue.

Fig. 5

Co-localization assays from ex vivo experiments. Mouse aorta tissue sections after ex vivo aptamer-Cy5 perfusion for 2 h, followed by a 22-h washout. The asterisk indicates the vessel lumen. (a) Co-localization of the aptamer-Cy5 signal with the Bodipy lipid staining. (b) Co-localization of the aptamer-Cy5 signal with the leukocyte marker CD45. These data indicate intra-plaque, leukocyte-specific localization of aptamer-Cy5.

Fig. 6

Co-localization assays from in vivo experiments. Mouse aorta tissue sections after in vivo application of aptamer-Cy5. The asterisk indicates the vessel lumen. (a) Co-localization of the aptamer-Cy5 signal with the Bodipy lipid staining. (b) Co-localization of the leukocyte marker CD45 with the aptamer-Cy5 signal, confirming specific targeting of the leukocyte population inside plaques.

We then wanted to test whether the aptamer-Cy5 inside the plaques is bound to leucocytes. The plaque sections were treated with an antibody for CD45, which is a membrane protein present on leucocytes. Fluorescent imaging revealed a high co-localization of the CD45 signal with the aptamer-Cy5 signal, with the signals localizing to single cells. This confirms specific binding of aptamer-Cy5 to leucocytes.

3.5.

In vivo Comparison Experiments

In order to corroborate the biological validity of the data generated with the novel ex vivo model, we performed classical (i.e., one endpoint per animal after sacrificing) in vivo experiments as a reference. To this purpose, ${\mathrm{apoE}}^{(-/-)}$ mice were treated by intravenous injection of aptamer-Cy5 in PBS to reach a blood concentration of $\sim 1.5\mathrm{\mu M}$. After 24 h, the mice were sacrificed, and the aortas were harvested, fixed, and cryo-sectioned. Tissue sections were processed and stained in identical manner as the sections from the ex vivo experiments. Imaging of these in vivo sections revealed co-localization of the aptamer-Cy5 signal with the Bodipy and the CD45 signal (Fig. 6). These in vivo results are in accordance with the ex vivo results and therefore corroborate the suitability of the ex vivo model for biological experiments.