2.1.

Experimental Setup

A Fourier domain OCT (FD-OCT)¹¹ system was developed with a center wavelength of $840\mathrm{nm}$ , a spectral width (FWHM) of $45\mathrm{nm}$ , and an optical power of $1\mathrm{mW}$ . A broadband superluminescent diode (SLD) served as the light source. Light was transmitted by a fiber coupler to the scanner head (Fig. 1 ), in which the light was collimated into a free-space beam of $2.4\mathrm{mm}$ in diameter $(1∕{\mathrm{e}}^2)$ and divided by a beamsplitter cube into a reference and a sample beam. The sample beam was deflected by two galvanometer scanners in the $x$ and $y$ direction. Deflected light was focused onto the sample by an achromatic lens (L2) with a focal length of $25.4\mathrm{mm}$ and a diameter of $15\mathrm{mm}$ . Because of the much larger diameter of L2 compared to the sample beam diameter, the effective NA of 0.05 of the objective is determined by the focal length of the collimator, the NA of the fiber, and the focal length of L2. Backscattered light from the probe was superimposed with the reference beam and again coupled into the single-mode fiber. Interfered light was coupled into the spectrometer by a collimator with a focal length of $40\mathrm{mm}$ , and was spectrally resolved by a diffraction grating with $1200\text{lines}∕\mathrm{mm}$ before being focused onto the complimentary metal oxide semiconductor (CMOS) line detector ( $1024\text{pixels}$ , $7.8\times 125\mu \mathrm{m}$ ) by an achromatic lens with a $75\text{-}\mathrm{mm}$ focal length. This OCT system allows the measurement of complex biological tissue in a range of $2\mathrm{mm}$ , with an axial resolution of $8\mu \mathrm{m}$ in air ( $6\mu \mathrm{m}$ in tissue with refractive index $\mathrm{n}=1.33$ ), a lateral resolution of about $7\mu \mathrm{m}$ , and a scanning rate of $1.2\mathrm{kHz}$ (A-scan rate). Although the measurement range is $2\mathrm{mm}$ , its sensitivity range is limited to a range of $370\mu \mathrm{m}$ by the Rayleigh length of the sample beam. A schematic of the experimental setup and beam course in the scanner head is shown in Fig. 1. The camera, which is coupled by a dichroic mirror, allowed orientation on the sample surface, and can be used with additional optics for microscopy. Optimal position in relation to the sample was achieved by mounting the scanner head on a three-axis translation stage.

Fig. 1

Experimental setup to investigate in-vivo vasodynamics in the mouse model. Near-infrared (NIR) light is transmitted by fiber optics [single mode fiber (SMF)] and coupled by a collimator (C) with a focal length of $12\mathrm{mm}$ into the scanner head (SH). The light beam is divided by a beamsplitter (BS) cube into a reference and sample beam. The sample beam is then deflected by two galvanometer scanners ( $xy$ GU) into $x$ and $y$ directions for plane scanning of the sample. The sample beam is focused on the sample by an achromatic lens (L2) with a focal length of $25.4\mathrm{mm}$ . The reference light is focused L1 to the reference mirror (RM). Backscattered light from the sample is superposed with the reference light at the beamsplitter cube. This interference signal is transmitted into the spectrometer and spectrally resolved by a diffraction grating $(1200\text{lines}∕\mathrm{mm})$ . The interference spectrum is acquired by a silicon line detector. Using a fast Fourier transformation (FFT), the interference signal is converted into the corresponding depth information. Additionally, in the course of the sample beam, a dichroitic mirror (DM) is inserted for adaptation of a video camera (VC). Further abbreviations: IU is illumination unit for VC; $\lambda$ is the wavelength; $z$ is the depth; $I$ is the Intensity; and $A$ is log. Amplitude is in decibels.

2.2.

Mouse Model

All experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, 7th edition, National Academy Press, Washington, D.C., 1996). The study was approved by the governmental animal care and use committee. The murine saphenous artery and vein of the right hind leg in C57BL/6 mice aged 6 or $20\text{weeks}$ was investigated. Anesthesia was delivered by peritoneal application of a mixture of 95% Ketanest $(10\mathrm{mg}∕\mathrm{ml})$ and 5% Xylazin $(20\mathrm{mg}∕\mathrm{ml})$ using a dose of $10\text{-}\mu \mathrm{l}∕\mathrm{g}$ body weight. Before imaging, the skin covering the saphenous artery was incised and a $5\times 5\text{-}{\mathrm{mm}}^2$ skin flap was removed to obtain access to the vessel area. Phosphate-buffered saline (PBS) containing $80\text{-}\mathrm{mM}$ ${\mathrm{K}}^{+}$ was used to induce vasoconstriction (VC). Vasodilation (VD) was induced with $1\text{-}\mathrm{mM}$ sodium nitroprusside solution $(\mathrm{PBS}+\mathrm{SNP})$ . The solutions were applied on the outer vessel wall. During investigation, the mouse was fixed to a temperature-controlled operation table for small rodents.

2.3.

Two-Dimensional Image Acquisition—Vasodynamics

Throughout the image acquisition procedure, the vessel-containing area was covered with buffer (PBS) to prevent the tissue from drying out. To measure vasodynamics, images of the inner vessel diameter were taken at identical lateral positions and recorded by performing B-scans consisting of 300 A-scans. An image acquisition rate of 4 B-scans per second was used. VC and VD were documented by initial 30 B-scans to determine the unaffected vessel diameter. Next, image acquisition was stopped and the buffer was exchanged for a buffer solution containing vasoactive stimuli (VC: $\mathrm{PBS}+{\mathrm{K}}^{+}$ ; VD: $\mathrm{PBS}+\mathrm{SNP}$ ). Resultant changes in diameter were imaged by additional 300 B-scans. The final values of the inner diameter after VC and VD were determined using the last 50 images (equivalent to the last $12\sec$ of the time-resolved image stack) of the OCT image stack acquired during VC and VD. The time-resolved image stacks of the static and dynamic portion of each stimulus represent a period of approximately $75\sec$ .

2.4.

Three-Dimensional Image Acquisition—Morphology

The investigated area was also mapped by 3-D image stacks $(2\times 2\times 2{\mathrm{mm}}^3)$ , under control conditions and after completion of VC and VD.

2.5.

Image Quantification

Flowing blood was seen in these OCT images as a specific black shadow, caused by the movement of scattering objects, e.g., blood cells. This movement reduced the interference signal by fringe washout.^{12, 13} The saphenous artery was seen in OCT images as a black semicircular structure (Figs. 2 and 3 ). The pronounced border between the dark blood and bright vessel wall was used to quantify the inner diameter of the artery. Quantification was carried out using existing algorithms from commercially available software, National Instruments Vision Assistant 8.2.1 (NIV), in combination with National Instruments LabVIEW protocols. In the first B-scan of the time-resolved stack, the user defined the region of interest (ROI), which included the saphenous artery, by a rectangle. The border between the dark lumen and bright vessel wall was detected with NIV routine cycle detection. The protocol was modified to detect the visible semicircular structure representing the saphenous artery lumen. The inner diameter of the semicircle and the coordinates of the center point detected in the first B-scan diameter were used to define the ROI in the following B-scan. The ROI was defined by the user only in the first B-scan. All subsequent B-scans in the time-resolved image stack were analyzed automatically. The detected inner diameters and the corresponding times were saved as spreadsheet files. To characterize the dynamics of inner diameter change in response to vasoactive stimuli, a sigmoid function [Eq. 1] was fitted to the measured dataset of time-dependent inner diameter $d\left(t\right)$ . The coefficient $\Delta d$ describes the total change of inner diameter, $t_{1∕2}$ the elapsed time for half of inner diameter change, $\tau$ the time constant of diameter change, $d_0$ the initial diameter, and “ $t$ ” the time.

Fig. 2

(a), (b), and (c) 2-D cross sectional OCT images and (e) and (f) volume rendering illustrations of the investigated area under (a) and (d) control conditions, (b) and (e) VC, and (c) and (f) VD. The saphenous vein (white arrow), the saphenous artery (black arrow), and the nervus vastus medialis (gray arrow) are indicated. Vessel wall thickness $\mathit{wt}$ of the saphenous artery is highlighted by double arrows. The resolution of our OCT system is sufficient to image the lipid-filled vacuoles of perivascular adipocytes (lv).

Fig. 3

Temporal changes of the inner diameter after induction of vasoconstriction $\left({\mathrm{K}}^{+}\right)$ and vasodilation (SNP). Left side: 2-D cross sectional OCT image taken at resting conditions of saphenous vein (white arrow) and saphenous artery (black arrow). Right: cross sectional images of the acquired time-resolved stacks. The white line in the left image marks the cutting plane through the time-resolved stacks. The temporal resolution of the inner diameter of the saphenous artery and vein are visible as black bands in the right image. After application of ${\mathrm{K}}^{+}$ , the inner diameter decreases, whereas SNP application enlarges the vessel diameter.

Vascular function dynamics were characterized by the time to half maximum $t_{1∕2}$ (inflection point of the fitted function) during VC and VD, and the velocity of inner diameter change at the inflection point. The maximum velocity $v_{\max }$ of VC and VD corresponds to the slope at $t=t_{1∕2}$ , where the second time derivative of Eq. 1 is null and is given by:

3.1.

Morphology

To investigate in-vivo vascular function in small vessels, we chose the murine saphenous artery and vein of the right hind leg because of its advantageous anatomical position beneath the skin on the musculus vastus medialis, its suitable inner diameter of approximately $100\mu \mathrm{m}$ , and its rapid response to vasoactive stimuli. Figure 2a is a representative OCT grayscale cross section of a murine artery and vein, as well as the surrounding tissue and other structures including nerves. The surrounding tissue consists of perivascular fat and appears as a foamy structure. Lipid containing vacuoles (lv; left arrow) in perivascular adipocytes (approximately $80\mu \mathrm{m}$ in diameter) are visible beneath the surface, which is represented by connecting tissue of muscle fascia. The distinctive ring structures encircling the vessel lumens are the arterial and venous vessel walls. Vessel wall thickness $\mathit{wt}$ is indicated by a double arrow. The lumen and vessel wall of the saphenous vein (white arrow) and artery (black arrow), as well as the nerve (nervus vastus medialis; gray arrow), are shown under control conditions [Fig. 2a], and after VC [Fig. 2b] and VD [Fig. 2c]. Morphology is revealed by 3-D volume rendering of murine tissue containing the same structures before [Fig. 2d] and after VC [Fig. 2e] and VD [Fig. 2f].

3.2.

Vasodynamics

The morphological vessel parameters used to characterize vessel dynamics of mice aged 6 (6w) and 20 (20w) weeks were: 1. the inner diameter $d$ as determined from the blood-filled lumen confined by the vessel wall, and 2. the vessel wall thickness $\mathit{wt}$ of the saphenous artery and vein. During vasoconstriction (VC) and vasodilation (VD), these parameters ( $d_{\mathrm{VC}}$ , $d_{\mathrm{VD}}$ , ${\mathit{wt}}_{\mathrm{VC}}$ , and ${\mathit{wt}}_{\mathrm{VD}}$ ) differ from resting conditions ( $d_{\mathrm{res}}$ , ${\mathit{wt}}_{\mathrm{res}}$ ). To define vessel reactions to vasoactive stimuli, we determined the dynamic range $dr=\mid d_{\mathrm{VD}}-d_{\mathrm{VC}}\mid$ and dynamic ratio $(d_{\mathrm{VC}}∕d_{\mathrm{VD}})$ as a degree of vessel stiffness. The relative wall thickness ${\mathit{wt}}_{\mathrm{rel}}={\mathit{wt}}_{\mathrm{res}}∕(d_{\mathrm{res}}∕2)$ characterized the vessel type. In previous studies, the relative wall thickness of the aorta was characterized by a value of 0.2, which increased to 0.5 in arteries and to 1.0 in arterioles.¹⁴ This increase was found to be more moderate in the venous bloodstream, starting at 0.1 for the vena cava, 0.2 in veins, and 0.3 in venules. The maximum velocity $v_{\max }$ of inner diameter change provides information about how fast the vessel reacts to vasoactive stimuli, and the time to half maximum diameter change $t_{1∕2}$ indicates how long it takes until these reactions are finished.

Changes in the inner diameter of the saphenous artery and vein during VC and VD were recorded in a quantitative manner using the FD-OCT technique (Table 1 ). Analyzing the saphenous vein, we detected only slight inner diameter and wall thickness changes between both age groups in response to vasoactive stimuli. In contrast, the inner diameter of the saphenous artery of $6\text{-}\text{week}$ -old mice altered from $122\pm 6\mu \mathrm{m}$ under resting conditions to $10\pm 3\mu \mathrm{m}$ after VC $\left(d_{\mathrm{VC}}\right)$ and $227\pm 4\mu \mathrm{m}$ after VD $\left(d_{\mathrm{VD}}\right)$ . At $20\text{weeks}$ of age, the inner diameter of the saphenous artery had increased due to the growth of the animal to $172\pm 9\mu \mathrm{m}$ (resting conditions), $38\pm 7\mu \mathrm{m}$ $\left(d_{\mathrm{VC}}\right)$ , and $245\pm 8\mu \mathrm{m}$ $\left(d_{\mathrm{VD}}\right)$ . While the dynamic range of the vessel diameter did not change with age ( $6\mathrm{w}=217\pm 5\mu \mathrm{m}$ and $20\mathrm{w}=207\pm 9\mu \mathrm{m}$ ), the vessel wall stiffness increased significantly by a factor of 4 from $0.04\pm 0.01$ (6w) to $0.16\pm 0.03$ after $20\text{weeks}$ of age. In both age groups, the arterial vessel wall was always thicker than the venous vessel wall. Arterial wall thickness in $6\text{-}\text{week}$ -old mice increased from $24\pm 1\mu \mathrm{m}$ at resting conditions to $38\pm 2\mu \mathrm{m}$ after VC, and decreased to $18\pm 1\mu \mathrm{m}$ after VD. In $20\text{-}\text{week}$ -old mice, $\mathit{wt}$ changed from $22\pm 1\mu \mathrm{m}$ during resting conditions to $46\pm 2\mu \mathrm{m}$ after VC and $18\pm 1\mu \mathrm{m}$ after VD. The ${\mathit{wt}}_{\mathrm{VC}}$ and ${\mathit{wt}}_{\mathrm{rel}}$ of $6\text{-}\text{week}$ -old mice showed significant differences compared to $20\text{-}\text{week}$ -old animals. The relative wall thickness ${\mathit{wt}}_{\mathrm{rel}}$ of the arteries changed from a value of 0.44 (6w) to 0.27 (20w). The measured ${\mathit{wt}}_{\mathrm{rel}}$ for arteries is in the range of the relative wall thickness described in the literature.¹⁴

Table 1

Quantification of the inner diameter d and vessel wall thickness wt of the saphenous artery and vein under resting condition dres and wtres , after vasoconstriction (VC) dVC and wtVC , and after vasodilation (VD) dVD and wVD . Dynamic range dr=∣dVD−dVC∣ and dynamic ratio dVC∕dVD , and the relative wall thickness wtrel=wtres∕(dres∕2) were compared between mice aged 6 (6w) and 20 (20w) weeks. p*<0.05 ; p**<0.01 ; p***<0.001 versus 6w.

Figure 3 shows a representative image of changes in the saphenous artery inner diameter over time (black arrow). Here, cross sections from the time-resolved image stacks under control, VC, and VD conditions were sequentially combined to clearly demonstrate inner diameter changes after the application of vasoactive stimuli. In the saphenous artery lumen, an alternating light and dark pattern was observed after VD. This likely represents the detection of blood cells with reduced flow velocity during the diastolic part of the cardiac cycle by the OCT system.

Figure 4 shows an example of changes in the saphenous artery inner diameter during VC and the corresponding graph of the fitted sigmoid function. Inner diameter was determined by the algorithm described in the image quantification section of the methods. The dynamic parameters $v_{\max }$ and $t_{1∕2}$ during VC and VD for the saphenous arteries in each age cohort are shown in Table 2 . Dynamic responses were not detectable for the saphenous vein. Although the reaction velocity $\left(v_{\max }\right)$ was faster in response to VD than VC in both age groups, comparison of $t_{1∕2}$ indicates that VC progressed more rapidly ( $6\mathrm{w}=4.5\pm 0.4\mathrm{s}$ ; $20\mathrm{w}=6.5\pm 0.4\mathrm{s}$ ) than VD ( $6\mathrm{w}=8.5\pm 1.2\mathrm{s}$ ; $20\mathrm{w}=9.8\pm 0.5\mathrm{s}$ ).

Fig. 4

Time course of inner diameter change during vasoconstriction (VC) measured by FD-OCT (gray circles) and the fitted sigmoid function (black line). At time $t_0$ , the vasoconstrictor ${\mathrm{K}}^{+}$ was applied. $\Delta \mathrm{d}$ indicates the change from the initial inner diameter $d_0$ to the final inner diameter $d_{\mathrm{VC}}$ during VC. The dynamics of inner diameter change were quantified using the time of half VC $t_{1∕2}$ and the maximum velocity $v_{\max }$ of inner diameter change, reflecting the slope at the inflection point. VD dynamics were analyzed in a similar manner.

Table 2

Quantification of maximum velocity vmax in μm∕s of inner diameter change and time to half maximum VC or VD t1∕2 in s during VC and VD in 6-week (6w) and 20-week -old (20w) mice. No dynamic behavior was detected in the saphenous vein (n.d.). p**<0.01 versus 6w.