Key words: Anesthetics, volatile: enflurane; halothane; isoflurane. Endothelium, vascular: endothelin; endothelium relaxing factor. Oxygen; hypoxia. Veins: mesenteric.

SYSTEMIC hypoxia produces multiple cardiovascular adaptations that lead to the redistribution of blood flow to organs with high oxygen demand. [1–4 ] Numerous in vitro studies have demonstrated that hypoxia also may exert a direct effect on blood vessels and contribute locally to changes in peripheral vascular tone. Inhalational anesthetics interfere with systemic hemodynamic and direct vascular responses to hypoxia, resulting in altered vascular reactivity [5 ] and impaired blood flow to various organs. [2,3,6 ] Anesthetic agents therefore compromise homeostatic adjustments during hypoxia.

Splanchnic capacitance vessels function as an active blood mobilization system under sympathetic control. Acute systemic hypoxia reduces splanchnic capacitance via mesenteric venoconstriction because of combined influences on the central nervous system, peripheral chemoreflex activation, and sympathetic efferent transmission. [4,5 ] This response is attenuated by volatile anesthetics, [7 ] which may affect the redistribution of circulating blood volume essential for homeostatic adjustments to hypoxia. Alternatively, hypoxia evoked mesenteric vasoconstriction may be in part responsible for nonocclusive mesenteric ischemia, [8 ] and volatile anesthetics may provide a protective effect against this phenomenon.

We have reported previously that mesenteric venoconstriction occurring during acute systemic hypoxia in intact rabbits is markedly reduced by inhaled halothane. [7 ] However, direct effects of hypoxia, or hypoxia in the presence of volatile anesthetics, on mesenteric veins are unknown.

The purpose of the current in vitro study was to determine whether short-term acute hypoxia has a direct vasoconstrictive effect on isolated small mesenteric veins of the rabbit, to evaluate the mechanism of the hypoxic response, and to examine the effects of halothane, isoflurane, and enflurane.

Experiments were performed on 268 rings of small (500–800 micro meter in diameter) veins isolated from mesentery of the terminal ileum of the rabbit. All experimental protocols and the animal use were approved by the Animal Care Committee of the Medical College of Wisconsin. Seventy-five male New Zealand white rabbits, weighing 1.53 plus/minus 0.02 kg were anesthetized with thiamylal (20 mg/kg, intravenous). The mesentery was excised and placed in a cold physiologic salt solution (PSS) of the following composition (in mM): NaCl 119, KCl 4.7, MgSO⁴1,17, CaCl²1.6, NaHCO³27.8, NaH²PO⁴1.18, EDTA (ethylenediaminetetraacetic acid) 0.026, HEPES (4-[2-hydroxyethyl]-1-piperazine-ethanesulfonic acid) 5.8, and glucose 5.5. The veins were dissected, cleaned of adherent tissue, and cut into 4-mm long segments (rings). In some rings, the endothelium was removed mechanically by inserting a 25-G hypodermic needle into the vascular lumen and gently rolling the vessels over a soft silastic plate moistened with PSS. Individual rings were suspended on Tungsten stirrups in 15-ml organ baths (Radnoti, Monrovia, CA) containing oxygenated (95% Oxygen²-5% CO²) PSS, maintained at pH 7.37–7.4 and 37 degrees Celsius. The P^O2(500–600 mmHg) and P^CO2(35–40 mmHg) in the bath solution were monitored frequently (ABL3, Radiometer, Copenhagen, Denmark). The isometric tension was measured via FTO3 force transducers (Grass, Quincy, MA) and recorded using Grass model 7 polygraph.

A passive force that permitted maximum contraction of veins to 80 mM potassium (50 mg, equivalent to 0.5 mN and defined here as basal resting tension) was placed on each ring. The rings were allowed to equilibrate for 2 h and then contracted twice with potassium (80 mM KCl), to stabilize the vascular preparations. In some denuded veins, effectiveness of endothelium removal was verified histologically using silver nitrate and cobalt bromide staining. [9 ] In all vein rings, the presence or absence of functional endothelium was tested by determining the response to 10⁴M acetylcholine (Sigma, St. Louis, MO) in 10⁶M norepinephrine (Sigma, St. Louis, MO) precontracted rings. Twenty percent to 50% relaxation was seen in endothelium-intact vessels, whereas no relaxation or a slight increase in tension was observed in the vessels without endothelium (Figure 1(c)).

Hypoxia (10 min) was produced by replacing the gas mixture aerating organ baths with 95% Nitrogen²/5% CO², decreasing P sub O²from 600 plus/minus 8.3 mmHg to 26 plus/minus 0.8 mmHg within 1 min. P^CO2(34 plus/minus 1.0 mmHg) and pH (7.42 plus/minus 0.01) of the bath solution remained unchanged. Oxygen was reintroduced after 10 min in the continued presence of the precontracting agent, and the posthypoxic responses were recorded for an additional 3 min. After washouts, a 1-h recovery period (normoxia) was used to separate successive hypoxic episodes.

Unstimulated (Quiescent) Veins. Veins with or without endothelium were exposed to two or three episodes of hypoxia (10 min), separated by 1-h intervals of normoxia.

Stimulated Veins. Veins with and without endothelium were precontracted with 3 x 10⁷M norepinephrine (EC⁴⁰). Hypoxia (10 min) was initiated during, arbitrarily chosen, third minute of contraction to agonist. In some experiments, hypoxia was conducted in absence of extracellular calcium, after 30-min preincubation of rings in Calcium²+-free PSS. Phenylephrine (3 x 10 6 M), prostaglandin PGF sub 20 (5 x 10⁶M), or KCl (60 mM) also were used as precontracting agents.

Endothelium-intact veins precontracted with 3 x 10 sup -7 M norepinephrine were used to further characterize the hypoxic response. After control hypoxia and a period of recovery, the effects of the following pharmacologic agents were examined: alpha-adrenergic antagonist, phentolamine (5 x 10 sup -6 M, Regitine mesylate, Ciba-Geigy, Summit, NJ); beta-adrenergic antagonist, propranolol (10 sup -5 M, Sigma, St. Louis, MO); inhibitor of cyclooxygenase, indomethacin (10⁵M, Sigma); inhibitor of lipoxygenase, nordihydroguaiaretic acid (NDGA, 2 x 10⁵M, Sigma); blocker of cell membrane voltage-gated calcium channels, nicardipine (10⁵M, Sigma); a depleter of Calcium²+ in sarcoplasmic reticulum, ryanodine (5 x 10 sup 6 M, Sigma); inhibitor of nitric oxide synthase, N^G-nitro-L-arginine methyl ester (L-NAME: 5 x 10⁵M, Sigma); inhibitor of endothelin-converting enzyme, phosphoramidon (10⁵to 10³M, Sigma); antagonist of endothelin ET^Areceptors on vascular smooth muscle, BQ-123 (10⁶M, PED-3512-PI, Peptides International, Louisville, KY); and endothelin-I (10⁹to 10⁷M, Sigma). Hypoxia was repeated in the presence of individual drugs after 15–20-min exposure to these compounds. Stock solutions of all drugs, except indomethacin and NDGA, were prepared daily in distilled water. Indomethacin was dissolved in ethyl alcohol and NDGA in propylene glycol. The concentrations of all drugs are expressed as the final molar (M) concentrations in the bath solution.

The effects of halothane (Halocarbon, North Augusta, SC), isoflurane, and enflurane (Anaquest, Madison, WI) on the responses to hypoxia were examined in endothelium-intact veins. After control hypoxia, recovery, and 15-min equilibration with 1.0 MAC doses of individual anesthetics, hypoxia was repeated in the continued presence of either halothane, isoflurane, or enflurane. The 1.0 MAC values of volatile anesthetics for male New Zealand white rabbits as well as concentration of anesthetics in the bath solution are listed in the Table 1. Concentration of anesthetics in the baths was determined using gas chromatography (Sigma 3B gas chromatograph, Perkin-Elmer, Norwalk, CT). In some experiments involving anesthetics, the responses to hypoxia were examined after increasing calcium concentration in the baths from 1.6 to 3.2 mM.

Tension (mg) was measured and reported as an increase above the initial basal resting tension. Data are expressed as mean plus/minus SEM. In all experiments n refers to the number of rabbits. The number of vascular rings also is reported. Data were analyzed by multiple analysis of variance for repeated measures with individual contrasts using the Super ANOVA statistical software produced by Abacus Corporation (Berkeley, CA) for Macintosh computers. Arc-sin transformations were performed on all percentage changes to ensure normal distribution of these values. The differences were considered to be statistically significant when P was less than 0.05.

Unstimulated Veins. Quiescent veins, with and without endothelium, showed no change in basal resting tension during repeated episodes of hypoxia.

Stimulated Veins. Denuded veins (n = 9 rabbits, 16 rings) demonstrated a greater contractile response to 3 x 10 sup - M norepinephrine compared with endothelium-intact veins (196 plus/minus 35 mg vs. 104 plus/minus 12 mg) but were unresponsive to hypoxia (Figure 1and Figure 2).

In endothelium-intact veins (16 matching rings from the same group of nine rabbits) norepinephrine time controls (15 min) demonstrated time-dependent decay of contraction to 3 x 10⁷M norepinephrine (gradual decrease from 104 plus/minus 12 mg to 69 plus/minus 8 mg;Figure 1and Figure 2). Ten minutes hypoxia augmented contraction to 3 x 10⁷M norepinephrine causing an increase in tension (from 140 plus/minus 10 mg to 343 plus/minus 36 mg). Hypoxic response had a characteristic triphasic profile (Figure 1). At the onset of hypoxic challenge, there was a transient (1–2 min) relaxation (decrease in tension from 140 plus/minus 10 mg to 98 plus/minus 13 mg), which was followed by augmentation of the contraction to norepinephrine sustained throughout the period of hypoxia. Reoxygenation in the continued presence of norepinephrine induced rapid posthypoxic relaxation (from maximum 343 plus/minus 36 mg to 128 plus/minus 19 mg). This pattern was reproducible during successive episodes of hypoxia (up to 6), separated with periods of normoxia. A similar pattern of hypoxic response was observed after preconstriction with phenylephrine (3 x 10⁶M). Hypoxic response was inhibited by phentolamine (5 x 10⁶M) and did not occur in the absence of calcium (incubation in Calcium²-free PSS). In veins precontracted with prostaglandin PGF²⁰, hypoxia resulted in a transient (1–3 min) increase in contraction followed by a gradual decrease to the prehypoxic level. Both endothelium-intact and denuded veins precontracted with 60 mM KCl showed no further increase in contraction during hypoxia.

Hypoxic facilitation of contraction to norepinephrine was unaffected by propranolol (n = 5 rabbits, 14 vessels, data not shown), indomethacin (n = 5 rabbits, 18 vessels), NDGA (n = 5 rabbits, 14 vessels), or ryanodine (n = 5 rabbits, 13 vessels). Nicardipine (n = 5 rabbits, 13 vessels) depressed the hypoxic response by 86 plus/minus 5%. Indomethacin, NDGA, ryanodine, and nicardipine data are summarized in Figure 3.^L-Name induced a threefold increase in basal contractile response to 2 x 10 sup -7 M norepinephrine (n = 5 rabbits, 24 vessels). However, in the presence of Sub L-NAME, the magnitude of hypoxic augmentation of contraction to norepinephrine was similar to that in control (Figure 4). Phosphoramidon depressed the hypoxic response by 82 plus/minus 8%(n = 8 rabbits, 16 vessels;Figure 5), and BQ-123 by 47 plus/minus 10%(n = 3 rabbits, 16 vessels;Figure 5). The exogenous endothelin-1 dose-response curves (10⁹to 10⁷M) were identical in the veins with and without endothelium (data not shown). Under normoxic conditions, precontraction of vein ring with either phenylephrine, norepinephrine, or KCl had no effect on the contraction to exogenous endothelin-1 (data not shown). BQ-123 attenuated contraction to exogenous endothelin-1 by 60 plus/minus 9%(data not shown).

In endothelium-intact veins, 1.0 MAC halothane (n = 6 rabbits, 21 vessels), isoflurane (n = 6 rabbits, 19 vessels), or enflurane (n = 5, 14 vessels) had no effect on basal contraction to norepinephrine. However, the anesthetics inhibited all phases of the hypoxic response (Figure 6and Figure 7). These effects were reversible after discontinuing anesthetic delivery. Increasing calcium concentration in the baths from 1.6 to 3.2 mM partially reversed the inhibitory effect of anesthetics (data not shown). Halothane and isoflurane also inhibited hypoxic contractions in endothelium-intact veins treated with^L-NAME (n = 6 rabbits, 26 vessels;Figure 8). Under normoxic conditions, anesthetics produced relaxation of the endothelium-intact veins precontracted with 2–5 x 10⁹M endothelin-1 (n = 5 rabbits): 39 plus/minus 3% relaxation in halothane group (12 rings), 28 plus/minus 3% in isoflurane group (15 rings), and 20 plus/minus 3% in enflurane group (14 rings).

Vascular responses to hypoxia differ between species and depend on type of vessels (arteries or veins), tissue origin of vessels under study, and the function of a specific circulatory bed. Hypoxia augments contractile responses to norepinephrine in isolated canine pulmonary and splenic veins via an endothelium-dependent mechanism. [10 ] After removal of the endothelium, hypoxic contraction is reduced in splenic veins and reversed to relaxation in pulmonary veins. [10 ] Endothelium-dependent hypoxic contraction of isolated porcine and canine coronary veins is mediated by cyclooxygenase metabolites of arachidonic acid. [11 ] Contractile responses to norepinephrine in isolated canine saphenous and femoral veins with or without endothelium are unaffected by hypoxianous and femoral veins with or without endothelium are unaffected hypoxia. [10,12 ] However, acute hypoxia causes relaxation of norepinephrine precontracted rings of rabbit saphenous veins (unpublished observations from our laboratory). In isolated canine saphenous veins, acute hypoxia increases endogenous norepinephrine release but decreases norepinephrine metabolism, depletes norepinephrine tissue content, and depresses smooth muscle responses to nerve stimulation. [5 ].

Our study demonstrates that mesenteric capacitance veins of rabbits respond to short-term episodes of acute hypoxia with an increase in contraction to alpha-adrenergic agonists norepinephrine and phenylephrine. Hypoxic contraction is inhibited by phentolamine but is unaltered in the presence of beta-adrenergic antagonist propranolol. The response is transient when precontraction is achieved with prostaglandin PGF²alpha and absent during precontraction with potassium. The hypoxic contraction is endothelium-dependent. It does not occur in the absence of extracellular calcium and is markedly depressed by nicardipine, a blocker of voltage-gated calcium channels in the cell membrane. It is, however, unaffected by ryanodine, a depleter of calcium stores in the sarcoplasmic reticulum. These characteristics indicate a strong dependence on extracellular calcium availability and the importance of transmembrane calcium ion influx into the endothelial and/or smooth muscle cells of veins. Lack of indomethacin and NDGA effects rule out a contribution of cyclooxygenase and lipoxygenase products of arachidonic acid metabolism.

It had been proposed that endothelium-dependent responses to hypoxia might be due to the modulation of vascular endothelium relaxing factor (EDRF)/nitric oxide production. [13 ] Hypoxia may inhibit basal release of EDRF and thus induce or augment vasoconstriction in rabbit pulmonary artery [14 ] and human mammary artery. [15 ] In spontaneously hypertensive rats, hypoxia causes contractions of aorta and coronary artery via progressive inhibition of EDRF synthesis. [16 ] However, hypoxia also may stimulate EDRF release, producing vasodilation of rabbit thoracic aorta and femoral arteries. [17,18 ] In our study,^L-NAME enhanced the sensitivity of mesenteric veins to norepinephrine but had no effect on hypoxic increase in contraction to norepinephrine. This suggests a hypoxic mechanism not involving EDRF.

Endothelium-dependent hypoxic facilitation of contraction to vasoactive agents alternatively can be attributed to the action of endothelium-derived contracting factors (EDCFs), primarily endothelin-1 (ET 1), a potent vasoconstrictor peptide. In vascular endothelial cells, a precursor (big ET-1) is converted to ET-1 by endothelin converting enzyme, which is selectively blocked by phosphoramidon. [19,20 ] In vascular smooth muscle, ET-1 initiates contraction by binding to specific ET^Areceptors. [21,22 ] Cultured endothelial cells isolated from bovine coronary artery [23 ] and human umbilical veins, [24 ] as well as the endothelium of perfused rat mesenteric arteries, [25 ] synthesize ET-1 in response to hypoxia. Rabbit mesenteric veins demonstrate high sensitivity to ET-1. [26–28 ] In our study, the hypoxic augmentation of contraction to norepinephrine occurs exclusively in the presence of intact endothelium, is inhibited by phosphoramidon, and is suppressed by BQ-123, all of which suggest that endogenous endothelin may contribute to the hypoxic contraction of rabbit mesenteric veins. Considering a former model of endothelial cell-dependent vasoconstriction [29,30 ] and a recent hypothesis of two-way communication between the vascular endothelium and vascular smooth muscle, [31 ] we suggest the following mechanism of endothelium-dependent hypoxic contraction of rabbit mesenteric veins. Under hypoxic conditions, a specific adrenoceptor agonist acting directly on the venous smooth muscle cell receptors may evoke a response that could be transferred to the endothelium and activate endothelin-1, which, in turn, may alter/increase the contraction of the smooth muscle. Endothelin-1 may also specifically attenuate the effects of EDRF in the venous smooth muscle. [32 ] Although the production of both endothelin-1 and EDRF depends on extracellular calcium availability, the initial steps of EDRF/nitric oxide synthesis also strongly depend on molecular oxygen availability. [33 ] Thus, under hypoxic conditions, a decrease in oxygen tension may impair EDRF/nitric oxide synthesis. Reduced production of EDRF/nitric oxide and accelerated synthesis of endothelin-1 may enhance the adrenergic responsiveness of the smooth muscle.

Endothelium-dependent hypoxic augmentation of contraction to alpha-adrenergic agonists is not limited to rabbits (Lagomorpha). We have found (unpublished observations from our laboratory) the same type of response to hypoxia in isolated mesenteric veins of guinea pig (Rodentia). However, additional studies involving other species are necessary to draw a general conclusion about possible clinical relevance of these observations.

Volatile anesthetics influence the vascular system via a variety of mechanisms involving the alteration of autonomic function, [34,35 ] modulation of endothelium-dependent and EDRF-mediated processes, [36–38 ] or direct, endothelium-independent actions on vascular smooth muscle. [39–41 ] Although volatile anesthetics are known to attenuate the autonomic responses to hypoxia in vivo, [2,3,6,7 ] their direct vascular effects in the presence of hypoxia are not clarified. It has been demonstrated recently that hypoxia and halothane have direct but opposite effects on metabolism and disposition of endogenous norepinephrine in isolated canine saphenous veins. Separately, hypoxia increased whereas halothane decreased endogenous norepinephrine release. However, when present together, hypoxia and halothane inhibited endogenous norepinephrine metabolism and release, resulting in an overall suppressant effect. [5 ] Our study demonstrates a similar depressant effect of halothane, isoflurane, and enflurane on endothelium-dependent hypoxic contractions of rabbit mesenteric veins. This effect is partially reversed by increasing extracellular calcium. Furthermore, the hypoxic contraction is inhibited by halothane and isoflurane in the absence as well as presence of L-NAME, which suggests a mechanism largely independent of EDRF. The mechanism of volatile anesthetic actions on hypoxic contractions of mesenteric veins may include modulation of transmembrane ion channel activity, [42–44 ] inhibition of calcium influx into the endothelial and/or smooth muscle cells, inhibition of endothelin synthesis, [45 ] and attenuation of endothelin effects in the smooth muscle cells.

To better understand the mechanism of mesenteric venoconstriction during hypoxia and explain the inhibitory actions of volatile anesthetics, further studies are necessary to examine the effects of hypoxia and volatile anesthetics on ion currents and ion channel activity, particularly calcium and potassium, in the venous endothelial and smooth muscle cells.