The Impact of Enteral Nimodipine on Endothelial Cell Apoptosis in an Animal Subarachnoid Hemorrhage Model

Objective: Enteral nimodipine is the most evidence-based and widely used drug for the treatment of delayed cerebral ischemia (DCI) and is known to have various neuroprotective functions. However, the neuroprotective mechanism of nimodipine still remains unclear and the effects of nimodipine remain ambiguous. Herein, we studied the effect of enteral nimodipine on endothelial apoptosis after subarachnoid hemorrhage (SAH). Method: SAH was experimentally introduced in white rabbits (n=42) that were grouped: enteral nimodipine (group N, n=14); a control that received normal saline (group S, n=13); and a control without hemorrhage (group C, n=15). On the third day after SAH induction, the brain stem, including the vertebrobasilar vascular system, was extracted. The effects of enteral nimodipine were analyzed by group using histopathologic analysis, including immunohistochemical staining of apoptosis-related proteins (Bcl2 [anti-apoptotic] and Bax [pro-apoptotic]). Result: Cytoplasmic vacuolation of smooth muscle cells was observed in groups S and N, and was more prominent in group S. Endothelial desquamation was observed only in group S. For the basilar artery, expression of Bcl2 and Bax in group N was lower than in group S, but signi�cant differences were not observed (p = 0.310 and p = 0.710, respectively). In penetrated arterioles, the expression of Bax in group N was signi�cantly lower than that of group S (p < 0.001). The thickness of the tunica media in the basilar artery was thinner in group N than in group S (p < 0.001). Conclusion: This study suggests that enteral nimodipine may have a neuroprotective function by inhibiting endothelial apoptosis in small arterioles and preventing smooth muscle cell proliferation in large arteries.


Introduction
The development of microsurgical or neurointerventional treatment after SAH due to rupture of a cerebral aneurysm has reduced morbidity and mortality at the time of rupture but has not yet reduced the incidence of DCI following hemorrhage.[1,2] DCI is caused by cerebral vasospasm, microthrombosis, and cortical spreading depression, etc [3][4][5][6] and is closely related to endothelial damage immediately after the rupture of a cerebral aneurysm.[7] Endothelial damage in the early phase after an aneurysmal rupture increases coagulability, blood-brain barrier permeability, microthrombosis and induced nitric oxide downregulation.[7] Cerebral vasospasm after rupture of a cerebral aneurysm is associated with endothelial apoptosis.[7][8][9][10][11] Blockers of apoptosis (e.g., caspase and p53 inhibitors) ameliorated the delayed phase of cerebral vasospasm in an experimental SAH animal model.[12][13][14] Enteral nimodipine is the most evidence-based and widely used drug for the treatment of DCI and is known to have various neuroprotective functions, such as improving neurologic outcomes after SAH and reducing the ischemic area in ischemia.[15][16][17][18] However, the neuroprotective mechanism of nimodipine still is unclear and the effects of nimodipine remain ambiguous.[19] It was shown that nimodipine inhibited endothelial cell apoptosis by reducing intracellular calcium in ux in animal models of ischemia.[20][21][22] However, the route of administration of nimodipine in these studies was not enteral (intravenous infusion) and endothelial cell death was also not shown.Herein, we studied the effect of enteral nimodipine on endothelial apoptosis in the early phase after experimental SAH using immunohistochemical staining of apoptosis-related proteins (Bcl2 [anti-apoptotic] and Bax [proapoptotic]) in large arteries and small arterioles.Histologic change and vessel parameters were also evaluated.The purpose of this study was to elucidate the immunohistochemical effect of enteral nimodipine on the cerebral artery in the early stage after experimental SAH.

Methods 1. Experimental animal selection and preparation
This study was approved by the Animal Ethics Committee in our institution and all methods were performed according to ARRIVE guidelines.A total 50 New Zealand white rabbits weighting 2.5-3.0 kg were used for the study without regard to gender.Experimental animals were randomly divided into three groups: a control group (group C) that did not have induced SAH; a hemorrhage group orally administered normal saline after induced SAH (group S); and a hemorrhage group orally administered nimodipine (group N).

Induction of SAH in animal model
Ketamine (35 mg/kg) and xylazine (5 mg/kg) were intramuscularly injected into the thigh muscle of rabbits in a fasted state (6 h prior) for anesthesia.The occipital-cervical area and both ears were epilated using depilation cream.The animals were secured in a supine position.The head of each animal was positioned in place and the occipital area was sterilized (75% alcohol).For a dura puncture, 1% lidocaine was administered into the suboccipital area, and the dura was punctured under uoroscopy guide using a 23-gauge spinal needle.To identify the subarachnoid space, cerebrospinal uid (CSF) was identi ed using a spinal needle, and contrast agent was identi ed in the subarachnoid space after the cisternal administration of contrast by uoroscopy.After about 2 mL of CSF was slowly drained, the same amount of unclotted arterial blood in groups H and N was obtained from the central ear artery using a 24-gauge needle and administered into the subarachnoid space through the spinal needle.Thereafter, the head was maintained in the Trendelenburg position for approximately 20 minutes to allow the administered arterial blood to diffuse the basal cistern.The Xper computed tomography of angiography equipment (Allura Xper FD20/10; Philips Healthcare, Amsterdam, Netherlands) was used to verify SAH in the basal cistern.
Each animal with induced SAH, except for those in group C, was randomly placed into groups S and N.After the animals fully recovered 3 h after SAH induction, each group was given each drug six times a day through a gastric tube.The dosage of nimodine was calculated based on the body surface area.[23] Animals in group S were each administered normal saline (0.010ml/cm 2 /day) and those in group N were administered enteral nimodipine (0.032mg cm 2 /day) dissolved in 3 mL of pH 4.5 acetate buffer containing 0.3% of sodium dodecyl sulfate.
All animals were sacri ced on the third day after SAH induction through air embolism of the central artery in the ear using the same anesthetic method.The scalp of the suboccipital area was ayed, suboccipital muscles were dissected, and the occipital bone and posterior arch of atlas were exposed.Occipital craniectomy and posterior laminectomy of atlas was performed, the dura was widely opened, and the brain stem, including the vertebrobasilar vascular system, was extracted.

Staining of specimen
Harvested biopsy specimens were kept in a tubes of 10% formalin and embedded in para n.Para nembedded 5-um tissue microarray sections were stained with hematoxylin-eosin (H&E) and immunochemical stains.Stained tissue sections were shown to contain the basilar artery in the basilar groove.
For each antibody, immunochemical staining was undertaken according to each manufacturer's protocol.A monoclonal mouse anti-human bcl2 antibody (Clone 124, code no.M0887; Dako, Glostrup, Denmark) was used for bcl2 immunochemical staining.Tissue sections were depara nized with 20 min of heatinduced epitope retrieval in Dako target Retrieval Solution (Code no.S1700; Dako, Glostrup, Denmark)).Tissue sections did not dry out during the following immunohistochemical staining procedure: Primary antibody was diluted by approximately 1:100 and applied to pre-treated tissue sections for 30 min at room temperature.The negative control lacked primary antibody and was made up of Dako Mouse IgG1 (Code no.X0931; Dako, Glostrup, Denmark), diluted to the same mouse IgG concentration as the primary antibody.The slides were stained in automated immunostaining kits (Dako EnVision+/HRP kits, code no.K4005; Glostrup, Denmark).
A monoclonal mouse Bax antibody (Clone 2D2, code no.33-6400; Thermo Scienti c, Waltham, MA, USA) was used for Bax immunochemical staining.The tissue sections were depara nized using 10 mM sodium citrate (pH 6.0), microwaved for 10 min, blocked in 3% H2O2-methanol for 15 min at room temperature, and then washed with double-distilled water and phosphate buffered saline (PBS).Tissue sections were probed with primary antibody diluted in 3% bovine serum albumin (BSA)-PBS (Blocker BSA in PBS, code no.37525; Thermo Scienti c, Waltham, MA, USA) at a dilution of 1:100 overnight at 4ºC in a humidi ed chamber.Tissue sections were washed in PBS with Tween 20 (Pierce Concentrated Buffer Stocks, code no.28352; Thermo Scienti c, Waltham, MA, USA) and incubated with a horseradish peroxidase-conjugated secondary antibody followed by colorimetric detection (Piece DAB substrate kit, code no.34002; Thermo Scienti c, Waltham, MA, USA).Finally, tissues were counterstained with hematoxylin and dehydrated with ethanol and xylene.

Histopathologic analysis
An independent neuropathologist, who was blinded to information on the groups, analyzed the scanned images using Aperio ImageScope (Version 12.4.6;Leica Biosystems, Deer Park, US).In H&E staining, structural changes of the basilar artery and penetrated arterioles were observed.The thickness of the tunica media in the basilar artery was measured from the internal elastic lamina to the outermost layer of the smooth muscle cell layer; the average value of the thickness at four different locations (at 90-degree intervals) of the artery was used.
The expression of Bcl2 and Bax was observed in each endothelial cell and smooth muscle cell of the basilar arteries and penetrated arterioles.The expression levels of Bcl2 and Bax were analyzed according to the intensity and continuity of staining in the endothelium using the following phenotypic scoring system: 3, strong staining or encircled staining; 2, moderate staining or partially continuous staining; 1, weak staining or scattered staining; 0, absence of brown staining.The intensity of brown color was based on von Luschan's Chromatic Scale: a scale of 14 or less was negative, 15 or more to 19 was weak, 20 or more to 26 was moderate, and 27 or more was strong staining.[24] The degree of expression was indicated by summing the scores for each of intensity and continuity.

Statistical analysis
Statistical analysis was performed using SPSS, version 22.0 (IBM corp., Armonk, NY, USA).Kruskal-Wallis analysis was used for comparative analysis of the lumina area of basilar arteries, the thickness of the tunica media, and scores for Bcl2 and Bax between groups.To check the causality between each variable, multiple linear regression analysis was used by classifying the luminal area and thickness of the tunica media in the basilar artery as the dependent variable and the degree of Bcl2 and Bax expression as the independent variable.All tests were two-sided with 0.05 denoting statistical signi cance.

Results
Of the 50 rabbits, 42 were included in this study (group C, n=15; group S, n=13; group N, n=14).Eight rabbits were excluded because of death during SAH induction.For specimens obtained from experimental animals, 42 basilar arteries (group C, n=15; group S, n=13; group N, n=14) and 64 penetrated arterioles (group C, n=19; group S, n=21; group N, n=24) were investigated.All animals with SAH exhibited reduced movement and food intake compared to group C.

Histopathologic change of the vasculature and brain stem
A thin smooth muscle layer, a at internal elastic layer, and elongated smooth muscle nuclei were observed in the basilar arteries of animals of group C (Figure 1).Hemorrhage around the basilar artery, a shrunken vessel diameter, the proliferation of smooth muscle cells, a corrugated internal elastic lamina, and con gurational change of the nuclei in smooth muscle cells due to tonic contracture of smooth muscle ber were observed in the basilar arteries of animals of groups S and N showing hemorrhage.Cytoplasmic vacuolation of smooth muscle cells was observed in animals of groups S and N, but, notably, this was more prominent in group S. Endothelial desquamation was observed in animals of group S only.A discontinuity of the internal elastic lamina was not observed all groups.In arterioles in the subarachnoid space, a proliferation of smooth muscle cells was observed in all groups showing hemorrhage.However, in penetrated arterioles, a proliferation of smooth muscle cells was not observed even in animals of the group showing hemorrhage.

Immunohistochemical expression of the vasculature and brain stem
In basilar arteries, the expression of Bcl2 in the endothelium was observed in all groups (Table 1).However, the degree of expression in tissues of animals of group C was very weak or non-existent, and no continuous expression was observed in the adjacent endothelium.For groups S and N, the intensity and continuity of staining for Bcl2 expression was more marked compared to tissues of animals of group C. Bax was expressed only in tissues from animals in groups S and N. Expression scores for Bcl2 and Bax in the endothelium were signi cantly higher for both groups S and N with hemorrhage than for group C without hemorrhage (p < 0.001, both).In a groupwise comparison, the expression of Bcl2 and Bax in the tissues of rabbits of group N was lower than for those of group S, but no signi cant differences were noted (p = 0.310 and p = 0.710, respectively).
In penetrated arterioles, the expression of Bcl2 in the endothelium was similar to that in the basilar artery.However, the expression of Bax in penetrated arterioles from tissues from animals belonging to group N was signi cantly lower than that of group S (p < 0.001).Comparing the expression of Bcl2 and Bax between basilar artery and penetrated arteriole, no statistical difference in the expression of Bcl2 was noted between the two vessels for groups S and N: Bax was not very expressed in penetrated arterioles in group N only, which was statistically signi cant (p< 0.001).

Comparative analysis of measured vessel parameters in basilar artery
The luminal areas of the basilar artery were signi cantly smaller in both groups S and N compared to group C (p < 0.001).In a groupwise comparison, areas for group N were wider than for group S, with a marginal signi cance (p = 0.098).The thickness of the tunica media in the basilar artery was signi cantly higher in both groups S and N than in group C (p < 0.001).In a groupwise comparison, the thickness for group N was signi cantly thinner than that for group S (p = 0.012).Data are summarized in Table 2.
In an analysis of independent variables affecting the luminal area (R 2 =0.631, corrected R 2 =0.612,F=33.36, p < 0.001) and thickness of the tunica media (R 2 =0.713, corrected R 2 =0.698,F=48.32, p < 0.001) in the basilar artery, a multiple linear regression model showed Bcl2 expression affected the luminal area narrowing (β=-0.559)and thickness of the tunica media (β=0.598) in the basilar artery, while Bax did not affect the luminal area (p = 0.077) but did affect the thickness of the tunica media in the basilar artery (β=0.292).

Discussion
Several randomized controlled trials have demonstrated improvement in clinical outcomes with enteral nimodipine after SAH due to cerebral aneurysm but the mechanisms remain unknown.[16-18]Although controversy exists concerning a difference in e cacy according to the route of administration of nimodipine, [25] the e cacy of nimodipine is clear.Several studies have described the neuroprotective mechanisms of nimodipine.Nimodipine improves microcirculation by reducing the release of thromboxane A2 from platelets,[26] or prevents a deleterious vascular response to cortical spreading depression in the ischemic brain.[27][28][29] In studies to date, it is known that the mechanism of DCI after rupture of a cerebral aneurysm is very multifactorial, and nimodipine, which is used as a treatment, is judged to improve clinical results by reducing secondary cerebral ischemia through various mechanisms.
Endothelial dysfunction immediately after rupture of a cerebral aneurysm is associated with endothelial apoptosis.[7,10] Endothelial dysfunction causes an imbalance of vascular regulators, disturbance of the microcirculation, and blood-brain barrier breakdown, which leads to DCI. [7] In animal studies, blockade of apoptosis improved cerebral vasospasm.[12][13][14] The Bcl2 family, including pro-and anti-apoptotic proteins, is directly involved in calcium homeostasis in the cytoplasm between mitochondria and the endoplasmic reticulum, and controls bioenergetics, adenosine triphosphate production, and reactive oxygen species through a calcium pool of the mitochondria to in uence cell-fate decisions including apoptosis, cell survival, and cell migration.[30,31] Bcl2, an anti-apoptotic protein, interacts with the inositol 1,4,5-trisphosphate receptor to reduce the release of calcium from the endoplasmic reticulum to the mitochondria, thereby maintaining mitochondrial bioenergetics and protecting against calciuminduced apoptosis.[30,31] In contrast, Bax, a pro-apoptotic protein, promotes the release of calcium from the endoplasmic reticulum and apoptotic cascade.[30,31] In a pathological situation, such as SAH, the sudden extracellular in ux of calcium through L-type calcium channels increases Bcl2 expression to transport cytoplasmic calcium to the endoplasmic reticulum.[30] However, when the calcium store in the endoplasmic reticulum exceeds the limit, Bax is expressed, thereby releasing calcium from the endoplasmic reticulum, triggering an apoptotic series.
In our immunohistochemical analysis, tissues from rabbits in group C barely expressed Bcl2 and showed no expression of Bax in both large arteries and arterioles.Tissues from rabbits in Group S showed a higher expression of Bcl2 and Bax in both arteries and penetrated arterioles.However, for group N, the expression of Bcl2 in both large arteries and penetrated arterioles, and Bax expression in large arteries were lower than those of tissues from animals in group S, despite a lack of statistical signi cance.Of note, Bax in penetrated arterioles was signi cantly less expressed.The reason why nimodipine has a greater effect on the penetrated arteriole than on large arteries is that the latter are thought to be more directly affected by extrinsic factors such as oxyhemoglobin in the subarachnoid space.[32]It is already known that nimodipine has a better vasodilator effect on smaller arteries.[33,34] That is, nimodipine in the penetrated arteriole with few extrinsic factors partially blocks the in ux of extracellular calcium to maintain cytoplasmic calcium homeostasis and prevent a calcium-dependent apoptotic cascade.As a result, Bax is less expressed.Multiple linear regression analysis revealed that the expression of Bcl2 in the basilar artery had a negative effect on the luminal area and a positive effect on the thickness of the tunica media, showing a greater effect than Bax.It is judged that the level of Bcl2 expression in endothelial cells of the basilar artery can affect the proliferation of smooth muscle cells.
Histological changes in the cerebral arteries after SAH are already well known, our ndings are also similar, [7,14,35] however, several peculiarities existed.First, cytoplasmic vacuolization in the smooth muscle cells of the basilar artery was observed in both groups S and N.This nding was previously reported.[14] Cytoplasmic vacuolization is a known morphological phenomenon observed in mammalian cells exposed to various natural and arti cial low molecular weight compounds.[36]Cytoplasmic vacuolization is associated with cell death[36] in a mechanism that differs from apoptosis because it was not stained in our immunohistochemical studies.It is unclear whether the vacuolization seen in our study was caused by SAH or other situations (administrated drugs or cerebral damage due to air embolism), and therefore further research is needed.However, notably, vacuolization was more frequently and prominently observed in group S than group N. Second, the thickness of the tunica media for group N was signi cantly thinner than for group S and the luminal area of basilar arteries for group N was larger than for group S with a marginal signi cance.That is, nimodipine suppressed the proliferation of smooth muscle cells, and dilated the luminal area of the vasoconstricted basilar artery.Vasoconstrictors, such as endothelin-1, are released after SAH increases the reactivity of endothelial cells, activates the contraction and proliferation of smooth muscle cells, and activates other growth factors.[37,38] Endothelial desquamation meant that smooth muscle cells were more easily exposed to neurotransmitters, toxins, and vasoactive agents.[14] In our study, endothelial desquamation was observed only in group S but not group N. Therefore, in addition to indirectly assisting the production of vasodilators, nimodipine is considered to contribute to the inhibition of smooth muscle cell proliferation by maintaining the functional integrity of endothelial cells through calcium homeostasis.Several studies have been made on the inhibition of vascular smooth muscle cell proliferation by nimodipine.[35,39] Compared to group C, histopathological changes and an arterial area reduction in groups S and N suggest that the experimental SAH model was well implemented.Nevertheless, this study has several limitations.First, the number of animals participating in this study was marginal for a three-group comparison.Second, the analysis of immuno uorescence staining was performed only qualitatively.If quantitative analysis such as cytometry was performed, more reliable data would have been obtained.Capillaries were also not included in our study, and therefore further research on the endothelium of the capillary is needed.

Conclusion
This study suggests that enteric nimodipine may have a neuroprotective function by inhibiting endothelial apoptosis in small arterioles and preventing smooth muscle cell proliferation in large arteries.It is believed that these ndings solidify the effect of enteral nimodipine after SAH through immunohistochemical staining of apoptosis-related proteins.However, other functions of nimodipine may exist and additional research is needed.agni cation; scale bar at bottom left of gure).Group C showed weak Bcl2 expression and no Bax expression.Bcl2 and Bax in the basilar artery were similarly expressed in groups S and N.However, in penetrated arterioles, Bax was more expressed in group S of Bax (white arrow) than in group N; the expression of Bcl2 (black arrow) was similar in the two groups.

Figures Figure 1
Figures

Table 1 .
32. Ogihara K, Zubkov AY, Bernanke DH, Lewis AI, Parent AD, Zhang JH.Oxyhemoglobin-induced apoptosis in cultured endothelial cells.J Neurosurg.1999;91(3):459-65.33.Brandt L, Ljunggren B, Andersson KE, Edvinsson L, MacKenzie E, Tamura A, Teasdale G. Effects of topical application of a calcium antagonist (nifedipine) on feline cortical pial microvasculature under normal conditions and in focal ischemia.J Cereb Blood Flow Metab.1983;3(1):44-50.Comparative analysis of immunohistochemical staining scores of each group ‡ is the value of the luminal area of the basilar artery and p § is the value of the thickness of the tunica media.
39. Ko Y, Totzke G, Graack GH, Heidgen FJ, Meyer zu Brickwedde MK, Düsing R, Vetter H, Sachinidis A. Action of dihydropyridine calcium antagonists on early growth response gene expression and cell growth in vascular smooth muscle cells.J Hypertens.1993;11(11):1171-8.Tables C: Control group, S: Hemorrhage group, N: Nimodipine group.p † value by Kruskal-Wallis test.Each groupwise comparison was analyzed by a Kruskal-Wallis test and adjusted with a Bonferroni correction.p

Table 2 .
Microscopic comparative analysis of luminal area and thickness of tunica media in the basilar C: Control group, S: hemorrhage group, N: Nimodipine group.p † value by Kruskal-Wallis test.p ‡ value by Kruskal-Wallis test and adjusted with Bonferroni correction for multiple comparisons in the luminal area of the basilar artery.p § value by Kruskal-Wallis test and adjusted with Bonferroni correction for multiple comparisons in the thickness of the tunica media.