INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

UPWARD OF SEVENTY PERCENT of patients with sleep apnea have systemic hypertension (9). Often, the hypertension is reversed with effective apnea treatment (4, 30). Transient blood oxygen desaturation below 70% along with recurrent acute elevation of blood pressure (BP), tachycardia, and increased sympathetic nervous system (SNS) activity accompanies human apnea and is well described in acute animal models (3, 6, 17). Several studies in humans have demonstrated chronic, elevated diurnal SNS activity in awake sleep apnea patients (22, 31). Despite the clinical association between sleep apnea and chronic hypertension, no causative relationship nor mechanism has been identified linking the acute sleeping apnea BP changes with chronic daytime hypertension. Although "sympathetic overactivity" remains a catchword describing this clinical relationship, it is unclear how sympathetic overactivity might be translated into daytime increased vascular resistance. The limitation of BP changes in response to episodic hypoxia (EH) by SNS blockade is compatible with the concept of sympathetic overactivity but may simply be an "early step" in the mechanism leading to more permanent differences in vascular reactivity. Changing vascular tone from interactions between the endothelium and vascular smooth muscle in response to chronic EH may be later stages in the mechanism, either through the effect of direct, prolonged sympathetic overactivity or through local effects of recurrent EH on the vascular endothelium. Recent publications in humans with obstructive sleep apnea point to differences in vascular reactivity that could involve the endothelial nitric oxide (NO) system (7, 8).

During the last decade, it has become increasingly apparent that the endothelial cell is capable of releasing substances that regulate vascular smooth muscle tone (31). Among the vasoactive substances produced by the endothelium are endothelin and endothelium-derived relaxing factor (or NO), prostaglandins (PFG₂, PGE₂), thromboxanes, and leukotrienes, all likely to be important (16). Agents such as ACh, bradykinin, and substance P work through the NO system by activating NO synthase (NOS) in the vascular endothelium, leading to generation of NO from L-arginine (26). NO also modulates sympathetic neurotransmission by decreasing both the release of and smooth muscle contractions to endogenous norepinephrine (NE) (14). Microvascular studies using animal models of hypertension have suggested that an increase in the reactivity of arterioles to endogenous and exogenous vasoconstrictors occurs as well as impairment in the action of endothelium-derived vasoactive factors (16, 31). Endogenous basal NO as well as endothelin activity may play a role in vascular reactivity of small resistance vessels in hypertension (28). For example, in spontaneously hypertensive rats (SHR), elevated endogenous endothelin has been shown to potentiate constriction to angiotensin II, and administration of BQ-123 (endothelin antagonist) produces a fall in systemic BP (1, 24).

We have developed a normobaric rat preparation that mimics the hypoxic changes seen in sleep apnea patients (12). With the use of individual cylindrical cages, with a rapid (12 s) exchange of the inspired fractional concentration of oxygen (FI_O2) to as low as 2-3%, arterial blood oxygen saturation acutely falls to levels around 70-75%. Such intermittent hypoxia or EH, when administered repetitively in 30-s cycles for 7 h/day for 35 days, increases resting, diurnal mean arterial pressure (MAP) by 8-13 mmHg (10-12). Using this model, we tested the hypothesis that recurrent EH may cause a chronic daytime BP increase through changes in basal tone and microvascular responsiveness to endogenous vasoconstrictors and/or vasodilators such as endothelin and NO.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Surgery. Thirty-eight 12- to 14-wk-old (350 ± 50 g) Sprague-Dawley rats were purchased from Harlan Sprague Dawley (Indianapolis, IN). Anesthesia for all surgery was achieved by intraperitoneal injection of 1 ml of a cocktail containing ketamine (37.5 mg/ml) and xylazine (5 mg/ml). In 16 rats, the catheter portion of a radiotelemetry probe (Data Sciences, St. Paul, MN) was introduced transluminally at the iliac bifurcation of the abdominal aorta, with the tip resting just distal to the renal arteries. The telemetry unit was attached to the anterior abdominal wall as the incision was closed. BP was not measured in the remaining 22 rats to preserve both cremaster arteries for subsequent microvascular studies.

Study groups. Rats were divided into six groups as follows: 1) telemetry rats (n = 8 EH rats and n = 8 unhandled normoxia controls); 2) nontelemetry rats (n = 6 EH rats with dose response curves and n = 6 unhandled normoxia control rats with dose response curves); 3) EH with N^G-nitro-L-arginine methyl ester (L-NAME) (n = 5 EH rats with L-NAME; n= 6 unhandled normoxia controls with L-NAME). Previous studies in this model using sham (same chambers, compressed air in 30-s cycles)-stimulated and handled rats have demonstrated no change in BP over time compared with unhandled nonstimulated rats (10-12).

Hemodynamic measurements. The eight telemetry-probe rats exposed to EH as well as the eight normoxia control rats were placed in their chambers on the morning of days 0, 8, 14, 20, 26, 32, and 37 for measurement of resting (nonhypoxia stimulated) BP. On these days, from 0900 to 1200, EH was withheld for 3 h while MAP and pulse data were collected continuously and averaged.

Hypoxic chambers. As described previously, animals exposed to EH (n = 19) were housed in identical cylindrical Plexiglas chambers (length = 28 cm, diameter = 10 cm, volume = 2.4 liters) with snug-fitting lids (10-12). With the use of a timed solenoid valve, pure nitrogen was distributed to each chamber for 12 s at a flow that was adjusted to reduce the ambient FI_O2 to 2-3% for ~3-6 s. This was followed by infusion of compressed air, allowing gradual return (over 15-18 s) of ambient air to FI_O2 of 20.9%. The cycle was repeated twice per minute for 6-8 h on 35 consecutive days. A dampening device at the air/nitrogen end of the chamber was used to dissipate the air stream so that no direct jets of gas disturbed the animal. Each day of the 35-day experiment, the rats were placed in the same chamber in the morning, and nitrogen flow was adjusted to reach the above specified concentrations. The minimal FI_O2 in each chamber was assessed at least twice daily (and adjusted) throughout the 35-day exposure period by sampling ambient nadir oxygen (MiniOX I, Catalyst Research, Owings Mills, MD). Mean daily nadir FI_O2 was calculated for each cage.

In vivo cremaster preparation. Rats were anesthetized with pentobarbital sodium (50 mg/kg), and the trachea were cannulated (PE-240). The carotid artery was cannulated (PE-50 tubing). The core temperature of the rat was maintained at 35°C with a servo-controlled heating mat. The scrotum was incised medially and gently dissected (to protect neural connections) free from the testicle. The cremaster was suspended with silk sutures in a solution of (in mM) 112.9 NaCl, 25.5 NaHCO₃, 11.6 dextrose, 4.7 KCl, 1.19 KH₂PO₄, and 1.19 MgSO₄ · 7H₂O. Both PO₂ (35-45 Torr) and pH (7.4 ± 0.05) were controlled by changing the amount of carbon dioxide and nitrogen bubbled through the warm bath (35°C). The microvascular preparations were observed using closed-circuit television microscopy on the stage of Leitz trinocular microscope with light transmitted through the optical port of the bath, tissue, and microscope. The image of the microvessels was transferred to a television monitor and simultaneously stored on a video recorder. The cremaster arteriole was selected for observation on the basis of the vascular branching pattern. The largest arteriole entering the cremaster muscle was termed first order (1A) with next two branching levels termed second (2A) and third (3A) order. A single third-order vessel from each rat was used for this study. Vessel diameters were measured directly from the calibrated television monitor (magnification ×1,000-1,500) at baseline and after topical application of the vasoactive agents. All portions of this protocol were approved by the animal studies committee of the University of Louisville School of Medicine.

Microvascular protocols. Each protocol was begun with a 5-min control period followed between protocols by a 30-min wash with warm physiological saline in the cremaster bath. For protocol 1, topical application of NE (10¹⁰ to 10⁶ M) was given with concentrations increased 10-fold at 5-min intervals. For protocol 2, a 5-min control period was followed by one topical dose of NE (10⁷ M); a 5-min equilibration was followed by topical application of ACh (Sigma Chemical) in increasing 10-fold concentrations (10¹⁰ to 10⁶ M) at 5-min intervals. For protocol 3, endothelin-1 (10¹² to 10⁸ M; Peptides International, Louisville, KY) was applied topically in concentrations (10¹² to 10⁸ M) increasing 10-fold at 10-min intervals. For protocol 4, L-NAME (Sigma Chemical) was administered. After 5-min control period, L-NAME was applied in a single dose (10⁵ M) with response measured at 20 min. After the NE study, an NE dose of 10⁷ M was given before application of ACh doses because this amount constricted the arterioles ~50% in control and EH groups.

RNase protection assays. At study termination, telemetry EH and control rats were decapitated, the chest and abdomen were incised, a large-bore catheter was inserted into the lower thoracic aorta, the aorta was clamped below the renal arteries, and the kidneys were flushed with iced physiological saline until blanched. The kidneys were removed and flash frozen in liquid nitrogen. Tissue was stored at 70°C.

Statistical methods. In microvascular studies, each data point was expressed as a percentage of the average baseline (control period) resting diameter. Only one vessel was studied in each animal. The ED₅₀ (dose of drug that produces 50% of maximal effect) was determined using a curve-fitting program, and the values were converted to pD₂ values (negative log of ED₅₀) as an indication of microvascular reactivity. Student's t-test was used to compare difference between two groups of data. Weekly MAP values from the telemetry probe rats were compared by one-way ANOVA for repeated measures with post hoc Bonferroni's and Student's t-tests when applicable. The null hypothesis was rejected at P < 0.05. Deviation from mean is reported as ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The baseline MAPs were not different between the two telemetry groups, and there was no significant change in MAP for the control group from baseline to end of study (Fig. 1). In the EH rats, resting, unstimulated MAP levels days 8 through 37 (end-of-study) were significantly greater than day 0 results (baseline) (P < 0.01-0.05, repeated measures ANOVA). EH MAP at day 37 was significantly greater than that for control at day 37 (P < 0.03, ANOVA, repeated measures for multiple groups).

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Resting, nonstimulated mean arterial pressure (MAP) for episodic hypoxia (EH; ) and controls (Cont; ) measured on days specified throughout the 35-day exposure to EH. *MAP for EH rats was significantly greater than day 0 (baseline) results (P < 0.01-0.05, repeated measures ANOVA). +EH MAP on day 37 was significantly greater than that of controls on day 37 (P < 0.03, ANOVA, repeated measures for multiple groups).

There was no difference in baseline arteriolar constriction (10⁷ M NE) between the EH and control groups (Table 1). However, the EH rats (n = 6) showed less responsiveness to ACh in the doses tested than did the control rats (n = 6) (Fig. 2; P < 0.05). There was no significant difference in vasoconstriction to increasing doses of NE nor to increasing doses of endothelin-1 compared with controls, with complete arteriolar constriction at the last dose (10⁸ M) for both EH and control groups (Table 1). A single dose of L-NAME (10⁵ M) resulted in greater vasoconstriction in control (n = 5) rats (diameter = 61.2 ± 7.4% of baseline) compared with EH (n = 5) rats (diameter = 83.2 ± 6.4% of baseline) (P < 0.05; Fig. 3).

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Dose-response curve for ACh starting with vessels ~50% constricted with norepinephrine (107 M). Horizontal axis shows increasing doses of ACh administered every 5 min, and vertical axis is mean diameter as a percentage of baseline or "percent relaxation." *Values for ACh vary from controls by P < 0.05 for concentrations of 108 to 106 M.

View larger version (10K): [in this window] [in a new window]   Fig. 3.   Mean percent constriction for all control rats vs. EH rats given a single dose of NG-nitro-L-arginine methyl ester (L-NAME). After a 5-min control period, L-NAME was applied in a single dose (105 M), with response measured at 20 min. The means differ at P < 0.05 level.

There was no significant difference between the groups in renal tissue eNOS mRNA normalized to -actin (Fig. 4).

View larger version (68K): [in this window] [in a new window]   Fig. 4.   Semiquantitative analysis (bottom) of rat endothelial nitric oxide synthase (eNOS) mRNA expression in the kidneys of EH and non-EH rats (top). "Housekeeping" gene was -actin. The mean eNOS values expressed are the ratio of Phospho-Imager values of the test mRNA over -actin mRNA expressed in arbitrary units. There was no significant difference between groups with regard to eNOS mRNA.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study was designed to examine microvascular reactivity in Sprague-Dawley rats exposed to intermittent hypoxia or EH for 35 days vs. age-matched, unhandled control rats. As in previous studies in this model, the EH-exposed rats showed a substantial increase in resting, nonhypoxic MAP, in this case, 16 mmHg, whereas the control rats showed no significant change (10-12). Other important findings of this study are as follows: 1) EH rats showed less in vivo vasorelaxation in response to increasing doses of ACh than non-EH-exposed control rats, 2) there was significantly greater constriction in the control rat vessel diameter with L-NAME compared with the EH group, and 3) there was no difference between groups in the vasoconstrictor response to endothelin-1. These results are most likely due to a significant difference in the posthypoxic, basal endothelial NO release between the control and EH group.

The cremaster muscle responds similarly to striated muscle elsewhere in the body and is representative of a vascular bed that is a major site of peripheral resistance and where increases in vascular reactivity have been observed during hypertension. It is for this reason that we choose this model to investigate the mechanism of the microvascular resistance with intermittent hypoxia-induced hypertension. In the NE protocol, we found no difference between the control and EH groups, which suggests that the NE effect was directly on the vascular smooth muscles and endothelial NO had no modulating effect on NE. This supports similar findings in the unrubbed femoral artery in Wistar Kyoto rats vs. SHR exposed to NE, although endothelial function may have a compensatory mechanism to enhance contractile responsiveness of smooth muscle cells in rubbed hypertensive rat arteries (2).

L-NAME is a stereospecific inhibitor of NOS, and the vasoconstrictor response to L-NAME is reduced in patients with hypertension (5). We tried to block NO production using L-NAME and found significantly more arteriolar constriction in the control group, suggesting greater eNOS blocking effect (increased basal activity) compared with the EH rats.

It is nearly impossible to measure eNOS mRNA content of cremaster vessels because it would require pooling of tissue from hundreds of rats to achieve sufficient mRNA levels. We were thus forced to measure eNOS in some other vascular tissue; in this case, we chose the kidney. The importance of continuous hypoxia ranging from 6 to 24 h on NO activity in the kidney has been measured by various means, including examination of NO metabolites in urine (23) and in cultures of rat proximal tubular epithelial cells (25), by examination of the effects of hypoxia on proximal tubular cell function in NOS knockout mice (19), and by measuring eNOS mRNA in whole organs (13). The results of such studies are confusing or contradictory in that NO activity in some is upregulated whereas in others eNOS activity is downregulated. Nevertheless, without literature on changes in kidney eNOS in the setting of EH, we were forced to measure this directly in control vs. EH rats. We were unable to demonstrate an effect of EH on eNOS (Fig. 4), with several possible explanations. The eNOS mRNA level simply reflects the rate of eNOS formation but does not reflect absolute levels of eNOS protein, which in EH animals could be low due to increased eNOS degradation, decreased formation due to lack of substrate, or block at the receptor level. Biosynthesis of NO from L-arginine is an oxygen-dependent process, and hypoxia could influence NO formation in vascular beds. It is possible that the 48-h lapse after the last hypoxia exposure could allow restoration of eNOS mRNA. Also, the kidney tissue was not separated into cortex vs. medulla, and there could well be local differences within organs as well as between organs (kidney vs. cremaster microvasculature) in eNOS mRNA levels. We also cannot exclude the release of other hypoxia-stimulated, endothelium-derived contracting substances such as adenosine and PGF₂, reduced substrate levels (L-arginine), or the presence of reactive oxygen species. Finally, high hemoglobin levels may increase blood NO-carrying capacity. Although we did not measure hematocrit in these rats, we have generally found only small or no differences in hematocrit levels between EH and control rats in previous studies. However, if NO activity were substantially influenced by hemoglobin content, the resulting changes would have been opposite to the findings in our study, that is, higher hematocrit (EH animals) would show more NO vasodilator activity and greater ACh responsivity.

Important to the combination of hypertension and hypoxia is the interaction between O· and NO· (18). Because O· and NO· are both free radicals, they undergo extremely rapid, diffusion-limited radical/radical reaction that markedly alters biological availability of NO· for many cell functions. For example, a major product of this reaction is peroxy nitrite anion (OONO), which is only a weak vasodilator compared with NO·, markedly impairing vasodilator function. In normal vessels, the balance between O· and NO· favors net production of NO·, permitting a basal state of vasorelaxation and normal BP (18). Disease states such as atherosclerosis, hypertension, and diabetes may alter this balance. Nakazono et al. (21) have demonstrated that superoxide dismutase, a ROS scavenger, acutely lowers BP in SHR but has no effect on BP in normotensive rats.

There is precedent for decreased vascular reactivity to ACh in patients with intermittent or episodic hypoxemia from sleep apnea. Using forearm blood flow with strain gauge plethysmography, Carlson et al. (7) examined vasodilation responses to infused ACh in eight normotensive subjects, eight hypertensive sleep apnea subjects, and appropriate normal controls. Decreased forearm vascular relaxation to ACh was shown in all of the apnea subjects, regardless of BP status, compared with controls. Duchna et al. (8) recently studied 12 normotensive sleep apnea subjects vs. 12 matched nonapneic controls to measuring bradykinin (endothelium-dependent NO) and nitroglycerine (NO-independent) vasodilation dose-response curves. These authors found that the bradykinin curve was depressed (less NO-dependent vasodilation) compared with controls and that 2 mo of treatment with nasal continuous positive airway pressure allowed return of the curves to control levels. The degree of abnormality was related to the severity of hypoxemia during sleep.

One study has implicated endothelin-1 as having a possible role in the etiology of apnea-induced hypertension. Phillips et al. (27) found elevated circulating levels of endothelin-1 in 22 patients with obstructive sleep apnea vs. 12 healthy controls. Endothelin-1 normalized in the apnea subjects after the use of nasal continuous positive airway pressure therapy. A subsequent study found exactly the opposite results (15). We did not measure plasma endothelin-1 levels in our EH rats. However, to investigate the possible role of endogenous endothelin-1 in the mechanism of microvascular resistance in EH vs. normoxic rats, we calculated endothelin-1 dose-response curves without adding blocker or modulating substance. We found no difference between the control and EH group.

From these results, we conclude that NO most likely plays an important role in the basal tone of microvascular resistance arterioles in rats exposed to EH. In EH hypertensive rats, we found less response to ACh, which is an NO stimulator, and less response to the NO inhibitor L-NAME compared with the control rats. Our results are consistent with findings in humans with hypertension and in human apnea subjects. We are unable to conclude from our data whether the lower NO activity was due to decreased release or production of nitric oxide or by interference in NO activity by other substances such as peroxy nitrite molecules.