The response of the vestibulosympathetic reflex to linear acceleration in the rat

The vestibulosympathetic reflex (VSR) in the anesthetized rat causes increases in blood pressure (BP) that were maximal during upward and forward translation and decreases in BP that were largest during downward accelerations. Single backward accelerations were ineffective as were lateral translations in producing regular, predictable increases in BP. A comparison of results in alert and anesthetized animals suggests that the VSR primarily produces changes in BP, not heart rate, and is relatively insensitive to changes in state of alertness.

Abstract

The vestibulosympathetic reflex (VSR) increases blood pressure (BP) upon arising to maintain blood flow to the brain. The optimal directions of VSR activation and whether changes in heart rate (HR) are associated with changes in BP are still not clear. We used manually activated pulses and oscillatory linear accelerations of 0.2–2.5 g along the naso-occipital, interaural, and dorsoventral axes in isoflurane-anesthetized, male Long-Evans rats. BP and HR were recorded with an intra-aortic sensor and acceleration with a three-dimensional accelerometer. Linear regressions of BP changes in accelerations along the upward, downward, and forward axes had slopes of ≈3–6 mmHg · g⁻¹ (P < 0.05). Lateral and backward accelerations did not produce consistent changes in BP. Thus upward, downward, and forward translations were the directions that significantly altered BP. HR was unaffected by these translations. The VSR sensitivity to oscillatory forward-backward translations was ≈6–10 mmHg · g⁻¹ at frequencies of ≈0.1 Hz (0.2 g), decreasing to zero at frequencies above 2 Hz (1.8 g). Upward, 70° tilts of an alert rat increased BP by 9 mmHg · g⁻¹ without changes in HR, indicating that anesthesia had not reduced the VSR sensitivity. The similarity in BP induced in alert and anesthetized rats indicates that the VSR is relatively insensitive to levels of alertness and that the VSR is likely to cause changes in BP through modification of peripheral vascular resistance. Thus the VSR, which is directed toward the cardiovascular system, is in contrast to the responses in the alert state that can produce sweating, alterations in BP and HR, and motion sickness.

NEW & NOTEWORTHY

The vestibulosympathetic reflex (VSR) in the anesthetized rat causes increases in blood pressure (BP) that were maximal during upward and forward translation and decreases in BP that were largest during downward accelerations. Single backward accelerations were ineffective as were lateral translations in producing regular, predictable increases in BP. A comparison of results in alert and anesthetized animals suggests that the VSR primarily produces changes in BP, not heart rate, and is relatively insensitive to changes in state of alertness.

the vestibulosympathetic reflex (VSR) is critical for maintaining stable blood flow to the brain when arising. This is accomplished by constricting peripheral arteries and veins in the muscle beds (Paulev 2000). Additionally, there is constriction of blood vessels in the legs produced by muscle sympathetic nerve activity (MSNA) in humans that inhibits blood pooling in the legs (Eckberg et al. 1985; Macefield et al. 1999; Robertson 2008, 2012; Wallin and Sundlöf 1982; Wallin et al. 1975). Although heart rate (HR) and blood pressure (BP) are interconnected through the baroreflex, only BP is increased, not HR, when the VSR is activated upon standing (Borst et al. 1982) or head-up tilt (Woodring et al. 1997; Yates and Miller 1998). The converse condition, where there are drops in BP of 20 mmHg or more when standing from a supine or prone position, orthostatic hypotension, is also not associated with changes in HR (Kanjwal et al. 2015; Schatz et al. 1996; Task Force for the Diagnosis and Management of Syncope et al. 2009). Thus there is a characteristic dissociation between BP and HR when the changes in BP are driven by the VSR.

There have been relatively few studies that characterize the directionality and frequency response of the VSR. Yates and Miller (1994) studied effects of oscillation of the head in pitch in decerebrate, cerebellectomized cats with denervation of the cervical nerves. Changes in splanchnic nerve activity occurred when the head was rotated in pitch toward the spatial vertical at a rate of 25°/s with a latency of 1.4 s, activating the vertical semicircular canals and the otolith organs. With the use of the same reduced cat preparation, which involved intercollicular separation of the cerebral hemispheres from the brain stem and a section of the cervical nerves to reduce cervical input, Woodring et al. (1997) showed that there were significant increases in BP but not in HR when the head was moved upward along the spatial vertical. We have also demonstrated that the VSR is activated in anesthetized rats when they are tilted up 70° (0.91 g) (Cohen et al. 2013; Yakushin et al. 2014). However, with the use of small, linear accelerations (0.2 g), alert humans were translated forward, back, and side to side in the horizontal plane (Yates et al. 1999). Subjects had their heads upright or flexed forward or back. The directions of the accelerations were randomized and unexpected. Changes in BP and HR were elicited at latencies of 1.5 s in all subjects, regardless of the direction of movement or whether their heads were upright or flexed. BP increased maximally by 7.9 mmHg and HR by 1–2 beats/min at 3–6 s after accelerations. The initial increases in BP and HR were followed by decreases, which occurred 6–9 s after acceleration. Thus the changes in BP and HR occurred in each subject, regardless of the direction of translation or the position of the head. Findings were similar in alert rats after unexpected, randomized, lateral, and forward-back accelerations of 0.1–0.3 g. The accelerations caused a short-latency increase in BP of 8.27 mmHg and ≈5 beat/min increases in HR, ≈520 ms after stimulation, followed by decreases in BP and HR (Zhu et al. 2007). In summary, both studies on alert subjects found similar activation of BP and HR in response to unexpected, weak linear accelerations in alert subjects regardless of the direction of oscillation. The changes in HR disappeared when the rats were put under anesthesia (Zhu et al. 2007). The initial changes in BP were maintained under anesthesia, however.

There have been relatively few studies of BP or HR responses to linear acceleration along the body z-axis in humans. Watenpaugh and Hargens (2010) reported changes in BP and HR in humans during parabolic flight. There were increases in both BP and HR in response to increases and decreases in acceleration along gravity that occurred over periods of 9–15 s. Thus exposure to linear acceleration along the gravitational axis can induce changes in both BP and HR.

These studies pose a fundamental question: is the VSR, activated by acceleration of the head and body, similar in all directions in the horizontal plane, or is it maximally oriented to accelerations and decelerations along gravity and/or the gravitoinertial acceleration (Woodring et al. 1997; Yates and Miller 1994, 1998)? Additionally, does activation of the VSR with 0.2 g, which is equivalent to a rise in head and body of only 12°, cause changes in both BP and HR, or are the changes due to the state of alertness? Thus given the fact that there is relatively little information about the directions of acceleration that activate the VSR, particularly in the rat, the goal of this research was to determine the best directions of acceleration that caused increases in BP in this rodent and whether both BP and HR were activated by linear translation in this preparation.

METHODS

Eight adult, male Long-Evans rats (Envigo, Cambridgeshire, UK), weighing between 400 and 500 g, were used in this study. All experiments were approved by the Institutional Care and Use Committee of the Icahn School of Medicine at Mount Sinai.

Surgery and experiments were performed under isoflurane anesthesia, 4% induction, and 2% maintenance with oxygen. While in surgery, the animals were kept on a temperature-controlled heating pad at 37°C. An intra-aortic sensor (Data Sciences International, St. Paul, MN) was implanted in the abdominal aorta. Through an incision in the groin, the femoral artery was isolated and clamped. The transducer catheter was inserted into the vessel via a small arteriotomy and advanced into the abdominal aorta. The catheter was secured with ties around the artery, and the body of the sensor was placed into a subcutaneous pocket in the flank. Adequate pain medications were given so that the animals did not suffer postoperative pain, and there were no behavioral changes after the rats recovered from anesthesia. Further details of the implantation and surgery have also been given in earlier publications (Cohen et al. 2011, 2013; Holstein et al. 2012; Yakushin et al. 2014).

During the experiment, anesthetized animals were placed in a plastic box filled with soft packing material. The animals were placed on their bellies. The head was stabilized with packing material, approximately in the horizontal, stereotaxic position. The area around the nose remained open to allow for normal breathing. The rear of the animals was not covered to allow for visual control of breathing rate and body position during and after each translation. This type of fixation had a limitation for downward translation of 1 g.

In one experiment, an alert rat was placed in a tube inserted on a tilt table. Before testing, the animal was habituated over several days using a food reward so that it would remain calm during the test. During testing, the animal was rotated 70° in the nose-up position over 7 s. The axis of rotation was 13 cm below the center of the head.

The coordinate frame used in this study, shown in the prone rat (Fig. 1), consisted of the forward naso-occipital (x), left interaural (y), and dorso-ventral (z) axes. The upward (z) axis was close to the gravitational (g) axis. The rats were anesthetized with isoflurane during the experiments and were prone, with their noses pointed forward along the horizontal axis.

Fig. 1.

Coordinate system used in this study. The positive direction of head translation along the x-axis was forward, along the y-axis was leftward, and along the z-axis was upward. The rat's body was prone during these experiments.

Data analysis.

Computer programs developed in our laboratory were used to collect the data and for data analysis. BP data, as well linear acceleration along three axes, were sampled at 1 kHz with 12-bit resolution (Data Translation, Marlboro, MA). The data were converted to systolic BP and HR from intersystolic intervals, and HR was expressed in beats per minute (Yakushin et al. 2014). The average HR in the rats was ≈300 beats/min. Thus the systoles occurred at a rate of 1 per 200 ms, which was the effective sampling rate for BP-related events. This would make it impossible to determine latencies of activation that were <200 ms.

Single translations were performed manually over 1 s intervals, with peak acceleration ranging from 0.2 to 2.5 g in both directions along the x-, y-, and z-axes in five rats. Individual responses were synchronized to the peak acceleration and overlapped for intervals of 1 s before and 1 s after peak accelerations had occurred. Changes in linear acceleration, as well as in BP and HR, were determined as differences in these signals taken over 50 ms before the onset of translation against peak response values. Changes in BP and HR were plotted as a function of linear acceleration and fitted with regression lines. The sensitivity of the VSR to linear acceleration (in millimeters of mercury times gravitation⁻¹) was determined as a tilt of the regression line. The animals were tested two to six times on different days. Data from each experiment were analyzed for the variation in sensitivity. Since the variations were insignificant, data for each individual rat were combined. Changes in BP correlated with the translational accelerations, according to their magnitude when the animals were accelerated.

During the manual translations along the x-axis, there could be small accelerations along the y-axis and vice versa along the x-axis during y-axis accelerations, due to minor misalignment of the box orientation to the translational axis. On average, they were negligible compared with the accelerations induced along the translational axis. Maximal changes along the z-axis were present during deceleration periods, whereas the translational components along the z-axis during x-axis translation did not exceed 0.2 g.

The frequency characteristics of the VSR were also studied. The animals were oscillated along their long-body (x) and interaural (y) axes at frequencies ranging from 0.2 to 5.0 Hz. Linear accelerations along the axis of translation were 0.2 g at frequencies below 0.4 Hz, 0.6 g at frequencies below 1 Hz, and gradually increased to 1.5 g at 5 Hz. Three to 10 cycles of BP, HR, and the acceleration with similar periods of oscillation were fit with sinusoids at the stimulus frequency to determine their BP (in millimeters of mercury times gravitation⁻¹) and HR (in beats per minute times gravitation⁻¹) sensitivity. The phases of the responses were also determined. The zero phase represented the peak increase in BP or HR at the time of peak-positive acceleration. It was impossible to perform similar oscillations manually along the z-axis so that only single upward and downward accelerations were performed. One alert rat was tilted 70° nose up (0.91 g), and the change in BP was determined to compare with the changes in BP induced in the linear translations in the anesthetized rats. The HR was also studied in the alert rat during the nose-up tilt.

Effects of breathing rate on BP and HR.

Both BP and HR oscillate during breathing cycles, and the oscillation of BP and HR during breathing cycles was determined. Since the breathing cycles could be synchronized by substantial linear accelerations, we determined the amplitude of BP and HR oscillations with a sensor attached to a balloon placed under the belly of the anesthetized rats (Fig. 2A; breathing rate). The BP was lowest during inhalation and peaked during exhalation, but the magnitude of the changes varied considerably among rats. They were only substantial in two rats (R002 and R004). In these rats, minimal HR lagged minimal BP by ≈500 ms. Therefore, the measurements of changes in BP, HR, and linear acceleration were taken at the same time in the breathing cycle of all rats. Since the changes in BP and HR were less abrupt during exhalation, measurements of BP and HR were taken during the exhalation period.

Fig. 2.

Changes in blood pressure (BP) and heart rate (HR) during breathing cycle in the tested rats. A: example of BP recorded with an intra-aortic sensor and breathing (Br) rate with a balloon placed under the rat's belly. The data were processed to extract systolic BP (Sys BP) and HR. The pressure in the balloon increased during inspiration and decreased during expiration to indicate breathing rate. Each inspiration peak is marked by a vertical, dashed line (Inspiration). There was tight synchronization of inspiration with minimal values of systolic BP. There was no relationship of HR changes to breathing cycle. B–I: example of average systolic BP and HR changes during breathing cycle in each rat tested in this study. Approximately 100 intervals were synchronized by minimal systolic BP (Inspiration) and averaged. Black lines are average values; gray lines are individual values of the breathing ranges. Data are shown over 1.5 s before and after minimal systolic BP. Arrows point to inspiration peaks (Inspiration).

To determine the amplitude of the systolic BP and HR modulations, the data were synchronized from the lowest BP (peak inspiration) and overlapped at ±1.5 s intervals. This interval was equal to or more than two full cycles. Therefore, each third-lower BP peak was taken for data synchronization to avoid data overlap. Approximately 100 cycles were averaged (Fig. 2, B–I). The averaged data from Fig. 2A are shown in Fig. 2D. The duration of the breathing cycle varied between rats from 0.87 to 1.55 Hz (Table 1). The amplitude of the systolic BP modulations varied from 0.2 to 6.7 mmHg (Table 1). The changes in HR were significant (P < 0.05, F statistic) only in four rats and varied from 8.5 to 15.4 beats/min. The changes in HR were in phase with the changes in BP, lagging BP by 0.25 ± 0.01 s, as measured from the distance between the minimal systolic BP and the minimal HR. The lag time was independent of the duration of the breathing cycle (Table 1).

Table 1.

Changes in systolic BP and HR during the breathing cycle in anesthetized rats

Rat #	Breathing Frequency, Hz	Systolic BP, mmHg	HR Changes, beats/min	Minimal HR, beats/min
001	1.03 ± 0.33	0.27 ± 2.01	–	–
004	1.44 ± 0.15	2.24 ± 3.39	8.48 ± 4.30	250
005	1.55 ± 0.49	1.75 ± 4.27	–	254
006	1.10 ± 0.17	5.62 ± 2.12	11.51 ± 4.32	225
007	0.87 ± 0.03	0.36 ± 2.69	–	–
009	1.10 ± 0.26	0.23 ± 3.17	–	194
010	1.21 ± 0.31	6.67 ± 2.10	11.84 ± 4.19	249
011	1.11 ± 0.17	4.80 ± 3.71	15.42 ± 7.31	238

Only data at P < 0.05 are shown.

Statistical analyses.

The significance of BP and HR modulations induced by linear translations (see Fig. 6) was measured using Pearson's coefficient (Sockloff and Edney 1972). Groups of data were compared with a t-test with a Bonferroni correction. The significance of the fits through the data was determined using a reduced case of ANOVA (F statistic) (Yakushin et al. 1995, 2011). The significance of the regression lines for the functional relationships between the stimuli and responses was determined by the critical R value (Daniel 1995). Changes in BP or HR were considered to be significant if they exceeded 2 SD of the prestimulus data. The dominant peaks at the frequency of stimulation and twice the frequency of stimulation were compared statistically using a Student's paired t-test or a one-way ANOVA with repeated measures, applying a post-Bonferroni adjustment. Sensitivities to translations among different directions for each animal, as well as mean sensitivity obtained for individual animals, were compared with a one-way ANOVA, applying a post-Bonferroni adjustment. Changes in BP and HR were deemed significant at P < 0.05.

Fig. 6.

Changes in BP induced by translation at 1 g along x- and z-axes in rats R004 (A), R006 (B), R009 (C), R010 (D), and R011 (E). The arrows point to the average responses for individual rats (see Table 2). Black arrows point to responses induced by upward and forward translations; gray arrows point to responses induced by downward and backward translations.

RESULTS

When the rats were tested with short pulses of translation (<0.5 s) at various accelerations, there were no changes in BP for translations in any direction at high accelerations. Slower (1 s) translations induced changes in BP. The failure to induce BP changes at briefer accelerations is in agreement with the frequency characteristics of the VSR that are shown in the frequency analysis (see Figs. 8 and 9 below).

Fig. 8.

Effects of frequency of oscillation on activation of BP. A–E: changes in BP in response to linear oscillation at sinusoidal translation about the x-axis at ≈0.3 Hz (0.4 g). A: systolic and diastolic BP; B: systolic BP, fit with a sinusoid; C: sinusoidal changes in HR; D: the activating stimulus, i.e., the linear accelerations at ±0.4 g along the x-axis; E: lateral acceleration during the sinusoidal activation in D. F and G: the sensitivity (F) and phase (G) of the VSR induced by translation along the x-axis (forward, backward) as a function of stimulus frequency (abscissa). The 0 phases corresponded to forward motion. Note that the peak modulations were maximal at frequencies <1 cycle/s (G) and that the phases were 0 at these frequencies. H and I: there was no correlation with changes in BP when the oscillations were in the lateral (±y) direction (H), and the phases were widely distributed (I). Gray, horizontal bars in F and G are breathing-frequency range of this rat (average ± 1 SD).

Fig. 9.

The sensitivities (A and C) and the phases (B and D) of the VSR as a function of stimulus frequencies (abscissa) for the other 4 animals in this series. The animals were oscillated at different frequencies along the ±x-axis (forward-back; A and B) and the ±y-axis (side-to-side; C and D) by sinusoidal translation along the x-axis. The 0 phase corresponded to forward (B) and leftward (D) accelerations. In each rat, the sensitivities were maximal at oscillation frequencies below 1 Hz, and the phases were 0, i.e., in the forward direction. There was no similar relationship during ±y-axis oscillation (C and D). Gray, horizontal bars in A and B are breathing-frequency range of this rat (average ± 1 SD).

Upward-downward (z-axis) translations.

Upward translations with peak accelerations of 0.2–2.5 g along the vertical (z) axis induced large increases in BP in each of the 20 overlaid traces from each of the 5 rats (Fig. 3, A–E). The overlaid acceleration traces demonstrate the regularity and reproducibility of the stimuli for each animal. Peak BPs occurred close to the peak upward accelerations and were followed by reductions in BP that were associated with the decelerations. There were no consistent changes in HR associated with the changes in BP. As noted in methods, the changes in BP were calculated based on the appearance of the systoles, which occurred at HRs, ∼300/min, i.e., 1 every 200 ms. Since the central transmission time is likely to be <5–10 ms, the latencies of the stimuli appear falsely to overlap, although the former induced the latter. The same was true for BP changes during forward and back and lateral acceleration (see Figs. 4 and 5, respectively). There was a slight decrease in BP before the onset of acceleration related to respiration (see methods). The amplitudes of the induced changes in BP ranged from ≈5 to 20 mmHg. The decelerations that followed each of the upward accelerations were associated with a following depression in BP of ≈5 mmHg in each of the animals.

Fig. 3.

A–E: increases in systolic BP (bottom traces) in response to upward (+z) accelerations (top traces). Each increase in upward acceleration caused an associated increase in BP. Each set of traces is from a different rat. Note the regularity of the stimulus pulses and of the BP responses. F–J: decreases in BP (bottom traces) in response to downward (−z) accelerations. Downward accelerations also reliably evoked reductions in BP. The small drops in BP at the onset of the accelerations were due to inspiration. Approximately 20 traces are superimposed in each of the traces in A–J. K and L: averaged responses from A–J. There was an increase in BP associated with the upward acceleration in each rat (K) and a decrease in BP with downward acceleration in each animal associated with the downward acceleration (L). Only 1 animal had changes in HR associated with the upward acceleration (K). No changes in HR were observed in association with downward acceleration. Time in seconds on abscissa; BP in millimeters of mercury on the ordinate.

Fig. 4.

Changes in BP in response to forward (A–E) and backward (F–J) translation. Top traces: positive (A–E) and negative (F–J) x-axis linear accelerations; 20 trials in each. Bottom traces: BP. Forward (+x) acceleration generally produced increases in BP, whereas the backward (−x) accelerations did not cause regular changes in BP. K and L: averaged accelerations demonstrate that there were regular increases in BP when moving forward along the x-axis (K) but not when moving backward along the −x-axis (L). No changes in HR were observed in association with forward or backward accelerations. Time in seconds on abscissa; BP in millimeters of mercury on the ordinate.

Fig. 5.

Changes in BP during lateral (+y- and −y-axis) translation. Top traces: linear accelerations; bottom traces: BP. A–E: left (+y) axis translation; F–J: right (−y) axis translation. Twenty responses overlaid in each stimulus sequence. There were intermittent responses that were not synchronized with the stimuli. K and L: averaged responses for each of the 5 rats. There were increases (K) and decreases (L) in some rats, but they were not well synchronized with the linear accelerations. Time in seconds on abscissa; BP in millimeters of mercury on the ordinate.

Downward accelerations of approximately −1 g caused decreases in BP (Fig. 3, F–J). The magnitudes of the decreases in BP were smaller, due to the smaller accelerations that induce these changes, and the responses were not as well synchronized as those that followed the stronger, upward accelerations (cf. Fig. 3, F–J and A–E, respectively). There was an increase in BP at the end of the downward linear accelerations associated with the accelerations back to the prestimulus level.

The stimulus accelerations and the BP responses of each rat were averaged (Fig. 3, K and L). There was a close association of the peak acceleration and the maximal changes in BP in each rat (Fig. 3K), which ranged from 3 to 10 mmHg. The initial increases in BP were followed by a 3- to 10-mmHg drop that was associated with the deceleration after the initial acceleration. There was also a third increase in BP in response to the preceding decelerations. There were no consistent changes in HR during these accelerations and decelerations in four out of five tested rats (Fig. 3, K and L). Translation downward caused a 3- to 5-mmHg decrease in BP (Fig. 3L), followed by a 3- to 5-mmHg rise. These data demonstrate that there is a close association of increases and decreases in BP for acceleration and deceleration along the gravitational axis. The changes in BP were larger when the animals were translated up rather than down, but the linear accelerations were greater in the upward direction, and BP was systematically modified in each rat in both directions (Fig. 3, K and L).

Forward-backward (x-axis) translations.

Forward accelerations also induced changes in BP of ≈3 mmHg (Fig. 4, A–E), but the modulations were neither as large nor as well synchronized as those along the gravitational axis. In part, this could have been due to the smaller accelerations (≈1 g) along the x-axis. In four of the five rats, backward accelerations caused a significant decline and increase in BP (Fig. 4, F–I).

When averaged, the forward accelerations induced a 3- to 5-mmHg rise in each rat, followed by a smaller reduction in BP and a small, trailing BP increase (Fig. 4, K and L). The backward translations, however, did not produce consistent increases in BP, except in one rat (−2.62 ± 1.70 mmHg; Fig. 4L), and the peaks of the BP increases were widely spread during backward acceleration. Average changes to backward translations were 0.022 mmHg, with SD varying from 1.46 to 3.76 (P > 0.05). In summary, BP was altered in each of the rats by accelerating forward, although the changes were not as large or as consistent as the changes induced by translation along the gravitational axis. There were no consistent changes in HR during these accelerations and decelerations in four out of five rats (Fig. 4, K and L).

Side-to-side (y-axis) translations.

Lateral, linear accelerations along the y-axis induced small changes in BP, but their occurrence was variable in onset and amplitude (Fig. 5, A–J). On average, there were no consistent changes in BP in each of the rats: the increases and decreases varied among the animals, as did the onset and end of the responses. Average changes to leftward translations were 0.13 mmHg, with SD varying from 1.30 to 2.04 (P > 0.05). Average changes to rightward translations were −0.15 mmHg, with SD varying from 1.55 to 2.15 (P > 0.05). When individual responses were averaged, in two of the five rats, the effective acceleration was to the left and in the other three, to the right (Fig. 5, K and L, respectively). There were no consistent changes in HR during these accelerations and decelerations (not shown).

In summary, the changes in BP from activation of the VSR with linear acceleration were most coherent when displacements were upward and downward, less coherent when the rats were translated in the forward direction, and considerably less so when the rats were translated backward or laterally. There was no effect of the increase or decrease in BP on HR.

Magnitude of BP changes induced by single translations.

The relationship and magnitude of the BP changes induced by single translations with different accelerations up and down (Fig. 3), forward and backward (Fig. 4), and laterally (Fig. 5) were determined for each rat. To compare changes in BP induced by translation along different axes, individual responses were normalized to peak accelerations that induce these changes (Table 2). In rats R004 and R006, responses to downward and forward translations were similar in magnitude (Fig. 6, A and B, and Table 2; P > 0.05, ANOVA). Responses to upward translation were smaller but still significant [Table 2; P < 0.0001 (R004); P = 0.0043 (R006), ANOVA]. There were no responses to translation backward, left, and right (Table 2; P > 0.05). Responses of similar magnitude to forward and downward translations were recorded in R010 (Fig. 6D). Leftward and rightward translations in this animal also induced decreases and increases in BP, respectively (Table 2; P = 0.0009, ANOVA). Similar changes in BP magnitude were induced by downward, forward, and backward translations in rat R009 (Fig. 6C and Table 2; P < 0.0001), although the difference between downward and forward translations remained significant (P = 0.026, ANOVA). Forward translations in rat R009 induced increases in BP, whereas backward translations induced decreases in BP. Changes induced by upward translation were somewhat smaller but significantly different only from the magnitude of downward translation (Table 2; P < 0.0001, ANOVA). No responses were induced by leftward and rightward translations (Table 2; P > 0.05, ANOVA). The maximal changes in BP were induced by downward translation in rat R011 (Fig. 6E and Table 2; P < 0.0001, ANOVA). Changes induced by forward and upward translations were significant (Table 2; P < 0.0001) but comparable (Table 2; P > 0.05). Changes induced by backward translations were also significant (Table 2; P < 0.0001, ANOVA); however, they were comparable only with changes induced by upward translation (P = 0.697, ANOVA) and were in the same direction as changes induced by forward translation (see Fig. 6E). On average, for all five rats, the changes in BP, induced by translation up, down, and forward, were comparable (Table 2; P > 0.09, ANOVA of average values), whereas the changes induced by backward, left, and right translations were insignificant (Table 2; P > 0.05, ANOVA of average values). The magnitude (absolute value) of the average sensitivity for translations along the forward, upward, and downward accelerations, where changes were found significant, indicates that while the differences between them were significant (P = 0.009, ANOVA), the sensitivity to upward acceleration was smaller than the sensitivity to downward acceleration (P = 0.009), while differences between the sensitivities of the magnitude to the downward and forward accelerations (P = 0.903) and the upward and forward accelerations were not significant (P = 0.069).

Table 2.

Average (±1 SD) changes in BP induced by 1 g translations along different axes in 5 rats

Rat #	Upward	Downward	Forward	Backward	Leftward	Rightward
R004	2.72 ± 0.91	−5.64 ± 1.91	5.67 ± 2.59	−1.99 ± 1.98	1.52 ± 2.10	−0.78 ± 1.75
R006	2.23 ± 1.01	−3.90 ± 1.66	4.06 ± 1.73	−0.82 ± 3.93	−1.00 ± 1.41	0.82 ± 1.34
R009	1.32 ± 0.46	−4.67 ± 2.33	2.67 ± 1.07	−2.78 ± 1.84	1.00 ± 1.31	−1.61 ± 1.45
R010	1.43 ± 0.66	−2.67 ± 2.32	3.58 ± 1.56	−0.22 ± 2.82	−1.75 ± 1.76	2.30 ± 1.36
R011	2.52 ± 0.92	−6.04 ± 3.07	3.22 ± 1.33	1.52 ± 1.76	1.03 ± 1.35	−0.07 ± 1.24
Average	2.04 ± 0.64	−4.58 ± 1.36	3.84 ± 1.14	−0.86 ± 1.66	0.16 ± 1.44	0.13 ± 1.51

An alert rat was tilted 70° nose up (0.91 g) to compare with the results of the upward accelerations, which had peak values of ≈2.5 g. The three-times increment in upward acceleration, due to the tilt, induced a proportional increment in BP to ≈9 mmHg · g⁻¹. This indicated that whereas responses in alert rats are larger (see Table 2), VSR sensitivity under anesthesia remains significant.

Linear acceleration changes in BP.

The relationship of the BP changes to the magnitude of linear acceleration induced by translations along different directions was determined for each rat. Results from a typical animal are shown in Fig. 7 and for all five rats in Table 3. The sensitivity to upward translation in the rat whose data are shown in Fig. 7 was 3.1 mmHg · g⁻¹ (Fig. 7A). This was close to the average sensitivity of 3.2 mmHg · g⁻¹ for all of the animals (Table 3). The sensitivity to downward translation in this animal was 2.9 mmHg · g⁻¹ (Fig. 7B). It was smaller than the accelerations obtained in other animals, which, on average, were 4.6 mmHg · g⁻¹ (Table 3).

Fig. 7.

Scatter plots of changes in systolic BP as a function of peak linear acceleration in a (typical) rat. Each dot represents a single trial. Plots show results of upward (A), downward (B), forward (C), backward (D), leftward (E), and rightward (F) acceleration. The confidence levels of the linear regressions are shown below each graph. A–C: the linear regressions through the data had a significant correlation for 3 of the data sets [upward (A), downward (B), and forward (C)]. Despite the intermittent changes in BP when the animal was translated back (D), left (E), and right (F), the magnitude of the responses did not correlate with the magnitude of acceleration. Abscissa, magnitude of acceleration; ordinate, change in BP in millimeters of mercury.

Table 3.

BP with regard to translational acceleration in individual rats

Rat #	Upward, mmHg · g−1	Downward, mmHg · g−1	Forward, mmHg · g−1
004	3.3	4.0	8.7
006	3.6	4.0	3.6
009	2.0	3.4	3.9
010	3.1	2.9	4.2
011	4.1	9.0	7.1
Average	3.2	4.7	5.5

Values are in millimeters of mercury (P < 0.05).

The sensitivity to forward translation in this rat was 4.2 mmHg · g⁻¹ (Fig. 7C). This was close to the average sensitivity of all of the animals (5.5 mmHg · g⁻¹; Table 3). The sensitivity to backward translations was not significant in any animal (P > 0.05; Fig. 7D). There was no significant correlation with changes in BP with backward accelerations in all other animals (P > 0.05). The sensitivity to leftward and rightward translation could not be determined for this rat (P > 0.05; Fig. 7, E and F, respectively) as in the other animals, because the correlation of BP with acceleration was not a significant ANOVA. The pooling of the sensitivities to translations in upward, downward, and forward directions, where the linear regression fits were significant, did not reveal any differences (P = 0.227; Table 3), and the average sensitivity of the translational VSR to acceleration in these three directions was 3.8 mmHg · g⁻¹.

Oscillation along the x- and y-axes.

The rats were linearly oscillated at various frequencies to determine the frequency characteristics of the VSR. Oscillation of an animal along the x-axis at a frequency of ≈0.3 Hz with an acceleration of 0.40 g (Fig. 8D) caused significant modulations in both systolic (Fig. 8, A and B) and diastolic (Fig. 8A) BP at the frequency of oscillation. Oscillations of BP about the sinusoidal fit were caused by changes in BP due to breathing-rate modulation (Fig. 8B). The modulations in HR, however, were small and insignificant (Fig. 8C; P > 0.05, F statistic). The spread of acceleration to the y-axis was minimal (≈0.06 g; Fig. 8E), i.e., ∼10 times smaller than the modulations produced by acceleration along the x-axis (Fig. 8D). BP increased with the peak forward acceleration and decreased with backward acceleration. The sensitivity and phases of the systolic BP changed when the translation along the two axes was computed as a function of frequency (Fig. 8, F–I). The sensitivity of the x-axis translation was greater at low frequencies and gradually decreased to ≈1 mmHg · g⁻¹ at frequencies above 1.8 Hz (Fig. 8F). The phases of the responses at all frequencies, ≈0°, indicated that the peak increase of BP had occurred during forward acceleration (Fig. 8G). Modulation of systolic BP due to linear oscillations at 0.3 Hz was not affected much by oscillation of BP due to breathing rate, because the magnitude of the changes in BP due to breathing was relatively small (Fig. 8B). Furthermore, sensitivities to linear translation at breathing frequencies (Fig. 8F) were consistent with sensitivities at frequencies below and above. This indicates that modulation of BP due to linear translations was much stronger than the changes induced by breathing rate.

Similar results were obtained in four other animals (Fig. 9, A and B). That is, the peak sensitivity at low frequencies varied from 5 to 10 mmHg · g⁻¹, and the sensitivity was substantially larger at frequencies below 1 Hz. As the frequency of oscillation increased, the sensitivity decreased to 1 mmHg · g⁻¹ at 5 Hz. The phase of the response at all frequencies was 0° in all of the animals. This indicates that the peak increases in BP occurred at peak forward accelerations (Fig. 9B).

The BP responses to oscillations about the y-axis were small and variable. The sensitivity in one animal was ∼1 mmHg · g⁻¹ at all frequencies (Fig. 8H). The phases of the responses varied between 0° and −180°, indicating that the peak changes occurred randomly in phase with leftward and rightward accelerations (Fig. 8I). Two other animals had higher sensitivities at very low frequencies (Fig. 9C), but the changes were substantially smaller than those during translation about the x-axis (Fig. 9A); the sensitivity decreased to ≈1 mmHg · g⁻¹ at higher frequencies (Fig. 9C). The sensitivity of two other animals was ≈1 mmHg · g⁻¹ at all frequencies. The sensitivity of the systolic BP due to linear oscillations about the x-axis was not affected substantially by oscillation of BP (Fig. 9, A and B), indicating that the modulation of BP due to linear translations along the x-axis was much stronger than the changes induced by breathing rate.

DISCUSSION

The major findings of this study are that the VSR in the anesthetized rat responds to strong pulses of translational acceleration, causing increases in BP that were maximal during upward and forward translation and decreases in BP during downward accelerations. Single backward accelerations were ineffective, as were lateral translations in producing regular, predictable increases in BP. There were occasional surges in BP during backward and lateral translations, but the responses were inconsistent and were not regularly related to the accelerations. Therefore, they could have little functional significance for control of BP. With the use of the sensitivity to forward acceleration in an oscillatory analysis, the VSR was shown to be most sensitive at oscillations up to ∼1.8 Hz. This is consistent with the frequency response of the VSR to the changes in the gravitational axis on the head and body during the seconds it takes to arise from a prone or supine position. There was an obvious conflict with the breathing rate and the results of the oscillatory study at 1 Hz and above, but there was no conflict with the breathing rate at lower or higher frequencies. Therefore, we conclude that the VSR is predominantly active in the upward, forward, and downward directions and at the frequencies up to ∼1.8 Hz in the anesthetized rat.

HR was unaffected by single translations in these animals, although it could be recruited by oscillations in the forward-back direction (Fig. 8C). The finding that there was no change in HR during the translations that caused increases in BP is consistent with the lack of changes in HR during activation of the VSR in humans when arising from a prone or supine position (Borst et al. 1982). Similarly, there are no additional changes in HR during orthostatic hypotension when BP drops by at least 20 mmHg upon standing (Porter and Kaplan 2011; Schatz et al. 1996; Task Force for the Diagnosis and Management of Syncope et al. 2009). However, it could also be related to the depression of baroreflex function under anesthesia. Similar results were reported by Yates and colleagues (Woodring et al. 1997; Yates and Miller 1994) from activation of the VSR through the vertical semicircular canals and otolith organs in decerebrate cats in which there was a large increase in BP on upright flexion of the head without changes in HR. HR can be altered, however, with strong and/or continuous activation of the VSR [see Fig. 3 of Yakushin et al. (2014)], which also presumably occurred during the continuous oscillation shown in Fig. 6C. These findings indicate that the baroreflex sensitivity, which is a measure of changes in HR vs. changes in systolic BP, is close to zero when BP rises or falls as a result of VSR activation. The baroreflex sensitivity is also close to zero when BP and HR both fall during vasovagal oscillations or vasovagal responses in the anesthetized rat (Cohen et al. 2011, 2013; Yakushin et al. 2014).

As noted in methods, the changes in BP were calculated based on the appearance of the systoles, which occurred at HRs, on average, of ∼300/min, i.e., 1 every 200 ms. The distance from the labyrinths to the rostral ventral lateral medulla (RVLM) and the thoracic and lumbar spinal cord is short in the rat, and there are monosynaptic connections between the otolith portions of the vestibular nuclei and RVLM (Holstein et al. 2012, 2014) and between RVLM and preganglionic sympathetic neurons in the vicinity of the thoracic spinal cord (Destefino et al. 2011). Consequently, the central transmission time is likely to be <5–10 ms. As a result, the stimuli overlap in Figs. 3–5, as they are in the range of the resolution of the computation of changes in systolic BP. Consequently, the latency of the changes in BP could not be accurately determined from these measurements.

In studies of alert humans and rats, Yates et al. (1999) and Zhu et al. (2007) found that weak, 0.2 g translations along the x- and y-axes in the horizontal plane reliably produced elevations of both BP (≈4.0 mmHg · g⁻¹) and HR (1–5 beats/min) at 1–2 s, followed by inhibition of both at 8–10 s. The translations were given at random intervals and were unexpected. The latency of the changes in BP and HR did not change over the course of many repetitions, causing Yates et al. (1999) to conclude that the subjects had not adapted to the presentation of the stimuli, probably because they could neither know when nor in what direction the accelerations would occur. Similarly, Watenpaugh and Hargens (2010) also reported changes in BP and HR during parabolic flight when the aircraft plunged toward earth.

When the rats were anesthetized in the Zhu et al. (2007) study, the changes in HR disappeared, but the initial changes in BP were unaffected by the change in the state of alertness. Moreover, there was no change in response when the baroreceptors were ablated in these rats. The significance of this is that the initial change in BP in response to linear acceleration is not modified by a reduction in the state of alertness, suggesting that the results obtained in this study are similar to tho se in either the alert or anesthetized state.

The implication of the finding that linear acceleration induces changes in BP and not HR for pulses, but not sinusoidal stimulation, is that there are at least two neural subsystems that link the vestibular nuclei to the sympathetic system. One uses monosynaptic projections from otolith and possibly superior vestibular nuclei neurons to the RVLM using glutamatergic transmission (Holstein et al. 2012, 2014). This link is generally associated with activation of the VSR and in humans with activation of MSNA that functions to constrict vessels in the leg when BP drops to maintain blood flow to the brain (Yates et al. 2014). Another branch of the vestibulosympathetic pathway through the nucleus raphe pallidus and parabrachial nucleus, which uses serotonergic transmission into the nucleus tractus solitarius and RVLM (Balaban 1996a, b; Moazzami et al. 2010; Yates et al. 1994), is responsible for changes in HR, anxiety, sweating, nausea, vomiting, and motion sickness. Each of these pathways may have different response times. Thus we speculate that the VSR, which is responsible for BP changes and not changes in HR, is predominantly generated by the glutamatergic pathway and is relatively insensitive to the state of alertness. We further postulate that the activation of the vestibular system in alert humans and rats in the Yates et al. (1999) and Zhu et al. (2007) studies caused activation of both components of the sympathetic system, whereas under anesthesia and in conditions associated in alert humans during rising or in anesthetized rats, primarily the glutamatergic link was activated, although the other sympathetic component could be recruited with strong stimulation. The projections that predominantly function by changing BP alone are probably mediated largely through constriction of peripheral blood vessels. Moreover, the fact that the VSR functions to change BP, even in anesthetized animals, and has a specific direction specificity is quite informative about the signal processing in the VSR and should impact how this system could be modeled (Raphan et al. 2016).

Differences in the orientation of the VSR in rats and humans.

What is the significance of the finding that the VSR had the most sensitivity forward, upward, and downward in the rat, and can these results be applied to humans? The rat has been particularly valuable in studies of the baroreflex sensitivity of connections between the vestibular nuclei and the various components of the sympathetic system, and more recently, we have shown that we have the first small animal model of the vasovagal response, which has not been studied adequately in animals before (Cohen et al. 2011, 2013; Raphan et al. 2016; Yakushin et al. 2014). This justifies the use of the rat in studies of vascular phenomena that can be applied to humans. However, rats commonly only stand upright on their hind paws to reach for food so that changes in BP associated with movement of the head along the gravitational axis can occur but are not as common as movement of the body in other directions. The increase in BP in the erect position is also similar to the increase in activity in splanchnic nerves in the cat when the head was pitched up, activating the vertical semicircular canals and the otolith organs (Woodring et al. 1997; Yates and Miller 1994, 1998). Consequently, it is likely that an increase in BP, when moving up along the gravitational axis, may be a common feature of quadrupeds and humans. Rats can also move forward vigorously so that the increase in BP when accelerating forward and the decrease when halting seem reasonable. However, it is not known whether there is the same effect on BP in humans on sudden forward starts, although there is strong evidence that the autonomic system in quadruped and humans is activated by forward motion that induces vergence and pupillary constriction in rabbits and monkeys (Gilmartin 1986; Maruta et al. 2005; Yakushin et al. 2008, 2009). There is also similar vergence and constriction of the pupils and lens in humans when viewing near objects as would occur when moving forward (Yoshida and Watanabe 1969).

However, the long-body axis in humans is aligned with gravity, whereas the long-body axis in rats is parallel to the ground, i.e., 90° from that in humans. Additionally, humans are bipedal, as are various other animals, including some apes and birds (Alexander 2004), whereas most other higher-order vertebrates are quadrupedal. There is substantial input from the body and limbs to the vestibular nuclei that can sense the body's orientation to gravity (Mittelstaedt 1996; Wilson et al. 2006; Yates and Bronstein 2005; Yates et al. 2000). We speculate that the difference in orientation to the gravitational axis in bipedal and quadrupedal species has altered the VSR so it responds differently in quadrupeds and bipeds to downward head movement. Ray and colleagues (Carter and Ray 2008; Hume and Ray 1999; Ray and Monahan 2002; Shortt and Ray 1997) have shown conclusively that head-down tilt regarding the gravitational axis causes an increase in MSNA in humans, whereas it causes a decrease in BP in rats (Fig. 3), rabbits (Tsubota et al. 2012), and cats (Woodring et al. 1997). It would not make sense for MSNA to be inactive when the upright human or other bipeds look down to navigate over uneven surfaces, and presumably, the different input from the body-tilt receptors, which project to the vestibular nuclei (Fredrickson et al. 1966; Yates et al. 2000), has been influential in modifying the reflex. With the emphasis of this point, the increase in MSNA on head-down tilt is even more striking when subjects are erect and looking down (Carter and Ray 2008).

Thus intraspecies differences limit the assignment of mechanisms of vestibulosympathetic activation in quadrupeds to bipeds without further investigation. It would also be important to know whether expected accelerations induced similar changes in BP without affecting HR in the observed directions in the alert rat. Despite this, this study has shown the close relationship of activation of the VSR with the gravitational axis, as well as the relative inactivity of the heart in the anesthetized rat when activating the VSR. This appears to be a fundamental mechanism governing vestibulosympathetic excitation and could make these results useful in further studies to unravel the underlying cause for orthostatic hypotension, but they must also be verified in alert rats and humans who know when and where they will be translated.

GRANTS

Support for this work was provided by a National Institute on Deafness and Other Communication Disorders research grant (DC0 12573; to B. Cohen).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

S.B.Y. conceived of and designed research; S.B.Y. performed experiments; S.B.Y. analyzed data; S.B.Y., G.P.M., T.R., and B.C. interpreted results of experiments; S.B.Y. prepared figures; S.B.Y. and B.C. drafted manuscript; S.B.Y. and B.C. edited and revised manuscript; S.B.Y., G.P.M., T.R., and B.C. approved final version of manuscript.