The reactivity of cerebral arteries to different stimuli varies according to vessel size. Whether nitric oxide mediates hypoxic vasodilation is controversial. The authors considered this question by measuring the diameter of pial arteries and arterioles with or without exposure to the nitric oxide synthase inhibitor, N omega-nitro-L-arginine methyl ester (L-NAME).


The cranial window technique, combined with microscopic video recording, was used in an experiment involving 20 cats anesthetized with fentanyl and midazolam. The diameters of pial arteries and arterioles were measured under the following conditions: (1) normoxia (PaO2 > 100 mmHg); (2) hypoxia (PaO2 < 45 mmHg); (3) normoxia with L-NAME infusion; and (4) hypoxia with L-NAME infusion. Changes in vessel diameter were analyzed with respect to artery size.


Under hypoxic conditions, arteries and arterioles smaller than 200 microns were dilated significantly (P < 0.05). In arterioles smaller than 200 microns, L-NAME attenuated this hypoxic vasodilation (P < 0.05). In contrast, under normoxic conditions, L-NAME caused significant vasoconstriction in arteries larger than 100 microns but not in arteries smaller than 100 microns.


Arteries and arterioles smaller than 200 microns are dilated by hypoxia, and nitric oxide contributes to this process. Nitric oxide synthesis may also be related to the regulation of resting vascular tone in arteries larger than 100 microns.

Key words: Animals: cats. Arteries: cerebral. Gases: Nitric oxide. Microvasculation: arterioles. Nitric oxide synthase blockers: N omega-nitro-L-arginine methyl ester. Signs and symptoms, respiratory: hypoxia.

Cerebral blood vessels change their diameter in response to various kinds of physiologic and pharmacologic stimuli. This responsiveness is reported to depend on diameter; that is, changes in diameter of larger arteries differ from those of smaller arteries. [1,2] 

During hypoxia, cerebral blood vessels dilate and cerebral blood flow (CBF) increases. Several studies using nitric oxide (NO) synthase inhibitors examined the possible contribution of intrinsic NO to hypoxic vasodilation. [3-10]Although changes in CBF have been assessed in several studies, the role of NO in hypoxic vasodilation of cerebral vessels of various sizes remains controversial. Furthermore, it is unclear whether the reactivity of arteries to hypoxia differs among arteries with different diameters. To examine the contribution of NO to hypoxic vasodilation, we analyzed the responsiveness of cerebral pial arteries and arterioles to hypoxia before and after administration of an NO synthase inhibitor, N omega-nitro-L-arginine methyl ester (L-NAME), using the cranial window technique combined with microscopic video recording. [11,12] 

Surgical Procedures

The study was approved by the Animal Experiment Committee of Nara Medical University. Twenty cats weighing 1.4 to 4.3 kg were used. Anesthesia was induced by having the cats inhale ethyl ether and maintained with sevoflurane in 33% oxygen and 67% nitrogen until intravenous anesthetics were started. After tracheal intubation and intramuscular injection of 0.4 mg/kg pancuronium bromide, the cats' lungs were mechanically ventilated using a pump ventilator (Harvard Respirator; Summit Medical Co., Tokyo, Japan) to maintain end-tidal carbon dioxide at 30 to 35 mmHg. Endtidal carbon dioxide was measured continuously with a carbon dioxide analyzer (HP 14360A Carbon Dioxide Transducer; Hewlett Packard, Andover, MA). After the femoral vein was cannulated, 50 micro gram [centered dot] kg sup -1 [centered dot] h sup -1 fentanyl, 5 mg [centered dot] kg sup -1 [centered dot] h sup -1 midazolam, and 0.2 mg [centered dot] kg sup -1 [centered dot] h sup -1 pancuronium in lactated Ringer solution were infused at a rate of 3 ml [centered dot] kg sup -1 [centered dot] h sup -1. The femoral artery was cannulated for arterial pressure monitoring and blood gas analysis. The head of each animal was secured in a stereotaxic frame. A mid-sagittal incision was made from the occiput to the forehead. Two circular holes approximately 10 mm in diameter were made in the bilateral parietal bone with a power drill, and then dura and arachnoid were removed. The holes were filled with artificial cerebrospinal fluid (Shimizu Pharmacy Co., Shimizu, Japan) and covered with glass, the edges of which were sealed with an adhesive, alpha-cyanoacrylate monomer (Alon-alpha; Touagousei Co., Tokyo, Japan), to create closed cranial windows. The holes were located over the area supplied by the middle cerebral artery.

Measurements of Internal Diameter of Pial Arteries and Arterioles

Erythrocytes from 2 ml of blood were separated and rinsed with saline by centrifugation at 3,000 rpm for 5 min. three times. To label erythrocytes with fluorescein isothiocyanate, the cells were incubated for 1 h in 10 ml phosphate-buffered saline (pH 7.8) containing fluorescein isothiocyanate (Sigma Chemical Co., St. Louis, MO) at 1 mg/ml.

After intravenous injection of fluorescein isothiocyanate-labeled erythrocytes, pial arteries and arterioles were visualized through one of each pair of cranial windows using an intravital epifluorescence microscope (Nikon, Osaka, Japan), and images of the vessels were recorded with a silicon-intensified target-tube video camera system (SIT video camera system; Hamamatsu Photonics, Hamamatsu, Japan) linked to a video recorder. [12]Later the tape was analyzed using an image-analysis system (ARGUS-10; Hamamatsu Photonics). In observing the flow pattern of fluorescein isothiocyanate-erythrocyte vessels with smaller outflow, branches along the flow stream were identified as arteries or arterioles, and vessels that had smaller inflow branches along the flow stream were identified as veins. In this manner, pial vessels could be identified as arteries and arterioles according to the flow pattern of fluorescein isothiocyanate-erythrocytes. Thereafter, 1-2 mg fluorescein isothiocyanate-labeled dextran (fluorescein isothiocyanate-dextran; Sigma Chemical Co.) was administered intravenously, and the internal diameter of pial arteries and arterioles was measured. Diameters were measured in multiple vessels, moving the microscope at a magnification of 200 times with the image-analysis system (ARGUS-10). The coefficient of variance of measurements under this condition was less than 2%. To measure vessel diameter in the same arteries and arterioles under different conditions and times, distribution of branches, divisions, curves, and crossings of arteries and veins were carefully checked and recorded manually on paper to create a chart. Using this chart, measurements of vessel diameters under different conditions were obtained at the same sites.

Experimental Protocol

The other cranial window was cannulated with a polyvinyl chloride tube for continuous monitoring of intracranial pressure (ICP). During the experiment, heart rate, mean arterial pressure, and ICP were continuously recorded. Arterial blood pH, PaCO2, PaO2, HCO sup -3, and hemoglobin were measured with a blood gas analyzer (ABL2; Radiometer, Copenhagen, Denmark). Acidemia, when present, was corrected appropriately with sodium bicarbonate according to base deficit values. Pharyngeal temperature was maintained at 36 to 38.5 degrees Celsius with a heating pad.

After normoxia (PaO2> 100 mmHg) was confirmed, diameters of pial arteries and arterioles were measured for use as control baseline data. To induce hypoxia, the inspiratory oxygen concentration was reduced to 10% to 12% with increases of inspiratory nitrogen, ensuring a PaO2value less than 45 mmHg. Pial artery and arteriole diameter were measured 20 min after induction of hypoxic ventilation. Normoxia (Pa sub O2> 100 mmHg) was then achieved again by increasing inspiratory oxygen concentration. Thirty minutes after normoxia was reestablished, 30 mg/kg L-NAME (Sigma Chemical Co.) dissolved in saline was continuously infused intravenously for 30 min. The L-NAME dose, 30 mg/kg was chosen according to a previous report. [11]The continuous infusion duration, 30 min, was chosen because preliminary experiments showed that a 30-min infusion caused relatively smaller systemic hemodynamic changes than did 10- or 20-min infusions or bolus injections. The diameters of pial arteries and arterioles after L-NAME infusion were measured again under both normoxic and hypoxic conditions were induced as previously described. Hemodynamic parameters, arterial blood pH, PaO2, PaCO2, HCO sup -3, and ICP were recorded, and each pial arterial diameter was measured.

In an additional time-control experiment to examine the effect of two exposures to hypoxia, six additional cats were subjected to two exposures to hypoxia using the protocol already described, but without infusion of L-NAME during the second induction of normoxia and hypoxia. Briefly, after hypoxia was induced following the initial establishment of normoxia, a second normoxic period was again achieved and maintained for 30 min, and saline without L-NAME was continuously infused intravenously for 30 min. This was followed by the second induction of hypoxia. Diameters of pial arteries and arterioles were measured again as previously described.

Data Analysis and Statistics

The results for examined arteries and arterioles were divided into the following four groups according to diameter (D), which was measured under control normoxic conditions. Group 1 consisted of arteries and arterioles with diameters less than 50 micro meter; group 2, those with diameters greater than 50 micro meter but less than 100 micro meter; group 3, those with diameters greater than 100 micro meter but less than 200 micro meter; and group 4, those with diameters of 200 micro meter or larger. The percentage change in diameter under hypoxic conditions was expressed as the following equation: Equation 1where Dhypoxiaand Dnormoxiadenote arterial diameters under hypoxic and normoxic conditions, respectively.

For statistical analysis of the changes in diameter, one vessel from each animal of each group was chosen to represent several vessels. The representative vessel was that with a median diameter under control normoxic conditions in each group among the vessels that dilated under control hypoxic conditions. In this manner, the vessel number in each group for statistical analysis was the same or less than the number of animals, because, in some windows, the vessels in groups 3 and 4 could not be observed.

Values for vessel diameter and percentage change in diameter were statistically compared using Friedman's test and the Wilcoxon matched-pairs test. Statistical comparisons of hemodynamic parameters, arterial blood pH, PaO2, PaCO2, HCO sup -3, hemoglobin, and body temperature were done using two-way analysis of variance with repeated measurements followed by Scheffe's F test. All data were expressed as mean +/- SE, and differences were considered significant at P < 0.05.

Hemodynamic parameters, arterial blood pH, PaO2, Pa sub CO2, bicarbonate concentration, and ICP are summarized in Table 1. Under hypoxic conditions, heart rate and ICP increased, and the mean arterial blood pressure decreased both with and without L-NAME, although significant changes in mean arterial blood pressure were only seen with L-NAME. PaO2did not significantly differ after continuous infusion of L-NAME under either normoxia or hypoxia.

(Table 2) summarizes the results of a time-control study using another six cats, in which animals were exposed to hypoxia twice under the same protocol, but without L-NAME infusion. Hemodynamic parameters, data of arterial blood gas analysis, and ICP were similar to those in the experiments using 20 cats (data not shown). In the second induction of hypoxia without L-NAME infusion, vessels smaller than 200 micro meter (group 1, 2, and 3) were significantly dilated to a degree similar to that in the first induction of hypoxia, although those in group 4 were not significantly dilated. There was no significant difference in percentage changes in vessel diameter between the first and second induction of hypoxia.

Because the cranial window diameter was as small as 10 mm, we could not observe the vessels in group 4 in all animals using this technique. We could not collect sufficient data on vessels in group 4 to evaluate the effect of L-NAME. In the time-control study, vessels in group 4 were not significantly dilated under hypoxic conditions. For these two reasons, we examined vessels in groups 1, 2, and 3 in the following analysis.

(Table 3) summarizes data regarding vessel diameters in the three groups. Changes within 57 arterioles and arteries with diameters from 11.2 to 189.5 micro meter were examined. There were 20 vessels in group 1, 20 in group 2, and 17 in group 3. Under the hypoxic condition without L-NAME infusion, arterioles and arteries smaller than 200 micro meter (groups 1, 2, and 3) were significantly dilated (from 31.6 +/- 2 to 42 +/- 3.1 micro meter in group 1, from 67.7 +/- 1.9 to 82.4 +/- 6.1 micro meter in group 2, and from 133.1 +/- 5.8 to 149.2 +/- 7.2 micro meter in group 3).

In response to continuous L-NAME infusion 30 min after reestablishment of normoxia, arteries larger than 100 micro meter and smaller than 200 micro meter (group 3) were significantly constricted (from 133.1 +/- 5.8 to 117.4 +/- 6.2 micro meter). Vessel diameter in groups 1 and 2 did not significantly change (from 31.6 +/- 2 to 29.8 +/- 2.2 micro meter in group 1 and from 67.7 +/- 1.9 to 66.5 +/- 2.9 micro meter in group 2). Under hypoxic conditions after continuous L-NAME infusion, vessel diameters in groups 1, 2, and 3 increased, but dilation was significant only in group 1 (increase from 29.8 +/- 2.2 to 32.2 +/- 2.1 micro meter).

(Figure 1) shows the percentage changes in vessel diameter under hypoxic conditions with or without L-NAME infusion. Without L-NAME infusion, the percentage change in diameter in group 1 was the greatest, and that in group 3 was the smallest of the three groups. In all three groups, percentage changes in diameter with L-NAME were significantly lower than those without L-NAME. Attenuation of hypoxic vasodilation by L-NAME was observed in vessels with diameters smaller than 200 micro meter (groups 1, 2, and 3).

Several studies have used the cranial window technique to examine the responsiveness of cerebral blood vessels to various stimuli. In the present case, its application allows us to evaluate the in vivo effects of an NO synthase inhibitor, L-NAME, and hypoxia on cerebral vessels. The results are consistent with earlier reports showing that reactivity of cerebral vessels varies according to their size. Throughout our study, L-NAME caused vasoconstriction in the large vessel group with diameters ranging from 100 to 199 micro meter (group 3) and attenuated hypoxic vasodilation in all three groups. The results obtained are consistent with earlier reports showing that the reactivity of cerebral vessels varies with size. [1,2,5,13,14] 

Several investigations used NO synthase inhibitors to determine whether NO contributed to hypoxia-induced augmentation of CBF or hypoxia-induced dilation of cerebral blood vessels. [4-10]Most studies simply measured the CBF. [4,6-10]Iwamoto and colleagues [6]and Isozumi and associates [10]reported that an NO synthase inhibitor attenuated augmentation of CBF during hypoxia and suggested that NO helped augment CBF. Although regional CBF reflects the sum of the responses of the small arteries and arterioles, the so-called resistance vessels, it does not represent the flow of any single vessel. Thus it does not allow distinction between vessels of different sizes and lacks the ability to detect very minute changes in a small, limited population of arteries and arterioles. However, with our method using direct observation of vessels, changes in single vessels can be readily assessed and differences in responsiveness between large and small arteries and arterioles to hypoxia or L-NAME were clearly analyzed.

In a time-based control study using six cats, we confirmed that the second induction of normoxia and hypoxia caused similar changes in vessel diameter as the first one, as shown in Table 2. The changes in vessel diameter were not significantly different between the first and the second induction of hypoxia. From this observation, vessel diameter changes in the experiments with L-NAME infusion could be considered to be related primarily to L-NAME infusion rather than to the two inductions of normoxia and hypoxia. Because the vessels in group 4 were not significantly dilated under control hypoxic conditions in a time-based control study, we did not analyze the changes in group 4.

Pelligrino and coworkers [5]also directly observed changes in 30- to 60-micro meter diameter pial arterioles under graded-hypoxic conditions with L-NAME. They found that NO contributed to hypoxic vasodilation under a severely hypoxic condition of PaO2< 33 mmHg, but not under higher oxygen tensions in the rat. Although we used a different animal in our study, the results of the present study, which showed attenuation of hypoxic vasodilation by L-NAME in arteries smaller than 50 micro meter, seem to be similar to the findings of that report.

Nitric oxide contributes to physiologic and pathologic control of vascular tone and cerebral circulation. [2,13-17]Under normoxic conditions in our study, L-NAME infusion induced significant constriction in arteries with diameters larger than 100 micro meter (group 3), but not in smaller arteries (groups 1 and 2). From this result, NO synthesis and release seem to help regulate resting tone in vessels larger than 100 micro meter. However, we cannot preclude that autoregulatory vasodilatation of distal arteries smaller than 100 micro meter in response to vasoconstriction of upstream larger arteries may have masked any local vasoconstrictive response under L-NAME infusion. Faraci [2]reported that relatively large arteries (275 +/- 50 micro meter) were constricted more by NG-monomethyl-L-arginine infusion than were small ones (62 +/- 6 micro meter), and he concluded that a greater resting NO "tone" is present in large rather than in small vessels. He noted further that large vessels showed basal synthesis and release of NO to regulate resting tone, whereas small arteries showed little basal synthesis or release of NO despite an equivalent capacity for NO synthesis and release.

In the present study, significant attenuation of hypoxic vasodilation by L-NAME was observed in all arteries and arterioles smaller than 200 micro meter (Figure 1, groups 1, 2, and 3). Although percentage changes tended to be larger in vessels with a smaller diameter with and without L-NAME infusion, the contribution of NO to hypoxic vasodilation did not differ among the three group (Figure 1). Thus these findings suggest that vessels smaller than 200 micro meter exhibit some NO-dependent dilation, even though NO may not be the sole factor responsible for hypoxic vasodilation.

As possible other factors, prostaglandin E1[10]and adenosine [10,18,19]have been reported to participate in hypoxic vasodilation, although some investigators denied any contribution from prostaglandins to arteriolar responsiveness in hypoxia. [15,20,21] 

Several hypotheses have been proposed to explain the direct effect of hypoxia on NO production and metabolism, some mechanisms in endothelial cells have been reported. Hypoxia induces an increase in intracellular free Calcium2+ and activates Calcium2+ -dependent NO synthase. [22,23]It also suppresses generation of superoxide anions that inactivate NO. [24]Under anoxic conditions, NO production is suppressed because the synthesis of NO requires the incorporation of oxygen molecules. [25]Pohl and Busse, [26]however, reported that hypoxia at PaO2of 24 +/- 8 mmHg stimulated the release of NO from vessels and from cultured endothelial cells. At hypoxic levels in our study (PaO2= 36 to 37 mmHg), we also supposed that NO was synthesized and released, as in Pohl and Busse's study, although the role of endothelial cells or of Calcium2+ and superoxide anions must still be clarified.

In conclusion, under hypoxic conditions, vasodilation was observed with presumed involvement of intrinsic NO synthesis and release in pial arteries and arterioles with diameters smaller than 200 micro meter. Basal release of NO, probably participating in regulation of resting vascular tone, appears to reside in arteries between 100 micro meter and 200 micro meter in diameter.

Gotoh F, Muramatsu F, Fukuuchi Y, Amano T: Dual control of cerebral circulation: Separate sites of action in vascular tree in autoregulation and chemical control, Cerebral Circulation and Metabolism. Edited by TW Langfitt, LC MacHenry Jr, M Reivich, H Wollmam. New York, Springer-Verlag, 1975, pp 43-5.
Faraci FM: Role of endothelium-derived relaxing factor in cerebral circulation: Large arteries vs. microcirculation. Am J Physiol, 1991; 261:H1038-42.
Iadecola C, Pelligrino DA, Moskowitz MA, Lassen NA: Nitric oxide synthase inhibition and cerebrovascular regulation. J Cereb Blood Flow Metab 1994; 14:175-92.
Pelligrino DA, Koenig HM, Albrecht RF: Nitric oxide synthesis and regional cerebral blood flow responses to hypercapnia and hypoxia in the rat. J Cereb Blood Flow Metab 1993; 13:80-7.
Pelligrino DA, Koenig HM, Albrecht RF: The role of nitric oxide (NO) in cerebral arteriolar relaxation during hypoxia in the rat [Abstract]. Anesthesiology 1993; 79:A757.
Iwamoto J, Yoshinaga M, Yang SP, Krasney E, Krasney J: Methylene blue inhibits hypoxic cerebral vasodilation in awake sheep. Am J Physiol 1992; 73:H2226-32.
Koz'niewska E, Oseka M, Stys'T: Effects of endothelium-derived nitric oxide on cerebral circulation during normoxia and hypoxia in the rat. J Cereb Blood Flow Metab 1992; 12:311-17.
Buchanan JE, Phillis JW: The role of nitric oxide in the regulation of cerebral blood flow. Brain Res 1993; 610:248-55.
McPherson RW, Koehler RC, Traystman RJ: Hypoxia, alpha2-adrenergic, and nitric oxide-dependent interactions on canine cerebral blood flow. Am J Physiol 1994; 266:H476-82.
Isozumi K, Fukuuchi Y, Takeda H, Itoh Y: Mechanisms of CBF augmentation during hypoxia in cats: Probable participation of prostacyclin, nitric oxide and adenosine. Keio J Med 1994; 43:31-6.
Sandor P, Komjati K, Reivichi M, Nyary I: Major role of nitric oxide in mediation of regional CO2 responsiveness of the cerebral and spinal cord vessels of the cat. J Cereb Blood Flow Metab 1994; 14:49-58.
Kumano H, Furuya H, Yomosa H, Nagahata T, Okuda T, Sakaki Y: Response of pial vessel diameter and regional cerebral blood flow to CO2 during midazolam administration in cats. Acta Anaesthesiol Scand 1994; 38:729-33.
Rosenblum WI, Nishimura H, Nelson GH: Endothelium-dependent L-Arg- and L-NMMA-sensitive mechanisms regulate tone of brain microvessels. Am J Physiol 1990; 259:H1396-H1401.
Auer LM, Pucher R, Leber K, Ishiyama N: Autoregulatory response of pial vessels in the cat. Neurol Res 1987; 9:245-8.
Wei HM, Weiss HR, Sinha AK, Chi OZ: Effects of nitric oxide synthase inhibition on regional cerebral blood flow and vascular resistance in conscious and isoflurane-anesthetized rats. Anesth Analg 1993; 77:880-5.
Rees DD, Palmer RMJ, Schulz R, Hodson HF, Moncada S: Characterization of three inhibitors of endothelial nitric oxide synthase in vitro and in vivo. Br J Pharmacol 1993; 101:746-52.
Tanaka K, Gotoh F, Gomi S, Takashima S, Mihara B, Shirai T, Nogawa S, Nagata E: Inhibition of nitric oxide synthesis induces a significant reduction in local cerebral blood flow in the rat. Neurosci Lett 1991; 127:129-32.
Hoffman WE, Albrecht RF, Miletich DJ: The role of adenosine in CBF increases during hypoxia in young vs aged rats. Stroke 1984; 15:124-9.
Morii S, Ngai AL C, Ko KR, Winn HR: Role of adenosine in regulation of cerebral blood flow: Effects of theophylline during normoxia and hypoxia. Am J Physiol 1987; 253:H165-75.
Jackson WF: Prostaglandins do not mediate arteriolar oxygen reactivity. Am J Physiol 1986; 250:H1102-8.
Skabe T, Siesjo BOK: The effect of indomethacin on the blood flow-metabolism couple in the brain under normal, hypercapnic and hypoxic conditions. Acta Physiol Scand 1979; 107:283-4.
Luckhoff A, Pohl U, Busse R: Increased free calcium in endothelial cells in response to hypoxia and restitution of normoxia. Pflugers Arch 1986; 406(Suppl 1):R46.
Busse R, Mulsch A: Calcium-dependent nitric oxide synthesis in endothelial cytosol is mediated by calmodulin. FEBS Lett 1990; 265:133-6.
Rubanyi GM, Vanhoutte PM: Superoxide anions and hyperoxia inactivate endothelium-derived relaxing factor. Am J Physiol 1986; 250:H822-7.
Palmer RMJ, Ashton DS, Moncada S: Vascular endothelial cells synthesize nitric oxide from L-arginine. Nature 1988; 333:664-6.
Pohl U, Busse R: Hypoxia stimulates release of endothelium-derived relaxant factor. Am J Physiol 1989; 256:H1595-H1600.