Autonomic Neuroscience: Basic and Clinical 187 (2015) 45–49
Contents lists available at ScienceDirect
Autonomic Neuroscience: Basic and Clinical journal homepage: www.elsevier.com/locate/autneu
Segmental origins of cardiac sympathetic nerve activity in rats Natasha H. Pracejus a, David G.S. Farmer a, Robin M. McAllen a,b,⁎ a b
Florey Institute of Neuroscience and Mental Health, University of Melbourne, Parkville, Victoria 3010, Australia Department of Anatomy and Neuroscience, University of Melbourne, Parkville, Victoria 3010, Australia
a r t i c l e
i n f o
Article history: Received 6 October 2014 Received in revised form 20 November 2014 Accepted 26 November 2014 Keywords: Preganglionic White ramus Sympathetic chain
a b s t r a c t The segmental origins of cardiac sympathetic nerve activity (CSNA) were investigated in 8 urethane-anesthetized, artificially ventilated rats. The left upper thoracic sympathetic chain was exposed retropleurally after removing the heads of the second to fourth ribs. The preganglionic inputs to the chain from segments T1–T3 and the trunk distal to T3 were marked for later sectioning. CSNA was recorded conventionally, amplified, rectified and smoothed. Its mean level was quantified before and after each preganglionic input was cut, usually in rostro-caudal sequence. The level after all inputs were cut (i.e. noise and residual ECG pickup) was subtracted from previous measurements. The signal decrement from cutting each preganglionic input was then calculated as a percentage. CSNA in all rats depended on preganglionic drive from two or more segments, which were not always contiguous. Over the population, most preganglionic drive came from T3 and below, while the least came from T1. But there was striking inter-individual variation, such that the strongest drive to CSNA in any one rat could come from T1, T2, T3, or below T3. These findings provide new functional data on the segmental origins of CSNA in rats. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Sympathetic preganglionic neurons originate in the spinal cord, from where they emerge segmentally to join the sympathetic chain via their corresponding white ramus, and synapse in ganglia to supply their respective postganglionic neurons (Langley, 1892; Rubin and Purves, 1980). A ganglion receives its preganglionic input from multiple spinal segments, therefore multiple segments normally provide the supply destined for a target organ (Kamonsińska et al., 1991; Langley, 1892). Thus, each contributing spinal segment drives a proportion of the overall sympathetic traffic to a given target such as the heart (Lichtman et al., 1980; Njå and Purves, 1977). Previous physiological studies have established that the main segmental contribution to the cardiac sympathetic supply differs between species. In dogs, for instance, the strongest inputs appear to come from the first to third thoracic roots (Kostreva et al., 1977; Norris et al., 1974, 1977). Conversely, in cats the third and fourth thoracic segments contribute most to the cardiac sympathetic supply (Kamonsińska et al., 1991; Kocsis and Gyimesi-Pelczer, 1998; Ninomiya et al., 1993; Szulczyk and Szulczyk, 1987). As far as we are aware, no comparable data exist for the laboratory rat. Anatomical studies, analyzing transsynaptic retrograde labeling after pseudorabies virus (PRV) injection into the heart, indicate that the upper six or seven thoracic segments could contribute to the cardiac sympathetic supply (Ter Horst et al., 1993, 1996). However, those reports do ⁎ Corresponding author at: Florey Institute, University of Melbourne, Parkville, Vic. 3010, Australia. E-mail address: rmca@florey.edu.au (R.M. McAllen).
http://dx.doi.org/10.1016/j.autneu.2014.11.011 1566-0702/© 2014 Elsevier B.V. All rights reserved.
not specify the relative labeling within those segments, nor can such anatomical data reveal the strength of physiological drive from each segment. To resolve the latter issue, we adopted the method that Ninomiya et al. (1993) and Kocsis and Gyimesi-Pelczer (1998) used in cats, sequentially cutting the upper thoracic white rami, to determine the contribution of each spinal segment to the cardiac sympathetic nerve signal. 2. Methods 2.1. Animals Eight male Sprague–Dawley rats (320–440 g) were used. All animals were on a 12/12 light–dark cycle, with ad libitum access to food and water, and were housed with at least one other animal. Experimental approval was acquired from the Animal Ethics Committee of The Florey Institute of Neuroscience and Mental Health. 2.2. Anesthesia Before experimentation, each rat was anesthetized with Pentobarbital sodium (60 mg/kg, i.p.). A tracheostomy was then inserted to artificially ventilate the rat with 2% Isoflurane in oxygen, which was continued for the duration of preparatory surgery. After the surgery, anesthesia was switched to urethane (1.4 g/kg, i.v.) over approximately 30 min, while progressively discontinuing Isoflurane. Throughout the experiment, anesthesia was maintained at a depth that abolished corneal and withdrawal reflexes by giving supplementary doses of urethane if
46
N.H. Pracejus et al. / Autonomic Neuroscience: Basic and Clinical 187 (2015) 45–49
required (10% of initial dose). At the end of the experiment, the rat was killed with an overdose of Pentobarbital (160 mg/kg, i.v.).
2.3. Experimental preparation and monitoring Throughout the experiment, the animal was artificially ventilated with oxygen at 70 inflations per minute and a tidal volume of 3–4 ml (rodent ventilator, Ugo Basile, Italy). Respiratory pressure and expired CO2 levels were monitored continuously (Capstar 100 Carbon Dioxide Analyzer, USA). Minor adjustments of rate or volume were made to keep end-tidal CO2 levels close to 4%. Catheters were inserted into the right femoral artery and vein for measurement of blood pressure and administration of fluids, respectively. The venous catheter was filled with normal saline and the arterial catheter with 20 U/ml heparinized normal saline. On occasion, 1–2 ml of Gelofusine (B. Braun, Germany) was given i.v. to support blood pressure. A rectal probe monitored the animal's temperature, which was kept near 37 °C with a heating pad.
2.6. Experimental procedure For the experiment, cardiac sympathetic nerve activity (CSNA) was recorded until a stable level could be maintained for at least 5 min. Then, one by one, the preganglionic inputs were cut or, in a few cases, firmly crushed with watchmaker's forceps. Sufficient time (1–2 min) was allowed between each cut to establish a new level of activity. In 6 out of 8 cases, the sequence was rostro-caudal: T1, T2, T3, and then below T3. In the remaining 2 rats the T2 segment was cut first. 2.7. Data analysis The cardiac sympathetic whole nerve signal was rectified and smoothed (time constant 0.1 or 0.02 s). For each experimental segment, the mean rectified signal was measured over a period of at least 1 min. The signal drop after each preganglionic each nerve section was quantified as a percentage of the mean, rectified neural signal. The neural component of the measurement was determined by subtraction of the non-neural component (noise and residual ECG pickup), which was measured at the end of the experiment after all preganglionic inputs had been cut.
2.4. Surgery The dorsal spine of the first thoracic segment was exposed and clamped with a hemostat. The hemostat was supported by a clamp stand in order to raise the thoracic spine and reduce ventilatory movements. The dorsal parts of the first four ribs on the left side were exposed and scraped clean of muscle. Taking care to avoid the intercostal blood vessels, the heads of the second to fourth ribs (occasionally also the first rib) were removed with bone rongeurs. The distal rib ends were ligated and retracted. These and the wound edges were stitched to a rigidly held metal ring positioned over the exposed area. The intervening tissue, intercostal muscle, and inconvenient blood vessels were carefully removed to expose the stellate ganglion and the upper thoracic sympathetic chain. Special care was taken not to puncture the pleura. Further dissection exposed the preganglionic connections of the first, second, and third thoracic segments. These, and the sympathetic trunk below T3, were marked by placing loose black silk filaments underneath them for later sectioning. One or two cardiac sympathetic nerves were identified as they left the caudo-lateral aspect of the stellate ganglion, coursing through the pleura to the heart: if two, the largest was chosen for recording.
3. Results Each panel in Fig. 1 displays excerpts of the rectified and smoothed CSNA for each of the different conditions from one of the eight experimental animals. Ongoing CSNA took the form of bursts locked to the cardiac cycle, reflecting its strong barosensitivity. In comparison to the initial signal before any cuts, this particular rat showed no discernable change in CSNA in response to cutting the T1 input. However, after cutting the T2 input, the CSNA decreased by 21%. No change was observed after cutting T3. The remaining baseline activity after cutting the final input – the entire sympathetic trunk below T3 – was taken as the noise level (including ECG pickup). Strikingly, the contribution of each segmental input to CSNA varied considerably between animals (Fig. 2). In some, the strongest input came from T3, while in others it came from T4 and below. In every case two or more segments contributed materially to the signal. The mean contributions of each segment to CSNA from all experimental animals, displayed the following pattern: T1 contributed the least to the overall signal (9 ± 6.8%), T2 contributed slightly more (21 ± 6.8%), with the most contribution coming from T3 (30 ± 13.2%)
2.5. Cardiac sympathetic nerve recording The selected cardiac sympathetic nerve was separated from the surrounding connective tissue by blunt dissection with fine forceps. It was cut where it entered the pleura en route to the heart, leaving approximately 5 mm of the central end for recording. A very fine silk filament was tied to its cut end. The exposed region was then flooded with liquid paraffin. A stainless steel dental scraper with a 2 mm round end plate was placed on the pleura immediately next to the origin of the cardiac nerve leaving the stellate ganglion. This prevented the nerve recording from being displaced by cardiac and respiratory movements and minimized electrocardiogram (ECG) pickup, by serving as a reference ground for the recording. The cardiac sympathetic nerve was placed over a pair of 0.2 mm diameter silver wire electrodes, using the silk thread to anchor the nerve to the distal electrode. The signal was recorded differentially with a NeuroLog head stage and preamplifier (NL 900A, Digitimer Ltd, U.K.). It was amplified (×5000–20,000) and filtered (high-pass: 10–100 Hz; low-pass: 500–1000 Hz) before being recorded and displayed on computer (CED Micro1401 interface, Spike2 software Cambridge Electronic Design, Cambridge, U.K.), along with blood pressure, heart rate (derived from the blood pressure signal), respiratory pressure, end-tidal CO2, and rectal temperature.
Fig. 1. CSNA records (rectified, smoothed with 20 ms time constant) from a representative experiment. Excerpts are shown before and after cutting the segmental preganglionic inputs, as indicated on the left. The record at the bottom (0%) represents the noise and residual ECG pickup after all inputs were disconnected.
N.H. Pracejus et al. / Autonomic Neuroscience: Basic and Clinical 187 (2015) 45–49
47
100
Percentage (%)
75 Rat 1 Rat 2 Rat 3 Rat 4 Rat 5 Rat 6 Rat 7 Rat 8
50
25
0
T1
T2
Input
T3
T4+
Fig. 2. Calculated percentage segmental contributions to CSNA in each of the 8 animals, as indicated.
or from T4 and below (39 ± 11.8%). Error bars represent the standard error of mean (SEM) (Fig. 3). Examining arterial pressure during the cutting sequence revealed little change in 4 rats (−1 to + 3 mmHg) but a fall in pressure in the remaining 4 rats (–6 to –19 mm Hg). 4. Discussion The three new findings of this study are 1) ongoing CSNA originates from more than one spinal segment in rats, 2) the overall mean segmental contribution to CSNA was the least from T1 and the most from T3 and below, 3) despite this trend, and most importantly, there was a substantial difference between rats in the individual segmental contributions to CSNA. 4.1. Findings from other physiological studies Compared to physiological studies investigating the supply to the cardiac sympathetic nerves in other mammals, the present findings for the rat show both similarities and differences. In cats, information from different approaches has implicated the third and fourth thoracic segments (T3, T4) as playing the dominant role in sympathetic control
of the heart. Szulczyk and Szulczyk (1987) and Kamonsińska et al. (1991) successively stimulated the first five thoracic white rami (T1–T5) and measured the magnitude of the evoked activity within cardiac sympathetic nerves. Both these studies found that the T3 outflow made the largest contribution, closely followed by the T4 outflow. However, whereas this method sheds light into which segments contain neurons with synaptic connections to postganglionic cardiac sympathetic nerves, it does not reveal which of these are physiologically active. In order to measure the physiological contribution of different spinal segments, Ninomiya et al. (1993) quantified the decline in ongoing CSNA after successively cutting each of the first five thoracic white rami. They too found that the largest segmental contribution came from T3 and the smallest from T1 and T5. In a similar experiment, Kocsis and Gyimesi-Pelczer (1998) sequentially cut white rami T1–T8 and found the main contribution to CSNA from T3–T5. In cats, therefore, it is agreed that the main drive to CSNA comes from the T3 and T4 segments, with inter-individual differences being reported in only one of the studies (see Kocsis and Gyimesi-Pelczer, 1998). In dogs, Mizeres (1958) sequentially stimulated the T1–T6 white rami, and obtained marked cardioacceleration and cardioaugmentation from the right and left T2 and T3 rami. Norris et al. (1974, 1977) stimulated the T1–T5 ventral roots. They found that the left and right T1 and T2
Fig. 3. The mean percentage contributions to CSNA over all 8 rats. Error bars show SEM.
48
N.H. Pracejus et al. / Autonomic Neuroscience: Basic and Clinical 187 (2015) 45–49
roots caused the greatest contractile response, while the right T1–T3 roots generated the largest cardiac acceleration. In both these studies, the strongest responses were elicited after stimulation of the T2 root. However, all of these canine studies noted inter-individual variability. In pigeons, MacDonald and Cohen (1970) stimulated the lower three cervical ganglia and upper four thoracic ganglia, and found the strongest cardiac effects evoked from T1–T3. Inotropic effects were strongest after stimulating the left side and chronotropic actions were strongest from the right. Yet it should be noted that this approach does not directly expose the individual spinal segments responsible, as each ganglion receives preganglionic inputs from several segments. These experiments also noted inter-individual differences in response sizes. Overall, these physiological studies indicate that 1) more than one spinal segment provides the sympathetic supply to the heart, and 2) the dominant segments are always in the upper thoracic cord but they can differ between species and also within a species. 4.2. Findings from neuroanatomical studies Neuroanatomical studies with conventional retrograde tracers can only provide indirect information on our current question. In dogs, Armour and Hopkins (1981) found that the canine middle cervical ganglia received the largest amount of neuronal staining after injecting horseradish peroxidase (HRP) into previously physiologically identified cardiac nerves, with less neuronal labeling within the stellate, and very little in the superior cervical ganglion (Armour and Hopkins, 1981). Conversely, in rats, retrograde tracers injected into different areas of the heart labeled many cells in both the middle cervical and stellate ganglia (Guić et al., 2010; Pardini et al., 1989). Similar findings were made in cat (Kuo et al., 1984; Shih et al., 1985). For retrograde transport to preganglionic neurons, the stellate ganglion has generally been used as a surrogate for the cardiac sympathetic supply, but this approach is confounded by the fact that stellate ganglion cells supply many other targets besides the heart (see Kocsis and Gyimesi-Pelczer, 1998). Preganglionic neurons with inputs to the stellate ganglion were found ipsilaterally from C8–T9 in cats (Kuo et al., 1984; Pardini and Wurster, 1984), and from C8–T8 in rats, with the largest projection from T2 segment (Pyner and Coote, 1994; Strack et al., 1988). The most directly relevant anatomical data come from two studies using the retrograde transsynaptic tracer, PRV. Ter Horst et al. (1993) found preganglionic neurons in the T1–T6 segments of the spinal cord to be labeled after injecting PRV into the left ventricle. Additionally, after injection of this virus into several regions of the heart, labeled preganglionic neurons were found in segments T1–T7 (Ter Horst et al., 1996). Although these studies outline the range of segments with disynaptic connections to the heart, they do not give information on their functional strength, nor do they uncover which are used physiologically by the animal to provide cardiac sympathetic tone. Curiously also, these reports do not mention which segments contained the most labeled neurons (Ter Horst et al., 1993, 1996).
suppressed by anesthesia (Matsukawa et al., 1993), but no data exist on the specific effects of urethane on CSNA. Also, the circulatory state of the animal following extensive thoracic surgery and artificial ventilation is less than optimal, which might reflexly increase CSNA. Second, a possible relevant factor is the order in which segmental inputs were disconnected (usually from rostral to caudal). Most preganglionic inputs to neurons within sympathetic ganglia give rise to unitary EPSPs in the ganglion cell, which in turn either fire or fail to fire a postganglionic action potential: summation of two inputs is rare (Bratton et al., 2010; McLachlan, 2003). Presynaptic inputs can thus be considered to act independently of each other, and there should be no significant interaction between the inputs from different spinal segments, even if they contact the same ganglion cell. The sequence of disconnection would matter if there were some non-linear interaction between segments, but for the reasons discussed above, this is unlikely. On the other hand, BP fell during the cutting sequence in half of the rats, which could have reduced the ongoing level of baroreceptor inhibition of CSNA. In those cases, the level of CSNA later in the experiment may have been artificially high, reducing our estimate of inputs disconnected early in the sequence but inflating our estimate of the final input to be cut (T4 and below). Third, all recordings were made on the animal's left side, so only statements about the segmental supply to the left cardiac sympathetic nerve can be made. Other studies suggest that the left sympathetic supply predominantly influences cardiac inotropy rather than chronotropy (MacDonald and Cohen, 1970; Mizeres, 1958; Norris et al., 1974, 1977), presumably by its action on the ventricles. Whether the same principles apply to the right cardiac nerves and chronotropic control remains to be tested. Finally, these experiments were technically demanding and the possibility that the dissection sometimes damaged one or more of the white rami before they were formally cut cannot be excluded. However, any such damage would not have caused any consistent bias to the data, and the general finding that more than one segment contributes to CSNA remains clear. 5. Conclusion This study has found that laboratory rats of one strain, one sex, and on one side show a surprising variability in their segmental origins of the cardiac sympathetic supply. More than one segment always contributes materially to ongoing CSNA, but its strongest input may come from T1, T2, T3, or from T4 and below, although more commonly from the lower segments. These findings provide useful background information for future studies on the central control of the cardiac sympathetic supply in rats. Acknowledgments We thank David Trevaks for technical assistance. This study was supported by Australian Research Council grant DP 130104661 and by the Victorian Government Operational Infrastructure Support Program. RMcA was supported by a National Health and Medical Research Council Research Fellowship (566667).
4.3. Limitations First, for technical reasons, this study was conducted on urethaneanesthetized rather than conscious rats. Compared to other types of anesthetics, urethane seems to be the most suitable when it comes to investigating cardiovascular mechanisms. When looking at renal sympathetic nerve activity (RSNA), Matsukawa and Ninomiya (1989) found urethane anesthesia in conscious, instrumented cats to have no effect on heart rate (HR) or blood pressure (BP), but to cause a transient suppression of RSNA, while Shimokawa et al. (1998) noted an increase in RSNA, HR, and BP in chronically instrumented rats after urethane administration. In decerebrate rats, Sapru and Krieger (1979) reported a decrease in HR and BP after giving urethane. CSNA has been reported to be selectively
References Armour, J.A., Hopkins, D.A., 1981. Localization of sympathetic postganglionic neurons of physiologically identified cardiac nerves in the dog. J. Comp. Neurol. 202, 169–184. Bratton, B., Davies, P., Jänig, W., McAllen, R., 2010. Ganglionic transmission in a vasomotor pathway studied in vivo. J. Physiol. 588, 1647–1659. Guić, M.M., Košta, V., Aljinović, J., Sapunar, D., Grković, I., 2010. Characterization of spinal afferent neurons projecting to different chambers of the rat heart. Neurosci. Lett. 469, 314–318. Kamonsińska, B., Nowicki, D., Szulczyk, A., Szulczyk, P., 1991. Spinal segmental sympathetic outflow to cervical sympathetic trunk, vertebral nerve, inferior cardiac nerve and sympathetic fibres in the thoracic vagus. J. Auton. Nerv. Syst. 32, 199–204. Kocsis, B., Gyimesi-Pelczer, K., 1998. Spinal segments communicating resting sympathetic activity to postganglionic nerves of the stellate ganglion. Am. J. Physiol. Regul. Integr. Comp. Physiol. 275, R400–R409.
N.H. Pracejus et al. / Autonomic Neuroscience: Basic and Clinical 187 (2015) 45–49 Kostreva, D.R., Zuperku, E.J., Cusick, J.F., Kampine, J.P., 1977. Ventral root mapping of cardiac nerves in the canine using evoked potentials. Am. J. Physiol. Heart Circ. Physiol. 1, H590–H595. Kuo, D.C., Oravitz, J.J., DeGroat, W.C., 1984. Tracing of afferent and efferent pathways in the left inferior cardiac nerve of the cat using retrograde and transganglionic transport of horseradish peroxidase. Brain Res. 321, 111–118. Langley, J.N., 1892. On the origin from the spinal cord of the cervical and upper thoracic sympathetic fibres, with some observations on white and grey rami communicantes. Philos. Trans. R. Soc. B Biol. Sci. 183, 85–124. Lichtman, J.W., Purves, D., Yip, J.W., 1980. Innervation of sympathetic neurones in the guinea-pig thoracic chain. J. Physiol. 298, 285–299. MacDonald, R.L., Cohen, D.H., 1970. Cells of origin of sympathetic pre-and postganglionic cardioacceleratory fibers in the pigeon. J. Comp. Neurol. 140, 343–358. Matsukawa, K., Ninomiya, I., 1989. Anesthetic effects on tonic and reflex renal sympathetic nerve activity in awake cats. Am. J. Physiol. 256, R371–R378. Matsukawa, K., Ninomiya, I., Nishiura, N., 1993. Effects of anesthesia on cardiac and renal sympathetic nerve activities and plasma catecholamines. Am. J. Physiol. 265, R792–R797. McLachlan, E.M., 2003. Transmission of signals through sympathetic ganglia - modulation, integration or simply distribution? Acta Physiol. Scand. 177, 227–235. Mizeres, N.J., 1958. The origin and course of the cardioaccelerator fibers in the dog. Anat. Rec. 132, 261–279. Ninomiya, I., Malpas, S.C., Matsukawa, K., Shindo, T., Akiyama, T., 1993. The amplitude of synchronized cardiac sympathetic nerve activity reflects the number of activated pre- and postganglionic fbers in anesthetized cats. J. Auton. Nerv. Syst. 45, 139–147. Njå, A., Purves, D., 1977. Specific innervation of guinea-pig superior cervical ganglion cells by preganglionic fibres arising from different levels of the spinal cord. J. Physiol. 264, 565–583. Norris, J.E., Foreman, R.D., Wurster, R.D., 1974. Responses of the canine heart to stimulation of the first five ventral thoracic roots. Am. J. Physiol. 227, 9–12.
49
Norris, J.E., Lippincott, D., Wurster, R.D., 1977. Responses of canine endocardium to stimulation of the upper thoracic roots. Am. J. Physiol. 233, H655–H659. Pardini, B.J., Wurster, R.D., 1984. Identification of the sympathetic preganglionic pathway to the cat stellate ganglion. J. Auton. Nerv. Syst. 11, 13–25. Pardini, B.J., Lund, D.D., Schmid, P.G., 1989. Organization of the sympathetic postganglionic innervation of the rat heart. J. Auton. Nerv. Syst. 28, 193–202. Pyner, S., Coote, J.H., 1994. Evidence that sympathetic preganglionic neurones are arranged in target-specific columns in the thoracic spinal cord of the rat. J. Comp. Neurol. 342, 15–22. Rubin, E., Purves, D., 1980. Segmental organization of sympathetic preganglionic neurons in the mammalian spinal cord. J. Comp. Neurol. 192, 163–174. Sapru, H.N., Krieger, A.J., 1979. Cardiovascular and respiratory effects of some anesthetics in the decerebrate rat. Eur. J. Pharmacol. 53, 151–158. Shih, C.-J., Chuang, K.-S., Tsai, S.-H., Liu, J.-C., 1985. Horseradish peroxidase localization of the sympathetic postganglionic neurons innervating the cat heart. J. Auton. Nerv. Syst. 13, 179–189. Shimokawa, A., Kunitake, T., Takasaki, M., Kannan, H., 1998. Differential effects of anesthetic on sympathetic nerve activity and arterial baroreceptor reflex in chronically instrumented rats. J. Auton. Nerv. Syst. 72, 46–54. Strack, A.M., Sawyer, W.B., Marubio, L.M., Loewy, A.D., 1988. Spinal origin of sympathetic preganglionic neurons in the rat. Brain Res. 455, 187–191. Szulczyk, A., Szulczyk, P., 1987. Spinal segmental preganglionic outflow to cervical sympathetic trunk and postganglionic cardiac sympathetic nerves. Brain Res. 421, 127–134. Ter Horst, G.J., Van den Brink, A., Homminga, S.A., Hautvast, R.W., Rakhorst, G., Mettenleiter, T.C., De Jongste, M.J., Lie, K.I., Korf, J., 1993. Transneuronal viral labelling of rat heart left ventricle controlling pathways. Neuroreport 4, 1307–1310. Ter Horst, G.J., Hautvast, R.W., De Jongste, M.J., Korf, J., 1996. Neuroanatomy of cardiac activity-regulating circuitry: a transneuronal retrograde viral labelling study in the rat. Eur. J. Neurosci. 8, 2029–2041.
Contents lists available at ScienceDirect
Autonomic Neuroscience: Basic and Clinical journal homepage: www.elsevier.com/locate/autneu
Segmental origins of cardiac sympathetic nerve activity in rats Natasha H. Pracejus a, David G.S. Farmer a, Robin M. McAllen a,b,⁎ a b
Florey Institute of Neuroscience and Mental Health, University of Melbourne, Parkville, Victoria 3010, Australia Department of Anatomy and Neuroscience, University of Melbourne, Parkville, Victoria 3010, Australia
a r t i c l e
i n f o
Article history: Received 6 October 2014 Received in revised form 20 November 2014 Accepted 26 November 2014 Keywords: Preganglionic White ramus Sympathetic chain
a b s t r a c t The segmental origins of cardiac sympathetic nerve activity (CSNA) were investigated in 8 urethane-anesthetized, artificially ventilated rats. The left upper thoracic sympathetic chain was exposed retropleurally after removing the heads of the second to fourth ribs. The preganglionic inputs to the chain from segments T1–T3 and the trunk distal to T3 were marked for later sectioning. CSNA was recorded conventionally, amplified, rectified and smoothed. Its mean level was quantified before and after each preganglionic input was cut, usually in rostro-caudal sequence. The level after all inputs were cut (i.e. noise and residual ECG pickup) was subtracted from previous measurements. The signal decrement from cutting each preganglionic input was then calculated as a percentage. CSNA in all rats depended on preganglionic drive from two or more segments, which were not always contiguous. Over the population, most preganglionic drive came from T3 and below, while the least came from T1. But there was striking inter-individual variation, such that the strongest drive to CSNA in any one rat could come from T1, T2, T3, or below T3. These findings provide new functional data on the segmental origins of CSNA in rats. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Sympathetic preganglionic neurons originate in the spinal cord, from where they emerge segmentally to join the sympathetic chain via their corresponding white ramus, and synapse in ganglia to supply their respective postganglionic neurons (Langley, 1892; Rubin and Purves, 1980). A ganglion receives its preganglionic input from multiple spinal segments, therefore multiple segments normally provide the supply destined for a target organ (Kamonsińska et al., 1991; Langley, 1892). Thus, each contributing spinal segment drives a proportion of the overall sympathetic traffic to a given target such as the heart (Lichtman et al., 1980; Njå and Purves, 1977). Previous physiological studies have established that the main segmental contribution to the cardiac sympathetic supply differs between species. In dogs, for instance, the strongest inputs appear to come from the first to third thoracic roots (Kostreva et al., 1977; Norris et al., 1974, 1977). Conversely, in cats the third and fourth thoracic segments contribute most to the cardiac sympathetic supply (Kamonsińska et al., 1991; Kocsis and Gyimesi-Pelczer, 1998; Ninomiya et al., 1993; Szulczyk and Szulczyk, 1987). As far as we are aware, no comparable data exist for the laboratory rat. Anatomical studies, analyzing transsynaptic retrograde labeling after pseudorabies virus (PRV) injection into the heart, indicate that the upper six or seven thoracic segments could contribute to the cardiac sympathetic supply (Ter Horst et al., 1993, 1996). However, those reports do ⁎ Corresponding author at: Florey Institute, University of Melbourne, Parkville, Vic. 3010, Australia. E-mail address: rmca@florey.edu.au (R.M. McAllen).
http://dx.doi.org/10.1016/j.autneu.2014.11.011 1566-0702/© 2014 Elsevier B.V. All rights reserved.
not specify the relative labeling within those segments, nor can such anatomical data reveal the strength of physiological drive from each segment. To resolve the latter issue, we adopted the method that Ninomiya et al. (1993) and Kocsis and Gyimesi-Pelczer (1998) used in cats, sequentially cutting the upper thoracic white rami, to determine the contribution of each spinal segment to the cardiac sympathetic nerve signal. 2. Methods 2.1. Animals Eight male Sprague–Dawley rats (320–440 g) were used. All animals were on a 12/12 light–dark cycle, with ad libitum access to food and water, and were housed with at least one other animal. Experimental approval was acquired from the Animal Ethics Committee of The Florey Institute of Neuroscience and Mental Health. 2.2. Anesthesia Before experimentation, each rat was anesthetized with Pentobarbital sodium (60 mg/kg, i.p.). A tracheostomy was then inserted to artificially ventilate the rat with 2% Isoflurane in oxygen, which was continued for the duration of preparatory surgery. After the surgery, anesthesia was switched to urethane (1.4 g/kg, i.v.) over approximately 30 min, while progressively discontinuing Isoflurane. Throughout the experiment, anesthesia was maintained at a depth that abolished corneal and withdrawal reflexes by giving supplementary doses of urethane if
46
N.H. Pracejus et al. / Autonomic Neuroscience: Basic and Clinical 187 (2015) 45–49
required (10% of initial dose). At the end of the experiment, the rat was killed with an overdose of Pentobarbital (160 mg/kg, i.v.).
2.3. Experimental preparation and monitoring Throughout the experiment, the animal was artificially ventilated with oxygen at 70 inflations per minute and a tidal volume of 3–4 ml (rodent ventilator, Ugo Basile, Italy). Respiratory pressure and expired CO2 levels were monitored continuously (Capstar 100 Carbon Dioxide Analyzer, USA). Minor adjustments of rate or volume were made to keep end-tidal CO2 levels close to 4%. Catheters were inserted into the right femoral artery and vein for measurement of blood pressure and administration of fluids, respectively. The venous catheter was filled with normal saline and the arterial catheter with 20 U/ml heparinized normal saline. On occasion, 1–2 ml of Gelofusine (B. Braun, Germany) was given i.v. to support blood pressure. A rectal probe monitored the animal's temperature, which was kept near 37 °C with a heating pad.
2.6. Experimental procedure For the experiment, cardiac sympathetic nerve activity (CSNA) was recorded until a stable level could be maintained for at least 5 min. Then, one by one, the preganglionic inputs were cut or, in a few cases, firmly crushed with watchmaker's forceps. Sufficient time (1–2 min) was allowed between each cut to establish a new level of activity. In 6 out of 8 cases, the sequence was rostro-caudal: T1, T2, T3, and then below T3. In the remaining 2 rats the T2 segment was cut first. 2.7. Data analysis The cardiac sympathetic whole nerve signal was rectified and smoothed (time constant 0.1 or 0.02 s). For each experimental segment, the mean rectified signal was measured over a period of at least 1 min. The signal drop after each preganglionic each nerve section was quantified as a percentage of the mean, rectified neural signal. The neural component of the measurement was determined by subtraction of the non-neural component (noise and residual ECG pickup), which was measured at the end of the experiment after all preganglionic inputs had been cut.
2.4. Surgery The dorsal spine of the first thoracic segment was exposed and clamped with a hemostat. The hemostat was supported by a clamp stand in order to raise the thoracic spine and reduce ventilatory movements. The dorsal parts of the first four ribs on the left side were exposed and scraped clean of muscle. Taking care to avoid the intercostal blood vessels, the heads of the second to fourth ribs (occasionally also the first rib) were removed with bone rongeurs. The distal rib ends were ligated and retracted. These and the wound edges were stitched to a rigidly held metal ring positioned over the exposed area. The intervening tissue, intercostal muscle, and inconvenient blood vessels were carefully removed to expose the stellate ganglion and the upper thoracic sympathetic chain. Special care was taken not to puncture the pleura. Further dissection exposed the preganglionic connections of the first, second, and third thoracic segments. These, and the sympathetic trunk below T3, were marked by placing loose black silk filaments underneath them for later sectioning. One or two cardiac sympathetic nerves were identified as they left the caudo-lateral aspect of the stellate ganglion, coursing through the pleura to the heart: if two, the largest was chosen for recording.
3. Results Each panel in Fig. 1 displays excerpts of the rectified and smoothed CSNA for each of the different conditions from one of the eight experimental animals. Ongoing CSNA took the form of bursts locked to the cardiac cycle, reflecting its strong barosensitivity. In comparison to the initial signal before any cuts, this particular rat showed no discernable change in CSNA in response to cutting the T1 input. However, after cutting the T2 input, the CSNA decreased by 21%. No change was observed after cutting T3. The remaining baseline activity after cutting the final input – the entire sympathetic trunk below T3 – was taken as the noise level (including ECG pickup). Strikingly, the contribution of each segmental input to CSNA varied considerably between animals (Fig. 2). In some, the strongest input came from T3, while in others it came from T4 and below. In every case two or more segments contributed materially to the signal. The mean contributions of each segment to CSNA from all experimental animals, displayed the following pattern: T1 contributed the least to the overall signal (9 ± 6.8%), T2 contributed slightly more (21 ± 6.8%), with the most contribution coming from T3 (30 ± 13.2%)
2.5. Cardiac sympathetic nerve recording The selected cardiac sympathetic nerve was separated from the surrounding connective tissue by blunt dissection with fine forceps. It was cut where it entered the pleura en route to the heart, leaving approximately 5 mm of the central end for recording. A very fine silk filament was tied to its cut end. The exposed region was then flooded with liquid paraffin. A stainless steel dental scraper with a 2 mm round end plate was placed on the pleura immediately next to the origin of the cardiac nerve leaving the stellate ganglion. This prevented the nerve recording from being displaced by cardiac and respiratory movements and minimized electrocardiogram (ECG) pickup, by serving as a reference ground for the recording. The cardiac sympathetic nerve was placed over a pair of 0.2 mm diameter silver wire electrodes, using the silk thread to anchor the nerve to the distal electrode. The signal was recorded differentially with a NeuroLog head stage and preamplifier (NL 900A, Digitimer Ltd, U.K.). It was amplified (×5000–20,000) and filtered (high-pass: 10–100 Hz; low-pass: 500–1000 Hz) before being recorded and displayed on computer (CED Micro1401 interface, Spike2 software Cambridge Electronic Design, Cambridge, U.K.), along with blood pressure, heart rate (derived from the blood pressure signal), respiratory pressure, end-tidal CO2, and rectal temperature.
Fig. 1. CSNA records (rectified, smoothed with 20 ms time constant) from a representative experiment. Excerpts are shown before and after cutting the segmental preganglionic inputs, as indicated on the left. The record at the bottom (0%) represents the noise and residual ECG pickup after all inputs were disconnected.
N.H. Pracejus et al. / Autonomic Neuroscience: Basic and Clinical 187 (2015) 45–49
47
100
Percentage (%)
75 Rat 1 Rat 2 Rat 3 Rat 4 Rat 5 Rat 6 Rat 7 Rat 8
50
25
0
T1
T2
Input
T3
T4+
Fig. 2. Calculated percentage segmental contributions to CSNA in each of the 8 animals, as indicated.
or from T4 and below (39 ± 11.8%). Error bars represent the standard error of mean (SEM) (Fig. 3). Examining arterial pressure during the cutting sequence revealed little change in 4 rats (−1 to + 3 mmHg) but a fall in pressure in the remaining 4 rats (–6 to –19 mm Hg). 4. Discussion The three new findings of this study are 1) ongoing CSNA originates from more than one spinal segment in rats, 2) the overall mean segmental contribution to CSNA was the least from T1 and the most from T3 and below, 3) despite this trend, and most importantly, there was a substantial difference between rats in the individual segmental contributions to CSNA. 4.1. Findings from other physiological studies Compared to physiological studies investigating the supply to the cardiac sympathetic nerves in other mammals, the present findings for the rat show both similarities and differences. In cats, information from different approaches has implicated the third and fourth thoracic segments (T3, T4) as playing the dominant role in sympathetic control
of the heart. Szulczyk and Szulczyk (1987) and Kamonsińska et al. (1991) successively stimulated the first five thoracic white rami (T1–T5) and measured the magnitude of the evoked activity within cardiac sympathetic nerves. Both these studies found that the T3 outflow made the largest contribution, closely followed by the T4 outflow. However, whereas this method sheds light into which segments contain neurons with synaptic connections to postganglionic cardiac sympathetic nerves, it does not reveal which of these are physiologically active. In order to measure the physiological contribution of different spinal segments, Ninomiya et al. (1993) quantified the decline in ongoing CSNA after successively cutting each of the first five thoracic white rami. They too found that the largest segmental contribution came from T3 and the smallest from T1 and T5. In a similar experiment, Kocsis and Gyimesi-Pelczer (1998) sequentially cut white rami T1–T8 and found the main contribution to CSNA from T3–T5. In cats, therefore, it is agreed that the main drive to CSNA comes from the T3 and T4 segments, with inter-individual differences being reported in only one of the studies (see Kocsis and Gyimesi-Pelczer, 1998). In dogs, Mizeres (1958) sequentially stimulated the T1–T6 white rami, and obtained marked cardioacceleration and cardioaugmentation from the right and left T2 and T3 rami. Norris et al. (1974, 1977) stimulated the T1–T5 ventral roots. They found that the left and right T1 and T2
Fig. 3. The mean percentage contributions to CSNA over all 8 rats. Error bars show SEM.
48
N.H. Pracejus et al. / Autonomic Neuroscience: Basic and Clinical 187 (2015) 45–49
roots caused the greatest contractile response, while the right T1–T3 roots generated the largest cardiac acceleration. In both these studies, the strongest responses were elicited after stimulation of the T2 root. However, all of these canine studies noted inter-individual variability. In pigeons, MacDonald and Cohen (1970) stimulated the lower three cervical ganglia and upper four thoracic ganglia, and found the strongest cardiac effects evoked from T1–T3. Inotropic effects were strongest after stimulating the left side and chronotropic actions were strongest from the right. Yet it should be noted that this approach does not directly expose the individual spinal segments responsible, as each ganglion receives preganglionic inputs from several segments. These experiments also noted inter-individual differences in response sizes. Overall, these physiological studies indicate that 1) more than one spinal segment provides the sympathetic supply to the heart, and 2) the dominant segments are always in the upper thoracic cord but they can differ between species and also within a species. 4.2. Findings from neuroanatomical studies Neuroanatomical studies with conventional retrograde tracers can only provide indirect information on our current question. In dogs, Armour and Hopkins (1981) found that the canine middle cervical ganglia received the largest amount of neuronal staining after injecting horseradish peroxidase (HRP) into previously physiologically identified cardiac nerves, with less neuronal labeling within the stellate, and very little in the superior cervical ganglion (Armour and Hopkins, 1981). Conversely, in rats, retrograde tracers injected into different areas of the heart labeled many cells in both the middle cervical and stellate ganglia (Guić et al., 2010; Pardini et al., 1989). Similar findings were made in cat (Kuo et al., 1984; Shih et al., 1985). For retrograde transport to preganglionic neurons, the stellate ganglion has generally been used as a surrogate for the cardiac sympathetic supply, but this approach is confounded by the fact that stellate ganglion cells supply many other targets besides the heart (see Kocsis and Gyimesi-Pelczer, 1998). Preganglionic neurons with inputs to the stellate ganglion were found ipsilaterally from C8–T9 in cats (Kuo et al., 1984; Pardini and Wurster, 1984), and from C8–T8 in rats, with the largest projection from T2 segment (Pyner and Coote, 1994; Strack et al., 1988). The most directly relevant anatomical data come from two studies using the retrograde transsynaptic tracer, PRV. Ter Horst et al. (1993) found preganglionic neurons in the T1–T6 segments of the spinal cord to be labeled after injecting PRV into the left ventricle. Additionally, after injection of this virus into several regions of the heart, labeled preganglionic neurons were found in segments T1–T7 (Ter Horst et al., 1996). Although these studies outline the range of segments with disynaptic connections to the heart, they do not give information on their functional strength, nor do they uncover which are used physiologically by the animal to provide cardiac sympathetic tone. Curiously also, these reports do not mention which segments contained the most labeled neurons (Ter Horst et al., 1993, 1996).
suppressed by anesthesia (Matsukawa et al., 1993), but no data exist on the specific effects of urethane on CSNA. Also, the circulatory state of the animal following extensive thoracic surgery and artificial ventilation is less than optimal, which might reflexly increase CSNA. Second, a possible relevant factor is the order in which segmental inputs were disconnected (usually from rostral to caudal). Most preganglionic inputs to neurons within sympathetic ganglia give rise to unitary EPSPs in the ganglion cell, which in turn either fire or fail to fire a postganglionic action potential: summation of two inputs is rare (Bratton et al., 2010; McLachlan, 2003). Presynaptic inputs can thus be considered to act independently of each other, and there should be no significant interaction between the inputs from different spinal segments, even if they contact the same ganglion cell. The sequence of disconnection would matter if there were some non-linear interaction between segments, but for the reasons discussed above, this is unlikely. On the other hand, BP fell during the cutting sequence in half of the rats, which could have reduced the ongoing level of baroreceptor inhibition of CSNA. In those cases, the level of CSNA later in the experiment may have been artificially high, reducing our estimate of inputs disconnected early in the sequence but inflating our estimate of the final input to be cut (T4 and below). Third, all recordings were made on the animal's left side, so only statements about the segmental supply to the left cardiac sympathetic nerve can be made. Other studies suggest that the left sympathetic supply predominantly influences cardiac inotropy rather than chronotropy (MacDonald and Cohen, 1970; Mizeres, 1958; Norris et al., 1974, 1977), presumably by its action on the ventricles. Whether the same principles apply to the right cardiac nerves and chronotropic control remains to be tested. Finally, these experiments were technically demanding and the possibility that the dissection sometimes damaged one or more of the white rami before they were formally cut cannot be excluded. However, any such damage would not have caused any consistent bias to the data, and the general finding that more than one segment contributes to CSNA remains clear. 5. Conclusion This study has found that laboratory rats of one strain, one sex, and on one side show a surprising variability in their segmental origins of the cardiac sympathetic supply. More than one segment always contributes materially to ongoing CSNA, but its strongest input may come from T1, T2, T3, or from T4 and below, although more commonly from the lower segments. These findings provide useful background information for future studies on the central control of the cardiac sympathetic supply in rats. Acknowledgments We thank David Trevaks for technical assistance. This study was supported by Australian Research Council grant DP 130104661 and by the Victorian Government Operational Infrastructure Support Program. RMcA was supported by a National Health and Medical Research Council Research Fellowship (566667).
4.3. Limitations First, for technical reasons, this study was conducted on urethaneanesthetized rather than conscious rats. Compared to other types of anesthetics, urethane seems to be the most suitable when it comes to investigating cardiovascular mechanisms. When looking at renal sympathetic nerve activity (RSNA), Matsukawa and Ninomiya (1989) found urethane anesthesia in conscious, instrumented cats to have no effect on heart rate (HR) or blood pressure (BP), but to cause a transient suppression of RSNA, while Shimokawa et al. (1998) noted an increase in RSNA, HR, and BP in chronically instrumented rats after urethane administration. In decerebrate rats, Sapru and Krieger (1979) reported a decrease in HR and BP after giving urethane. CSNA has been reported to be selectively
References Armour, J.A., Hopkins, D.A., 1981. Localization of sympathetic postganglionic neurons of physiologically identified cardiac nerves in the dog. J. Comp. Neurol. 202, 169–184. Bratton, B., Davies, P., Jänig, W., McAllen, R., 2010. Ganglionic transmission in a vasomotor pathway studied in vivo. J. Physiol. 588, 1647–1659. Guić, M.M., Košta, V., Aljinović, J., Sapunar, D., Grković, I., 2010. Characterization of spinal afferent neurons projecting to different chambers of the rat heart. Neurosci. Lett. 469, 314–318. Kamonsińska, B., Nowicki, D., Szulczyk, A., Szulczyk, P., 1991. Spinal segmental sympathetic outflow to cervical sympathetic trunk, vertebral nerve, inferior cardiac nerve and sympathetic fibres in the thoracic vagus. J. Auton. Nerv. Syst. 32, 199–204. Kocsis, B., Gyimesi-Pelczer, K., 1998. Spinal segments communicating resting sympathetic activity to postganglionic nerves of the stellate ganglion. Am. J. Physiol. Regul. Integr. Comp. Physiol. 275, R400–R409.
N.H. Pracejus et al. / Autonomic Neuroscience: Basic and Clinical 187 (2015) 45–49 Kostreva, D.R., Zuperku, E.J., Cusick, J.F., Kampine, J.P., 1977. Ventral root mapping of cardiac nerves in the canine using evoked potentials. Am. J. Physiol. Heart Circ. Physiol. 1, H590–H595. Kuo, D.C., Oravitz, J.J., DeGroat, W.C., 1984. Tracing of afferent and efferent pathways in the left inferior cardiac nerve of the cat using retrograde and transganglionic transport of horseradish peroxidase. Brain Res. 321, 111–118. Langley, J.N., 1892. On the origin from the spinal cord of the cervical and upper thoracic sympathetic fibres, with some observations on white and grey rami communicantes. Philos. Trans. R. Soc. B Biol. Sci. 183, 85–124. Lichtman, J.W., Purves, D., Yip, J.W., 1980. Innervation of sympathetic neurones in the guinea-pig thoracic chain. J. Physiol. 298, 285–299. MacDonald, R.L., Cohen, D.H., 1970. Cells of origin of sympathetic pre-and postganglionic cardioacceleratory fibers in the pigeon. J. Comp. Neurol. 140, 343–358. Matsukawa, K., Ninomiya, I., 1989. Anesthetic effects on tonic and reflex renal sympathetic nerve activity in awake cats. Am. J. Physiol. 256, R371–R378. Matsukawa, K., Ninomiya, I., Nishiura, N., 1993. Effects of anesthesia on cardiac and renal sympathetic nerve activities and plasma catecholamines. Am. J. Physiol. 265, R792–R797. McLachlan, E.M., 2003. Transmission of signals through sympathetic ganglia - modulation, integration or simply distribution? Acta Physiol. Scand. 177, 227–235. Mizeres, N.J., 1958. The origin and course of the cardioaccelerator fibers in the dog. Anat. Rec. 132, 261–279. Ninomiya, I., Malpas, S.C., Matsukawa, K., Shindo, T., Akiyama, T., 1993. The amplitude of synchronized cardiac sympathetic nerve activity reflects the number of activated pre- and postganglionic fbers in anesthetized cats. J. Auton. Nerv. Syst. 45, 139–147. Njå, A., Purves, D., 1977. Specific innervation of guinea-pig superior cervical ganglion cells by preganglionic fibres arising from different levels of the spinal cord. J. Physiol. 264, 565–583. Norris, J.E., Foreman, R.D., Wurster, R.D., 1974. Responses of the canine heart to stimulation of the first five ventral thoracic roots. Am. J. Physiol. 227, 9–12.
49
Norris, J.E., Lippincott, D., Wurster, R.D., 1977. Responses of canine endocardium to stimulation of the upper thoracic roots. Am. J. Physiol. 233, H655–H659. Pardini, B.J., Wurster, R.D., 1984. Identification of the sympathetic preganglionic pathway to the cat stellate ganglion. J. Auton. Nerv. Syst. 11, 13–25. Pardini, B.J., Lund, D.D., Schmid, P.G., 1989. Organization of the sympathetic postganglionic innervation of the rat heart. J. Auton. Nerv. Syst. 28, 193–202. Pyner, S., Coote, J.H., 1994. Evidence that sympathetic preganglionic neurones are arranged in target-specific columns in the thoracic spinal cord of the rat. J. Comp. Neurol. 342, 15–22. Rubin, E., Purves, D., 1980. Segmental organization of sympathetic preganglionic neurons in the mammalian spinal cord. J. Comp. Neurol. 192, 163–174. Sapru, H.N., Krieger, A.J., 1979. Cardiovascular and respiratory effects of some anesthetics in the decerebrate rat. Eur. J. Pharmacol. 53, 151–158. Shih, C.-J., Chuang, K.-S., Tsai, S.-H., Liu, J.-C., 1985. Horseradish peroxidase localization of the sympathetic postganglionic neurons innervating the cat heart. J. Auton. Nerv. Syst. 13, 179–189. Shimokawa, A., Kunitake, T., Takasaki, M., Kannan, H., 1998. Differential effects of anesthetic on sympathetic nerve activity and arterial baroreceptor reflex in chronically instrumented rats. J. Auton. Nerv. Syst. 72, 46–54. Strack, A.M., Sawyer, W.B., Marubio, L.M., Loewy, A.D., 1988. Spinal origin of sympathetic preganglionic neurons in the rat. Brain Res. 455, 187–191. Szulczyk, A., Szulczyk, P., 1987. Spinal segmental preganglionic outflow to cervical sympathetic trunk and postganglionic cardiac sympathetic nerves. Brain Res. 421, 127–134. Ter Horst, G.J., Van den Brink, A., Homminga, S.A., Hautvast, R.W., Rakhorst, G., Mettenleiter, T.C., De Jongste, M.J., Lie, K.I., Korf, J., 1993. Transneuronal viral labelling of rat heart left ventricle controlling pathways. Neuroreport 4, 1307–1310. Ter Horst, G.J., Hautvast, R.W., De Jongste, M.J., Korf, J., 1996. Neuroanatomy of cardiac activity-regulating circuitry: a transneuronal retrograde viral labelling study in the rat. Eur. J. Neurosci. 8, 2029–2041.
