INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

SILENT MYOCARDIAL ISCHEMIA and cardiac arrhythmias are frequent and major complications of diabetes mellitus that place the patient at risk of sudden cardiac death. Autonomic neuropathy is thought to play an important role in the pathogenesis of these complications (7). Cardiac parasympathetic and sympathetic fibers are affected, leading to a decrease in heart rate variability (7, 26) and impaired baroreflex function (13). This impairment in autonomic control, in concert with other disease processes occurring in the diabetic heart, can contribute to life-threatening arrhythmias. Diabetic patients also have a blunted perception of cardiac pain (1). As a result, myocardial ischemia or infarction may be associated with only mild symptoms that go unrecognized. These observations suggest an impairment of cardiac afferent (sensory) as well as efferent autonomic function, which may compound the risk for cardiac death in diabetic patients.

Cardiac afferent fibers are responsible for initiating protective cardiovascular reflex adjustments in addition to the perception of pain during myocardial ischemia (11), but little is known about the impact of diabetes on sensory nerve function from the heart. The heart possesses a sensory innervation of mechanically and chemically sensitive endings that send signals to the central nervous system via vagal and sympathetic afferent pathways. Stimulation of cardiac vagal afferent endings evokes reflex hypotension and bradycardia (sympathoinhibitory), also known as a Bezold-Jarisch effect (6). By contrast, stimulation of cardiac sympathetic afferent endings evokes reflex hypertension (sympathoexcitatory effect) as well as the perception of cardiac pain (12).

There are several important factors that evoke these cardiac reflexes. Cardiac mechanoreceptors are stimulated by wall motion (22). However, cardiac chemosensitive endings, stimulated by a variety of metabolically active substances known to be produced by the stressed myocardium [e.g., bradykinin (BK), prostaglandins, adenosine, and H⁺], play a major role in mediating reflex adjustments and the perception of cardiac pain during myocardial ischemia (6, 11, 12, 22-25).

It is reasonable to suggest that the function of cardiac chemosensory endings is depressed, which is consistent with the generalized peripheral autonomic neuropathy characteristic of chronic diabetes (1, 7). Nevertheless, no direct evidence exists to describe how the sensory innervation of the heart is altered in diabetes. One of the factors known to contribute to cardiomyopathy and peripheral neuropathy in diabetes is oxidative stress (5). Because cardiac afferent endings are known to be sensitive to oxygen radicals (9, 23-25), it is possible that the elevated presence of oxygen radicals in cardiac tissue during diabetes may lead to alterations in cardiac afferent function.

The objectives of this study were to determine whether chemosensory function from the heart is altered in diabetes and whether antioxidant treatment can influence these alterations in cardiac afferent function.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Methods used in this study conform with the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH Publication No. 85-23, Revised 1996) and were approved by the Institutional Animal Care and Use Committee at the University of Nebraska Medical Center.

Induction of diabetes. Male Sprague-Dawley rats (180-200 g) were divided randomly into sham, diabetic [streptozotocin (STZ) rats], and insulin (Ins) replacement (STZ + Ins rats) groups. The rats were housed individually in cages and provided access to food and water ad libitum. Diabetes was induced by a single intraperitoneal injection of STZ (85 mg/kg) in a 2% solution of 0.1 M cold citrate buffer. Sham rats received a similar injection of vehicle.

Chronic antioxidant treatment. One-half of the animals in the sham and STZ groups received chronic treatment with vitamin E (VE); the other half received VE vehicle (mineral oil). Rats were treated with VE (-tocopherol acetate; 60 mg/kg per os daily) prophylactically starting 2 days before STZ/vehicle injection, and this continued until the day of the experiment. The -tocopherol acetate (1308 U/g; Sigma) was dissolved in mineral oil. About 0.1 ml of the solution was administered orally using a 30-mm (18 gauge) feeding tube.

Experimental preparation. Six to eight weeks after administration of STZ, rats were anesthetized intraperitoneally with a mixture of 2% -chloralose and 25% urethan in saline (5 ml/kg). The trachea was cannulated low in the neck, and the lungs were ventilated by a Harvard rat respirator (60 breaths/min) with air supplemented with O₂. Body temperature was maintained at 37°C by a heating pad. Polyethylene catheters were inserted into a femoral artery and vein for measurement of mean arterial blood pressure (MABP) and administration of drugs, respectively. The neck was opened in the midline to expose the cervical vagus and aortic nerves. The chest was opened via a sternotomy to allow application of chemicals on the surface of the heart. Heart rate (HR) was measured by a cardiotachometer triggered by the arterial pressure pulse. MABP were measured by strain gauges. Estimated fluid loss was replaced with intravenous administration of physiological saline at a rate of 4-6 ml · kg¹ · h¹.

Sinoaortic and cardiac denervation. Arterial baroreceptors were denervated by cutting the carotid sinus and aortic nerves. Baroreceptor denervation was confirmed by the absence of changes in renal sympathetic nerve activity (RSNA) and HR when MABP was increased by intravenous injections of phenylephrine (0.1-5 µg/kg). Cardiac vagal denervation was performed by cutting the cervical vagal nerves. Cardiac sympathetic denervation was performed by surgical transection of the stellate ganglia and T1-T4 sympathetic rami. Confirmation of vagal or sympathetic denervation was assessed by the absence of an inhibitory or excitatory effect, respectively, of epicardial capsaicin (Caps) on RSNA.

Recording of RSNA. The abdomen was opened at the midline to expose the renal nerves and kidneys. A branch of a sympathetic nerve running along the left renal artery was dissected free, and the sheath was removed. The nerve was cut distally, and the central end was placed on a bipolar silver electrode. The nerve and electrode were covered with mineral oil. Nerve signals were amplified (Grass P511), displayed on an oscilloscope (Gould 450), and fed into a rate meter (Frederick Haer). Impulses were counted by ratemeter in 0.1-s bins. Resting nerve discharge was normalized as 100% for each animal, and changes in nerve activity were expressed as a percent change from the resting (baseline) value.

Recording of cardiac afferent impulses. Vagal afferent impulses were recorded from "single fibers" split from fine slips dissected from the distal cut end of the left cervical vagus (3). Sympathetic afferent impulses were recorded similarly from nerve fibers split from fine slips dissected from the distal cut end of the left stellate nerve (9). Nerves were covered in a pool of mineral oil, and the nerve fibers were placed on a silver electrode. Impulses were amplified (Grass P511), displayed on an oscilloscope (Gould 450), and fed into a rate meter (Frederick Haer) whose window discriminators were set to accept potentials of a particular amplitude. Impulses were counted by rate meter in 1-s bins. Nerve fibers that had one, or at most two, easily distinguishable action potentials were used. We studied only spontaneously active fibers that had receptive fields in the heart that could be located precisely.

Measurement of lipid peroxides in plasma. Lipid peroxides in the plasma of the animals of all experimental groups were measured at the time of the experiment by the method described by Ohkawa et al. (14). Venous blood (0.05 ml) was withdrawn and put into 1 ml of physiological saline. After centrifugation at 4,000 rpm for 10 min, we transferred 0.5 ml of supernatant to another centrifuge tube, and plasma lipids were isolated by precipitating with phosphotungstic acid, followed by the reaction with thiobarbituric acid and fluorometric measurement of thiobarbituric acid-reactive substances (TBARS), which was made at 515 nm excitation and 553 nm emission. The results were expressed in terms of malonaldehyde (in nmol/ml of blood), with the use of the fluorescence intensity of the solution obtained by reacting 0.5 nmol of tetraethoxypropane with thiobarbituric acid as a standard. This method is well established for estimation of the intensity of free radical reactions in experimental and clinical conditions (19).

Experimental protocol for reflex studies. To test cardiac chemoreflexes in anesthetized rats with sinoaortic denervation, Caps (1.0, 5.0, and 10.0 µg in 10 µl of saline) and BK (1.0, 5.0, and 10.0 µg in 10 µl of saline) were applied on a circle of filter paper (3 mm in diameter) and placed on the anterior surface of the left ventricle for 30 s. Reflex responses of RSNA and hemodynamics to the stimulation of cardiac afferents by these chemicals were recorded. Arterial baroreceptors were denervated to eliminate any potential influence on reflex responses secondary to changes in hemodynamics. After each application, we removed the paper, and the surface of the heart was washed with warm saline. A recovery period of at least 10-15 min was observed between applications. After constructing the three-point dose-response relationship for each test chemical, we transected either the cervical vagi or cardiac sympathetic nerves, and the epicardial applications of the test chemicals were repeated. At the end of the experiment, the remaining afferent nerves (either cervical vagi or cardiac sympathetic nerves) were transected, and epicardial applications of the test chemicals were repeated to verify the reflex nature of the evoked responses.

Experimental protocol for cardiac afferent studies. To prevent reflex changes in hemodynamcis in response to epicardial application of test substances from potentially influencing afferent responses, the experiments were performed after vagotomy and cardiac sympathectomy. We recorded impulses from vagal and sympathetic afferent fibers with chemosensitive endings in the heart. The chemosensitivity of a cardiac afferent ending was affirmed by its response to topical application of Caps to the surface of the heart (10 µg). We considered the fiber to be activated when its activity increased by >25% above the basal level. Caps was chosen as a test chemical because it is known to directly stimulate only chemosensitive C fiber endings and not cardiac mechanoreceptors (3).

Analysis of data. HR, MABP, RSNA (reflex study), and cardiac afferent activity (afferent study) were recorded by a thermal recorder (Astro-Med MT 95000) and captured by a data-processing computer (PowerLab, ADInstruments). In the reflex study, resting RSNA was normalized as 100% for each animal, and changes were expressed as a percent change from the resting (baseline) value. In the afferent study, fiber activity was calculated as the average number of impulses per second (imp/s) over the 10 s of maximal activity occurring during the final 30 s of the control period and during the first 30 s of the experimental intervention (peak afferent responses always occurred within the first 30 s). Because of the individual variability in the control activity of different fibers, the neural responses were also expressed as a percent change from the baseline.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Body weights, blood glucose, and baseline hemodynamics are listed in Table 1 for each group of rats at the time of the experiments (6-8 wk post-STZ injection). STZ and STZ + VE rats gained less weight and exhibited higher blood glucose than sham rats at the time of the experiments. The STZ + Ins rats exhibited weight gain, blood glucose, and hemodynamics similar to sham rats. Chronic VE administration was without effect on any of these variables independently of STZ, although STZ + VE rats tended to gain more weight than those with STZ alone (not significant). Body weight and blood glucose were similar among all of the groups before STZ injection.

Cardiac chemoreflex responses in diabetic rats with intact cardiac innervation. Figure 1, A and D, illustrates representative tracings of the changes in RSNA in sham and STZ rats with intact cardiac innervation, respectively, in response to application of Caps (10 µg/10 µl) onto the anterior surface of the left ventricle. In sham rats, both Caps and BK (data not shown) caused a biphasic response: initial inhibition of RSNA and hypotension for 3-5 s followed by a more prolonged activation and hypertension. Changes in HR were small and inconsistent. Cardiac sympathetic denervation and vagotomy studies confirmed that the initial sympathoinhibitory phase was mediated by cardiac vagal afferents and the secondary sympathoexcitatory phase was mediated by cardiac sympathetic afferents (results described below).

View larger version (55K): [in this window] [in a new window]   Fig. 1.   Chart recordings illustrating changes in renal sympathetic nerve activity (RSNA), arterial blood pressure (ABP), and heart rate in response to epicardial application of capsaicin (Caps) to the left ventricle in sham and streptozotocin (STZ)-induced diabetic rats with intact cardiac innervation and those after either vagal (VD) or cardiac sympathetic denervation (SD). A: sham rat with intact cardiac innervation. B: sham rat after VD. C: sham rat after SD. D: diabetic rat with intact cardiac innervation. E: diabetic rat after VD. F: diabetic rat after SD. Recordings after VD and cardiac SD were obtained from separate animals.

Cardiac vagal afferent chemoreflex in diabetic rats. In this series, the cardiac sympathetic nerves were cut to confine reflex responses solely to activation of cardiac vagal afferent fibers (Bezold-Jarisch reflex). Transection of cardiac sympathetic afferent fibers did not affect baseline RSNA in either sham or STZ rats.

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Changes in peak RSNA in response to epicardial application of either Caps (A and B) or bradykinin (BK; C and D) in sham and STZ-induced diabetic rats before (open symbols) and after (closed symbols) cardiac SD. A and C: initial sympathoinhibitory phase of biphasic response as illustrated in Fig. 1. B and D: secondary sympathoexcitatory phase of the response. *P < 0.05, sham vs. STZ; P < 0.05, intact vs. SD.

Cardiac sympathetic afferent chemoreflex in diabetic rats. In this series, the vagus nerves were cut to confine reflex responses solely to activation of cardiac sympathetic afferent fibers. Before vagal denervation, the sympathoexcitatory phase of the RSNA response to the low dose of Caps (1 µg) was depressed in STZ rats, but the RSNA response to the higher doses of Caps (5 and 10 µg) (Fig. 3B) in STZ rats was similar to that of sham animals. The sympathoexcitatory phase of the RSNA response to BK was significantly depressed at all doses in STZ rats (Fig. 3D).

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Changes in peak RSNA in response to epicardial application of either Caps (A and B) or BK (C and D) in sham and STZ-induced diabetic rats before (open symbols) and after (closed symbols) VD. A and C: initial sympathoinhibitory phase of biphasic response as illustrated in Fig. 1. B and D: secondary sympathoexcitatory phase of the response. *P < 0.05, sham vs. STZ; P < 0.05, intact vs. VD.

Cardiac chemoreflex responses in diabetic rats treated with VE. Baseline RSNA in sham + VE and STZ + VE rats did not differ from its respective controls (sham, STZ rats) and was not affected by cardiac sympathetic denervation. In cardiac sympathectomized rats, reflex decreases in RSNA, MABP, and HR in response to both epicardial Caps and BK were significantly greater in STZ + VE rats compared with STZ rats and were not different from reflex decreases observed in either sham or sham + VE rats (Fig. 4 and Table 4). Chronic VE had no effect on reflex responses in sham rats (sham + VE vs. sham rats: Fig. 4 and Table 4). These results indicate that chronic VE treatment prevented the impairment of the vagal afferent cardiac chemoreflex in diabetic rats.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Effect of vitamin E treatment (VE) on the reflex inhibition of RSNA in response to epicardial application of Caps (left) and BK (right) in sham and STZ-induced diabetic rats with cardiac SD (cardiac vagal afferent reflex). *P < 0.05 compared with sham.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Effect of VE on the reflex activation of RSNA in response to epicardial application of Caps (left) and BK (right) in sham and STZ-induced diabetic rats with VD (cardiac sympathetic afferent reflex). *P < 0.05 compared with sham.

Cardiac vagal afferent activity in diabetic rats. We studied impulse activity from single vagal fibers arising from chemosensitive endings in the heart in response to epicardial application of Caps and BK. Examples of the cardiac vagal afferent recordings are illustrated in Fig. 6. All of the fibers had receptive fields in the left ventricle. Baseline activity of the vagal afferent fibers was not different in sham and STZ rats (0.6 ± 0.2 vs. 0.9 ± 0.6 imp/s, respectively; not significant). Cardiac vagal afferent responses to Caps and BK were depressed in STZ rats compared with sham rats (Fig. 7). Vagal afferent responses to epicardial Caps and BK were significantly greater in STZ + VE rats compared with STZ rats and were not different from afferent responses observed in either sham or sham + VE rats (Fig. 7). These results indicate that chronic VE prevented the impairment of cardiac vagal afferent discharge to chemical stimulation in the diabetic state. Chronic VE had no effect on the responses of cardiac vagal afferents to Caps and BK in sham rats (sham + VE vs. sham rats: Fig. 7).

View larger version (38K): [in this window] [in a new window]   Fig. 6.   Recordings of impulse activity from vagal afferent fibers arising from the left ventricle in a sham rat, a STZ-treated rat, and a STZ + VE-treated rat in response to epicardial Caps (1 µg in 10 µl). Bars indicate the duration of Caps application. AP, action potential.

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Effect of VE treatment on the impulse frequency (IF) of cardiac vagal afferent fibers in response to epicardial application of Caps (left) and BK (right) in sham and STZ-induced diabetic rats. *P < 0.05 compared with sham.

Cardiac sympathetic afferent activity in diabetic rats. Fig. 8 illustrates responses of cardiac sympathetic afferent fibers to epicardial Caps and BK. All of the fibers had receptive fields in the left ventricle. Baseline activity of the sympathetic afferent fibers was not different in sham and STZ rats (2.1 ± 1.4 vs. 2.3 ± 0.8 imp/s; not significant). Cardiac sympathetic afferent responses to Caps and BK were depressed in STZ rats compared with sham rats (Fig. 9). Sympathetic afferent responses to both epicardial Caps and BK were significantly greater in STZ + VE rats compared with STZ rats and were not different from afferent responses observed in either sham or sham + VE rats (Fig. 9). Chronic VE had no effect on the responses of sympathetic afferents to Caps and BK in sham rats (sham + VE vs. sham rats: Fig. 9). These results indicate that chronic VE prevented the impairment of cardiac sympathetic afferent discharge to chemical stimulation in the diabetic state.

View larger version (39K): [in this window] [in a new window]   Fig. 8.   Recordings of impulse activity from cardiac sympathetic afferent fibers arising from the left ventricle in a sham rat, a STZ-treated rat, and a STZ + VE-treated rat in response to epicardial Caps (1 µg in 10 µl). Bars indicate the duration of Caps application.

View larger version (18K): [in this window] [in a new window]   Fig. 9.   Effect of VE treatment on the IF of cardiac sympathetic afferent fibers in response to epicardial application of Caps (left) and BK (right) in sham and STZ-induced diabetic rats. *P < 0.05 compared with sham.

Comparison of cardiac vagal and sympathetic afferent responses. As exemplified in Figs. 6 and 8, there were no differences between cardiac vagal and sympathetic afferents with respect to the latency, time to peak activity, and duration of the afferent responses to the chemical mediators in any of the experimental groups. In general, both populations of afferent endings were activated within a few seconds of application of the chemical, followed by a rapid crescendo in activity that peaked and waned during the period of exposure. Differences in afferent responses between sham and STZ rats were confined to the frequency of impulse activity (Figs. 6-9).

Afferent responses to veratridine. Figure 10 illustrates responses of cardiac vagal and sympathetic afferent fibers to epicardial veratridine. Unlike afferent responses to Caps and BK, afferent responses to veratridine were not altered in the STZ rats compared with sham rats.

View larger version (15K): [in this window] [in a new window]   Fig. 10.   IF of cardiac vagal afferent fibers (left) and cardiac sympathetic afferent fibers (right) in response to epicardial application of veratridine in sham and STZ-induced diabetic rats.

Lipid peroxide levels in plasma. Figure 11 illustrates the plasma levels of TBARS in each of the experimental groups. The level of lipid peroxides in STZ rats was significantly greater than in sham rats. Chronic treatment with VE had no significant effect on the level of TBARS in sham animals but prevented the accumulation of TBARS in diabetic animals.

View larger version (21K): [in this window] [in a new window]   Fig. 11.   Effects of STZ and VE treatments on the concentration of lipid peroxides in plasma measured as thiobarbituric acid-reactive substances (TBARS). *P < 0.05 compared with sham.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Results from the present study indicate that both the vagal- and sympathetic-afferent limbs of the cardiac chemoreflex are markedly depressed in STZ-induced diabetic rats. The results also indicate that this functional impairment in the diabetic state can be prevented by chronic treatment with VE. From these results, we believe that oxidative stress is the major underlying mechanism causing impairment of the chemosensitive and reflex properties of cardiac sensory nerve endings in diabetes.

It is well documented that diabetes is associated with increased oxidative stress, as evidenced by the increased accumulation of lipid peroxides in plasma of rats (10) and humans with diabetes mellitus (15). Oxidative stress is thought to be a major contributor to cardiovascular disease in diabetes mellitus (8). Treatment of diabetic animals with antioxidants diminishes indexes of the oxidative stress and improves endothelium-dependent relaxation of coronary vessels (21). A morphological study (21) has also shown that chronic treatment with VE largely prevents the degeneration of autonomic nerve fibers in the hearts of diabetic rats. However, to our knowledge, this is the first study to document changes in the functional properties of cardiac afferent nerve endings in diabetes and the important role of oxidative stress in these changes.

STZ is a glucose molecule with a highly reactive nitrosourea side chain. The mechanism of the cytotoxic action of the drug and its selectivity for pancreatic -cells is thought to involve binding to a glucose transporter on the plasma membrane of the -cell and subsequent methylation, free radical generation, and nitric oxide production (20). It is unlikely that STZ could have exerted a toxic influence directly on cardiac vagal afferent endings. We found that cardiac chemoreflex function was not impaired in STZ rats with insulin implants to maintain normoglycemia. This result links cardiac afferent dysfunction to insulin depletion and hyperglycemia rather than to the drug itself. Hyperglycemia-driven metabolic changes in diabetes increase generation of free radicals through several mechanisms, including glucose autoxidation, protein glycation, increased substrate flux through the polyol pathway, and depression of natural antioxidant systems (5). One or more of these potential mechanisms is likely to be involved in the impairment of cardiac afferent function in the diabetic rats.

In the present study we administered VE before STZ injection to minimize oxidative stress from the onset of hyperglycemia. Because the cytotoxic effect of STZ is thought to involve free radical generation, it is possible that the efficacy of VE could be related to a potential protective effect of the antioxidant on pancreatic -cells to prevent STZ-induced cell death. Although we did not assess -cell viability or plasma insulin levels directly in our study, we found that STZ rats treated with VE became hyperglycemic to the same degree as rats treated with STZ alone and exhibited similar signs of diabetes, such as polydipsia, polyuria, and impaired weight gain. Furthermore, in other unpublished studies (4), we have found that administration of VE several weeks after the induction of diabetes also is effective in reversing the impairment in cardiac chemoreflex function.

An important observation from the present study is that epicardial application of mediator substances, such as Caps and BK, to the heart of normal, anesthetized rats evokes a biphasic sympathetic reflex (with respect to RSNA) consisting of an initial, short-lasting, sympathoinhibitory phase mediated by cardiac vagal afferents and a secondary, longer-lasting, sympathoexcitatory phase mediated by cardiac sympathetic afferents. The afferent pathways and reflex effects of these two opposing reflexes from the heart have been recognized for many years (6, 11-12, 22). However, epicardial application of chemical mediators in other species, such as the dog or cat, generally evokes a prominent sympathetic afferent reflex with little or no opposing input from cardiac vagal afferents (11). On the basis of these observations, it is generally assumed that cardiac sympathetic afferent endings are located primarily in the epicardial regions of the heart, whereas the vagal afferent endings more extensively innervate the subendocardial regions, not accessible to epicardial applications.

In the present study in rats, epicardial application of these test chemicals was capable of stimulating both populations of afferents to a degree sufficient to cause mutual antagonism of the reflex renal sympathetic nerve responses. This mutual antagonism was evidenced by an enhancement of the cardiac sympathetic afferent reflex after vagotomy and similar enhancement of the cardiac vagal afferent reflex after cardiac sympathectomy. The biphasic nature of the integrated reflex response appeared to be mediated principally by differences in the latency and duration of central integration of the two reflex pathways. We could find no evidence of distinct differences in the two pathways with respect to the dynamics of the afferent discharge to the chemical mediators. Our results may suggest that cardiac sympathetic and vagal afferent endings are distributed more uniformly in the epicardium of the rat heart than those of larger species. It is also possible that the small size of the rat heart allows chemicals to diffuse rapidly to endings below the epicardial surface, allowing more uniform activation of vagal and sympathetic afferent endings.

In diabetic rats, we found that both the sympathoinhibitory and sympathoexcitatory phases of the integrated cardiac chemoreflex response were blunted. This effect could conceivably occur through an enhancement of the opposing cardiac chemoreflexes, resulting in greater mutual occlusion of the reflex pathways. On the contrary, we found that there was a significant impairment in both the cardiac vagal and sympathetic afferent chemoreflexes, as revealed after cardiac sympathetic denervation and vagotomy, respectively. We also found that in sham rats, vagotomy increased resting RSNA in addition to potentiating sympathetic activation caused by chemical stimulation of cardiac sympathetic afferents. However, in diabetic rats, vagotomy did not affect resting RSNA or potentiate the sympathetic afferent reflex. These results correlate with our observation that resting RSNA appeared to be elevated in the diabetic state. Furthermore, chronic VE treatment prevented this impairment of vagal restraint of sympathetic outflow in diabetic rats. These data suggest that oxidative stress in diabetes impairs vagal influences that limit tonic sympathetic activity. This observation may have important implications for the study of heart rate variability and the pathogenesis and treatment of arrhythmias in diabetic patients.

Impairment of both the vagal- and sympathetic afferent limbs of the cardiac chemoreflex in diabetic animals may be the result of oxidative injury at any of the many levels of the neural reflex arc: afferent, central, and efferent. Because cardiac chemosensitive afferents are stimulated acutely by oxygen radicals (9, 23), chronic exposure to oxygen radicals may desensitize or injure the afferent endings and diminish the reflexes that arise from their activation. We tested this possibility by directly recording action potentials from cardiac vagal and sympathetic chemosensitive afferents in response to stimulation by Caps, BK, and veratridine. Our data demonstrate that the responses of both cardiac vagal and sympathetic afferent fibers to Caps and BK were depressed in diabetic rats. Chronic treatment with VE completely abolished the detrimental effect of diabetes on the chemosensory properties of the cardiac afferent nerves. At the same time, VE prevented the accumulation of free radical oxidation products, lipid peroxides, in the plasma of the diabetic animals. This defect of the sensory properties of the cardiac afferent nerve endings appears to be mediated by oxidative stress and provides a primary explanation for the depressed cardiac chemoreflexes in our diabetic animals.

One aspect of our afferent and reflex results that is discrepant is the correlation between resting RSNA and resting cardiac chemosensitive afferent activity in sham and diabetic rats. Baseline nerve activities from both cardiac vagal and sympathetic afferent fibers did not differ between sham and diabetic rats, whereas resting RSNA was elevated and vagal restraint of resting sympathetic outflow was blunted in diabetic animals. The functional significance of this paradox cannot be resolved in the present study, but several possible explanations can be considered. An impairment in central integration of the vagal afferent limb of the cardiac chemoreflex may also occur in the diabetic state, causing a reduction in tonic restraint of sympathetic outflow in the face of normal resting afferent activity. In addition, other vagal reflexes that may play a comparatively larger role in the tonic control of sympathetic outflow also may be impaired in the diabetic state. It is well documented that the neural component of the volume reflex is blunted in STZ diabetic rats (16). It is possible that cardiopulmonary mechanoreceptors are similarly blunted in the diabetic state and that a reduction in afferent input from this vagal reflex pathway can account for the alterations in resting renal sympathetic outflow observed in diabetic rats (17).

We found that the responses of cardiac vagal and sympathetic afferent chemosensitive endings to veratridine, unlike Caps or BK, were not attenuated in diabetic rats. This observation may provide some insight into the mechanism of the oxidative impairment of the afferent endings. Most chemical mediators of chemosensitive endings, including Caps and BK, evoke depolarization of the afferent neuron by binding to membrane receptors or ligand-gated ion channels in the nerve terminal, leading to depolarization (18). Unlike Caps and BK, veratridine bypasses the receptor-mediated complex and evokes action potentials by direct activation of the voltage-gated Na⁺ channel (2) at the level of the spike-initiating zone. As we have shown, the impairment of cardiac vagal and sympathetic afferent discharge in diabetic rats did not extend to the general ability of the afferent fibers to propagate action potentials, because veratridine responses were normal. Thus the impairment in cardiac afferent responses to Caps and BK in the diabetic rats appears to involve a dysfunction in the chemical transduction process before spike initiation. However, we cannot deduce from the present study whether the duration or severity of the diabetic state can influence the degree and mechanism(s) of oxidative impairment of afferent function. It seems reasonable to suggest that a more prolonged exposure to oxidative injury may lead to axonal injury and more generalized neuropathy.

Chemosensitive nerve endings in the heart play an important role in mediating cardiovascular adjustments to myocardial ischemia and the perception of cardiac pain (6, 11-12, 22-25). The conclusions of this study are that cardiac chemoreflexes are markedly depressed in diabetic rats and, correspondingly, that chronic antioxidant treatment can effectively prevent this impairment. Our results show that cardiac chemosensory nerves are desensitized in diabetes as a result of a dysfunction in the chemical transduction process in the sensory terminal. The efficacy of VE in preventing this afferent dysfunction is consistent with a mechanism mediated by free radicals. An impairment in cardiac chemosensory function may play an important role in exacerbating diabetic cardiac complications, such as silent ischemia and arrhythmias. In addition, these results support evidence that chronic antioxidant therapy may be effective in reducing the risk of these complications in diabetic patients. Given the uncertainties in extrapolation of our results from diabetic rats to the human condition, further studies are needed to assess their clinical implications.