2.1.

Unilateral Femoral Artery Ligation

Three-month-old C57BL/6 male mice, weighing 20 to 30 g, were housed in a temperature-controlled room, fed with mouse food and tap water. All experimental animal procedures performed in this study conform to the guidelines of the US National Institutes of Health. Mice were anesthetized with isoflurane (0.2 l/min O₂, 0.8 l/min air) and placed in the supine position on a circulating warm water blanket, which maintained body temperature at 37°C. After shaving and disinfection, a 1 cm transversal skin incision was made in the groin. The femoral artery was identified, as shown in Fig. 1, and aparted from the femoral vein. An artery ligation thread was put through the opening of 1 to 2 mm between the femoral artery and vein proximally, just below the joint point between the profunda femoral artery and epigastric artery (see Fig. 1). Then, the animal was moved to a custom-made stereotaxic stage under isoflurane anesthesia and then subject to UHS-OMAG and LSI imaging. After control images were captured, the femoral artery was occluded using triple surgical knots without changing imaging position. After acquiring follow-up imaging data, the incision was closed with silk sutures.

Fig. 1

Schematic diagram illustrating the occlusion site and imaging site for unilateral hind-limb ischemia. UHS-OMAG C-scan (2.5×2.5 mm²) and B-scan are shown by solid line box and dashed line, respectively.

2.2.

UHS-OMAG In Vivo Imaging

The system setup used to achieve UHS-OMAG is similar to that used in our previous work.^{11, 17} Briefly, a superluminescent diode with a central wavelength of 1310 nm and a bandwidth of 65 nm was used as the light source providing a ∼8.9 μm axial resolution in the tissue. In the sample arm, a 30 mm focal length objective lens was used to achieve ∼9 μm lateral resolution. The output light from the interferometer was sent to a spectrometer with spectral resolution of ∼0.141 nm, producing a total depth range of ∼2.2 mm in the tissue. To achieve ultrahigh sensitive imaging of the flow, we used a high speed camera with 47,000/s of line rate and applied a novel scanning protocol in this system, as described in Ref. 17. With this upgrade, a whole 3D data volume covering ∼2.56 mm in both the X and Y directions was captured in 10 s. The detectable blood flow ranged from ∼4 μm/s to ∼23 mm/s, a range sufficient to image the blood supply in muscle tissue.

Before UHS-OM3AG imaging, the right hind-limb of the mice was shaved and depilated. To avoid the influence of skin on imaging quality, a 5 mm longitudinal skin incision was made to expose our imaging objective, the gastrocnemius muscle (see Fig. 1); this area was kept moist under a piece of plastic foil. During the imaging, the animal was immobilized in a custom-made stereotaxic stage and was lightly anesthetized with isoflurane. The body temperature was kept at 37°C by use of a warming blanket; temperature was monitored by a rectal thermal probe throughout the experiment. Before ligating the femoral artery, a control C-scan (3D) and repeated B-scan (2D, M-B mode) of 10 s were acquired as the baseline for future comparison. The imaging sites of C- and B-scan are shown by a solid line box and dashed line, respectively, in Fig. 1. The two ends of ligating thread were then held by hand. Once ligation started, M-B scan was initiated and continued for 1 min. Then, the final C-scan recorded the perfusion at 5 min after onset of ligation at the same site as the control. Follow-up UHS-OMAG images were taken at 48 h, and 1 week after the femoral artery was banded.

2.3.

Laser Speckle-Contrast Imaging for Comparison

A laser diode (808 nm) and CMOS camera were positioned above the gastrocnemius muscle such that a 4×4 mm² area was imaged. Raw speckle images were acquired at 300 Hz for 5 s and a speckle-contrast image was computed from each raw speckle image according to its definition, the ratio of the standard deviation of the intensity to the mean intensity of speckle pattern in the imaging region.²⁰ Speckle contrast is a measure of speckle visibility, which is related to the motion of the scattering particles, in this case blood flow. Each set of 20 speckle-contrast images was averaged together. A relative blood flow image was then obtained by using speckle-contrast values. Blood flow was imaged before ligation introduction, during ligation, at 5 min, and 48 h after ligation.

3.1.

Imaging of Muscle Microvascularization

In Fig. 2, the UHS-OMAG in vivo imaging results produced by one volume dataset [2.5×2.5×2 (x-y-z) mm³] are shown. Figure 2a shows one typical cross-sectional image within the UHS-OMAG structural volume, which is identical to the conventional optical coherence tomography (OCT) image, where the borders between different muscle groups are visualized (see dashed lines, for example). Figure 2b gives the corresponding blood flow image obtained from the UHS-OMAG algorithm,¹⁷ where the capillary flows within the cross section of skeletal muscle are abundant from the surface to a depth of 700 μm, as shown by a bar. Due to strong light attenuation in the muscular tissue, the capillaries are not readily apparent below 700 μm, but the arterioles with faster flow velocity are still identified as illustrated by arrows. Because the flow and structural images are co-registered in UHS-OMAG, they can be reconstructed in one 3D image [Fig. 2c], allowing the observation of detailed information regarding microvascular blood perfusion. For example, within this volume data set, the 3D view [Fig. 2c] can be rotated, cut from any angle to examine, in detail, the blood flows at different depths within tissue. With the application of the volume segmentation algorithm²¹ and by the reference to the typical schematic [Fig. 2g] of a vascular network in skeletal muscle,²² the blood perfusion in the muscle tissue was separated into three layers, shown in Figs. 2d, 2e, 2f, respectively. As articulated in Fig. 2g, the first layer [∼200 μm in thickness, Fig. 2d] gives a segment of the venules (V) and a thinner portion of capillary network, in which the lateral capillary connections exist besides the typical longitudinal orientation of the capillaries (shown by arrows). The second layer [with a thickness of ∼500 μm, Fig. 2e] depicts a few branches of arterioles or venules providing connections between dense capillary bundles, which correspond to the transverse arterioles (TA) or collecting venules (CV) designated in Fig. 2g. Another consistent finding in our segmented images was that arterioles seem to lie in the deeper layer [thickness of ∼500 μm, Fig. 2f], connecting with their neighboring venules in the superficial layer through the interconnected branches in the middle layer.

Fig. 2

Typical in vivo UHS-OMAG images of detailed microcirculation within skeletal muscle in mice are compared with the typical schematic obtained from histological work. (a) is one UHS-OMAG cross-sectional image (B-scan) of microstructures, and (b) is the corresponding blood flow image where the capillary flows in a layer shown by a bar are readily apparent. (c) shows 3D volumetric rendering of blood perfusion within scanned muscle tissue volume. After segmentation, the 2D x-y projection views of muscle blood microflows in three layers [(d)–(f)] are in agreement with the vascular network described in schematic (g). Venule (V); arteriole (A); collecting venule (CV), and transverse arteriole (TA).

3.2.

Dynamic Imaging of Skeletal Muscle Perfusion

The changes in blood perfusion in skeletal muscle in response to femoral artery ligation were characterized in anesthetized mice. Figure 3 shows a set of representative time-course blood perfusion patterns from the gastrocnemius muscle area of the same hind-limb. Compared to the perfusion image before ligation [Fig. 3a], the image at 5 min [Fig. 3b] reflected the compromised blood supply to the limb caused by the acute femoral artery ligation. In our preliminary trials, the reduction in blood supply was always pronounced in the limbs within 1 h after ligation; here, only the image at 5 min is presented. Soon after, the collateralization process was initiated in the ligated limb. Figure 3c shows that the imaging tissue volume was collaterally perfused at 48 h. By comparing Figs. 3a and 3c, we find that, although the muscle perfusion was quickly re-established, significant changes in perfusion distribution persisted during this period. It was also evident that the microvessels were dilated after their blood flow was restored. As illustrated in Fig. 3d, a significant restoration of perfusion in the affected skeletal muscle was observed one week after ligation. It seems that the blood perfusion at this time point surpassed that of the baseline, despite the flow stoppage in the acute period after ligation. This suggests that the collaterally dependent perfusion reserve in the hind-limb has the capacity to reach, or even to exceed, the normal resting perfusion level at a higher demand, which agrees with the results previously published in Refs. 23 and 24. Specifically, UHS-OMAG is able to identify the collateral vessels [arrows in Fig. 3d], that were developed from an arteriole, at 48 h after ischemic onset, which was thought to be associated with the augmenting blood supply in this period.

Fig. 3

Representative functional microcirculation network in the muscle in response to femoral artery ligation. (a)–(d) show time course of perfusion changes before ligation (a), at 5 min, (b) 48 h and, (c) 1 week (d) after ligation. Acute perfusion reduction and restoration can be observed. The brighter microvessels observed in (c) are indicative of increased blood flow associated with collateral perfusion. In (d), arrows indicate that new patterns of collateral vessels are developed during perfusion recovery. Cross-sectional images (e)–(h) detect the variation of blood perfusion before ligation (e) at 5 min, (f) 48 h and, (g) 1 week (h) after ligation along tissue depth.

Detected variation in blood perfusion along tissue depth, as seen in the cross-sectional UHS-OMAG images [Figs. 3e, 3f, 3g, 3h], provides more useful information as to the collateral perfusion after hind-limb ischemia. By comparing the baseline [Fig. 3e] to the 5 min image [Fig. 3f], we find that the ligation had a more pronounced effect on the flow in the capillaries than that on the flow in arterioles. At 48 h after ligation, the superficial tissue layer recovers the blood flow much faster than the deep layer, and vessel diameters seemed to become larger than those in the baseline. One week after ligation, no further difference can be observed in the capillary blood flow within the superficial layer; however, the arteriole blood flow in the deep layer was seen to become higher, and some newly developed collateral vessels were evident. Besides providing the perfusion evolution during the ligation, both projection and cross-section UHS-OMAG maps (Fig. 3) also show that the recovery response was inhomogeneous across different types of vessels over time.

The current temporal resolution (180 fps) of the UHS-OMAG system allowed us to observe the acute flow response and high-frequency blood flow changes in the muscle suffering from ligation. By normalizing the results from all time points to that of the baseline preligation period, the fractional flow signal changes were computed based on consecutive B-scan images. Figure 4a is a representative cross-sectional B-scan at the baseline, showing the muscular microcirculation involved during the time-course analysis. Figures 4b and 4c demonstrate the values of fractional flow signal change in each B-scan from two individual vessels [the arrows in Fig. 4a], along with a fitting line showing the flow reaction trends 45 s after ligation. Figure 4d shows the averaged time course for the flow signal changes over the region denoted by the dashed box in Fig. 4a. The temporal sequences of flow response for this representative region revealed a clear decrease that deviates from the baseline up to ∼60% within the first 10 s, indicating an acute phase following ligation. This gradually reached a constant level near 70% during the following 30 s, suggesting a chronic phase after the acute one.

3.3.

Comparison of UHS-OMAG Imaging With LSI Imaging

The comparison between UHS-OMAG and LSI imaging was performed on the same type of animal and under the same surgical conditions. The LSI images provided the maps of relative blood flow. In the baseline image [Fig. 5a], blood vessels were evident, but capillary bundles were difficult to distinguish, due to the limited sensitivity of LSI imaging to the slow blood flows. The ligation was performed in the same limb, causing a rapid reduction in muscle perfusion [Fig. 5b] and a significant reduction at 5 min after ligation [Fig. 5c]. Forty-eight hours after ischemic onset, the restoration of flow could not be identified in the same field of view. Although the time course of UHS-OMAG and LSI maps were not co-registered, the results clearly suggest that UHS-OMAG has greater sensitivity to the slow blood flows than the LSI does, in addition to its ability to depth-resolve the restored blood flow.

Fig. 5

Mapping blood perfusion after femoral artery ligation using LSI. (a)–(d) show LSI blood perfusion maps before, during, 5 min and 48 h after ligation. It is noteworthy that a rapid flow reduction initiated during ligation (b) and continuously developed immediately after ligation (c) and persisted at 48 h after onset (d) by use of LSI where the sensitive analysis of the capillary flow is currently difficult.