Angiotensin II has been implicated in the progression of diabetic retinopathy, which is characterized by altered microvasculature, oxidative stress, and neuronal dysfunction. The signaling induced by angiotensin II can occur not only via receptor-mediated calcium release that causes vascular constriction, but also through a pathway whereby angiotensin II activates NADPH oxidase to elicit the formation of reactive oxygen species (ROS). In the current study, we administered the angiotensin II receptor antagonist candesartan (or vehicle, in untreated animals) in a rat model of type 1 diabetes in which hyperglycemia was induced by injection of streptozotocin (STZ). Eight weeks after the STZ injection, untreated diabetic rats were found to have a significant increase in tissue levels of angiotensin converting enzyme (ACE; p < 0.05) compared to non-diabetic controls, a 33% decrease in retinal blood flow rate (p < 0.001), and a dramatic increase in p22phox (a subunit of the NADPH oxidase). The decrease in retinal blood flow, and the increases in retinal ACE and p22phox in the diabetic rats, were all significantly attenuated (p < 0.05) by the administration of candesartan in drinking water within one week. Neither STZ nor candesartan induced any changes in tissue levels of superoxide dismutase (SOD-1), 4-hydroxynonenal (4-HNE), or nitrotyrosine. We conclude that one additional benefit of candesartan (and other angiotensin II antagonists) may be to normalize retinal blood flow, which may have clinical benefits in diabetic retinopathy.

ObjectiveThe purpose of this study is to investigate the retinal effects of methamphetamine.IntroductionMETH, a highly addictive stimulant of neurotransmitters is associated with cardiovascular and neurological diseases. METH‐induced ophthalmic complications are also present but have been insufficiently investigated.Materials and methodsC57BL/6J mice were administrated progressively increasing doses of METH (0 to 6 mg/kg) by repetitive intraperitoneal injections for 5 days (4 times per day). Norepinephrine levels in plasma were measured by ELISA. Cross‐sections of the central retinal artery (CRA) and retina were stained with hematoxylin and eosin. Expression of various proteins was determined by immunostain and immunoblot in the CRA and retina.ResultsElevated norepinephrine levels in plasma were induced by METH (6.49±1.72 ng/mL, n=5), compared with saline treatment (1.38±0.48, P<0.001, n=5). After treatment of METH, the ocular pupil was dilated and retinal thicknesses of the outer nuclear layer, outer plexiform layer, inner nuclear layer, and inner plexiform layer were decreased. Platelet endothelial cell adhesion molecule‐1 (PECAM‐1) protein expression level was decreased in the CRA and retina of METH‐treated mice, along with the endothelial proteoglycans glypican‐1 and syndecan‐1. Moreover, a regulator of the extracellular matrix, membrane type 1 matrix metalloproteinase (MMP‐14), but not MMP‐2 and MMP‐9, was increased in the retina by METH treatment.ConclusionThese results suggest that METH administration is involved in retinal degeneration with a vascular loss of PECAM‐1 and the glycocalyx, and an increase of MMP‐14 in the CRA and retina.Support or Funding InformationThis work was supported by funding from the National Institute of Health (NIH) EY025632.This abstract is from the Experimental Biology 2019 Meeting. There is no full text article associated with this abstract published in The FASEB Journal.

Diabetic retinopathy is known as a microvascular complication of hyperglycemia, with a breakdown of the blood-retinal barrier, loss of pericytes, formation of microhemorrhages, early decreases in perfusion and areas of ischemia, with the latter speculated to induce the eventual proliferative, angiogenic phase of the disease. Our animal models of diabetic retinopathy demonstrate similar decreases in retinal blood flow as seen in the early stages of diabetes in humans. Our studies also show an alteration in the retinal distribution of red blood cells, with the deep capillary layer receiving a reduced fraction, and with flow being diverted more towards the superficial vascular layer. Normal red blood cell distribution is dependent on the presence of the endothelial surface layer, specifically the glycocalyx, which has been reported to be partially lost in the diabetic retina of both humans and animals. This review addresses these two phenomena in diabetes: altered perfusion patterns and loss of the glycocalyx, with a possible connection between the two.

Patients with inflammatory bowel disease suffer not only from gut inflammation, but also from extraintestinal manifestations of the disease, including ocular pathology. The mechanisms causing ocular inflammation in these patients are unknown. The purpose of the current study was to investigate the possible vascular changes occurring in the retina using a mouse model of acute colitis, that is, ingestion of dextran sodium sulfate (DSS). Intravital microscopy of anesthetized mice revealed that DSS caused a significant 30–40% decrease in retinal red blood cell velocities, and a 45% decrease in total retinal blood flow, but no changes in intraocular pressure. To determine whether the decreases in retinal perfusion could be inhibited by an angiotensin II receptor antagonist, losartan was administered by eye drops in a subset of the mice prior to the intravital microscopy measurements. Topical losartan was able to largely attenuate the altered hemodynamics induced by DSS. We conclude that angiotensin II might be a possible target for reducing the vascular changes occurring distantly in the eye during colitis.

The objective of this study was to evaluate potential changes in platelet endothelial cell adhesion molecule‐1 (PECAM‐1) and capillary density occurring early in the diabetic retina. In these experiments, streptozotocin (STZ; or vehicle alone, for controls) was injected into rats (30 mg/kg for 3 days) to induce a model of type I diabetes. Eight weeks later, a fluorescent plasma tracer (high molecular weight FITC‐dextran) was injected i.v. into the anesthetized rats, and the eyes were excised just prior to euthanasia. One eye was used in a flatmount of the retinal tissue, with capillary density quantified using images captured through a microscope camera. The other eye was used for western blot analysis. Levels of PECAM‐1 were found to decrease by 44% (p<0.05), but in contrast, no changes were observed in the number of perfused capillaries (that is, perfused with FITC‐dextran). Further studies will be needed to examine the possibility, and implications, of diabetic retinal capillaries expressing lower levels of PECAM‐1. Funded by NIH EY017599 (NRH).

BACKGROUND:Approximately one-half of patients with inflammatory bowel disease (IBD) suffer from anemia, with the most prevalent cause being iron deficiency. Accompanying the anemia are increases in erythropoietin, a plasma protein that can initiate the feedback production of new red blood cells. Anemia also occurs in animal models that are used to investigate the mechanisms of IBD; however, the extent to which iron deficiency produces the anemia in these animal models is unknown. Also unknown in the different animal models of IBD is whether the anemia upregulates the production of erythropoietin or, alternatively, whether a decrease in erythropoietin contributes to the induction of anemia.METHODS:Two mouse models of colitis were used in this study: (1) acute 6-day ingestion of dextran sodium sulfate and (2) T-cell transfer into lymphopenic recipient mice. Measurements included indices of colitis severity, hematocrit, blood hemoglobin, plasma erythropoietin, serum iron concentration, plasma iron-binding capacities, transferrin saturation, and tissue iron concentrations.RESULTS:Both models of colitis induced significant decreases in hematocrit, blood hemoglobin, and transferrin saturation, with the spleen and liver showing a decrease in iron content in the T-cell transfer model. Additionally, both models of colitis demonstrated significant increases in plasma erythropoietin and plasma iron-binding capacities.CONCLUSIONS:The measurements of iron, whether in acute (dextran sodium sulfate) or chronic (T-cell transfer) models of colitis, were generally consistent with iron-deficient anemia, with large increases in erythropoietin indicative of tissue hypoxia. These changes in animal models of colitis are similar to those found in human IBD.

BACKGROUND:Ocular disease is known widely to occur in a subset of patients experiencing inflammatory bowel diseases. Although this extraintestinal manifestation has been recognized for a number of years, the pathogenetic mechanisms responsible for this distant organ inflammatory response are unknown.METHODS:In the current study, we used a T-cell transfer model of chronic colitis in mice in which we quantified colonic inflammation, ocular function (electroretinography), ocular blood flow (intravital microscopy of the retina), intraocular pressure, and retinal hypoxia.RESULTS:Ocular function in colitic mice was significantly impaired, with decreases in retinal b-wave amplitudes and oscillatory potentials. Moreover, retinal a waves and oscillatory potentials were delayed. Retinal blood flow was significantly reduced in the colitic mice, and this decrease in perfusion coupled with significant decreases in hematocrit would decrease oxygen delivery to the eye. Accordingly, mice with severe colitis showed increased levels of immunostaining for the hypoxia-dependent probe pimonidazole. Finally, intraocular pressures were found to be reduced in the colitic mice.CONCLUSIONS:Ocular disease occurs in a mouse model of chronic colitis, with retinal dysfunction seeming to be related to insufficient perfusion and oxygen delivery.

The objective of this study was to clarify whether retinal blood flow (RBF) rates are altered in rats during streptozotocin (STZ)‐induced diabetes. Reports of this possibility in the literature are mixed, possibly due to the different techniques used to measure RBF. STZ (or vehicle alone, in controls) was injected into the rats (30 mg/kg for 3 days) to induce a model of type I diabetes. Eight weeks later, RBF rates were measured in the anesthetized rats using intravital microscopy and a fluorescent tracer (FITC‐dextran) in a mean circulation (transit) time index of flow. In separate rats, RBF was measured using a microsphere infusion technique. The induction of diabetes resulted in a 30–35% decrease in RBF (***p<0.001) in the STZ rats as determined by the microsphere measurement. However, in a striking contrast to these results, the mean transit times of injected FITC‐dextran were ~3× faster in the STZ rats compared with the controls (**p<0.01). We hypothesize that this discrepancy is due to shunting of RBF away from the deep retinal capillary layers in the STZ rats, with the altered flow pattern allowing the rapid transit of tracer from the arterioles into the venules. Funded by NIH EY017599 (NRH).

PURPOSE:Reports of altered retinal blood flow in experimental models of type I diabetes have provided contrasting results, which leads to some confusion as to whether flow is increased or decreased. The purpose of our study was to evaluate early diabetes-induced changes in retinal blood flow in diabetic rats, using two distinctly different methods.METHODS:Diabetes was induced by injection of streptozotocin (STZ), and retinal blood flow rate was measured under anesthesia by a microsphere infusion technique, or by an index of flow based on the mean circulation time between arterioles and venules. Measurements in STZ rats were compared to age-matched nondiabetic controls. In addition, the retinal distribution of fluorescently-labeled red blood cells (RBCs) was viewed by confocal microscopy in excised flat mounts.RESULTS:Retinal blood flow rate was found to decrease by approximately 33% in the STZ rats compared to controls (P < 0.001) as assessed by the microsphere technique. However, in striking contrast, the mean circulation time through the retina was found to be almost 3× faster in the STZ rats (P < 0.01). This contradiction could be explained by flow redistribution through the superficial vessels of the diabetic retina, with this possibility supported by our observation of significantly fewer RBCs flowing through the deeper capillaries.CONCLUSIONS:We conclude that retinal blood flow rate is reduced significantly in the diabetic rat, with a substantial decrease of flow through the capillaries due to shunting of blood through the superficial layer, allowing rapid transit from arterioles to venules.

Objective In the current study of murine colitis, the potential roles of thromboxane and the thromboxane-prostanoid (TP) receptor were investigated, in as much as thromboxane signaling has been implicated in human inflammatory bowel disease.Methods Colitis was induced in C57BL/6 mice via ingestion of dextran sodium sulfate (DSS), with or without co-administration of the thromboxane synthase inhibitor ozagrel (25 mg/kg/day) or the TP receptor antagonist vapiprost (2.5 mg/kg/day).Results Immunohistochemistry of colonic tissue demonstrated a DSS-induced increase in TP receptor expression, but not of thromboxane synthase. Moreover, tissue levels of the metabolite thromboxane B-2 were unchanged by DSS. Vapiprost, but not ozagrel, partially attenuated histologic signs of inflammation induced by DSS, with vapiprost allowing a smaller increase in colon weight per unit length than ozagrel. Vapiprost also tended to attenuate DSS-induced alterations in intestinal transit.Conclusions In summary, TP receptor antagonism was more effective than thromboxane synthase inhibition in alleviating DSS-induced colitis in mice.

The mechanisms by which microvascular alterations contribute to the pathogenesis of the inflammatory bowel diseases (IBDs; Crohn's disease, ulcerative colitis) have not been clearly delineated. The purpose of the current study was to characterize the inflammatory events, microvascular alterations, and blood cell changes that occur in a mouse model of IBD. In this model, CD4(+) T-lymphocytes obtained from interleukin-10-deficient mice were injected intraperitoneally into lymphopenic, recombinase-activating gene-1 deficient (RAG(-/-)) mice. Two groups of control mice were also included: RAG(-/-) mice and C57BL/6 mice that were injected with phosphate-buffered saline but did not receive the T-cells. Four weeks later, the RAG(-/-) mice that had received the T-cell transfer showed significant signs of colonic inflammation, but without significant decreases in either body weight or mean arterial blood pressure. T-cell transfer increased the volume % of circulating platelets, while decreasing the number of circulating red blood cells. Additionally, the T-cell transfer tended to increase the circulating numbers of both lymphocytes and neutrophils when compared to unmanipulated RAG(-/-) mice. First-order colonic arterioles and venules tended to dilate in the colitic mice; however, the dilation was considerably more substantial with higher numbers of circulating leukocytes. The possibility that circulating inflammatory cells initiate the microvascular alterations in colitis warrants further investigation.

BACKGROUND:Hypoxia has been reported to be associated with the colonic inflammation observed in a chemically induced mouse model of self-limiting colitis, suggesting that low tissue oxygen tension may play a role in the pathophysiology of inflammatory tissue injury. However, no studies have been reported evaluating whether tissue hypoxia is associated with chronic gut inflammation. Therefore, the objective of the present study was to determine whether hypoxia is produced within the colon during the development of chronic gut inflammation. METHODS:Adoptive transfer of CD4(+) T cells obtained from interleukin-10-deficient (IL-10(-/-)) mice into lymphopenic recombinase-activating gene-1-deficient (RAG(-/-)) mice induces chronic colonic inflammation, with the inflammation ranging from mild to severe as determined by blinded histological analyses. Colonic blood flow, hematocrit, and vascular density were determined using standard protocols, whereas tissue hypoxia was determined using the oxygen-dependent probe pimonidazole. RESULTS:Adoptive transfer of IL-10(-/-) CD4(+) T cells into RAG(-/-) recipients induced chronic colonic inflammation that ranged from mild to severe at 8 weeks following T-cell transfer. The colitis was characterized by bowel wall thickening, goblet cell dropout, and inflammatory infiltrate. Surprisingly, we found that animals exhibiting mild colonic inflammation had increased hypoxia and decreased systemic hematocrit, whereas mice with severe colitis exhibited levels of hypoxia and hematocrit similar to healthy controls. In addition, we observed that the extent of hypoxia correlated inversely with hematocrit and vascular density. CONCLUSIONS:Changes in hematocrit, vascular density, and inflammatory state appear to influence the extent of tissue oxygenation in the T-cell-mediated model of chronic gut inflammation.

Adoptive transfer of naïve CD4+ T‐cells into lymphopenic mice induces chronic small and large bowel inflammation similar to Crohn's disease. Although much is now known regarding the immunopathology in this model of inflammatory bowel disease, virtually nothing is known about the microvascular hemodynamic changes in the induction and perpetuation of chronic gut inflammation. In this study, CD4+CD45RBhigh T‐cells obtained from healthy C57BL6 donors were transferred into lymphopenic recombinase activating gene‐1 deficient (RAG ko) mice, and in the following weeks, intravital microscopy was used to examine the ileum and proximal colon. Measurements included arteriolar diameters and the velocity of labeled platelets (indicator of blood velocity), with the results compared to unreconstituted RAG ko controls. In < 1 wk following reconstitution, velocity and wall shear rate of the arterioles decreased by > 50% compared to controls, with this decrease also observed at 4‐5 and 7‐9 weeks post‐reconstitution. At 7‐9 weeks, arteriolar diameters were found to be ~15% larger than in controls, but despite this dilation, flow rates in the individual vessels were decreased by ~30%. Whether the overall intestinal flow rate is decreased in this model may be dependent on the reported angiogenesis, which could explain the lower individual vessel flows despite the late vasodilation. Supported by Projects 1 & 2 of NIH P01DK043785.

Background:Adoptive transfer of naive T‐lymphocyte subsets into lymphopenic mice initiates chronic gut inflammation that mimics several aspects of inflammatory bowel disease (IBD). Patients with IBD can have profound alterations in intestinal blood flow, but whether the same is true in the T‐cell transfer model has yet to be determined. Methods:In the current study, chronic intestinal inflammation was induced in recombinase‐activating gene‐1‐deficient (RAG−/−) mice by adoptive transfer of CD4+ T‐lymphocytes obtained from interleukin‐10 deficient (IL‐10−/−) mice. Results:Four weeks later, widespread colonic inflammation was observed in the reconstituted recipients, in contrast to 2 control sets of mice injected with a different subset of lymphocytes or with vehicle alone. We observed that the resulting pathology induced in the reconstituted RAG−/− mice was divided distinctly into 2 subsets: 1 with blood flow near normal with very high inflammation scores, and the other with severely attenuated blood flow but with much lower signs of inflammation. Colonic and ileal blood flow rates in the latter subset of CD4+ mice averaged only ≈30% compared to the mice with higher inflammation scores. The lower blood flow rates were associated with greatly reduced red blood cell concentrations in the tissue, suggesting a possible loss of vascular density. Conclusions:In this model of chronic intestinal inflammation, mild inflammation was associated with significant decreases in blood flow. Inflamm Bowel Dis 2009

Adoptive transfer of naïve CD4+ T cells into lymphopenic mice induces chronic small and large bowel inflammation similar to Crohn's disease. Although much is now known regarding the immunopathology in this model of inflammatory bowel disease, virtually nothing is known about the microvascular hemodynamic changes during the induction and perpetuation of chronic gut inflammation. In this study, CD4+CD45RBhigh T cells obtained from healthy C57BL/6 donor mice were transferred into lymphopenic recombinase-activating gene-1-deficient (RAG knockout) mice, which induced small and large bowel inflammation. At various time points following reconstitution (3 days-9 wk), intravital microscopy was used to examine the microvessels in the submucosa of the ileum and proximal colon following infusion of fluorescently labeled platelets and injection of rhodamine 6G (to label leukocytes). Hemodynamic measurements and the extent of blood cell adhesion to the venular wall were compared with measurements in unreconstituted RAG knockout controls. In <1 wk following reconstitution, velocity and wall shear rate of the arterioles decreased by >50% compared with controls, with this decrease also observed at 4-5 and 7-9 wk postreconstitution. At 7-9 wk, arteriolar diameters were found to be approximately 15% larger than in controls, but, despite this dilation, flow rates in the individual vessels were decreased by approximately 30%. Venular platelet and leukocyte adherence were not significantly elevated above controls; however, an association was found between platelet adherence and venular shear rate. In summary, significant decreases in arteriolar velocity and shear rates are observed in this model of chronic gut inflammation.

Background: Ingpstion by mice of dextran sodium Sulfate (DSS) induces colonic vasoconstriction and inflammation, with some of the effects potentially mediated by the vasoconstrictor endothelin-1 (ET-I).Methods: fit this study, mice given 5% 40 k-D DSS for -5-6 days had elevated colonic immnostaining for ET-I and platelet endothelial cell adhesion molecule-1 (PECAM-1). Increased ET-1 can induce microvascular constriction; however, the increase in PECAM-1 is consistent with angiogenesis that could decrease How resistance.Results: Our measurements of intestinal blood flow. via infused microspheres, SLI-ests that these 2 factors may offset each other. with only a nonsignificant tendency for a DSS-induced decrease in flow. Daily administration of the endothelin converting enzyme inhibitor SM-19712 (15 mg/kg) attenuated DSS-induced increases in colonic immunostaining of ET-I and PECAM-1.Conclusions: SM-19712 attenuated histologic signs of tissue injury and inflammation induced by DSS. and decreased the extent of loose stools and fecal blood. However, the inhibitor did not significantly decrease DSS-induced colon shortening Or tissue levels of myeloperoxidase (all indicator Of neutrophil infiltration). (Inflamm Bowel Dis 2009:15:1007 1013)

Ingestion by mice of dextran sodium sulfate (DSS) induces colonic vasoconstriction and inflammation, with some of the effects potentially mediated by the vasoconstrictor endothelin‐1 (ET‐1). In this study, mice given 5% 40 kD DSS for 5‐6 days had elevated colonic immunostaining for ET‐1 and platelet endothelial cell adhesion molecule‐1 (PECAM‐1). Increased ET‐1 can induce microvascular constriction that could limit blood flow; however, the increase in PECAM‐1 is consistent with previous reports of DSS‐induced angiogenesis that could decrease flow resistance. Intestinal blood flow measured with microspheres revealed a tendency for a DSS‐induced decrease, indicating that the vasoconstriction may be a more dominant factor than angiogenesis in the overall microvascular resistance. DSS mice given daily 15 mg/kg doses of the endothelin converting enzyme inhibitor SM‐19712 had decreased tissue ET‐1, and a tendency for decreased PECAM‐1. SM‐19712 attenuated histologic signs of tissue injury and inflammation, and decreased the extent of loose stools and fecal blood. However, the inhibitor did not significantly decrease DSS‐induced weight loss, colon shortening, or tissue levels of myeloperoxidase (an indicator of neutrophil infiltration). Supported by Project 2 of NIH P01DK043785.

Thromboxane has been implicated in the constriction of intestinal submucosal arterioles in mice rendered colitic with dextran sodium sulfate (DSS). In the present study, C57BL/6 mice given 40 kD DSS were administered one of three doses of the thromboxane receptor antagonist vapiprost: 0.3, 0.5, and 1.0 mg/kg/day, or no vapiprost as a control. DSS and vapiprost were given in drinking water (5% w/v for DSS) for a period of 6 days, concluding with an assessment of a disease activity index comprising stool consistency, rectal bleeding, and weight loss. The presence of loose stools decreased by 30–40% with all doses of vapiprost. Rectal bleeding decreased by ~25% with the lower dose of vapiprost, and by ~40% with the two higher doses. However, vapiprost was unable to prevent DSS‐induced weight loss. The overall disease activity index (0–4 scale) decreased from 2.9 ± 0.1 (no vapiprost) to 2.3 ± 0.3 (p<0.05) with the higher vapiprost doses of 0.5 and 1.0 mg/kg/day. In summary, vapiprost attenuates the disease activity index of DSS‐induced colitis in mice, primarily by decreasing the presence of loose stools and rectal bleeding. Supported by CCFA and NIH (P01DK043785).

BACKGROUND:Dextran sodium sulfate (DSS) induces submucosal arteriolar constriction that reduces blood flow to the intestine, and the relevance of this decrease in flow needs further investigation. In the present study we examined the effects of a vasoconstrictor (pseudoephedrine) and a vasodilator (papaverine) on the outcome of DSS-induced colitis.METHODS:Mice were given DSS in drinking water for 6 days, with enemas on days 0, 1, 3, and 5 containing pseudoephedrine, papaverine, or no drug. At the conclusion of the 6-day protocol a disease activity index comprising weight loss, stool consistency, and rectal bleeding was evaluated, along with intravital microscopy observations of submucosal venular leukocyte and platelet adherence in the proximal colon and terminal ileum.RESULTS:Pseudoephedrine and papaverine had several contrasting effects on the outcome of DSS ingestion: pseudoephedrine induced the highest levels of weight loss, loose stools, venular platelet adherence, and overall disease activity index, while papaverine induced the highest levels of venular leukocyte adherence, but the lowest levels of rectal bleeding, loose stools, and overall disease activity index.CONCLUSIONS:The results suggest that vasoconstriction worsens the pathological consequences of DSS in the mouse model of colitis.

Hypercholesterolemia is associated with an attenuation of endothelium-dependent dilation in arterioles and an increase in leukocyte and platelet adhesion in venules. The proximity of closely paired arterioles and venules is thought to facilitate heat and mass transport between the two and could be involved in transport of inflammatory and/or vasoactive mediators from venule to arteriole. In the current study, we tested the hypothesis that the impaired arteriolar dilation associated with hypercholesterolemia might be dependent on P-selectin-dependent blood cell adhesion in the closely paired venules. Leukocyte and platelet recruitment in venules and the endothelium-dependent response to bradykinin in second-order arterioles were observed in the mouse intestinal submucosa using intravital microscopy. Four weeks of a high-cholesterol diet decreased bradykinin-induced arteriolar dilation more dramatically in closely paired arterioles than in distantly paired arterioles. The dysfunctional arteriolar dilation of closely paired arterioles in hypercholesterolemic mice was significantly improved when the experiments were repeated in P-selectin-deficient mice (given the high-cholesterol diet) or in hypercholesterolemic mice injected with a P-selectin monoclonal antibody. A similar improvement in dilation of closely paired arterioles was attained in hypercholesterolemic mice given the superoxide dismutase mimetic Tempol. These findings indicate that hypercholesterolemia-induced increases in venular leukocyte and platelet adhesion might contribute to the impaired endothelium-dependent dilation of closely paired arterioles via a mechanism that is distance limited and dependent on P-selectin and superoxide.