INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

PRIOR STUDIES (9, 16) related to nitric oxide (NO) formation at rest and during intestinal functional hyperemia have indicated that small arteries and larger arterioles, rather than the terminal microvasculature, were the major contributors to intestinal vascular regulation. The intestinal vasculature is not unique in this regard. During both skeletal muscle contractions (19, 23) and autoregulation of brain blood flow (14), dilation of the feed arteries and the largest arterioles is essential to normal regulation. At the present time, three major mechanisms have been proposed to coordinate the responses of large resistance vessels with those of the microvessels. First, Segal and Duling (23) found that large arterioles transmitted both constrictor and dilatory signals to the upstream feed arteries through cell-to-cell conduction between vascular smooth muscle cells. Second, dilation of small arteries and large arterioles may occur in response to increased NO production as flow shear rates are elevated by the dilation of downstream smaller arterioles (17, 18). Finally, Hester and co-workers (15, 20, 22), Falcone and Bohlen (13), Boegehold (2), and Steenbergen et al. (25) have shown that chemical communication can occur between venules and arterioles. This communication can include both diffusion of vasoactive chemicals from venous blood to resistance vessels and formation of NO in the venous wall. The problem to be solved in this study was whether the suspected increase in NO formation during intestinal absorptive hyperemia primarily originated from arterial or venous vessels. The key issues are to what extent venous vessels form sufficient NO to influence arterial vessels and if diffusion of materials from venous blood to the nearby resistance vessels activates both endothelial and non-endothelial-dependent vasodilation.

Resolving the role of NO in intestinal vascular regulation, particularly if venous communication of NO and endothelial-dependent vasodilators are involved, required that the arterial and venous perivascular NO concentration ([NO]) be measured. This is the best approach to determine if arterioles generated the NO that caused the observed vasodilation. Much of what is known about the role of endothelial NO in regulation of the in vivo microvasculature is based on the consequences of impairing production of NO. Increases in resting vascular resistance and deficits in vascular reactivity are used to infer the role of NO in regulation. The major problem with this investigative approach is that vascular smooth muscle may be excessively activated due to the absence of the relaxing effects of the basal [NO]. For example, Saito et al. (21) found that suppression of NO formation attenuated vasodilation during skeletal muscle contractions to a major extent only if the resting vasoconstriction was allowed to persist. In effect, the role of NO in vascular regulation may be dependent on how initial conditions are compromised when basal NO formation is suppressed. This is a significant concern for the intestinal vasculature. Bioassay studies indicated that vascular resistance at rest and during absorptive hyperemia was highly dependent on NO (9, 16). Furthermore, the increased sodium concentration in interstitial fluid, lymph, and venous blood during intestinal absorption of both carbohydrates and lipids has been implicated as a mechanism to stimulate endothelial NO formation (10, 24). To verify that the concentration of NO does in fact substantially increase during absorptive hyperemia and in response to sodium hyperosmolarity, the perivascular [NO] was measured using NO-sensitive recessed-tip microelectrodes. These measurements provided a unique perspective from which to judge the relative role of NO formation in the in vivo regulation of intestinal absorptive hyperemia.

In preliminary studies, I found that, during glucose absorption, the submucosal interstitial [NO] increased by 30-60% around both the largest arterioles and venules. The primary source of NO was not clear because NO can be released from arterioles, venules, and lymphatic vessels (6, 13, 24). A simpler tissue environment was needed to study venous to arterial communication. This was achieved by changing the site of measurement to the small feed arteries and veins immediately outside the intestinal wall where there was no appreciable parenchymal tissue, and areas were chosen where lymph vessels were not close to the artery or vein. The results demonstrated that venous to arterial communication primarily occurred through diffusion of endothelium-dependent and -independent dilator substances from the vein to artery and that NO formed in veins can influence the [NO] near the artery.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Animal and tissue preparation. A total of 46 male Sprague-Dawley rats (300-400 g; Harlan, Indianapolis, IN) were studied. Each animal was anesthetized with thiopental sodium (200 mg/kg; Abbott, Chicago, IL) injected subcutaneously at four locations over the thighs and lower back. One-fourth of the original dose was given if supplemental anesthesia was needed. The trachea was cannulated, and a femoral artery cannula was used to measure arterial pressure and administer fluids. The animal was placed on a water-heated support. The water temperature was adjusted as necessary to maintain the core temperature of 37-37.5°C, but the heater-skin interface temperature did not exceed 36°C. Core temperature was measured by passing a thermistor (Yellow Springs Instruments, Yellow Springs, OH) through the mouth to the stomach.

The jejunal region of the small intestine was used because mesenteric fatty tissue did not obscure small arteries and veins near the bowel wall. The bowel was slit longitudinally along the antimesenteric border with a microcautery and draped over a translucent pedestal (5). The muscular layer, which was facing upward, was superfused with 4-5 ml/min of a bicarbonate-buffered physiological saline equilibrated with 5% oxygen, 5% carbon dioxide, and 90% nitrogen (5). The temperature of the medium over the tissue was held at 37.5 ± 0.25°C by heating both the flowing superfusion fluid and the tissue support device. Isoproterenol (10⁸ M; Sigma Chemical, St. Louis, MO) and phenytoin (20 mg/l; Parke-Davis Pharmaceutical, Morris Plains, NJ) were added to the bathing medium to partially suppress bowel motility. The concentrations of drugs used have minor effects on resting vascular resistance (5). The mucosal surface was superfused at 0.5 ml/min with the same medium used for the muscular surface. Isotonic glucose solution was added to just the mucosal medium to establish final isotonic glucose concentrations of 100 or 300 mg/dl. The 300 mg/dl glucose solution would damage endothelial cells (7) if it reached the muscular surface of the bowel wall. However, the admixture of the mucosal and muscular layer fluids at their respective flow rates would have yielded a glucose concentration of not more than 60 mg/dl, which is benign (7).

Microvascular observation and measurement. Small arteries and microvessels were observed with an Olympus BMHJ microscope using long-working-distance Nikon ×10 and ×20 objectives. The image recorded by a charge-coupled device camera (Hammamatsu model XC-77) was digitized with a Matrox MPV video digitizer and processed using Image 1 image analysis software (Universal Imaging, West Chester, PA). The final magnification and XY distance calibration of video images was ×730 and ×460, based on images of a stage micrometer marked in 10- and 100-µm sections. Vessel and tissue dimensions were measured with a virtual caliper superimposed on the digital image.

The blood flow velocity was measured in first-order arterioles just after they entered the submucosal layer of the bowel wall. Each first-order arteriole was the extension of the feed artery into the wall of the small intestine, and both vessels have the same total flow. The flow velocity was measured using an Instrumentation for Physiology and Medicine model 102 velocity tracking instrument (San Diego, CA). The system was calibrated in units of millimeter per second linear velocity using a clear plastic strip coated with red blood cells. The strip was pulled through the optical field at various known velocities. Flow was calculated from the red blood cell velocity and the inner cross-sectional area of the first-order arteriole. Both the feed artery and companion vein have too large a diameter to directly measure flow velocity with transmitted light. However, because both of these vessels have the same total flow as the first-order arterioles, the flow velocities in the feed artery and companion vein could be calculated from the first-order arteriole flow divided by the cross-sectional area of either vessel.

Relative changes in oxygen consumption by the intestine were calculated from the blood flow and difference in percent oxygen saturation of hemoglobin (8) in the second-order arterioles and venules. The first-order vessels were not used for saturation measurements because their optical densities were too large for reliable measurements of hemoglobin saturation. However, prior studies (8) indicated essentially identical hemoglobin saturations in first- and second-order arterioles or first- and second-order venules. If the arteriolar percent oxygen saturation of hemoglobin was not at least 90% with natural ventilation, the animal was mechanically ventilated. The parameters of ventilation were 70 breaths/min, the natural ventilation frequency of conscious rats, and a tidal volume was adjusted to the minimum that would produce a 90-95% arteriolar saturation.

Microvascular pressures were measured with an Instrumentation for Physiology and Medicine Servo Nulling System (model 4A). Pipettes sharpened to a tip diameter of 4-8 µm at a 20° angle were used for the pressure measurements. The individual pipettes were calibrated against a mercury manometer and were accurate to within ±1 mmHg at pressures of 10-100 mmHg.

NO measurements. NO-sensitive recessed-tip microelectrodes were made using the guidelines developed by Buerk et al. (11). The microelectrodes had a 4- to 8-µm outer tip diameter after sharpening and a 50- to 70-µm open tip recess, and 50-100 µm of gold were plated distal to the recess. The microelectrodes were soaked in 5% Nafion (Aldrich, Milwaukee, WI) for 4-6 h before being dried at room temperature. A World Precision Instruments (Sarasota, FL) carbon fiber reference electrode was used in the fluid of the calibration cell and tissue support device. The microelectrode current was measured with either a Keithley (Cleveland, OH) model 610B electrometer or custom-designed electrometer based on an Analog Devices AD515 microchip. The microelectrodes were calibrated in a World Precision Instruments tonometer held at 37.5 ± 0.5°C. The microelectrodes were polarized at 0.8 V, with the microelectrode positive. Nitrogen-equilibrated saline was used as a zero reference for NO, and ~400 and ~800 parts per million NO in nitrogen were used for two other [NO] to establish an ~0-600-1,200 nM calibration range. The microelectrode currents were measured after gas bubbling had stopped and microelectrode current was stable. All microelectrodes used for data collection were found to have a linear relationship of electrical current to [NO] for the three calibration gases. In general, microelectrodes had a background current of 12-15 pA in nitrogen-equilibrated saline and generated an additional 1-2 pA/1,000 nM [NO]. This translates to an increase in output voltage of >300 mV per 1,000 nM increase in current for the Keithley electrometer at the gain setting used. Given the resolution of the electrometers used, the electronic noise and current drift during in vivo conditions, and sensitivity of the microelectrodes, the overall system could resolve 5-10 nM changes in [NO].

Reactivity of the microelectrodes with biological chemicals other than NO was a concern. The recessed-tip microelectrodes would respond with current increases equivalent to 50-100 nM NO to supraphysiological 1 mM norepinephrine or 1 mM ascorbic acid, two prevalent biological agents that react at about the same voltage as NO (11). However, the approximately 10-fold lower physiological concentrations of norepinephrine and ascorbic acid would not be detectable at typical tissue and perivascular [NO]. There may be intracellular compounds other than catecholamines and ascorbic acid that can react with the microelectrodes. To test for this possibility, tissue was excised, blotted to remove surface water, and immediately frozen to break most of the cell membranes. The tissue was then thawed to 37°C, and as soon as the tissue was warmed, the microelectrode tip was pushed into the tissue. In such permeabilized tissue, a current signal equivalent to an [NO] of 20-30 nM could be detected. Therefore, for biochemicals that would survive freezing and thawing, there did not appear to be a significant population which at biological concentrations could appreciably interact with the microelectrodes.

The microelectrodes were extremely sensitive to temperature. To avoid consequences of the thermal sensitivity, the bath temperature for a depth of at least 2-3 mm above the tissue was the same as the tissue. In addition, the microelectrodes were held at a 15-20° angle to the tissue so that the entire tapered portion of the microelectrode and ~5 mm of the shaft were submerged in heated fluid over the tissue.

To establish a 0 nM reference for NO during in vivo conditions, the microelectrode tip was moved 0.5-1 mm above the tissue surface. As the microelectrode was lowered to the tissue surface, the microelectrode current did not increase until the microelectrode tip was within 50 µm of the tissue surface. Within 25-50 µm of the tissue surface, microelectrode currents equivalent to 25-100 nM NO can be detected, with the higher currents consistently occurring over major arterioles and venules or small arteries and veins. Once the microelectrode tip penetrated the tissue, the parenchymal [NO] ranged from ~100 nM at minimum to approximately the same concentration as on the surface of nearby microvessels. To measure the vessel wall [NO], the microelectrode tip was lightly pressed against the outer vessel wall. As the intestine routinely moved, the pipette shaft was oriented at a 10-20° angle to the vessel longitudinal axis. This orientation minimized the risk of vessel wall penetration by the microelectrode tip during movement of the intestinal wall. The slight bending of the microelectrodes did not generate artifactual currents. Care was taken to avoid penetration of the vessel wall because if a platelet-leukocyte clot formed or leakage of red blood cells occurred, the [NO] decreased and did not recover fully. After the pipette tip was properly located, the vessel and microelectrode were allowed to stabilize for 2-5 min. The most frequent occurrence was a 10-30% increase in current during this time and stable currents thereafter. In ~10% of the cases, pressing the NO microelectrode tip against the vessel during the initial tissue penetration increased the current for 2-3 min and presumably reflected a transient increase in NO production.

Protocols. Two major types of protocols were used, plus a third protocol to test conducted vasodilation. In the first protocol, the mucosal layer was either superfused (0.5 ml/min) with standard bathing medium or the same medium with an isotonic glucose concentration of 100 or 300 mg/dl. The standard medium was a bicarbonate-buffered physiological solution that was equilibrated with 5% oxygen, 5% carbon dioxide, and 90% nitrogen to equal the gas tensions used in the bathing fluid over the muscle layer (5). The measurements made included inner vessel diameter, centerline red blood cell velocity, the [NO], and microvascular pressures, although the latter measurements were not made in each experiment. The mucosa was exposed to the glucose solutions for a minimum of 15 min before data collection began to ensure steady-state responses had developed. Occlusion of the small vein was necessary to test for effects of venous blood on the companion artery. A glass microtool with a 0.25-mm ball tip was pressed downward against the vein using a hydraulically controlled micromanipulator. This procedure generally was innocuous to the vein because both the preocclusion diameter and venous wall [NO] at an upstream site recovered after each occlusion. Care was taken to avoid mechanically interfering with the companion artery and NO microelectrode on the artery or vein wall during venous occlusion. The microvascular responses after vein occlusion were followed for ~10 min, but steady-state conditions were consistently reached within 5 min.

The second protocol was to locally superfuse a vasoactive agent of choice over the surface of a microvessel. A micropipette with a 100- to 150-µm tip diameter was used to deliver the solution containing the vasoactive agent. The vessel diameter and [NO] were measured before, during, and after the superfusion period. The fluid did not enter the parenchyma or mesenteric tissue over the vessels but simply flowed over the tissue surface and displaced the standard bathing fluid over about a 1-mm² area of tissue. The solution was pumped at 100 nl/min with a World Precision Instruments micropump. The solution used to carry drugs was Plasma-Lyte (Baxter Healthcare, Deerfield, IL). When the Plasma-Lyte alone was pumped over the vessels, it had no appreciable effect on vessel diameters or [NO] at the flow rate used.

The third protocol tested whether dilation of a large arteriole could be successfully communicated upstream to the small artery over a distance of ~2 mm. To test this possibility, acetylcholine was iontophoresed onto the first-order arteriole just within the bowel wall, and the small artery was observed 2 mm upstream, the typical site for observation of small arteries in all protocols. Acetylcholine was used because Delashaw and Duling (12) have found that the best-developed cell-to-cell communication occurred with this compound. The iontophoresis current used to release acetylcholine was increased until the first-order arteriole dilated ~20% to mimic natural absorptive hyperemia dilation.

Statistical analysis. All measurements included a control state and at least two perturbations plus an altered status, such as venous occlusion. Based on these conditions, one- and two-way analysis of variance was used to detect significant effects at P < 0.05, and a least significant difference test was used for individual comparisons. The SigmaStat statistical software system was used for all statistical tests (Jandel Scientific Software, San Rafael, CA). The data presented are means ± SE.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Support of increased metabolism by vascular responses and oxygen extraction. As shown by the data in Fig. 1, the aerobic metabolic rate increased 50-60% at a mucosal glucose concentration of 100 mg/dl and doubled at a glucose concentration of 300 mg/dl. About two-thirds of the increased metabolism were supported by elevated blood flow with the remainder of the oxygen provided by increased extraction. The decrease in vascular resistance was caused by dilation of both microvessels and the small arteries. To establish the relative changes in resistance of the arteries and microvessels, the pressure drop across each region and relative changes in blood flow were used to calculate the relative resistance in each region. Intravascular pressures in the femoral artery and in the feed arteries just as they entered the bowel wall defined the pressure drop across the arterial portion of the total intestinal vasculature. At rest, the arteries, up to vessels entering the bowel wall, dissipated 21.3 ± 2.5% of the total pressure drop across the entire intestinal vasculature. As shown in Fig. 2, during 100 and 300 mg/dl glucose absorption, the resistance of the arteries outside the intestinal wall decreased proportionately more than did that of the microvasculature. During 100 mg/dl glucose absorption, about one-third of the decrease in total intestinal resistance was due to the arterial vessels, and at 300 mg/dl, the arteries caused about one-fourth of the decrease in total resistance.

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Relative changes in intestinal blood flow, oxygen extraction, and metabolic rate during mucosal exposure to 100 and 300 mg/dl glucose are presented. Approximately 65% of the additional oxygen needed to support the increased metabolism during glucose absorption was provided by increased blood flow, and the remainder was obtained by increased extraction of oxygen. All of the values during glucose absorption were significantly greater (P  0.05) than during the resting state. Data are based on 12 sections of bowel studied in 8 rats. Data are means ± SE.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Relative changes in total intestinal vascular resistance, the resistance of arteries to the small intestine, and the microvascular resistance from the largest arterioles to smallest veins are shown for events during mucosal exposure to 100 and 300 mg/dl glucose. Arteries controlled 25-30% of the total resistance at rest. On a relative basis, the resistance of the arteries decreased proportionately more than did microvascular resistance during glucose absorption. All resistances during glucose absorption were significantly less (P < 0.05) than during the paired resting state. Data are based on the same 12 sections of bowel studied in 8 rats presented in Fig. 1.

Dilation of the small arteries during glucose absorption. The goal of this protocol was to determine if dilation of small arteries was influenced directly or indirectly by the composition of blood in the companion vein. The average resting inner diameter of the small arteries was 128 ± 80 µm, and the small veins had an average inner diameter of 283 ± 18 µm for the vessels depicted in Fig. 3. At rest, occlusion of the companion vein caused an insignificant constriction of the artery, as shown in Fig. 3. During glucose absorption when the arteries were dilated, venous occlusion negated ~65% of the arterial dilation during 100 mg/dl glucose and ~50% of the dilation during 300 mg/dl glucose. This suppression of the dilation of the small artery during venous occlusion required 2-5 min for full development.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Percent of control diameters of the small arteries just outside the intestinal wall are shown during mucosal exposure to 100 and 300 mg/dl glucose. The "natural flow" state refers to intact conditions, and "vein occluded" indicates the companion vein of the artery temporally was occluded just downstream from the area where the vessels were observed. One-half or more of the arterial vasodilation associated with glucose absorption was suppressed during occlusion of the vein. The decreased dilation of the artery was not due to myogenic effects of increased intravascular pressure during venous occlusion, as explained in RESULTS. + Significant increase in diameter relative to control conditions; * significant decrease in diameter relative to the diameter during natural flow conditions for the same glucose concentration. Data are based on 9 vessel pairs studied in 5 animals.

The suppressed dilation of the small arteries during occlusion of the companion vein while glucose was being absorbed was not due to a pressor myogenic response. Occlusion of the companion vein did not significantly change the pressure in small arteries. At rest, the pressure in small arteries after venous occlusion was 99.8 ± 1.4% of the full flow pressure and, during 300 mg/dl glucose, it was 101.6 ± 1.6% of the pressure during hyperemic blood flow and glucose exposure. The microvascular pressure changes were almost nonexistent because the highly interconnected venous system within the bowel wall allowed blood to freely drain through the nearest neighboring small veins on either side of the occluded vein.

As mentioned earlier, occlusion of the companion vein only suppressed 50-65% of the arterial dilation associated with absorptive hyperemia. A large fraction of the remaining dilation of the small artery during glucose absorption was likely influenced by cell-to-cell conduction from the downstream large arteriole, the first-order arteriole, within the bowel wall. During 300 mg/dl glucose absorption, the first-order arteriole typically dilated ~20%, as also reported in prior studies (3, 24). Results for conducted vasodilation from the large arteriole to the small artery were obtained from 12 vessels in 8 rats. The average dilation of the first-order arteriole was to 122.9 ± 1.8% of control, and the small artery dilated to 107.9 ± 1.4% of control. This magnitude of dilation by the small artery was about one-half that found during 300 mg/dl glucose exposure with natural venous outflow (Fig. 3). Although it may be a coincidence, during venous flow occlusion associated with absorption of 300 mg/dl glucose, the arterial diameter was 107.6 ± 2.6% of the control diameter and identical to the dilation caused by cell-to-cell communication. Part of the dilation by small arteries could be due to flow-mediated vasodilation. During 100 and 300 mg/dl glucose absorption, the calculated flow velocity in small arteries increased 8.7 ± 2.8 and 21.7 ± 4.2%, respectively.

NO as a vein-artery communication molecule. It is probable that NO generated in the vein wall diffused the 10- to 30-µm distance to the artery and potentially contributed to arterial dilation. The assumption is that venous blood during absorptive hyperemia would contain metabolites that induce the production of NO from the venous endothelium. A practical method to test these issues was to measure the [NO] on the external walls of veins and arteries. However, it was first necessary to confirm that the NO microelectrodes could predict changes in [NO] during in vivo conditions consistent with the current understanding of NO production as judged from bioassay experiments. To test the in vivo characteristics of the NO microelectrodes, topical 1 mM L-arginine was used to stimulate production of NO and topical 0.5 mM N^G-nitro-L-arginine (L-NNA) was used to partially suppress NO formation. The data are presented in Fig. 4 and are based on studies on seven large arterioles in five animals. The resting state in Fig. 4 is local superfusion over the vessel and overlying tissue with physiological medium from a micropipette. Superfusion flow of 100 nl/min did not influence the resting perivascular [NO], which was 334 ± 19 nM. Sustained superfusion with 1 mM L-arginine caused the [NO] to more than double to a concentration of 686 ± 53 nM within 2-3 min. After recovery, superfusion with 0.5 mM L-NNA over a 5- to 10-min period gradually decreased the [NO] by ~50% to 169 ± 22 nM. For about 10 min after exposure to L-NNA, topically applied 1 mM L-arginine had minimal effects on the [NO], which remained at 178 ± 48 nM. However, within 15-20 min of sustained topical L-arginine application, the [NO] gradually increased to 762 ± 147 nM. Within 5-10 min after a second L-arginine application was stopped, the [NO] returned to 296 ± 30 nM compared with 334 ± 19 nM at the beginning of the tests ~45-60 min earlier.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Nitric oxide (NO) concentration in nM was measured on the outer wall of 7 large arterioles within the intestinal submucosal layer of 5 rats during local superfusion with 1 mM L-arginine (L-ARG) or 0.5 mM NG-nitro-L-arginine (NNA). The initial resting concentration of NO was 338 ± 22 nM, and the final basal concentration was 296 ± 30 nM about 1 h later after the various tests, including suppression of NO formation with NNA. The initial L-ARG superfusion was associated with an increase in [NO], and NNA decreased the resting [NO] by about one-half in each artery. After NNA, the vessels would not respond to L-ARG (NNA + initial ARG), but within 15-20 min of sustained L-ARG, the vessels responded with an increase in [NO] (NNA + final ARG).

The [NO] on the arterial and venous walls during glucose absorption are shown in Fig. 5 and are based on 16 vessel pairs studied in 13 rats. The arterial [NO] was measured on the surface of the artery facing away from the vein, and the venous concentration was measured on the vein surface facing away from the artery. In each experiment, the [NO] was also measured in the tissue between the artery and vein. As shown in Fig. 5, the average resting arterial wall [NO] of 353 ± 29 nM was lower than that of 401 ± 48 nM on the wall of the companion vein. The averaged difference in concentration from vein to artery was 51.6 ± 43 nM, which was not significantly different from a presumed difference of 0 nM. The [NO] of tissue between the artery and vein was very close to the simple average of the artery and vein concentrations. Within 5 min after 100 mg/dl glucose was exposed to the mucosa, the arterial [NO] increased by ~50% to 540 ± 68 nM and did not increase further during the mucosal glucose concentration of 300 mg/dl. The increase in venous wall [NO] during absorptive hyperemia was ~40% or to 557 ± 60 nM for 100 mg/dl glucose and by ~57% or to 628 ± 70 nM for 300 mg/dl glucose. During 300 mg/dl glucose absorption, the difference in [NO] from vein to artery was 77.5 ± 30.5 nM and was significantly greater than a presumed concentration difference of 0 nM. The increase in [NO] for the tissue between the artery and vein during glucose absorption paralleled the increased concentration around the vein. The increase in arterial and venous wall [NO] could be sustained for at least 60 min during glucose exposure to the mucosa. Longer time frames of glucose exposure were not attempted, but there was no indication of a gradual decline of the increased [NO] on either arterial or venous vessels during sustained glucose absorption. When glucose exposure was stopped and superfusion of physiological saline over the mucosa was initiated, the [NO] returned to within ±50 nM of the previous resting concentrations within 10 min.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   NO concentration ([NO]) in nM was measured on the wall of 16 paired intestinal arteries and veins in 13 rats during rest and exposure of the mucosa to 100 and 300 mg/dl glucose. [NO] measurements were made on the side of each vessel facing away from its companion vessel. At rest, 13 of the 16 veins had [NO] that were higher by 50 nM than that of the arterial wall. During mucosal exposure to 100 mg/dl glucose, the arterial and venous [NO] increased significantly (P < 0.01) with respect to the rest concentration. Mucosal exposure to 300 mg/dl did not cause a further increase in the arterial [NO], but the venous [NO] increased significantly (P < 0.05) relative to the concentration recorded during 100 mg/dl glucose.

Mechanism of elevated arterial [NO] during glucose absorption. Suppression of the arterial dilation when the vein was occluded during absorptive hyperemia, as shown in Fig. 3, suggested that either a vasoactive species in venous blood or released from the vein wall contributed to dilation of the companion artery. To determine if something related to NO formation was involved, the [NO] was measured on the arterial and venous walls at rest and during 300 mg/dl glucose exposure with the vein open or occluded. As mentioned earlier, the arterial and venous wall [NO] was measured on the opposite surfaces of vessels. The data shown in Fig. 6 are based on nine vessel pairs in five rats. At rest, occlusion of the vein had no significant effect on either periarterial or perivenous [NO]. Absorption of 300 mg/dl glucose caused a 40-50% increase in both arterial and venous wall [NO], as was found for the animals studied for Fig. 5. When the vein was occluded during glucose absorption, both the arterial and venous wall [NO] were gradually reduced to near resting values over a period of 3-4 min. For a maximal reduction in [NO] to occur in both vessels, the flow in the vein must be almost completely stopped. Upon abrupt restoration of venous blood flow during glucose absorption, the venous wall [NO] increased to approximately the preocclusion concentration within 1 min, and the recovery of the arterial wall [NO] was somewhat slower by several minutes.

View larger version (33K): [in this window] [in a new window]   Fig. 6.   Percent changes in [NO] on the wall of paired intestinal arteries and veins are shown at the resting state, during mucosal exposure to 300 mg/dl glucose for the natural state (vein open), and during occlusion of the vein (vein closed). Data set is based on measurements of 6 paired vessels in 6 animals. Occlusion of the vein at rest did not significantly affect either the arterial or venous wall [NO]. As shown in Fig. 5 in a separate group of rats, mucosal absorption of glucose caused a large increase in both the arterial and venous [NO] when the vein was naturally open. Occlusion of the companion vein (vein closed) during glucose absorption resulted in a large decrease in the arterial and venous [NO], and the final concentrations were not different from those at rest or during venous occlusion at rest.

To determine if the mild sodium hyperosmolarity of venous blood during glucose absorption was a possible cause of increased NO formation, isotonic bathing fluid, bathing fluid made hypertonic with sodium chloride to 320 and 360 mosmol/l, and bathing fluid made hypertonic to 360 mosmol/l with mannitol were superfused (100 nl/min) from large-tip-diameter (100 µm) micropipettes over the vessel walls. Localized superfusion of medium was used so that any increases in the arterial wall [NO] would be the result of responses from the vessel studied. The order of sodium and mannitol solutions was varied between experiments, and at least a 10-min recovery was necessary between superfusions to change the fluid used. The 320 and 360 mosmol/l NaCl hypertonic solutions represented the osmolarities measured in venous blood in a prior study (10) during intestinal mucosal exposure to 100 and 300 mg/dl glucose in physiological saline. As shown in Fig. 7, the arterial wall [NO] during 360 mosmol/l was significantly higher than during 320 mosmol/l NaCl. However, the increase in arterial wall [NO] during 320 mosmol/l NaCl was more than twice that which occurred in response to 360 mosmol/l mannitol. The increase in [NO] with sodium hyperosmolarity was typically fully developed within 2 min after superfusion began and was sustained for the 10- to 15-min period of superfusion. In every experiment, the superfusion pipette position was adjusted to give the maximum arterial diameter response, which also corresponded to the maximum NO responses with sodium solutions. In Fig. 7B, the percent increase in vessel diameters is presented. The 320 mosmol/l NaCl was associated with dilation equivalent to that during 360 mosmol/l mannitol, and the 360 mosmol/l NaCl caused an approximately three times greater percent increase in diameter than did 360 mosmol/l mannitol.

View larger version (34K): [in this window] [in a new window]   Fig. 7.   Percent increase in the arterial wall [NO] (A) and inner diameter (B) are shown during isolated superfusion of bathing medium over just the vessel of interest. Medium was the normal bathing fluid whose composition had been increased to 320 mosmol/l with NaCl, to 360 mosmol/l with NaCl, or to 360 mosmol/l with mannitol; isotonic medium had no effect on either [NO] or arterial diameter. All of the hypertonic media caused a significant increase (* P  0.05) in both the arterial [NO] and inner diameter; however, the NaCl hypertonicity was much more effective in increasing the [NO] and thereby in increasing vessel diameter. Compared with the increase in [NO] to 360 mosmol/l mannitol, 320 mosmol/l NaCl increased [NO] ~2.5 times, and 360 mosmol/l NaCl increased [NO] about 5 times. Data set is based on 9 arteries studied in 8 rats.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The major goals of this study were to demonstrate that the concentration of NO is increased in the wall of intestinal resistance vessels during absorptive hyperemia, establish whether the primary source of the NO is arterial or venous vessels, and determine the mechanisms involved in causing this increase in [NO]. One of the first problems encountered with measurements of tissue and perivascular NO is whether a microelectrode system that had a reliable in vitro calibration could be trusted for in vivo studies. To address this issue, the [NO] was measured during pharmacological tests of the type routinely used in countless bioassay studies of the effects of endothelium-dependent vasodilators. As shown in Fig. 4, topical L-arginine caused a large increase in periarteriolar [NO] that could be blocked at least temporarily with L-NNA. In addition, topical L-NNA caused an average 50% reduction in the perivascular [NO], which reflected locally suppressed production of NO. Although these data do not confirm the validity of the in vivo [NO] measured, they do demonstrate that changes in perivascular [NO] consistent with the current understanding of the regulation of NO formation can be followed with NO-sensitive recessed-tip microelectrodes.

The next major problem to be resolved was to identify the sources of NO production. Preliminary studies indicated that, during mucosal glucose absorption, the [NO] around arterioles and venules within the bowel submucosal layer was increased ~200-300 nM. This "glow" of NO was too uniform between paired arterioles and venules to decide the major origin of NO. Therefore, the small artery and vein about to enter the bowel wall, which are actually extensions of the largest arteriole and venule, were used to simplify the measurements. In this simplified model, the source of vasoactive materials from the intestine would be venous blood and any agents released by the wall of the vein. This is reasonable because Steenbergen et al. (25) have established that, even after lymphatic occlusion with mineral oil, absorptive hyperemia proceeded normally as long as venous outflow was intact. The assumption in this earlier study (25) was that materials in venous blood reached the arterioles and caused part of the dilation. This assumption was based on the demonstration by Hester and co-workers (15, 20, 22) that materials, including non-endothelium-dependent vasodilators, in venous blood can diffuse to nearby arterioles and cause dilation in the cremasteric muscle vasculature. In this same context, Unthank and Bohlen (26) have shown that low-molecular-weight materials diffuse from venous blood to the interstitial environment beside arterioles. In addition, Falcone and Bohlen (13) have shown that intestinal venules release sufficient NO to cause dilation of nearby arterioles. As shown in Fig. 3, at least one-half or more of the dilation of arterial vessels during glucose absorption was caused by something related to the venous blood. The basis of this opinion is that, when the venous outflow beside the companion artery was rerouted during glucose absorption, the small arteries dilated much less. Furthermore, the large increase in arterial and venous wall [NO] during glucose absorption was decreased by 70-80% during venous occlusion, as shown in Fig. 6. Occlusion of the vein at rest had no appreciable effect on arterial diameter or arterial and venous wall [NO]. In addition, the diminished arterial vasodilation during glucose absorption when venous occlusion occurred was not of myogenic origin because arteriolar microvascular pressure did not change appreciably.

The source of the increased NO for both the artery and vein during glucose absorption appeared to be from their respective walls because the micropipette tip was on the vessel surface facing away from the companion vessel. Therefore, it is reasonable to propose that venous endothelium-dependent release of NO was stimulated by chemicals that originated in the venous blood. The same scenario likely existed for the small arteries; NO formation was stimulated by chemicals that originated in venous blood. For the average small artery, the minimum distance for NO to diffuse from the vein to the opposite side of the artery was ~150 µm, which is an unlikely distance for NO to survive during diffusion (1) in the oxygen-rich environment around large resistance vessels (25). For the arterial wall [NO] to increase during absorptive hyperemia, chemicals must have diffused from blood in the companion vein and initiated increased NO production in the arterial wall. However, on the surfaces of the two vessels facing each other, there would be a region of arterial wall whose [NO] was influenced by the usually higher [NO] of the companion vein, as shown in Fig. 5. At this point, the evidence obtained supports chemical communication from venous to arterial vessels through diffusion of vasoactive materials from venous blood and a sharing of NO from both vessels if they are in close proximity.

Even though absorptive hyperemia was associated with a substantial increase in the arterial wall [NO], there are clearly other important mechanisms that contributed to arterial dilation. After venous occlusion, about one-half of the absorptive hyperemia dilation by the small artery was retained, yet the [NO] was not significantly increased (Figs. 5 and 6). This remaining vasodilation is likely due to a combination of cell-to-cell communication from downstream arterioles back to the small artery, flow-mediated vasodilation, and diffusion of non-endothelial-dependent vasodilators from the venous blood. The possibility of cell-to-cell conduction from large arterioles upstream to small arteries was confirmed. The large arterioles in the bowel wall typically dilated ~20% during absorption of 300 mg/dl glucose. It was assumed that some fraction of this dilation could be transmitted upstream to the feed arteries, as Segal and Duling (23) have shown for a skeletal muscle vasculature. As documented in RESULTS, when ~20% dilation of downstream arterioles was induced with microiontophoretic application of acetylcholine, the small artery at a distance 2 mm outside the bowel wall dilated by ~8%, or about one-half of the dilation during 300 mg/dl glucose absorption.

The last issue to be discussed is a mechanism that caused the arterial and venous wall [NO] to increase during absorption. A potential contributor to NO released during absorptive hyperemia is the previously documented 20-60 mosmol/l increase in venous blood osmolarity during 100-300 mg/dl glucose absorption (10). The increase in osmolarity is primarily due to mucosal absorption of sodium and an anion, based on periarteriolar and perivenous measurements with sodium-sensitive microelectrodes (4). Steenbergen and Bohlen (24) found that an increase in osmolarity due to sodium caused much greater intestinal vasodilation than did a comparable increase in osmolarity due to mannitol. The difference in dilation responses between sodium and mannitol hyperosmolarity was linked to NO formation, as judged by ~50% suppression of vascular responses to sodium hyperosmolarity but a minor effect on dilation during mannitol hyperosmolarity after blockade of NO formation with an L-arginine analog. As shown in Fig. 7, topical sodium hyperosmolarity over just the vessel being studied caused an osmolarity-dependent increase in both the [NO] and artery diameter. By comparison, 360 mosmol/l mannitol solution caused only ~20% of the increase in [NO] associated with 360 mosmol/l NaCl and a comparatively small amount of dilation. In fact, 320 mosmol/l NaCl caused both a greater increase in [NO] and equivalent dilation of the same vessel during 360 mosmol/l mannitol. These results have demonstrated just how potent sodium hyperosmolarity is in terms of stimulating NO formation during absorptive hyperemia.

In conclusion, intestinal absorptive hyperemia is a coordinated vasodilation of both the intestinal wall microcirculation and small arteries leading to the bowel wall. The coordination in part is accomplished through venous to arteriolar and vein to artery communication of NO generated by the venous vessel and endothelium-dependent vasodilators in venous blood diffusing to the resistance vessels. One likely material used as an endothelium-dependent vasodilator is the hypertonic sodium concentration in venous blood. Simulation of sodium hyperosmolarity caused as large an increase in [NO] as did natural absorptive hyperemia and a much larger increase in [NO] than occurred with equivalent nonspecific mannitol hyperosmolarity. Even though the NO mechanism is a key process in absorptive hyperemia, it only accounted for about one-half of the maximum vasodilation response to glucose absorption. For small arteries near the bowel wall, a large fraction of the remaining vasodilation appeared to be arteriolar to arterial cell-to-cell conduction of vasodilation.