Inhibition of nitric oxide synthase transforms carotid occlusion-mediated benign oligemia into de novo large cerebral infarction

Abstract

Rationale: It remains unclear why unilateral proximal carotid artery occlusion (UCAO) causes benign oligemia in mice, yet leads to various outcomes (asymptomatic-to-death) in humans. We hypothesized that inhibition of nitric oxide synthase (NOS) both transforms UCAO-mediated oligemia into full infarction and expands pre-existing infarction. Methods: Using 900 mice, we i) investigated stroke-related effects of UCAO with/without intraperitoneal administration of the NOS inhibitor (NOSi) N_ω-nitro-L-arginine methyl ester (L-NAME, 400 mg/kg); ii) examined the rescue effect of the NO-donor, molsidomine (200 mg/kg at 30 minutes); and iii) tested the impact of antiplatelet medications. To corroborate preclinical findings, we conducted clinical studies. Results: UCAO alone induced infarction rarely (~2%) or occasionally (~14%) in C57BL/6 and BALB/c mice, respectively. However, L-NAME+UCAO induced large-arterial infarction in ~75% of C57BL/6 and BALB/c mice. Six-hour laser-speckle imaging detected spreading ischemia in ~40% of C57BL/6 and BALB/c mice with infarction (vs. none without) by 24-hours. In agreement with vasoconstriction/microthrombus formation shown by intravital-microscopy, molsidomine and the endothelial-NOS-activating antiplatelet cilostazol attenuated/prevented progression to infarction. Moreover, UCAO without L-NAME caused infarction in ~22% C57BL/6 and ~31% ApoE knock-out mice with hyperglycemia/hyperlipidemia, which associated with ~60% greater levels of symmetric dimethylarginine (SDMA, an endogenous NOSi). Further, increased levels of glucose and cholesterol associated with significantly larger infarct volumes in 438 UCAO-stroke patients. Lastly, Mendelian randomization identified a causative role of NOS inhibition (elevated SDMA concentration) in ischemic stroke risk (OR = 1.24; 95% CI, 1.11-1.38; P = 7.69×10^-5). Conclusion: NOS activity determines the fate of hypoperfused brain following acute UCAO, where SDMA could be a potential risk predictor.

Keywords: carotid artery occlusion; cerebral infarction; nitric oxide synthase; oligemia; stroke.

Figure 1

L-NAME+UCAO in mice can induce acute large cerebral infarction, which could be rescued by the NO donor molsidomine. (A) Top, experimental timeline for Experiment 1; TTC staining was performed after premature death or pre-planned sacrifice on day 1 or 7 for the sham surgery, L-NAME only, and UCAO only groups, and day 1, 2, 3, or 6 for the L-NAME+UCAO group. Middle, TTC staining of the brains from representative C57BL/6 mice (except for UCAO - lethal infarction of a BALB/c mouse) with no, small (≤ 100 mm³), or large (> 100 mm³) infarction in each of the four (sham surgery, L-NAME, UCAO, and L-NAME+UCAO) groups. Bottom-left, incidence of cerebral infarction by a pre-specified sacrifice time-point in the L-NAME+UCAO group. Light, intermediate, and dark colors represent small, large, and lethal infarction, respectively. Bottom-middle, stroke-related mortality curves in the L-NAME+UCAO group. Bottom-right, infarct volumes of the mice that survived until the pre-specified sacrifice time-points in the L-NAME+UCAO group. (B) Top, experimental timeline for L-NAME post-treatment (UCAO+L-NAME): L-NAME administration at a different post-UCAO time-point, followed by sacrifice 24 h later. Bottom-left, incidence of cerebral infarction. Bottom-right, infarct volumes of the mice that survived until the 24 h sacrifice time-point. For easier comparison with the corresponding L-NAME pre-treatment (L-NAME+UCAO) data in (A), the 1 d infarct-incidence and -volume bars are presented here again (-1 min). (C) Top, timeline for molsidomine experiments. Bottom-left, incidence of cerebral infarction by 24 h sacrifice time-point in L-NAME+UCAO mice treated with a single intraperitoneal dose of molsidomine or saline. Bottom-right, infarct volumes of the surviving mice. Infarct volumes are presented as box-and-whiskers plots. ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001. L-NAME = N_ω-nitro-L-arginine methyl ester; N/A = not applicable; NO = nitric oxide; TTC = 2,3,5-triphenyltetrazolium chloride; UCAO = unilateral proximal carotid artery occlusion.

Figure 2

LSCI for 6 h after L-NAME+UCAO detected SI in ~40% of mice with cerebral infarction assessed at 24 h, but in none without. (A) Timeline for experiments and persistent SI in the hemisphere ipsilateral to UCAO in a BALB/c mouse. SI began, with a substantial drop in rCoBF from the oligemia (> 50%, black arrow) to the severe ischemia level (< 30%, white arrows), in the core region (ROI-1), spreading anteromedially to encompass the majority of the ipsilateral hemisphere. Except for the medial portion (*), there was no rCoBF recovery (white arrow-head). See Video S1. (B) Transient SI (black and white arrows) in the ipsilateral hemisphere of another BALB/c mouse. See Video S2. (C) Transient SI with CoBF recovery (white and red arrows) in the contralateral hemisphere of another BALB/c mouse. See Video S3. (D) LS mean rCoBFs with 95% CIs, calculated after stratifications by the occurrence of SI [with (+) or without (-) rCoBF recovery up to 6 h] and cerebral infarction (up to 24 h). Gray, black, and blue ^* indicate P < 0.05 vs. the SI(-)·Non-infarcted group, SI(-)·Infarcted group, and SI(+)·Recovery(+)·Infarcted group, respectively, at each time period. Green and pink ^* indicate P < 0.05 vs. 10-30 and 30-330 min, respectively. Additionally, rCoBF values (mean ± SE) from 3 min before to 3 min after the last SI are displayed for each SI-positive group (upper graphs in the shaded areas). CI = confidence interval; L-NAME = N_ω-nitro-L-arginine methyl ester; LS = least squares; LSCI = laser speckle contrast imaging; rCoBF = regional cortical blood flow; ROI = region of interest; SE = standard error; SI = spreading ischemia; TTC = 2,3,5-triphenyltetrazolium chloride; UCAO = unilateral proximal carotid artery occlusion.

Figure 3

L-NAME+UCAO-mediated SI and infarction were not associated with systemic hemodynamic instability. (A) Timeline for LSCI and heart rate/BP experiment. (B-E) Heart rate, systolic and diastolic BPs, and rCoBF (all as mean ± SE) in the oligemic core (ROI-1): (B) saline-injected control group, (C) L-NAME only group, (D) UCAO only group, and (E) L-NAME+UCAO group, with stratifications by the occurrence of SI and cerebral infarction (up to 24 h); data for minor groups with sample sizes below 5 are shown in Figure S6. LS mean values with 95% CIs are presented in the graphs. Gray ^* indicates P < 0.05 vs. the SI(-)·Non-infarcted group at each time period. Orange and green ^* indicate P < 0.05 vs. baseline and 0-10 min. In addition to the statistical analyses, rCoBF values for the last SI during the 10-70 min period are displayed for each SI-positive group (as group mean ± SE) in order to show SI-related rCoBF drop to a trough level (inset graph in the shaded area). Corresponding heart rate and BP data (at the time-point of the lowest rCoBF) are also presented in inset graphs. BP = blood pressure; CI = confidence interval; L-NAME = N_ω-nitro-L-arginine methyl ester; LS = least squares; LSCI = laser speckle contrast imaging; rCoBF = regional cortical blood flow; ROI = region of interest; SE = standard error; SI = spreading ischemia; TTC = 2,3,5-triphenyltetrazolium chloride; UCAO = unilateral proximal carotid artery occlusion.

Figure 4

Abundant microthrombi were observed in infarcted mouse brain tissue 24 h post L-NAME+UCAO, corroborated by additional experiments in which the vasodilatory antiplatelet cilostazol reduced the occurrence of severe infarction, probably, in part, via its antithrombotic effects. (A) Representative high-resolution in vivo microCT images showing that fibrin-targeted gold nanoparticles can clearly identify sizeable thrombi and their evolution (yellow arrows and arrow-heads) after tPA therapy for embolic stroke. (B and C) Assessment of the presence of macrovascular and microvascular thrombi: (B) no evidence of macrovascular thrombosis on microCT imaging (serially for 6 d) vs. (C) histologic (H&E staining) evidence of abundant microvascular thrombosis (black arrows) predominantly in infarcted brain tissue (collected following sacrifice with cardiac perfusion). Venous thrombi are also apparent (blue arrows), whereas large arterial thrombi are not (red arrows). (D) Top, timeline for experiments to test if 8 d pre-treatment with an antiplatelet drug can show protective effects against L-NAME+UCAO-mediated infarction in C57BL/6 mice. Bottom-left, stroke-related mortality curves for the saline, aspirin, clopidogrel, and cilostazol groups. The mortality is lowest in the cilostazol group (P = 0.018 vs. saline group, log-rank test with post-hoc pairwise comparisons). Bottom-middle, infarction incidence by 6 d after stroke. Light, intermediate, and dark colors represent small (≤ 100 mm³), large (> 100 mm³), and lethal infarction, respectively. Severe (large or lethal) infarction was significantly less frequent in the cilostazol group (P = 0.046 vs. saline group, chi-square test). Bottom-right, infarct volumes (box-and-whiskers plots) of the mice that survived until 6 d. H&E = hematoxylin and eosin; L-NAME = N_ω-nitro-L-arginine methyl ester; microCT = micro computed tomography; tPA = tissue-type plasminogen activator; TTC = 2,3,5-triphenyltetrazolium chloride; UCAO = unilateral proximal carotid artery occlusion.

Figure 5

Intravital microscopy through a cranial window showed microvascular constriction and thromboembolism to impede distal microcirculation after L-NAME+UCAO. (A-C) Intravital images for the squared regions within the circles (far-right). (A) Stack images before and after L-NAME+UCAO in a BALB/c mouse. At 1 h, Texas-red-Dextran-positive vascularity is markedly decreased. At 4 h, post-mortem autofluorescence imaging and TTC staining (far-right) reveal whitish ischemic lesions (red arrows). (B) Transient constriction of the Y-shape arteriole in real-time (white arrow-heads) ~4 h after L-NAME+UCAO in a different BALB/c mouse (Video S4; See also Figure S7 for an additional case). Also note the congruent drop and rise in the blood flow of an adjacent arteriole (yellow arrow-heads), and Rhodamine-6G-positive platelets and leukocytes in the nearby venules (blue arrows). Post-mortem evidence of hemispheric infarction (bottom-row, far-right) suggests additional instances of vasoconstriction by ~24 h. (C) Stack images before and after L-NAME+UCAO in another BALB/c mouse. At 4 h, Rhodamine-6G-positive platelets and leukocytes (green and black arrows) are abundant in venules (top-row, far-right). Moreover, the magnified views (middle-row) of the blue squared regions (top-row) show that capillary plugging (white arrows) accompanies distal flow reduction (cyan and yellow arrows). Also note decreased vascular density at 4 h (* in the bottom-row). See Figure S8 for grouped quantification data for vascular diameter and density in all animals (n = 16). A representative non-infarcted mouse with no notable vascular changes is presented in Figure S9. L-NAME = N_ω-nitro-L-arginine methyl ester; TTC = 2,3,5-triphenyltetrazolium chloride; UCAO = unilateral proximal carotid artery occlusion.

Figure 6

UCAO without L-NAME could induce acute large artery cerebral infarction in mice with hyperglycemia and hyperlipidemia. (A) Top, timeline for experiments on the effects of prior HFD and/or STZ treatment in C57BL/6 (n = 75) and ApoE^-/- (n = 59) mice receiving UCAO. Middle, blood levels of glucose and cholesterol (total cholesterol, HDL cholesterol, LDL cholesterol, and TG) after HFD and/or STZ treatment: ^*P < 0.05, ^**P < 0.01, and ^***P < 0.001 (black and blue: inter-group difference; gray: inter-strain difference), Two-way ANOVA and Sidak's multiple comparisons for post-hoc tests. Bottom-left, representative TTC staining images in each group, stratified by lesion volume (no, small [≤ 100 mm³], and large [> 100 mm³] infarction). Bottom-right, incidence of cerebral infarction and infarct volumes (box-and-whiskers plots). Infarctions occurred in the STZ+HFD groups only, regardless of strain. Lesion volumes did not differ significantly between the strains. (B) Top, timeline for experiments to examine the effects of HFD+STZ treatment (n = 10 HFD+STZ-treated and 10 saline-treated C57BL/6 mice) on two endogenous NOSi (ADMA and SDMA) as well as on glucose and cholesterol. Bottom, measurement results (box-and-whiskers plots): ^***P < 0.001, Mann-Whitney U test. ADMA = asymmetric dimethylarginine; ApoE = apolipoprotein E; HDL = high-density lipoprotein; HFD = high-fat diet; LDL = low-density lipoprotein; L-NAME = N_ω-nitro-L-arginine methyl ester; NOSi = nitric oxide synthase inhibitor; SDMA = symmetric dimethylarginine; STZ = streptozotocin; TG = triglyceride; TTC = 2,3,5-triphenyltetrazolium chloride; UCAO = unilateral proximal carotid artery occlusion.

Figure 7

Mendelian randomization analysis showed a causative role of SDMA, an endogenous NOSi, in human ischemic stroke. (A) Characteristics of SNPs used in the Mendelian randomization study. # indicates an excluded SNP as a cross-correlated SNP, determined by linkage disequilibrium clumping (R² < 0.01). EA, OA, and EAF indicate effect allele, other alleles, and effect allele frequency, respectively. SNP effect was calculated as increase in exposure (μmol/L) per EA. (B) Forest plot for the effect of exposures (L-arginine, ADMA, and SDMA) on risk of ischemic stroke, assessed by using different Mendelian randomization methods. OR was estimated using all SNPs for each exposure (per 1 SD increase in L-arginine [25.2 μmol/L], ADMA [0.13 μmol/L], and SDMA [0.12 μmol/L]). The error bars represent the 95% CIs. (C) Left, Forest plot for the effect of each SDMA SNP on risk of ischemic stroke. The error bars represent the 95% CIs. Right, Scatter plot for the effect of SDMA on risk of ischemic stroke. The error bars represent SE. Regression lines for four Mendelian randomization methods are shown. ADMA = asymmetric dimethylarginine; CI = confidence interval; IVW = inverse variance weighted; MR-Egger = Mendelian randomization-Egger; NOSi = nitric oxide synthase inhibitor; OR = odds ratio; SD = standard deviation; SDMA = symmetric dimethylarginine; SE = standard error; SNP = single nucleotide polymorphisms.