INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ARTERIAL BARORECEPTORS (BRs) convey systemic hemodynamic information to the central nervous system (CNS), where it is integrated and evokes reflex responses responsible for moment-to-moment cardiovascular regulation. Two major classes of BRs exist based on axonal conduction velocity: those connected to myelinated (A)-type and unmyelinated (C)-type axons. As with somatic sensory neurons, these two classes of BR have very different dynamic sensory discharge characteristics. A-type BRs generally have higher mean discharge rates, lower pressure thresholds, and higher sensitivities (24, 36). Largely on the basis of electrical stimulation studies (1, 2, 10-12, 15, 16, 21, 28), these two classes of BR have long been suggested to evoke very different baroreflex-response relationships separated by activation frequencies. Maximal electrical activation of all BRs evokes substantial reflex decreases in mean arterial pressure (MAP) or sympathetic efferent activity at frequencies as low as 1 or 2 Hz (1, 2, 10-12, 15, 16, 21, 28). However, selective activation of A-type BRs alone is consistently reported to require higher frequencies (10-fold) for reflex responses equivalent to the activation of all aortic depressor nerve (ADN) axons (A+C; e.g., see Ref. 15). Despite extensive experiments, the mechanism responsible remains unclear. At least two relatively simple hypotheses can be suggested. The first arises from the fact that the overwhelming majority (as much as 90%) of all BR (32) axons in the rat ADN is unmyelinated (3). Thus the low-frequency reflex efficacy of C-type BRs could be due simply to their overwhelming numerical predominance in the total BR neuron population. An alternative is that C-type synaptic input to the CNS, despite its relatively lower mean frequency, could evoke proportionately more powerful responses than A-type BR inputs centrally.

The focus of the present studies was to evaluate important aspects of these two hypotheses. As a first step, we studied responses to electrical stimuli to the ADN sufficient to activate both A- and C-type BR inputs (16, 31) and compared these to responses in which transmission to the CNS of A-type BR inputs was selectively and reversibly blocked by anodal current (I_an). Our studies are the first to report successful use of anodal block to isolate the reflex contribution of C-type BRs of ADN in the rat, a species with an extensive experimental base of BR and baroreflex studies. We then tested graded levels of activation of A- and C-type BRs to assess responsiveness to activation of portions of the BR population within BR subtypes with the use of baroreflex frequency-response relationships. Our findings suggest a differential integration of A- and C-type BR information and support the alternative hypothesis, the concept that C-type information has proportionately greater reflex impact than A-type information. Additionally, only baroreflex responses to A-type BR activation were found to show dynamic enhancement to temporally patterned stimulus trains. Thus our data suggest that the integration of A- and C-type BR input differs importantly in the regulation of heart rate (HR) and systemic arterial pressure. Such differences may indicate pathway-specific differentiation of BR subtype sensory processing. Overall, these studies indicate an interesting match in absolute frequency ranges for each BR subtype between sensory encoding and their respective reflex characteristics.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experiments were conducted in accordance with protocols approved by the University Animal Care and Use Committee with adult male Sprague-Dawley rats (B & K, Kent, WA; 250-450 g). Rats were anesthetized with a combination of urethan (800 mg/kg) and chloralose (80 mg/kg) with supplements of chloralose (10 mg/kg) and breathed spontaneously. At the end of experiments, rats were killed with an overdose of pentobarbital sodium (300 mg/kg iv).

Baroreflex responses. For reflex studies, the cervical portion of the left ADN was isolated, and the nerve was placed on stimulating electrodes (bipolar, Teflon-insulated Pt-Ir wires) and covered with a petroleum jelly-mineral oil mixture. All other nerves were intact. The stimulating electrodes were connected to a computer-controlled programmable stimulator (AMPI Master-8) through a stimulus isolation unit. The stimulus parameters selected for this study were based on direct observations of the ADN electroneurogram (ENG) in companion studies in which intensity-action potential recruitment was measured (15). From the range of intensities studied in that series we selected two stimulation intensities, a low and a high level for most detailed study in the present work with ADN stimulation. Within each such reflex experiment, these two levels were held constant: low voltage (1.5-3 V) and high, supramaximal activation voltages (18-20 V). All shocks were 0.1 ms in pulse duration, long enough to activate C-type axons and short enough to avoid repetitive A-type axon activation within a single pulse (31). The high-voltage level was sufficient to activate all fiber types in the ADN (e.g., Ref. 15), whereas low-voltage stimuli evoked only A fiber volleys. We used a functional test to set the low intensity level in each experiment. In this procedure, intensity was increased until it reached a level at which a test train at 100 Hz elicited marked depressor responses, but trains at 1-2 Hz evoked no measurable changes in blood pressure. This level corresponded to an intensity level in the ENG experiments that evoked no C volley and a maximal A volley and assured a functional normalization of stimulus intensity across experiments. In a limited series of additional experiments, graded stimulation intensities were used to activate portions of the A- and C-type population of axons. In these experiments, intensities <3 V activated variable portions of the A fiber ADN axon population, and 6 and 8 V activated all A-type axons and only small portions of the C-type population based on averages for ADN-ENGs (15).

For determination of the frequency-response relationship, stimulus trains lasting 5 s were tested for selected fixed frequencies between 1 and 200 Hz: 1, 2, 5, 10, 20, 50, 100, and 200 Hz. Such 5-s trains made these relatively complex stimulus protocols of repeated brief-stimulus trains with and without I_an more pragmatic and thus more feasible. The resulting stimulus-frequency vs. peak-response relationships for these 5-s trains were quite similar to the relationships constructed for responses measured in the last 10 s of a 60-s train in previous studies (15). The order of application of frequencies was random. Stimulus trains were given 3-5 min apart, which was sufficient time to allow full recovery. MAP signals from the pressure processor were displayed on the pen recorder and digitized at 20 Hz for offline analysis.

Patterned stimulation Intermittent and constant modes of stimulation were tested. MAP responses to bursts of stimuli were compared with those evoked by constant stimulation of ADN at equal, 1-s-average frequencies (11, 12, 15, 16). Burst-mode stimulation consisted of fixed 1-s cycles in which a 250-ms period of stimulation (40-120 Hz) was followed by 750 ms of no stimulation. These cycles of stimulation were repeated for 60-s-total periods. To directly compare these intermittent or burst-mode stimulus patterns with constant patterns, stimulation was switched without pause between constant frequency to burst mode with an equal number of shocks delivered in each mode on a 1- to 1-s basis. The instantaneous frequencies used for mode comparisons were constant frequencies of 10, 20, and 30 Hz paired with burst frequencies of 40, 80, and 120 Hz. Thus the burst modes contained a fourfold higher absolute frequency delivered for only one-fourth of the 1-s cycle. Stimuli were switched between burst and constant mode twice for a total of 4 min of stimulation in each test. The order of the mode that presented first was randomized. All patterns were implemented via the programmable stimulator by downloading the protocols from files on the computer and triggered by a control pulse to the stimulator.

Anodal nerve block. To assess the reflex responses to activation of C fiber BRs alone, anodal block of A fiber conduction was used. Test stimuli were delivered to the ADN for 5 s at the fixed frequencies described in Patterned stimulation so that full frequency-response relationship could be constructed in each experiment. Anodal block of A-type nerve conduction was produced by a steady, direct current I_an applied to a portion of the ADN several millimeters central to the section being stimulated for reflex responses. I_an was delivered for 10 s, and its intensity (12-100 µA) was adjusted in each experiment to completely block the reflex response to a near-maximal A fiber stimulus (5-s train of 3-V stimuli at 100 Hz). Twenty percent increases above the minimal level required to effectively block responses to the low-intensity, high-frequency stimuli did not impair responses to low-frequency, high-intensity stimuli. Similar degrees of separation, "safety factors," were reported for anodal block in other species (27, 35).

It was not possible to directly assess A fiber nerve block by recording ADN-ENG waves, stimulating the nerve, and delivering I_an in these slender and fragile nerve trunks. We, as others did (35), based our assessment of anodal-blockade efficacy on indirect means. In our case, we used 1- to 2-V, 100-Hz stimuli for reflex tests that we previously showed to activate only A fiber axons (15). Stimuli for our functional A fiber tests were timed to start 3 s after the beginning of the 10-s blocking-current period and thus occurred in the middle of the anodal-block period. For these blocking-current adjustments, a test stimulus train was delivered before, during, and after the blocking current in successive trials. In all cases, the A fiber reflex response had both to be completely blocked by I_an and to completely recover after the anodal block for an experimental data set to be included in the summary analysis. This assured that the anodal block did not irreversibly damage A fibers. The fact that low-frequency (1-2 Hz) reflex responses before and during block were similar further suggests that I_an did not damage C-type axons either. In each case, once the adequate I_an level had been determined, the high-intensity reflex 5-s tests at the various fixed-stimulation frequencies were tested. In another series of experiments, more prolonged anodal blocking periods were used to determine the effects of intermittent burst-stimulation patterns. In these experiments, the reflex-test volleys lasted 1 min each and included switching from constant to burst mode twice for a total of 4 min of reflex-test stimulation. This required much more prolonged anodal blocks, but, in all cases, to be accepted for analysis, complete reversibility of the block was required as tested with the low-intensity, high-frequency test.

Analysis. To construct reflex relationships, MAP responses were calculated as changes in MAP (MAP) relative to the prestimulation value for each test and expressed as a percentage of that prestimulation value (MAP%). The control value was the average of MAP values sampled (20 Hz analog/digital rate) for 10 s preceding the stimulus. HR values were likewise determined from the cardiac beat intervals over identical periods and otherwise processed in a similar manner to the MAP values. Thus HR responses were calculated as changes relative to the prestimulation value for each test (HR%). For patterned trains (MAP only), the final 10 s of a particular stimulation pattern were averaged for a mean response and expressed as a change from control MAP. Across experiments, average reflex-response relationships were plotted as the mean response against stimulation frequency for low- and high-voltage stimuli and in the presence of I_an. Slopes (or gains) of baroreflex relationships were measured as least-squares linear regression fits to the submaximal portions of the log-frequency vs. reflex-response relationships, and a reflex "threshold" frequency was approximated by where these fits extrapolated an x-axis crossing. Comparisons were made using analysis of variance and, in some cases, post hoc, pairwise comparisons were done using Scheffé's test (P < 0.05 was considered significant).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Low-intensity electrical stimulation selectively activates a volley of A-type BR action potentials in the ADN ENG (15) as in all other peripheral nerves with myelinated axons, including the vagus (16, 36). Resting, prestimulation MAP averaged 102.0 ± 4.9 mmHg, and HR averaged 320.4 ± 4.5 beats/min (n = 34). Brief trains (5 s) of these low-intensity, A-type selective stimuli to the ADN evoked no reflex-MAP responses until the stimulus frequency exceeded 10 Hz (Fig. 1B). At intensities sufficient to activate A+C, stimulation frequencies as low as 1 Hz evoked substantial decreases in MAP (Fig. 1B). Thus, during combined activation of both A- and C-type axons, low-frequency BR action potential volleys are sufficient to activate MAP baroreflex responses, whereas these same low frequencies of activation of A-type BRs alone are not.

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Reflex mean arterial pressure (MAP) responses to electrical activation of myelinated (A), unmyelinated (C), or both types of baroreceptors (BR) in aortic depressor nerve (ADN). Example of responses to anodal block (A) of low-intensity (3 V), high-frequency (100 Hz) stimulation (). Low-intensity stimuli delivered for 5 s evoked reflex decreases in MAP (top trace, A) that were blocked by continuous anodal current (10 s, horizontal bar, middle trace, A). Reflex responses promptly recovered after anodal current was turned off (bottom trace, A). Resting control MAP was 92 mmHg. Average (n = 6) relationship between peak normalized baroreflex response (MAP%) and stimulus frequency (B) were constructed for low intensity (3 V), high intensity (18 V), and high intensity during anodal block. Note that anodal block reduced reflex responses only at relatively high frequencies of stimulation (>10 Hz). Points are normalized means ± SE. *Significant difference between high intensity and high intensity plus anodal block (P < 0.00001).

To isolate the reflex-MAP responses to C-type BR activation alone, we repeated our test-stimulus frequency-response series in the presence of I_an (Fig. 1A). Application of continuous I_an to the ADN completely and reversibly blocked MAP reflex responses to low-intensity, high-frequency test stimulation (see Fig. 1A). Such stimulus intensities activate A-type BRs only (15, 16). Note that I_an alone on this centrally connected but peripherally cut ADN had no effect on MAP (Fig. 1A; preceding ADN stimulation). In each experiment, we used no more than 20% greater than the minimum I_an required to functionally block the MAP responses to high-frequency, A fiber activation of ADN (Fig. 1A). During anodal block, low frequencies of high-intensity stimulation evoked MAP reflex responses that were not different from preblock levels at frequencies between 1 and 10 Hz (Fig. 1B). These C-type selective baroreflex MAP responses did not increase further for the highest frequencies (i.e., between 5 and 200 Hz; n = 6, P > 0.69). Overall, C-selective maximal MAP responses between 50 and 200 Hz (Fig. 1B) were significantly smaller than preblock responses (C alone, 24.1 ± 1.1 mmHg; A+C, 35.3 ± 0.7 mmHg, P < 0.00001) but were equivalent to A-selective stimulation (25.6 ± 1.4 mmHg, P = 0.24). One-half-maximal activation of the C fiber BR reflex MAP responses occurred at ~2 Hz, whereas the A fiber component occurred at somewhat >20 Hz. Such results suggest that in the A+C responses, the C-type BRs are required for the low-frequency (<10 Hz) baroreflex MAP responses and that the reflex contribution of A-type BRs is significant only at high frequencies (10 Hz).

Because A- and C-type BRs physiologically conduct such different action potential discharge patterns during pressure stimulation, we tested whether intermittent patterns of activation of BR subtypes might be more effective in evoking reflex-MAP responses. Experiments were designed so that constant numbers of stimulus shocks were delivered each second. For intermittent or burst patterns, this meant that in this protocol all stimuli were delivered in the first 250 ms of a 1-s cycle. Thus only the temporal distribution of the stimulus shocks was changed between burst and constant modes of stimulation and not the total number of shocks in a test or on a second-to-second basis.

Low-intensity, high-frequency (A selective) stimuli evoked brisk reflex decreases in MAP (Fig. 2A). Switching without delay from constant stimulation to bursts of intermittent stimuli reproducibly evoked clear increments in reflex-MAP responses in all animals tested (Fig. 2A). Thus at a 20-Hz second-to-second average stimulus frequency, grouping those A-type activating stimuli into 80-Hz bursts for one-fourth of the time substantially augmented the reflex-depressor responses. These augmented reflex-MAP responses quickly disappeared on switching back to constant-mode stimulation. On average, switching to this 80-Hz burst mode while presumably stimulating only A-type BRs increased MAP responses by ~35% (n = 6, P = 0.003; Fig. 2A). During high-intensity A+C-type activation in these same animals (Fig. 2B), minimal MAP changes were apparent on switching to burst patterns and average reflex responses were equivalent when fully activating both A- and C-type axons (n = 6, P = 0.185). The degree of augmentation of the A-type BR reflex response on switching from constant- to burst-patterned stimulation was relatively constant across the linear, submaximal response range for fixed frequencies of 10, 20, and 30 Hz, which used 40, 80, and 120 Hz in burst mode, respectively (n = 6, P < 0.01; Fig. 3). Thus baroreflex responses to A-type BR inputs are clearly very sensitive to the temporal distribution of action potentials in their input trains, and efferent mechanisms were clearly not saturated in these frequency ranges. This augmentation of A-type BR-reflex MAP responses is obscured, however, when all C-type BR axons are simultaneously activated.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Influence of burst and constant modes of ADN-BR stimulation on reflex changes in MAP. A: baroreflex MAP responses to low-intensity stimulation (which selectively activates A-type axons) were augmented by switching from constant, fixed-frequency pattern of stimulation to bursts of intermittent stimuli. Number of stimulus shocks delivered each second was constant both for bursts and constant patterns. In this example (top), switching from a constant frequency of 20 Hz to a burst mode consisting of 80-Hz stimulation for 250 ms followed by no stimuli for 750 ms resulted in larger reflex decreases in MAP during burst stimulation. All mode stimulation intervals were 1 min in duration and cycles within burst mode were 1 s. On average (n = 6; bottom), normalized responses MAP% for low intensity tests were significantly greater (**P = 0.003) during burst-mode stimulation. Points are means ± SE. B: at high intensities of stimulation (which activates all BRs in ADN), switching from constant stimulation to bursts of intermittent stimuli evoked similar MAP responses. A trace (top) from same animal as in A, but using high-intensity stimuli, shows little change in reflex response after switching from a constant frequency of 20 Hz to a burst mode that consisted of 80-Hz stimulation for 250 ms followed by no stimuli for 750 ms. All mode-stimulation intervals were 1 min in duration, and cycles for burst mode were 1 s. On average (n = 6), MAP% for high-intensity tests were not different (P = 0.185) during burst mode stimulation. Points are means ± SE.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Average normalized reflex responses for burst and constant stimulation of only myelinated ADN axons. Increases in mean frequency for low-intensity stimuli in constant mode resulted in successively larger reflex decreases in MAP. Switching to burst mode of stimulation resulted in a significant enhancement (**P < 0.01, n = 6) of these normalized reflex responses (MAP%) at each frequency tested. At each frequency, average total number of shocks per second was held constant. Note that 10, 20, and 30 Hz are frequencies on maximum sensitivity region of frequency-response relationship for A-type BRs (see Fig. 1). Points are means ± SE.

To isolate C-type baroreflex responses to patterned stimulation in the absence of A-type BR activation, anodal blocking currents were used. Prolonged periods (4.5 min) of I_an application were required for completion of repeated trials of constant- and burst-mode stimulation. In all cases, we bracketed these anodal protocols with tests of A-type MAP baroreflexes to assure that the anodal blocks were reversible and no nerve damage had occurred. Any experiments in which reflex responses to low-intensity, high-frequency stimulation were not equivalent to preanodal tests (see Fig. 1A) were discarded. Brief trains of presumed A-type selective stimuli (low intensity, high frequency) evoked substantial, reproducible depressor responses that were reversibly eliminated by I_an (Fig. 4A). When the same level of I_an was applied, high-intensity trains evoked large depressor responses that were not altered by switching between burst- and constant-mode patterns (Fig. 4B). Reversibility of the anodal block was indicated by the return of the low-intensity, high-frequency baroreflex MAP response (Fig. 4). This lack of augmentation when switching to phasic stimulation was true for a wide range of activation intensities (Fig. 5). Note in these experiments that test burst-stimulus patterns (Fig. 5) that the burst frequency was fourfold higher than the constant stimuli and that these absolute frequencies were on the segment of the maximum slope of the subtype frequency-response relationship (1-8 Hz for C type and 10-100 Hz for A type). Thus, unlike for A-type sensory inputs, baroreflex MAP responses to C-type BR inputs appear to be insensitive to the temporal patterning of action potential input trains.

View larger version (30K): [in this window] [in a new window]   Fig. 4.   Influence of stimulus patterning on baroreflex MAP responses to selective, maximal C-type BR activation in a single, representative experiment. Anodal current level reversibly blocked reflex response to a low-voltage, high-frequency (A fiber) test before (A), during (B), and after (C) prolonged anodal blocks. In B, stimulus intensity was increased (20 V) to activate all BR axons during application of anodal current. During anodal current, switching from constant fixed-frequency (C20; 20 Hz) to an intermittent burst pattern (B80; 80 Hz for 250 ms followed by no stimuli for 750 ms) produced no discernible change in MAP response. Each mode was tested for 1 min before switching. Note that 20 Hz is at maximum of frequency-response relationship (see Fig. 1) for C-type BRs.

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Influence of stimulus patterning on baroreflex MAP responses to graded, selective C-type BR activation in a single, representative experiment. Anodal current level reversibly blocked reflex response to a low-voltage, high-frequency test before and after prolonged anodal blocks as in Fig. 4 (results not shown). Constant fixed frequency (C2; 2 Hz) within high sensitivity range for C-type BR reflex responses (Fig. 1) was tested at graded, submaximal stimulus intensities [6 (A) and 8 V (B)] and compared with burst-mode stimulation (B8; 8 Hz for 250 ms followed by no stimuli for 750 ms). As with supramaximal intensity [20 V (C)], switching from constant to burst did not discernibly alter reflex response.

Both the low- and high-intensity electrical stimuli evoke synchronous and nearly maximal activation of axons (A and A+C type, respectively) as evidenced in the respective volleys recorded in the ADN EDG (15). To test partial recruitment within BR fiber types, we tested graded intensities in the selective A-type range (0.5-3 V, Fig. 6) as well as graded higher intensities (6 and 8 V) during anodal block to recruit variable proportions of the C-type population of ADN fibers (Fig. 7). Five-second stimulus trains were used in this anodal block protocol. At the lowest intensities (A selective), substantially higher frequencies (up to 50 Hz at 0.5 V) were required to elicit a significant reflex-depressor response (10 Hz, Fig. 6). As the intensity was increased within the A-selective range, however, there was a roughly parallel shift in the log frequency-response relationship to lower frequencies for the MAP baroreflex as a successively greater population of A-type BRs was recruited. Thus there was no significant change in the slope of these A-selective baroreflex MAP-response relationships (P > 0.05), only a shift in the intercept or threshold.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Frequency-response relationships for average normalized reflex decreases in MAP to graded, selective activation of A-type BRs. Intensities ranging from 0.5 to 3 V recruited greater BR reflex MAP responses at any given frequency. Increasing stimulus intensity did not change slopes of frequency-response relationships (P > 0.05) but appeared to shift threshold to lower frequencies of stimulation. MAP% points are means ± SE (n = 7 for 3 and 2 V; n = 4 for 1 V; n = 3 for 0.8 V; n = 2 for 0.5 V). *Significant changes in MAP (from baseline).

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Frequency-response relationships for normalized reflex decreases in MAP to graded activation of C-type BRs. Anodal block to select C-type axons lowered reflex response at any frequency and decreased slope of frequency-response relationships (C only) compared with A+C responses. Increasing stimulus intensity did not change (P > 0.05) threshold frequency of stimulation. MAP% points are means ± SE [n = 5 for 6 V (A); n = 6 for 8 V (B)]. All test points >1 Hz are significant changes in MAP (from baseline).

At 6 V (A+C responses) even 1- to 2-Hz stimulus trains evoked significant depressor responses that increased with increasing frequency up to ~10-20 Hz (Fig. 7). During anodal block, all frequency-response relationships for C-selective stimuli were depressed compared with A+C-preblock relationships (P < 0.03, n = 5; Fig. 7). At 6 V, significant depressor responses were evoked by 2-Hz trains. These are well below the minimum frequencies required for the reflex-MAP responses evoked by activation of A-type BR axons alone. Unlike for fully maximal stimuli (Fig. 1), anodal block significantly reduced the slope of the reflex-MAP responses (P < 0.03, n = 5) for submaximal intensities, which recruit only part of the total number of C-type axons but are still maximal for A-type ADN axons (see Figs. 1 and 2 in Ref. 15). Thus activating all A-type BR axons required substantially greater input frequencies to evoke a similar magnitude baroreflex depressor response than small fractions (<20%) of the C-type population of BRs. Such results make it unlikely that the greater MAP baroreflex efficacy of C-type BR inputs is simply due to the number of BR axons activated.

Another prominent A- and C-type difference between the baroreflex responses to BR activation is evident in the control of HR. For the 60-s, sustained-stimulation protocols (Fig. 8), A-selective trains were remarkably ineffective in producing bradycardia even at the highest frequencies of activation (200 Hz). With high intensities (20 V, A+C), however, increases in stimulus frequency progressively augmented the reflex bradycardia. Responses to maximal (20 V) C-selective activation of ADN BRs during anodal block overlaid the high-intensity A+C relationship from 1 to 10 Hz. Above 10 Hz, maximal C-selective responses evoked a plateau in HR response extending from 10 to 200 Hz. This maximum bradycardia (Fig. 8) reached during C-selective stimulation was much greater than the A-selective responses (P < 0.001) but only about one-half the full control (A+C) response level. Such observations are consistent with a supra-additive summation process and may indicate another basic difference in the organization of pathways for the baroreflex regulation of HR compared with MAP.

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Frequency-response relationships for reflex decreases in HR to near maximal activation of A- and C-type BRs. Maximal ADN activation at 20 V produced significant bradycardia responses at 2 Hz and maximal responses at 100-200 Hz (High A+C). Lower A-selective intensities (3 V; low) evoked only modest bradycardia at near-maximal frequencies. Anodal block to select C-type axons only decreased reflex-HR response at frequencies >10 Hz (High + Anodal  C). HR% points are means ± SE (n = 6). *Significant changes in MAP (from baseline).

This notion of pathway-dependent differences in A- and C-type central BR integration is reinforced by the experiments using graded C-type BR recruitment protocols with anodal block. With the use of these shorter train protocols, maximal and submaximal A+C-BR activation evoked qualitatively similar frequency-response relationships (compare Fig. 9 with Fig. 8). ENG studies of ADN (see Fig. 2 of Ref. 15) suggest that stimulus intensity can be decreased to a range that activates a reduced portion of C-type BRs while still activating all A-type BRs. Such intensities in the present studies evoked decreased reflex bradycardia at any given stimulus frequency (Fig. 9). Surprisingly, however, neither of these submaximal intensities (6 and 8 V) was sufficient to evoke significant reflex bradycardia when delivered as C-selective stimuli during anodal blockade. Note that although these C-selective stimulus trains did not evoke reflex bradycardia, they did produce very substantial reflex decreases in MAP in these same animals (Fig. 7). Under these stimulation protocols, the relative increment in C-type activation from 6 to 8 V appears to be amplified by the presence of A-type BR activity in the frequencies >10 Hz (Fig. 9). Thus the graded HR data are consistent with an important modulatory interaction between A- and C-type BRs in mediating enhanced, synergistic reflex responses in the cardiac pathways during combined A- and C-type BR inputs.

View larger version (20K): [in this window] [in a new window]   Fig. 9.   Frequency-response relationships for reflex decreases in HR to graded, selective activation of C-type BRs. Maximal ADN activation at 20 V without anodal block produced significant bradycardia at 5 Hz (A+C). Lower intensities (6 or 8 V) evoked smaller reflex HR responses at each frequency. Anodal block to select C-type axons blocked reflex bradycardia at any frequency. HR% points are means ± SE (without anodal block: n = 6 for 20 and 8 V; n = 5 for 6 V; with anodal block: n = 5 for 6 and 8 V). Baseline MAP = 96.1 ± 1.5 mmHg and HR = 352 ± 14.9 beats/min, n = 6. *Significant changes in MAP (from baseline).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Simultaneous activation of A- and C-type BRs evoked reflex decreases in blood pressure at much lower stimulation frequencies than with activation of A-type BRs alone (15). This finding is very similar to baroreflex responses previously reported for HR, blood pressure, and sympathetic activity to ADN activation in rabbit (1, 12, 27) and in rat (15, 28, 31). Such studies alone, however, cannot resolve the independent roles of A- and C-type BRs in this frequency-dependent behavior of the baroreflex. The present studies were undertaken to allow a direct and independent comparison of A- to C-type BR reflex responses. The use of electrically graded ADN stimulation enabled us to activate graded portions of the population of BR axons within the ADN. We left all other "buffer" nerves intact so that compensatory responses will also impact the absolute magnitude of the measured reflex responses. Combining graded activation with anodal blockade of A fiber conduction allowed us to isolate C-type BR responses from contributions of coactivation of A-type BRs. Such systematic studies have never been done in the rat, and the rat is probably the most widely studied species for broad aspects of autonomic regulation.

Excitation of C-type BRs alone elicited substantial reflex depressor responses even at stimulation frequencies as low as 1 Hz (Fig. 1). The frequency-response relationships for the MAP reflexes to C-selective stimuli diverged from the A+C-MAP relationships only at frequencies similar to those required to elicit a reflex decrease in MAP with A-selective stimuli (>10 Hz). Because of their technical difficulty, previous anodal block studies with ADN have generally been limited to larger species such as the rabbit, which possesses an ADN analogous to that in the rat. In the rabbit, inhibition of renal sympathetic activity by A-type BR activation with low-stimulus intensities occurs only at frequencies >10 Hz and overall frequency-response relationships were similar to our rat-MAP relationships (2).

These fundamental differences in frequency-response baroreflex relationships across BR subtype are consistent with the hypothesis that there are distinct differences in CNS processing of A- and C-type BR inputs. Among these differences, the first appears to be related to the absolute frequency of BR activation. We found that C-type BR inputs are required for low-frequency (1-5 Hz) reflex-MAP responses (Fig. 1), and over this low-frequency range reflex-MAP responses are not significantly altered by addition of conducted A-type BR inputs (i.e., A+C). In addition, stimulation of A-type BR inputs alone in this frequency range was ineffective in evoking reflex decreases in pressure, although such stimuli clearly evoke conducted action potentials in ADN (15) and visceral myelinated fibers are capable of higher frequency axonal transmission (8). Similar results were reported for ADN-evoked inhibition of renal sympathetic nerve activity in the rabbit (2). From these observations, we conclude that the low-frequency MAP responses require conducted C-type BR inputs. Interestingly, driving C-type BRs to higher frequencies (10-100 Hz) of discharge failed to further increment C-selective MAP reflex responses (Fig. 1). C-type BRs are capable of conducting action potentials along the ADN afferent pathway within this frequency range in which the reflex responses were constant (100 Hz; see Fig. 1 in Ref. 16). Within the CNS, activation of C-type axons to the spinal cord (26) or in the solitary tract (13) reliably evokes synaptic responses in this frequency range in second-order neurons in the dorsal horn or in the nucleus of the solitary tract (NTS), respectively. Thus the C-type baroreflex central pathways beyond the second-order neurons may be operating at their maximum in this 10- to 100-Hz frequency range. In contrast, this same 10- to 100-Hz stimulus range evoked frequency-graded MAP reflex responses from A-type BRs, indicating that A-type pathways require higher frequencies of input (Fig. 1) and thus integrate frequency information over a higher range of action potential frequencies compared with C-type pathways. The mechanisms for such a central processing difference in A- and C-type BR information are not known.

Afferent summation is a second major reflex characteristic that differs across A- and C-type pathways. Reflex-MAP responses to A- and C-type BR inputs had similar average maxima when stimulated separately (Fig. 1). A+C-MAP responses were modestly greater than either subtype, and this summation was clearly occlusive. Together such findings suggest that the plateau of response saturation in the frequency-response relationships at the highest frequencies with selective A- and C-type stimulation does not lie at the effectors in the periphery and may, therefore, represent some limitation within the CNS.

Interestingly, the augmentation of combined stimulation occurred only within the limited frequency range in which A-type BR inputs showed graded MAP responses (i.e., 10-50 Hz; Fig. 1). Thus, overall, absolute action potential frequency is the most distinguishing difference between A- and C-selective baroreflex MAP response relationships. The ranges in which baroreflex responses were graded with frequency were thus quite characteristic of BR subtype: 1-5 Hz for C-type BRs and 10-50 Hz for A-type BRs. Interestingly, these subtype-selective baroreflex frequency windows coincide with the typical discharge ranges found with pressure activation of A- and C-type single BR neurons in the rat (4, 24).

Physiologically, not only are the prevailing rates of discharge substantially different across these BR subtypes, but their distributions in time within each cardiac cycle are characteristically different. Impulses from both A- and C-type BRs fire phasically in the intact animal in response to the pulsatile arterial pressure excursions during each cardiac systole. The dynamic discharge capacities of A- and C-type BRs are well known to differ greatly (e.g., Refs. 4, 16, 24). Anodal block studies of the carotid sinus nerve in dogs suggest that dynamic BR information is carried by A-type BRs (3). However, with the use of maximal-intensity (A+C) stimulation of the ADN or the carotid sinus nerve, we found no differences in baroreflex MAP responses between constant and intermittent patterns of electrical stimulation (16). The comparable efficacy of phasic vs. constant patterns of BR activation has been widely debated over the years with varying experimental results and conclusions. For example, pulsatile pressure inputs to the carotid sinus cause a greater reduction in blood pressure than do steady pressures (6, 14, 23). Burst patterns of carotid sinus nerve stimulation evoked modestly larger sympathoinhibition than did constant patterns in dogs (29, 30). Burst activation of A-type BR axons in rabbits, however, produced no additional response over constant patterns (11). Our intermittent burst tests with the rat ADN augmented baroreflex MAP responses only during selective, A-type BR activation despite an equal number of stimuli on a second-to-second basis (Figs. 2 and 3). Thus the wide dynamic discharge range of A-type BR neurons is associated with a greater reflex sensitivity to phasic patterning and again this correspondence is consistent with a general functional matching of A-type BR discharge characteristics. Thus A-type CNS inputs may have more dynamic reflex modulation than that associated with the more sparse C-type BR discharge. C-type BRs may therefore give rise to CNS baroreflex responses that are slower, perhaps more steady state in character. This functional subtype separation has many similarities to that suggested for types 1 and 2 BR inputs in the dog (35).

In our subtype-selective BR stimulation studies, two important factors should be noted. First, these reflex responses are based on synchronous electrical activation of all axons of a given category and, second, C-type axons outnumber A-type axons in the rat ADN by a nine-to-one ratio (3). Thus the substantially greater numbers of BR neurons activated in C-type stimulation might influence the form of the baroreflex frequency-response relationship. In an attempt to test whether this large numerical discrepancy between A- and C-type BRs was a critical factor, we greatly reduced the C-type BR input by using submaximal stimulation intensities to test recruitment of different numbers of BR axons. In consideration of our earlier ENG studies of ADN (15), we reduced stimulus intensity during anodal block to levels that activate a small fraction of the C fiber population. This was based on integration of compound action potential C-waves from ADN (15). If these time integrals of the ENG represent the fraction of the population activated and we combine this with total axon counts of A- and C-type axon profiles in histological sections of ADN (7), we estimate that 6 V should activate a total number C-type BRs similar to the total number of A-type BRs in the ADN (Fig. 7). Despite this experimental adjustment in the total number of fibers activated, the frequency-dependent differences in the A- and C-selective MAP reflex relationships remained (compare Figs. 6 and 7). These distinctions persisted during graded activation of A- or C-type BRs, suggesting fundamental differences in central summation and integration. For C-type recruitment, the slope of the reflex-MAP relationship increased substantially as more C-type BRs were activated, whereas the threshold index for A-type baroreflex relationships shifted to lower activation frequencies with relatively constant slope as more BRs were recruited.

The HR baroreflex to selective A- and C-type BR activation in our rats showed surprisingly different patterns of integration than found in the reflex-MAP relationships. Previous work with ADN in rabbits had suggested that brief, high-frequency (100-150 Hz) trains of activation were equipotent for P-R interval modulation with A- and C-selective stimuli (22) but that C-type BR input was critically required for maximum bradycardia. In contrast, we found in the rat that selective A-type stimulation was surprisingly ineffective at changing HR even at stimulation frequencies of 200 Hz. Just as with MAP, low frequencies of C-type activation produced substantial bradycardia and these became maximal at activation rates of 10 Hz. However, A+C-HR responses were greater than the individual responses in the frequency range of 20-200 Hz. This supra-additive HR summation contrasts the occlusive pattern in MAP responses. We conclude from these results that subtype differences in BR reflex responses extend not only within but also across reflex output pathways (in this case, MAP and HR).

Perspectives