In Vivo Imaging of the Mouse Neurovascular Unit Under Chronic Cerebral Hypoperfusion
Background and Purpose—Proper brain function is maintained by an integrated system called the neurovascular unit (NVU) comprised cellular and acellular elements. Although the individual features of specific neurovascular components are understood, it is unknown how they respond to ischemic stress as a functional unit. Therefore, we established an in vivo imaging method and clarified the NVU response to chronic cerebral hypoperfusion.
Methods—Green mice (b-act-EGFP) with SR101 plasma labeling were used in this experiment. A closed cranial window was made over the left somatosensory cortex. To mimic chronic cerebral hypoperfusion, mice were subjected to bilateral common carotid artery stenosis operations using microcoils. In vivo real-time imaging was performed using 2-photon laser-scanning microscopy during the preoperative period, and after 1 day and 1 and 2 weeks of bilateral common carotid artery stenosis or sham operations.
Results—Our method allowed 3-dimensional observation of most of the components of the NVU, as well as dynamic capillary microcirculation. Under chronic cerebral hypoperfusion, we did not detect any structural changes of each cellular component in the NVU; however, impairment of microcirculation was detected over a prolonged period. In the pial small arteries and veins, rolling and adhesion of leukocyte were detected, more prominently in the latter. In the deep cortical capillaries, flow stagnation because of leukocyte plugging was frequently observed.
Conclusions—We established an in vivo imaging method for real-time visualization of the NVU. It seems that under chronic cerebral hypoperfusion, leukocyte activation has a critical role in microcirculation disturbance.
The neurovascular unit (NVU) is a conceptual framework that integrates responses in all cell types, including neuronal, glial, inflammatory, and vascular elements.1–5 This cell–cell integrated response is an indispensable factor used to maintain brain function and homeostasis. In fact, dysfunction of the NVU is the basis for many diseases.2–4 The concept of NVU emphasizes that maintenance of integrated cellular function is more important than just salvaging an individual cell alone. Although the intricate molecular pathway of neuronal death has been dissected in detail, the mechanisms of how the entire NVU responds to cerebral ischemia are not completely understood. Understanding this concept may provide a comprehensive framework for investigating mechanisms and therapies of ischemic brain damage.5
The NVU is a dynamic structure assembled by endothelial cells, basement membranes, perivascular astrocytes, pericytes, and neurons. Therefore, it is difficult to understand the whole structure (including the anatomic relationship between cells) and the dynamic changes that occur within a single specimen. To understand the NVU more easily, an imaging method that can detect the whole NVU component at one time in vivo should be introduced. Thus, the aim of the present experiments was (1) to establish an in vivo imaging method for the NVU, which has extraordinarily spatial and time-dependent features. A spatial feature would require in-detail, 3-dimensional (3D) observation of the intricate NVU structure, whereas a time-dependent feature would require repeated longitudinal observation to capture real-time events, such as dynamic microcirculation in capillaries and (2) to clarify the stress response of the NVU in the bilateral common carotid artery stenosis (BCAS) model of chronic cerebral hypoperfusion.
Materials and Methods
All procedures were performed in accordance with the guidelines for animal experimentation from the ethical committee of Mie University. Male green fluorescent protein transgenic mice (C57BL/6 TgN [b-act-EGFP] Osb) were used in this experiment (aged 9–14 weeks; Japan SLC, Inc, Shizuoka, Japan).6
Two-Photon Laser-Scanning Imaging Experiment
Mice were initially anesthetized with isoflurane (1%–2%), and a custom-made attachment device for holding the head in place was fixed to the skull. A closed cranial window was made by removing part of the skull (4 mm in diameter) over the left somatosensory cortex while leaving the dura intact, and the exposed cortex was sealed with a cover glass (Figure 1A).7 These animals were then allowed to recover from anesthesia and were housed in a cage with free access to food and water. Mice were also kept in the cage in-between scheduled experiments. For plasma labeling, sulforhodamine 101 (SR101) dissolved in saline (0.01 mol/L) was injected intraperitoneally (8 μL/kg body weight) just before beginning the imaging experiment.8 Each animal was placed on a custom-made apparatus under 1.2% isoflurane, and imaging was conducted with a 2-photon laser-scanning microscope (FV1000-2P; Olympus, Japan; For more detailed methodology see in the online-only Data Supplement).
Surgical Procedure of Chronic Cerebral Hypoperfusion
To replicate chronic cerebral hypoperfusion, mice were subjected to BCAS using microcoils.9 In brief, under anesthesia with 1.3% isoflurane, the common carotid arteries (CCA) were exposed through a midline cervical incision, and a microcoil (inner diameter: Rt CCA, 0.18 mm; Lt CCA, 0.16 mm) was applied to the bilateral CCA. We applied a 0.18 mm×0.16 mm combination coil (purchase from SAMINI Co, Ltd., Japan) to elicit neurovascular response as much as possible. This combination has been selected because that of 0.16 mm×0.16 mm results in many mice dying over the observation period.9 Sham-operated animals underwent bilateral exposure of the CCA without applying the microcoil.
Measurement of Cerebral Blood Flow
Cerebral blood flow (CBF) was determined by a laser speckle flowmetry (Omegazone, laser speckle blood flow imager, Omegawave), which obtains high-resolution 2D images in a matter of seconds, as previously described.10 Briefly, we measured CBF through the cranial window in the same physical condition in terms of anesthesia and body temperature as during 2-photon laser-scanning. A 780-nm laser semiconductor illuminated the area of interest, and light intensity was accumulated in a charge-coupled camera device and transferred to a computer for analysis. Image pixels were then analyzed to produce average perfusion values.
To exclude the effect of damage from the procedure involved in creating the cranial window, mice were allowed to rest for 1 week after the surgery. Laser speckle flowmetry and 2-photon laser-scanning were then performed as part of the preoperation period (Pre). Mice were randomly assigned to a sham (n=7) or BCAS (n=7) group, and 3 to 5 days after the preoperation period, sham or BCAS operations were performed (3 mice from the Sham group and 2 mice form the BCAS group were excluded because the cranial window became dim during the observation period.). Laser speckle flowmetry and 2-photon laser-scanning were then performed 1 day after BCAS or shame operations, and repeated again after 1 and 2 weeks. Anatomic morphology of the NVU (vessel structures, astrocytes, and pericytes) and microcirculation profiles (kinetics of erythrocytes and leukocytes) were evaluated at each time point for both sham and BCAS groups (for more detailed Methods see in the online-only Data Supplement).
Results are expressed as the mean± SD. The Mann–Whitney U test was used to evaluate differences between groups, and an ANOVA followed by a post hoc Tukey–Kramer test was used to evaluate differences over time. Two-sided P<0.05 was considered statistically significant.
NVU in the Normal Brain
In the operated cranial window, we were able to observe and clearly identify small pial vessels (see Figure 1A for a representative image). We identified volumes of interest and were then produced a maximum intensity projection from the surface of the brain to an imaging depth of 550 μm (Figure 1B). We 3D reconstructed these volumes of interests and then visualized whole vascular compartments (from pial small arteries to pial small veins) as is demonstrated in Figure 1B. In addition, zoom scanning of each vessels allowed us to identify the shapes of entire vessels (Figure 1C), and the NVU compartment from arterioles to capillaries and from capillaries to postcapillary venules (Figure 1D). In the transgenic mouse, only 4 cell types including astrocytes, pericytes, leukocytes, and platelets could be detected as cells positive for green fluorescent protein fluorescence under 2-photon laser-scanning microscope; therefore, each cell type could be identified based on their morphology (Methods and Figure I–III in the online-only Data Supplement).6,11 Thus, astrocytes, which extend their foot processes around the arterioles and capillaries, could be clearly identified (Figure 1D, arrows), and pericytes, which wrap around the capillaries, could also be clearly observed (Figure 1D, arrowheads). Plasma labeling with SR101 allowed us to visualize erythrocytes, which were detected as nonfluorescent particles within the plasma stream. The kinetics of erythrocytes in microcirculation could be detected dynamically by repeated volume scanning (Movie I in the online-only Data Supplement).
Regional CBF Under Chronic Cerebral Hypoperfusion
Figure 2 shows the mean CBF values, as measured by laser speckle flowmetry, in the both sham and BCAS groups (Figure 2A). In the sham group, CBF was not changed significantly throughout the entire observation period. In contrast, CBF values after the BCAS operation were significantly decreased when compared with those of preoperative baselines and sham controls (Figure 2B).
Figure 3 shows the inflow (pial artery) and outflow (pial vein) velocity changes in the 2 groups. Specifically, we measured the slope of line scan images of pial arteries (Figure 3A) and veins (Figure 3B) to indicate inflow and outflow velocity, respectively, at each time point.12 In the sham group, inflow velocity was not changed significantly throughout the entire observation period. In contrast, the inflow velocity after BCAS operation was significantly decreased when compared with those of preoperative baselines and sham controls (Figure 3C). Similarly, outflow velocity was not changed in the sham group, and significantly decreased after BCAS operation when compared with those of preoperative baselines and sham controls (Figure 3D).
Anatomic Morphology Under Chronic Cerebral Hypoperfusion
Under chronic cerebral hypoperfusion, the small pial vessels and the deep cortical capillaries did not show obvious morphological changes such as vascular loss or angiogenesis throughout the entire observation period (Figure 4A). Vascular shape such as tortuosity and size for arterioles (Figure 4B) and venules was not changed throughout the observation period, and obvious changes of wall thickness for arterioles (Figure 4B, Lt lower inset) and venules were also not detected. Although the arteriolar lumen (Figure 4B, Rt lower inset) slightly increased in post-BCAS operation, there was no significant difference in luminal area over time (1 day, 119.9±49.5%; 1 week, 102.2±14.1%; 2 weeks, 104.9±23.3% compared with the preoperative baseline). The venular lumen in the BCAS group also remained unchanged throughout the entire observation period (1 day, 95.2±12.5%; 1 week , 95±12.4%; 2 weeks, 106.6±15.1% compared with the preoperative baseline). In the sham group, wall thickness and luminal area did not show obvious changes throughout the observation period. Figure 4C show that astrocytes extended their foot process around the vessels but showed no obvious morphological differences between pre- and post-BCAS operative periods. In addition, pericytes were found to wrap around capillaries at the both periods with no morphological changes. Taken together, astrocytes and pericytes remained unchanged morphologically between BCAS and sham groups during the entire observation period.
Microcirculation in Deep Cortical Capillaries Under Chronic Cerebral Hypoperfusion
Figure 5A and Movies II to V in the online-only Data Supplement show representative microcirculation in the deep cortical capillaries under chronic cerebral hypoperfusion. Under low magnification, stagnation of capillary flow, which was detected as high fluorescent plasma segments without nonfluorescent erythrocytes (Figure 5A, white arrows; Movie III, in the online-only Data Supplement), was observed under chronic cerebral hypoperfusion, whereas it was rarely found under normal microcirculation (Movie II in the online-only Data Supplement). We also observed leukocyte plugging (Figure 5A, blue arrow) in the high fluorescent plasma stagnated capillary (Figure 5A, blue arrowheads) under higher magnification (Movies IV and V in the online-only Data Supplement). Plugged leukocytes did not tightly obstruct capillaries, but slowly moved downstream (Figure 5A, purple arrows) and eventually were released (plugging time was usually <120 s), resulting in reflow of plasma streams with erythrocytes (Figure 5A, yellow arrow). Long-term and complete leukocyte plugging, as well as leukocyte infiltration, in the perivascular parenchyma could not be identified during the observation period. On the contrary, platelet activation (Methods and Figure III in the online-only Data Supplement), which is a pivotal mechanism of focal cerebral ischemia, could not be found during the observation period. Frequency of capillary flow stagnation in the BCAS group was markedly higher than that of the sham group throughout the entire observation period (Figure 5B and 5C).
Leukocyte Rolling and Adhesion in the Cortical Vessels Under Chronic Cerebral Hypoperfusion
In the sham-operated group, rolling and adhesion were rarely detected throughout the entire observation period (Figure 6A). After BCAS operation, however, rolling and adhesion of leukocytes were typically observed in the pial venules although rarely in the pial arterioles under chronic cerebral hypoperfusion (Figure 6A, arrows). From 1 day to 2 weeks after the BCAS operation, the frequency of rolling and adherent leukocytes decreased in the both small pial veins and pial arteries (Figure 6B and 6C). However, 2 weeks after BCAS operation, the frequency of rolling and adhesion was still higher significantly in the pial veins.
Neuron-targeted therapies against brain ischemia have been uniformly unsuccessful in clinical trials. This is thought to be a consequence of lack of understanding of the integrated response among neuronal, glial, and vascular elements.1–5 Each of these elements responds to an ischemic event by activation and individual features of specific cell responses are well studied, but it remains unknown how they respond as a NVU.13 In this study, we developed an in vivo imaging method for the NVU. This method allows repeated longitudinal observation of the NVU in 3D, which is helpful to understand anatomic morphology of microvessels with varying sizes as a whole. In addition, this method allows identification of cellular components (astrocytes, pericytes, circulating leukocytes, and platelets) and allows investigation of the anatomic relationship between these cells. This imaging approach also provides semiquantitative information on the kinetics of erythrocytes and leukocytes to understand microcirculation alteration.
Chronic cerebral hypoperfusion induces subcortical white matter lesions and vascular cognitive impairment. This condition is often an initial trigger of a cascade of pathophysiological changes in the NVU2–4 and can be experimentally replicated in animal models of chronic cerebral hypoperfusion such as BCAS model.9,14 In these models, gray matter exhibits more slight pathophysiological changes than white matter under chronic cerebral hypoperfusion. In this study, we could not detect any morphological change in the cortical gray matter. The morphological changes as stress response of NVU in the cerebral cortex may be limited and under the detection level of 2-photon laser microscopy method. However, additional studies will be required to address this point finally.
As for the dynamic microcirculation, there was stagnation of capillary flow and plugging by leukocytes under chronic cerebral hypoperfusion. The phenomenon of capillary plugging has been postulated to explain no-reflow phenomenon during early reperfusion after ischemia.15,16 After several hours of ischemia, an incomplete restoration of blood flow, no-reflow phenomenon, has been reported also in other organs.17–19 Many pathophysiological mechanisms have been implicated in this process, however, the exact mechanisms responsible for this phenomenon have been unclear. Leukocytes are large stiff cells, which physiologically adhere to vascular endothelium, and are known to express adhesion molecules under a variety of conditions.20 The present study has provided the first evidence that leukocyte transiently plugs capillaries under chronic cerebral hypoperfusion, and this plugging has induced stagnation of capillary flow. In this stagnated plasma, no oxygen-transporting erythrocytes were found, implying a viscous cycle for further exacerbation of parenchymal hypoxia. On the contrary, platelet activation was not observed throughout the observation period. Antiplatelet treatment has been the target of therapeutic strategies for chronic cerebral hypoperfusion, but this strategy should be substituted for the agents that have the potential to suppress plugging of leukocytes and restore insufficient CBF. Indeed, beneficial effect of induced immune tolerance against adhesion molecules has been revealed in chronic cerebral hypoperfusion previously.21
Leukocyte rolling and adhesion in the pial vessels have been reported in various organs under a variety of conditions, appearing more prominently in the venules than in the arterioles.22–24 In accordance with these observations, leukocyte rolling and adhesion were observed typically in the pial veins during chronic cerebral hypoperfusion. Our results may further indicate that leukocyte activation may be the first step eventually leading to stress response of the whole NVU to chronic cerebral hypoperfusion.
In this study, we established an in vivo imaging method for the NVU and provided evidence that leukocytes activation may play a critical role in microcirculation disturbance under chronic cerebral hypoperfusion.
The study was conceived by Dr Tomimoto, and the protocol was drafted by Drs Suzuki, Tanaka, Mizoguchi, and Tomimoto. Drs Yata, Nishimura, Unekawa, and Tomita performed the experiments. Dr Yata was responsible for data analysis, data interpretation, and preparation of the report. All authors contributed to data interpretation and approved the final version.
Sources of Funding
This work was supported by Grants-in-Aid for Scientific Research (KAKENHI) grant number 22590929.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.114.005891/-/DC1.
- Received June 24, 2014.
- Revision received September 12, 2014.
- Accepted October 1, 2014.
- © 2014 American Heart Association, Inc.
- Shibata M,
- Ohtani R,
- Ihara M,
- Tomimoto H
- Fernández-Klett F,
- Offenhauser N,
- Dirnagl U,
- Priller J,
- Lindauer U
- Nimmagadda A,
- Park HP,
- Prado R,
- Ginsberg MD
- del Zoppo GJ,
- Schmid-Schönbein GW,
- Mori E,
- Copeland BR,
- Chang CM
- Helmke BP,
- Bremner SN,
- Zweifach BW,
- Skalak R,
- Schmid-Schönbein GW
- Fukuda S,
- Yasu T,
- Kobayashi N,
- Ikeda N,
- Schmid-Schönbein GW