In vivo imaging of nitric oxide in the male rat brain exposed to a shock wave

While numerous studies have suggested the involvement of cerebrovascular dysfunction in the pathobiology of blast‐induced traumatic brain injury (bTBI), its exact mechanisms and how they affect the outcome of bTBI are not fully understood. Our previous study showed the occurrence of cortical spreading depolarization (CSD) and subsequent long‐lasting oligemia/hypoxemia in the rat brain exposed to a laser‐induced shock wave (LISW). We hypothesized that this hemodynamic abnormality is associated with shock wave‐induced generation of nitric oxide (NO). In this study, to verify this hypothesis, we used an NO‐sensitive fluorescence probe, diaminofluorescein‐2 diacetate (DAF‐2 DA), for real‐time in vivo imaging of male Sprague–Dawley rats' brain exposed to a mild‐impulse LISW. We observed the most intense fluorescence, indicative of NO production, along the pial arteriolar walls during the period of 10–30 min post‐exposure, parallel with CSD occurrence. This post‐exposure period also coincided with the early phase of hemodynamic abnormalities. While the changes in arteriolar wall fluorescence measured in rats receiving pharmacological NO synthase inhibition by nitro‐L‐arginine methyl ester (L‐NAME) 24 h before exposure showed a temporal profile similar to that of changes observed in LISW‐exposed rats with CSD, their intensity level was considerably lower; this suggests partial involvement of NOS in shock wave‐induced NO production. To the best of our knowledge, this is the first real‐time in vivo imaging of NO in rat brain, confirming the involvement of NO in shock‐wave‐induced hemodynamic impairments. Finally, we have outlined the limitations of this study and our future research directions.


| INTRODUC TI ON
Despite extensive efforts, the pathophysiology and mechanisms of blast-induced traumatic brain injury (bTBI) are not fully understood. Nevertheless, accumulating evidence shows the involvement of cerebrovascular dysfunction in the pathobiology of bTBI (Elder et al., 2015;Gama Sosa et al., 2019;Logsdon et al., 2020a;Rodriguez et al., 2018). Blast causes a complex injurious environment (Cernak & Noble-Haeusslein, 2010) both in human and experimental animal exposure scenarios; this makes identifying the mechanisms underlying blast-induced cerebrovascular and hemodynamic abnormalities challenging, and assessing their interactions as well as their role in bTBI outcomes quite difficult. Our previous study demonstrated the occurrence of cortical spreading depolarization (CSD) in the rat brain exposed to a laser-induced shock wave (LISW) (Sato et al., 2014).
CSD was accompanied by transient (3 ~ 4 min) hyperemia/hyperoxemia, followed by long-lasting oligemia/hypoxemia in the cortex (Kawauchi et al., 2019(Kawauchi et al., , 2022Sato et al., 2014). It would be crucial to examine the cause of these hemodynamic abnormalities to fully understand the mechanisms of bTBI. Since LISW is spatially well confined and not accompanied by dynamic pressure (jet), this model allows us to observe the effects of primary blast (the effects of a shock wave itself) locally, in a limited brain region (Jitsu et al., 2021;Kawauchi et al., 2019Kawauchi et al., , 2022Sato et al., 2014).
Brain functioning depends on optimal coupling of cerebral blood flow (CBF) to neural activity, thus well-regulated delivery of metabolic substrates to meet the energy demand of neural activity. This process is called neurovascular coupling (NVC). Impaired NVC has been implicated in a broad range of neuropathologies, including neurodegenerative diseases. Recently, nitric oxide (NO), a major regulating factor of cerebral hemodynamics (Nicholls, 2019;Toda et al., 2009), has been indicated as one of the most essential mediators in NVC. In physiological conditions, NO is produced in endothelial cells (ECs) in response to mechanical forces such as hemodynamic shear stress and increased intraluminal pressure (Sriram et al., 2016).
Additionally, NO is produced in neurons upon activation of the glutamatergic N-methyl-d-aspartate (NMDA) receptor by the neuronal isoform of nitric oxide synthase (nNOS) and promotes vasodilation by activating soluble guanylate cyclase in smooth muscle cells (SMCs) of the adjacent arterioles (Bredt & Snyder, 1994;Lourenço & Laranjinha, 2021). In both cases, the produced NO acts on SMCs, by which blood flow can be regulated (Iadecola, 1993;Lancaster, 1994).
We hypothesized that increased Ca 2+ influx in ECs could be caused by shock wave-induced excessive mechanical stress. Based on the observations that shock wave exposure also increased intracellular Ca 2+ in neurons (Ravin et al., 2016), we postulated that the shock wave-related hemodynamic abnormalities were associated with NO generation in the brain (The details are described in the following section.). In this study, to verify this hypothesis, we used an NO-sensitive fluorescence probe for in vivo real-time imaging of NO in the rat cortex exposed to an LISW. Such a methodological approach requires in situ observation of the rat cortex under a microscope, which would not be easy with other blast injury models utilizing conventional shock wave sources such as shock tubes or open field blast exposures. To the best of our knowledge, this study demonstrates the first in vivo real-time imaging of NO in the shock wave-exposed rat brain.

| OUR HYP OTHE S IS FOR NITRI C OXIDE-REL ATED HEMODYNAMIC AB NORMALIT Y IN THE B R AIN E XP OS ED TO A S HO CK WAVE
We hypothesized that NO plays a crucial role in the shock waveinduced impairments of NVC ( Figure 1). Our assumption is based on the nature of mechanosensing function in ECs, as well as the responsive characteristics of neurons to shock wave exposure. Thus, we posit that the pressure wave generated by blast, passing through blood vessels, stimulates mechano-receptors in the vessel wall. This will lead to increased intracellular Ca 2+ both in ECs and neurons (Ravin et al., 2016;Sriram et al., 2016), reactivating eNOS and nNOS, respectively, to generate NO. When CSD occurs in neurons, massive ion movements (Ca 2+ influx) as well as upregulated glutamate release can further increase Ca 2+ (Ayata & Lauritzen, 2015;Dreier, 2011), also stimulating nNOS to generate NO. Since NO has a relatively long lifetime of a few seconds, it can diffuse to vascular smooth muscle cells (SMCs) (Gally et al., 1990;Garthwaite & Boulton, 1995).

Significance
Due to increasing risks from terrorism using explosive devices worldwide, blast-induced traumatic brain injury (bTBI) has been a medical concern over the past two decades. However, its pathology and mechanisms are still not fully understood. Our imaging showed, for the first time to the authors' knowledge, generation of nitric oxide (NO) in the brain exposed to a shock wave in vivo. The generation of NO coincided with the early phase of hemodynamic dysfunction, suggesting the involvement of NO in bTBI pathology. Control of NO production and/or its related downstream events might be a key issue in bTBI treatment.
abnormalities. Moreover, since ONOO − can diffuse through biological membranes and has contractile effects, it can reach capillary pericytes and SMCs in cerebral arteries, causing contraction of the latter and thus leading to hemodynamic abnormalities (Elliott et al., 1998;Ferrer-Sueta & Radi, 2009;Yemisci et al., 2009).
Our previous study showed that LISW-induced persistent oligemia/hypoxia was mitigated by pharmacological inhibition of NOS (Inaba et al., 2018); these results support the feasibility of our current hypothesis. The main goal of the current study was to confirm our hypothesis through direct observation of shock wave-induced NO production in a rat brain.

| MATERIAL S AND ME THODS
All requests for animals and procedures used in this study were approved by the Ethics Committee of Animal Care and Experimentation, National Defense Medical College, Japan (permission numbers: 16036, 19035).

| Animals and study design
Nine-week-old male Sprague-Dawley rats (Japan SLC, Inc., Tokyo, Japan) were obtained and housed in standard laboratory cages on a 12:12-h light/dark cycle in the room maintained at 23 ± 2°C and 40%-60% humidity with free access to rodent diet (CE-7, CREA Japan, Inc., Tokyo, Japan) and water after arrival. Rats were housed in groups of 3 or 4 without physical enrichment, and their cages were cleaned twice a week. Rats were handled when the cage was changed to move them to a clean cage with new bedding. Rats at 10-11 weeks of age weighing 282-351 g (n = 30) were randomly assigned to undergo one of the following three experiments: (i) NO imaging without LISW exposure (sham control), (ii) NO imaging with LISW exposure, and (iii) pre-administration of an NO synthase inhibitor and NO imaging with LISW exposure. Since the occurrence of CSD depends on the shock wave intensity and we intended to examine the effects of CSD on NO production, we used a shock wave impulse (time-integrated positive pressure component) at which CSD probabilistically occurred (Sato et al., 2014). Thus, the rats for the second experiment were further divided into two groups: those with CSD and those without CSD. For the rats administered an NO synthase inhibitor, CSD also occurred probabilistically, but the rats without CSD were excluded from analyses since this condition was out of the scope of this study. Accordingly, this study consisted of the following four groups: Group 1 (sham control, n = 6), Group 2 (NO imaging with LISW exposure, CSD+, n = 6), Group 3 (NO synthase inhibitor administration and NO imaging with LISW exposure, CSD+, n = 6), and Group 4 (NO imaging with LISW exposure, CSD-, n = 6).
The rats used in this study were all male rats (n = 30) and no female rats were enrolled to maintain consistency with our previous studies (Jitsu et al., 2021;Kawauchi et al., 2019Kawauchi et al., , 2022Sato et al., 2014).

| Generation and characteristics of an LISW
A method for generating an LISW was described previously (Sato et al., 2014). Briefly, a laser target, which was a light-absorbing F I G U R E 1 Hypothetical mechanisms for shock wave-induced persistent cerebral hemodynamic abnormalities associated with nitric oxide (NO) generation. A shock wave increases intracellular Ca 2+ in both endothelial cells and neurons, which triggers generation of NO via eNOS and nNOS activations, respectively. Neuronal NO generation is enhanced by the occurrence of cortical spreading depolarization (CSD). Superoxide anions (O 2 − ), generated in parallel, react with NO; this leads to generation of highly cytotoxic peroxynitrite (ONOO − ) damaging smooth muscle cells in addition to endothelial cells and neurons. material (0.5-mm-thick natural black rubber disk) on which an optically transparent material (1.0-mm-thick polyethylene terephthalate sheet) was adhered, was placed on the tissue. At the interface between the rubber bottom surface and the tissue upper surface, ultrasound gel was applied for acoustic impedance matching. A highintensity, short-duration laser pulse was focused onto the laser target through the transparent sheet. The black rubber absorbed the laser energy to induce plasma, and its expansion was accompanied by generation of a shock wave (LISW).
Previously, we observed that transcranial application of an LISW with an impulse of ~14 Pa·s caused CSD in a probabilistic manner (9 out of 12 rats) (Sato et al., 2014). In the present study, the impulse was decreased to ~10 Pa･s to lower the probability of CSD occurrence to roughly 50%. As a laser source, we used a Q-switched ruby laser (pulse width, 200-ns full width at half-maximum [FWHM]; wavelength, 694 nm; RH356; JMEC Co., LTD., Tokyo, Japan) at a laser fluence of 1.0 J/cm 2 with a beam diameter of 4 mm. The corresponding peak pressure and positive pulse duration were ~43 MPa and ~1 μs, respectively.

| NO imaging and LISW exposure
Rats were anesthetized with an intraperitoneal injection of pentobarbital sodium (50 mg/kg bolus and ∼15 mg/kg/h maintenance) and placed in a stereotactic frame. After shaving the head, the scalp was incised at the midline and the parietal bone was exposed. A cranial window with a diameter of ∼4.5 mm was prepared in the left parietal bone with a trephine, and the dura and arachnoid mater were carefully removed under a surgical microscope. As described below, the NO-sensitive fluorescence probe used in this study was ~446 Da, and arachnoid mater is known to be permissive for molecules up to ~40 kDa (Roth et al., 2014). However, we removed the arachnoid mater to obtain efficient delivery of the probe to the tissues. The exposed cortical surface was superfused with artificial cerebrospinal fluid (ACSF; ARTCEREB Irrigation and Perfusion Solution for Cerebrospinal Surgery, Otsuka Pharmaceutical Co., Ltd., Tokyo, Japan). Pentobarbital was chosen as the anesthetic for imaging while isoflurane was used for other surgical procedures because of the reasons described below. In the preparation of a cranial window, the use of isoflurane frequently caused bleeding from the skull/dura, and coagulation with a bipolar coagulator was therefore needed for cortical imaging. To avoid the bleeding and heating effects due to the use of a bipolar coagulator, we chose pentobarbital for imaging as in our previous study (Kawauchi et al., 2019;Sato et al., 2014).
Fluorescence emission from DAF-2 increases when it is covalently modified by NO to yield a triazol fluorescein (DAF-2T). Thus, the fluorescent signal from DAF-2T (detection substance) produces an integrated measure of local NO concentration within the loaded cells. It should be noted that fluorescence originating from DAF-2T is affected by tissue pH of lower than 7 (Kojima et al., 1999). Since it is known that CSD decreases tissue pH to ~6.8 for ~10 min from its onset (Ayata & Lauritzen, 2015;Mutch & Hansen, 1984), this time frame was excluded from analyses in this study. DAF-2DA was diluted with ACSF at a concentration of 0.2 mM, and superfused on the exposed cortical surface for 1 min, and thereafter a disk of transparent polymethyl methacrylate with a diameter of ~4.4 mm and a thickness of ∼0.5 mm was inserted within the window and sealed with dental cement. For fluorescence probe uptake, the rat was left for ~45 min before starting imaging. We checked the distribution of the probe in the cortex by observing fluorescence on a tissue cross section, confirming that DAF-2T was distributed at depths up to ~800 μm. Considering the penetration depths of light for fluorescence excitation/emission described below, the maximum observation depth was estimated to be several hundreds of micrometers in the present imaging setup. Figure 2a shows an experimental setup for NO imaging of the rat brain. A motorized translation stage mounting a compact stereotaxic apparatus was placed on a base of a stereoscopic fluorescence microscope (Thunder Imaging Systems, Leica Microsystems Inc., Wetzlar, Germany) equipped with a monochromatic CMOS camera (2048 × 2048 pixels, 16 bit, DFC 9000 GT, Leica Microsystems) and a 1.0× magnification objective lens with 61.5-mm working distance (PLANAPO 1.0X, Leica Microsystems). After the window preparation as described above, the rat was moved onto the stereotaxic apparatus and fixed. The rat fixed in the stereotaxic apparatus could be horizontally translated between the positions for NO imaging and laser (and hence LISW) exposure. At the position for NO imaging, the cranial window was set under the objective lens of the microscope. The focus of the objective was adjusted at the rat cortical surface under the plastic sheet within the cranial window; the distance between the objective and the top surface of the window was about 60 mm. Images were acquired at 4.0× digital zoom using 2 × 2 binning. At the position for LISW exposure, the distal end of an articulated laser beam delivery arm was placed above the laser target. Figure 2b shows positions of the cranial window and the site of LISW exposure on the rat's skull. The distance between the center of the area of the LISW exposure and the center of the cranial window was ∼7 mm. An LISW was transcranially applied to the ipsilateral frontal region. The cortex was observed through the window under the microscope with a filter set (excitation, 450-490 nm; detection, 500-550 nm).
For NO imaging, baseline images were first acquired and then the rat was moved to the position for LISW exposure. Immediately after LISW exposure, the rat was moved back to the imaging position, and imaging was started within 10 s after LISW exposure.
Importantly, we observed that the phototoxicity due to fluorescence excitation caused NO production, and the excitation light intensity and exposure time were therefore limited to be lower than ~46 mW/cm 2 and 50 ms, respectively. Time-lapse imaging was performed every 10 s for 1 min before exposure (baseline imaging) and for the first 10 min (t = 0-10 min) after exposure, every 1 min for the next 20 min (t = 10-30 min) post-exposure, and every 5 min for the last 30 min (t = 30-60 min) post-exposure. The occurrence of CSD was detected by observing distinct transient arteriolar dilatation, which took place within a few minutes after LISW exposure. During all of these imaging procedures, the body temperature of the rat was kept constant at 37.0 ± 0.5°C with a temperature-controlled body mat, and arterial oxygen saturation (SpO 2 ) was monitored at the hindlimb with a pulse oximeter (8600 V, Nonin Medical, Inc., Plymouth, Minnesota).

| Administration of an NO synthase inhibitor
At ~24 h before LISW exposure, Group 3 rats were anesthetized with isoflurane inhalation (5% for induction; 1.5%-2.0% for maintenance) and then a small incision was made in the groin and a polyurethane catheter (MRE-040; ID, 1.02 mm; OD, 0.64 mm) was inserted into the femoral vein. The NO synthase inhibitor nitro-L-arginine methyl ester (L-NAME, Dojindo Laboratories, Kumamoto, Japan) diluted with saline at a concentration of 4.0 mg/mL and a volume of 5.0 mL/kg animal body weight was intravenously injected through the catheter. With this protocol, NO synthase activity in the brain was expected to be inhibited by ~50% (Iadecola et al., 1994). After injection, the femoral vein was ligated with a suture, and the incision was closed with a suture. Rats received L-NAME administration under isoflurane anesthesia as long as ~24 h before imaging, and the injection volume was within the standard value for laboratory animals (Turner et al., 2011). Thus, we assumed that both the injection and anesthesia would not affect the results of imaging, and neither saline injection nor isoflurane anesthesia were applied to all groups of rats other than Group 3.

| Image analysis
Within the cranial window, typically, distal branches of the left pial middle cerebral artery could be observed, and from those, we chose one or two representative arteriolar branch(es). We analyzed spa-

| Statistical analysis
For the comparison of NO generations among groups, full time-course data may be used on the basis of a parametric test, but the uniformity of variance was not verified in our data, suggesting the use of a non-parametric test. However, it is uncommon to deal with two different independent variables in non-parametric tests. Thus, we chose two time points for the comparison: 20 min and 30 min post-exposure, which respectively corresponded to ~10 min after the end of tissue acidification due to CSD and to the time at which sustained hypoxemia was observed in our previous study (Kawauchi et al., 2019(Kawauchi et al., , 2022Sato et al., 2014).     showed an abrupt increase immediately after exposure, followed by drastically decreased and then again increased (by ~20%) intensities during CSD propagation (t = ~1 min to ~3 min). After that, the fluorescence intensity gradually increased and reached the level ~ 15% higher than the baseline at 40 min post-exposure, staying at that level until the end of the measurement. We assume that a decrease in the fluorescence intensity by ~20% during CSD could be caused by tissue acidification (Kojima et al., 1999).
While the changes in arteriolar wall fluorescence in Group 3 rats (with L-NAME, Figure 5c) showed a temporal profile similar to the temporal profile of changes observed in Group 2 rats, their intensity level was considerably lower than that in Group 2 rats. However, Group 4 rats showed a time-course of arteriolar wall fluorescence that was quite different to that in Group 2 and Group 3 ( Figure 5d): for the first 10 min post-exposure, fluorescence intensity showed an abrupt increase by ~2% and it gradually increased by ~10% until ~15 min post-exposure; after that, it plateaued until the end of the measurement.
Changes in fluorescence intensities measured in the parenchyma of Group 2 and Group 3 rats (Figure 5f,g) showed a temporal profile similar to that recorded in arterioles (Figure 5b,c); nevertheless, the intensity levels for the parenchyma were lower than those for the arterioles in both groups. Parenchymal fluorescence intensity measured in Group 4 rats showed an immediate slight increase after exposure, but the intensity did not greatly increase until the end of the measurement. In some cases, a minor decrease in the fluorescence intensity was observed in both the arteriolar wall and parenchyma during the last 20 min of observation, possibly due to leakage of the probe from the cells and its washout by extracellular fluid.

| DISCUSS ION
The present in vivo imaging using an NO-sensitive fluorescence probe showed increased fluorescence, indicative of NO production, in the rat brain exposed to an LISW. The most prominent increase in fluorescence was observed along the pial arteriolar walls in Group 2 (CSD+) rats (Figure 5b) during the period spanning from 10 min to 30 min post-exposure. This period of intense fluorescence coincided with the early phase of persistent cerebral oligemia/hypoxemia observed in our previous study (Kawauchi et al., 2019(Kawauchi et al., , 2022Sato et al., 2014), suggesting an association of NO production with shock wave-induced hemodynamic dysfunction in the brain. NOS inhibition (Group 3) caused no significant fluorescence increase at any timepoint analyzed (Figure 6), indicating that the observed NO production was due to NOS activation. Rats without CSD (Group 4) showed no significant increase in fluorescence intensity compared with that in CSD+ rats (Group 2) or sham control rats (Figure 7). These findings suggest that the observed NO production was due both to the shock wave exposure and the occurrence of CSD. Overall, the presented results seem to support our hypothesis (Figure 1).
We expected to observe NO-originating fluorescence both from arterioles and venules, but, interestingly, evident fluorescence was seen emerging only from the arterioles. We assume that the fluorescence emerging from the arterioles was emitted from SMCs over the pial arterioles, to which the eNOS-generated NO molecules in the ECs can be translocated (Lancaster, 1994). Although venular ECs also have eNOS, SMCs in venules are less contractile than arteriolar SMCs (Hartmann et al., 2021). Importantly, NO plays different role in arterioles than in venules (Pober & Sessa, 2007). NO produced in arterial ECs has a role in blood flow regulation by acting on arteriolar SMCs. On the other hand, venular ECs are responsible for trafficking leukocytes from blood into the tissues (Pober & Sessa, 2007), where NO molecules work mainly in the venular lumen without moving to SMCs. It should be noted, however, that some NO molecules in venules might remain under the detection limit of the imaging system used in this study.
The mechanisms shown in Figure 1 can be applied not only to cortical vessels but also to pial arterioles, which are the largest cerebrovascular component. Pial arterioles are heavily invested by perivascular nerves with multilayered (more than threelayered) SMCs with strong contractile capabilities (Bahr-Hosseini & Bikson, 2021). Thus, intense fluorescence observed along the pial arterioles would be generated both by eNOS in the ECs and nNOS in perivascular nerves. In addition, it is noteworthy that NO molecules generated in the parenchyma can be translocated to pial arteries through the pia mater and are involved in the regulation of blood flow in pial arterioles (Busija et al., 2008). Thus, NO on the pial arterioles can be closely associated with cortical neuronal activity.
Pial arterioles penetrate the brain parenchyma forming penetrating arterioles, which connect to the parenchymal arterioles leading to capillaries. Since pial arterioles and penetrating arterioles have common anatomical structure at the cellular level as well as vascular functions (Bahr-Hosseini & Bikson, 2021), the NO production in penetrating arterioles is expected to be similar to that in pial arterioles Indeed, we observed intense NO probe fluorescence along the penetrating arterioles on a tissue cross section (data not shown).
Fluorescence intensity in the parenchyma was greatly increased after CSD propagation in Group 2 (Figure 5f), but there was no significant difference in parenchymal fluorescence between Group 1 (control) and Group 2 either at 20 min or 30 min post-exposure ( Figures 6 and 7). We assumed that NO observed in the parenchyma would be generated mainly from the cortical neurons and capillaries, but it has been reported that NOS expression in the cortical capillaries is much less than that in the cerebral arteries/arterioles (Gabbott & Bacon, 1993). In addition, NO generated in the neurons can move to SMCs in the arteries/arterioles (Busija et al., 2008), possibly resulting in higher-intensity NO-originating fluorescence along the arteriolar wall than that in the parenchyma. Furthermore, arteriolar walls are contacted with the blood, and there is therefore an acoustic impedance mismatching at the interface. This would make the arterioles more susceptible to the shock wave stimulus than the parenchymal tissue. Vasoconstriction of the pial arterioles observed in Group 2 (Figure 4b) would be associated with ONOO − generated from NO (Figure 1) since ONOO − has been shown to act as a contractile substance for cerebral arteriolar SMCs (Elliott et al., 1998).
As the pial artery/arteriole is the most upstream vessel for the brain, its constriction should considerably affect the whole brain. It is also possible that the excessively produced NO in the arterioles penetrate into the surrounding parenchymal area, producing ONOO − as depicted in Figure 1.

F I G U R E 5
Average time courses of DAF-2T fluorescence intensity in the arteriolar walls (a-d) and their adjacent parenchyma (e-h) that were obtained from kymographs for all groups of rats (n = 30 from 6 rats in each group). Values are normalized by those before exposure. Vertical solid lines at t = 0 indicate the time of LISW exposure. Blue shadow bars represent the duration of CSD propagation. Time zones between the two vertical dashed lines indicate tissue acidification due to CSD; data during these time zones were excluded from the analysis.
The shock wave-induced vascular NO production observed in this study should closely be associated with vascular functional changes, including the blood-brain barrier (BBB) disruption, because eNOS activation has been shown to cause BBB disruption (Argaw et al., 2012;Cirino et al., 2003). This would be triggered by astrocytederived vascular endothelial growth factor (VEGF) as observed in the blast-induced mice TBI models (Ahmed et al., 2015;Logsdon et al., 2020b;Meabon et al., 2020). As described above, since blood vessels have fluid-tissue interfaces, there is an acoustic impedance mismatching (Nakagawa et al., 2011). Thus, its consequent tensile stress would also be involved in the mechanism of vascular changes.
Although only eNOS and nNOS appear in our hypothesis ( Figure 1), inducible NOS (iNOS) may also be involved in the shock wave-induced NO production. For example, since it has been reported that microglia can be activated by mechanical stress (Ayata & Schaefer, 2020), shock wave exposure may lead to microglia activation, possibly being followed by NO generation by iNOS. It is also known that pial macrophages can be activated by CSD (Schain et al., 2018), and shock wave-associated CSD may therefore result in NO production by iNOS. The experiment with the use of an iNOSselective inhibitor seems to be useful to analyze the role of iNOS in the shock wave-induced NO production. Group 3], total df = 89, between groups df = 2, within groups df = 87).
While our findings contribute to a better understanding of blast-induced TBI, we acknowledge certain limitations of our study. First, we used a pH-sensitive fluorescence probe; thus, the CSD-induced acidification could considerably affect fluorescence emission and negatively influence accurate NO imaging (Figures 4   and 5). Hence, we plan to use a pH-insensitive probe in our next study and compare our current and newly generated findings. Secondly, fluorescence measured in the parenchyma could emerge from cortical neurons and capillaries and, in certain situations, from microglia in which iNOS can produce NO (Corradin et al., 1993), as discussed above. To identify the exact NO sources in the parenchyma, higher resolution, depth-resolved imaging would be needed, perhaps using an in vivo confocal microscope with a selective NOS inhibitor. Thirdly, in this study, we used only one LISW condition. Our next task is to examine how different LISW conditions affect NO generation. The duration of the LISW used in this study was ~1 μs, which is considerably shorter than that of real-life explosion-generated shock wave (Courtney & Courtney, 2011). Considering the scaling law between the human brain and the rat brain (Ackermans et al., 2021), we assume that the optimum shock duration ranges from a few tens of microsecond to 100 μs. Thus, one of our future studies will focus on NO imaging of rat brains exposed to LISWs of longer duration. In our hypothesis, we posit that ONOO − formed via reaction of NO with O 2 − plays an important role. Thus, detection or imaging of ONOO − will be needed to fully complete our hypothesis.
The current study lacks studying sex differences. Since a female gonadal hormone, estrogen, has been shown to increase eNOS via cerebrovascular estrogen receptors, more NO may be produced by eNOS in female rats than in male rats (McNeill et al., 2002). Another sex difference may arise from sex-dependent characteristics of CSD.
Interestingly, it has been reported that the threshold concentration of extracellular potassium for eliciting CSD is 80% lower in female rats than in male rats (Adámek & Vyskočil, 2011). This may reduce calcium ions influx, resulting in decreased NO production by nNOS in female rats. These two potential sex differences are opposed, and it is not clear which is superior. Experimental analyses are needed to clarify the sex differences.
In conclusion, we demonstrated increased NO production in rat brains exposed to shock waves. Our in vivo imaging using a fluorescence probe allowed real-time observation of NO production, which coincided with the early phase of the occurrence of cerebral oligemia/hypoxemia, thus suggesting the involvement of NO production in the shock-wave-induced hemodynamic impairments.
We assume that shock wave-induced dysfunction of ECs (Hall et al., 2017;Logsdon et al., 2020a) leads to deterioration of NO production, which can counterbalance the shock wave-induced NO production in a time-dependent manner. Our future studies will address the temporal profile of shock wave-induced dynamical changes in NO production.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors have no conflict of interest to disclose.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data supporting the findings of this study are available from the corresponding author upon request.