- Email: [email protected]
Two types of rhythm in the respiratory network output in the isolated ventrolateral medulla in the neonatal rats
Neuroscience Letters 347 (2003) 53–56 www.elsevier.com/locate/neulet
Two types of rhythm in the respiratory network output in the isolated ventrolateral medulla in the neonatal rats Y.N. Shvareva,b,*, H. Lagercrantza, Y. Yamamotoa,c a
Department of Woman and Child Health, Q2: 07, Neonatal Research Unit, Astrid Lindgren Children’s Hospital, Karolinska Institutet, SE-171 76 Stockholm, Sweden b Behavioral Phenogenetics Laboratory, Institute of Cytology and Genetics, Siberian Division of the Russian Academy of Sciences, Novosibirsk 630090, Russia c Department of Anesthesiology and Intensive Care, Karolinska Hospital, SE-171 76 Stockholm, Sweden Received 12 March 2003; received in revised form 30 April 2003; accepted 27 May 2003
Abstract Effects of substance P and extracellular [Kþ]o on respiratory motor activity in the ventrolateral medulla in neonatal rat (0– 4 days old) brainstem-spinal cord preparation were studied. In addition to fictive eupneic rhythm (8 – 13 bursts/minute), the respiratory motor output was composed of biphasic bursts which might underlie the sigh pattern in vivo. These bursts had considerably lower frequency (0.15– 0.86 bursts/minute) and appeared when inspiratory neurons generated augmented biphasic discharges. The two rhythms were differently affected when the respiratory network excitability was increased by substance P or decreased by lowering external [Kþ]o, the effects on biphasic burst frequency being considerably greater. The augmented bursts could suppress inspiratory, but not pre-inspiratory neuron discharge, suggesting that pre-inspiratory neurons formed a supplementary rhythmic network which was not directly affected by biphasic burst generation. q 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Substance P; Respiration-related neurons; Respiration; [Kþ]o; Rhythm generation in vitro; Neonatal rat brainstem
During the last decades, the ventrolateral medulla (VLM), mainly its particular area called the pre-Bo¨tzinger complex (preBo¨tC), was suggested to be the crucial site for generation of respiratory rhythm in mammals [1,7]. However, the correspondence of the uniform rhythmic bursts of in vitro preparations to multiple rhythms underlying normal breathing in vivo has been controversial [1,10]. Recently, evidence was presented in slice preparation in mice that the respiratory network could generate multiple respiratory patterns. In population activity recorded in the preBo¨tzC and respiration-related neurons the patterns considered as fictive eupneic rhythm, sigh and gasp were demonstrated [3]. In this study we have found that respiratory output in the neonatal rat brainstem preparation in addition to eupneic rhythm, consisted of biphasic bursts which might underlie the sigh pattern in vivo. Both rhythms were differently affected when the respiratory network excitability was decreased by lower external [Kþ]o or increased by substance *
Corresponding author. Tel.: þ 46-8-517-77359; fax: þ46-8-517-77353. E-mail address: [email protected] (Y.N. Shvarev).
P (SP) which was earlier shown to activate respiratory system [1,2,8]. Experiments were performed on the brainstem-spinal cord preparation of newborn (0 – 4 days old) Wistar rat pup. The brainstem was dissected under deep ether anesthesia by cutting between the VIth cranial nerve roots and the lower border of the trapezoid body rostrally, and between 6th and 7th spinal nerves caudally. The isolated preparation was continuously perfused at the rate of 4.5 ml min21 in a 2-ml chamber with a solution containing (mM): 124 NaCl, 5 KCl, 1.2 KH2PO4, 2.4 CaCl2, 1.3 MgSO4, 26 NaHCO3, 30 glucose; equilibrated with 95% O2 and 5% CO2 at 288C. SP (Sigma) was applied by superfusion for 6 – 10 min in concentrations of 10 nM, 50 nM and 1 mM; pH was set at 7.4. In another set of experiments, the control perfusate with 6.2 mM [Kþ]o was replaced by 3.2 mM [Kþ]o for 30– 40 min. Respiratory activity was recorded from C4 ventral roots. Whole-cell records from respiratory neurons classified according to Onimaru and Homma [5] were obtained using the modified ‘blind’ patch-clamp technique, as described in detail previously [8]. The membrane potential was measured in a current-clamp mode using a CEZ-3100
0304-3940/03/$ - see front matter q 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0304-3940(03)00645-1
54
Y.N. Shvarev et al. / Neuroscience Letters 347 (2003) 53–56
voltage clamp amplifier (Nihon Hohden, Tokyo, Japan). Respiration-related neurons were found 50– 450 mm from the ventral surface, in the area between caudal and rostral part of the VLM [1]. The burst frequency was calculated as the number of C4 or respiratory neuron bursts per minute. Coefficient of variation was calculated over 9 min interval for eupneic rhythm and over five to seven cycles for type II bursts. Average data are given as mean ^ SEM. Significance level was set at P , 0:05. Two types of inspiratory pattern were observed under normal experimental conditions in C4 activity and in respiration-related neuron discharges in the VLM. The first pattern considered as in vitro eupneic rhythm [1] was represented by regular bursts with typical decrementing or bell-shaped pattern (Fig. 1) and frequency varying in a range of eight to thirteen bursts per minute (mean 10.8 ^ 2.3, coefficient of variation 0.026 ^ 0.002, n ¼ 24). In filtered intracellular recordings from inspiratory neurons (n ¼ 12) the eupneic pattern was represented by bell-shaped inspiratory drive potentials (Figs. 1D – F). The second pattern has not been described earlier in the brainstem preparation. In C4 activity it was characterized by biphasic shape including the first phase similar to eupneic burst and the second phase when the signal varied in amplitude. During this type of bursts the inspiratory drive potentials had a biphasic augmented shape, the later component having considerably bigger amplitude. In preinspiratory neurons (n ¼ 7) the excitatory drive potential during biphasic burst generation was not affected (Fig. 1H); inhibitory potentials in both pre-inspiratory and expiratory neurons (n ¼ 5) were clearly bigger (Figs. 1H,G). This type of respiratory activity was also regular, however the frequency was 10 – 30 times lower (Fig. 1I). The interburst interval varied in a range of 70 – 400 s corresponding to frequency of 0.15 – 0.86 bursts per minute with the coefficient of variation of 0.11 ^ 0.001 (n ¼ 11). The particular characteristic of the biphasic activity was its influence on the baseline burst periodicity. Visual data inspection distinguished four groups of observations in changing eupneic timing during biphasic burst generation (Fig. 1A). In order to analyze these observations we created scatter plots of the changes in eupneic burst cycle length around the biphasic bursts (Fig. 1B). Data seemed to group into four distinct clusters. We used cluster analysis to determine, whether these four groups were really separated or formed a diffuse single group. Cluster analysis showed that the eupneic cycle timing around biphasic bursts clustered into the discrete states (Figs. 1A,B; K-mean clustering, ANOVA, P , 0:0001). Cluster ‘b’ was formed by observations where the eupneic periods after the biphasic bursts, though being gradually delayed, were in a range of baseline period. At the same time other clusters consisted of observations where interburst periods following (‘a’), preceding (‘d’) or both following and preceding the biphasic
Fig. 1. Two types of respiratory rhythm in C4 roots and in respirationrelated neurons. (A) Here and throughout, asterisk labels the biphasic burst. In (a), traces represent the real and the integrated C4 activity. (b–d), only the integrated C4 data are depicted. T0 is the baseline eupneic cycle period. T1 is the duration of the eupneic cycle measured from the onset of the preceding eupneic burst until the onset of the biphasic burst. T2 is the time from the second peak of the biphasic burst to the onset of subsequent eupneic burst. T3 is the time from the onset of biphasic burst to the second peak of the same burst. (B, C) Scatter plots of the changes in the burst cycle length around the biphasic burst, represented as T2/T1 versus T1/T0 (B) and T2/T1 versus T3/T1 (C); T0 here is mean of two baseline eupneic cycles. Note several significantly separate groups of observations (K-mean clustering, ANOVA, P , 0:0001); the circles demonstrate the independent clusters. Fourteen records with three to seven biphasic bursts per record were pooled together giving 75 observations. ‘a’ and ‘d’ clusters in (C) represent the same observations as in (B), whereas the middle cluster in (C) is composed of the observations forming ‘b’ and ‘c’ clusters in (B). (D, H) Examples of C4 bursts and simultaneously recorded discharge patterns of inspiratory I (D), inspiratory II (E), inspiratory III (F), expiratory (G) and pre-inspiratory (H) neurons. The upper trace represents integrated C4 activity, the middle and lowest traces show membrane potential trajectories and their 2 Hz-low-pass filtered records, respectively. Note the presence of two distinct discharge patterns clearly visible in all pictures in integrated C4 activity and in filtered intracellular recordings. Inspiratory neurons discharge action potentials during inspiratory phrenic (C4) activity. Type I neurons receive excitatory postsynaptic potentials (EPSPs) prior to the onset and after the termination of inspiratory nerve bursts probably from pre-inspiratory neurons. Type II inspiratory neurons show EPSPs only during the inspiratory phase. Type III neurons are hyperpolarized by synchronized inhibitory postsynaptic potentials during both pre- and postinspiratory phase. Pre-inspiratory neurons are characterized by pre- and post-inspiratory excitation and inspiratory-related inhibition. Expiratory neurons discharge action potentials between the inspiratory phases, and are inhibited during the inspiratory phase. (I) Example of long-term whole-cell record from a type II inspiratory neuron.
bursts (‘c’) were divisible by two baseline periods. This means that one eupneic cycle was skipped. Biphasic bursts resemble the pattern recently described in rhythmic slice preparations in mice as in vitro analogue of
Y.N. Shvarev et al. / Neuroscience Letters 347 (2003) 53–56
sigh [3]. Authors reported that subsequent to sigh eupneic burst was gradually delayed by 1.5 –3 cycles and linearly depended on the phase of its occurrence within the eupneic cycle. We also obtained a significant correlation between the changes in baseline cycle around the biphasic burst and the phase of the second peak occurrence (T3/T1) (Fig. 1C, r ¼ 0:69, P , 0:0001, the regression line not shown). However, similar to Fig. 1B, the data grouped into three clusters (K-mean clustering, ANOVA, P , 0:0001) which again revealed existence of discrete cycle delays. Hence, the strong correlation seems to be defined by the shifting of the cluster centers. Thus, biphasic burst generation either slightly and gradually delays the following regular rhythmic burst, probably due to prolonged refractory period as consequence of augmented discharge or suppresses both preceding and following eupneic cycles. It is interesting that when eupneic cycles before and after biphasic bursts were considerably longer (‘a’, ‘d’ and ‘c’ clusters), six of seven pre-inspiratory neurons continued to discharge at the normal rate. However, four of five inspiratory III and four of five expiratory neurons received phasic inhibitory potentials (Fig. 2), which suggests that these neurons receive inhibitory synaptic inputs from preinspiratory neurons [5]. SP induced a biphasic effect on eupneic respiratory frequency: a pronounced decline (phase P1) followed by a weaker increase in burst rate (phase P2) (Fig. 3A). The transient frequency decrease coincided with increased excitability of respiration-related and tonic neurons and increased tonic discharges of motor C4 output, as described in detail earlier [8]. Frequency decrease was probably defined by phasic dissynchronization within respiratory network during mass activation of SP-sensitive neuronal groups in the VLM, while increase was associated with sustained membrane depolarization. At the same time, the biphasic burst frequency was monotonously and sharply increased, similar to changes in sigh frequency observed by Lieske et al. [3]. In three of eight cases normal respiratory rhythm was completely suppressed when SP of 1 mM was applied, so that respiration activity was represented exclusively by biphasic patterns (Fig. 3B). Frequency changes in both types of bursts depended on SP concentration (P , 0:001, one-way ANOVA), but the range of
Fig. 2. (A –C) Examples of activity of the respiration-related neurons and delay in respiratory cycle after biphasic burst with skipped eupneic cycle. From up to down, the traces represent the real and integrated C4 activity, membrane potential trajectories and their 2 Hz-low-pass filtered records, respectively. Arrows indicate the phasic inhibitory (A,C) and excitatory (B) drives at control frequency.
55
Fig. 3. Effects of SP on eupneic and biphasic burst frequency. (A) An example of changes in the C4 activity (upper panel) and in the membrane potential trajectory and discharge pattern (middle panel) in an inspiratory I neuron. The lowest panel represents the same intracellular record after 2 Hz low-pass filtration. ‘P1’ and ‘P2’ denote two phases in eupneic-like burst frequency changes, indicating a decrease in burst rate (P1) followed by its increase (P2). (B) In the inspiratory II neuron, example of switching from mixed respiratory rhythm to the rhythm represented exclusively by biphasic pattern, from up to down: C4 activity, integrated C4 activity and membrane potential trajectory.
changes was considerably bigger in biphasic burst (650%) as compared to eupneic pattern (42%) (P , 0:001, two-way ANOVA) (Fig. 4A). Decrease in external [Kþ]o from control level of 6.2mM to 3.2 mM produced significant frequency decrease in both types of rhythms (Fig. 4B). However, frequency changes in biphasic bursts were three times bigger than in eupneic ones (to 36 ^ 9% of control versus 81 ^ 3%, P , 0:001, t-test). It is interesting that during skipped cycle all inspiratory neurons ceased to charge, while pre-inspiratory ones were firing at the regular rate. These observations and the lack of biphasic burst influence on excitatory drive potential suggest that pre-inspiratory neurons compose a supplemen-
Fig. 4. The differential effects of SP application (A) and external [Kþ]o decrease (B) on eupneic and biphasic burst frequency. (A) During P1 phase (white columns), the eupneic burst frequency dose-dependently decreased to 81 ^ 9.4, 84 ^ 6, 42 ^ 5.8%, while during P2 phase (grey columns) it increased to 110 ^ 6.4, 119 ^ 4, 119 ^ 3%, when SP of 10 nM (n ¼ 9), 50 nM (n ¼ 7) and 1 mM (n ¼ 8) was applied (P , 0:0001, two-way ANOVA). At the same time, biphasic burst frequency (black columns) dose-dependently increased to 269.6 ^ 51, 356.8 ^ 33 and 651 ^ 76%, correspondingly (P , 0:0001, one-way ANOVA). (B) 3.2 mM [Kþ]o (n ¼ 8) induced significant decrease in eupneic rhythm frequency, to 81 ^ 3% of control and in biphasic rhythm frequency, to 36 ^ 9% (P , 0:001, t-test).
56
Y.N. Shvarev et al. / Neuroscience Letters 347 (2003) 53–56
tary network which is not directly affected by biphasic burst generation. Our results are consistent with the report that respiratory output in the brainstem preparation depends on interaction of two coupled rhythmically active networks, one composed of rhythmic PreBo¨ tC inspiratory and another of pre-inspiratory neurons [4]. Pre-inspiratory neurons were shown to supply periodic drive to the inspiratory network, and in a case of transmission failure skipped respiratory cycles were observed. Thus, we found that the eupneic and biphasic burst patterns have different frequencies and sensitivity to changes in respiratory network excitability. Our results suggest also that within the rhythmic inspiratory network in the neonatal brainstem, the biphasic rhythm generation interferes with eupneic rhythmogenesis, leading to transient uncoupling of two networks responsible for rhythm generation with a time window within two baseline cycles. The underlying ionic basis for these changes is not known, nor whether two rhythms are generated by the same neuronal network. These questions need further investigation. Biphasic bursts behavior resembles the sigh pattern in intact mammals represented by periodic biphasic augmented burst occurring at low frequencies, followed by brief apnea [6]. This suggestion is supported by evidence that local perturbation in the pre-Bo¨ tC alone was sufficient to change the respiratory output pattern inducing sigh and gasp patterns in normoxic deafferented animals [9]. In our experiments, biphasic bursts in the respiratory output had no stable augmented shape, and the eupneic cycle delay around them tended to be transposed into several discrete states. However, this pattern, as with eupneic bursts, seems to result from restricted experimental conditions when afferent inputs were removed [1,7]. In conclusion, the respiratory motor output in the neonatal brainstem preparation in addition to eupneic rhythm was composed of biphasic bursts which might underlie the sigh pattern in vivo. These biphasic bursts had considerably lower frequency and appeared when inspiratory neurons generated augmented biphasic discharges. Two rhythms were differently affected when the respiratory network excitability was changed, the effects on biphasic burst frequency being considerably greater. The augmented
bursts could suppress inspiratory, but not pre-inspiratory neuron discharges, suggesting that pre-inspiratory neurons formed a supplementary rhythmic network which was not directly affected by biphasic burst generation.
Acknowledgements We thank Drs Julia Skrinskaya, David Parker and Zoltan Nagy for help with the manuscript. This study was supported by grants from Sa¨ llskapet Barnava˚ rd Foundation (1999), Stiftelsen Frimurare Orden (2001) and Swedish Medical Research Council (MFR, 19X-05234-21DK).
References [1] K. Ballanyi, H. Onimaru, I. Homma, Respiratory network function in the isolated brainstem-spinal cord of newborn rats, Prog. Neurobiol. 59 (1999) 583–634. [2] P.A. Gray, J.C. Rekling, M. Bocchiaro, J.L. Feldman, Modulation of respiratory frequency by peptidergic input to rhythmogenic neurons in the pre-Bo¨ tzinger complex, Science 826 (1999) 1566–1568. [3] S.P. Lieske, M. Thoby-Brisson, P. Telgkamp, J.M. Ramirez, Reconfiguration of the neural network controlling multiple breathing pattern: eupnea, sighs and gasps, Nat. Neurosci. 3 (2000) 600–607. [4] N.M. Mellen, W.A. Janczewski, C.M. Bocchiaro, J.L. Feldman, Opioid-induced quantal slowing reveals dual networks for respiratory rhythm generation, Neuron 37 (2003) 821–826. [5] H. Onimaru, I. Homma, Whole cell recordings from respiratory neurons in the medulla of brainstem-spinal cord preparations isolated from newborn rats, Pflu¨ gers Arch. 420 (1992) 399– 406. [6] J. Orem, R.H. Trotter, Medullary respiratory neuronal activity during augmented breaths in intact unanesthetized cats, J. Appl. Physiol. 74 (1993) 761 –769. [7] W. Richter, K.M. Spyer, Studying rhythmogenesis of breathing: comparison of in vivo and in vitro models, Trends Neurosci. 24 (2001) 464 –472. [8] Y.N. Shvarev, H. Lagercrantz, Y. Yamamoto, Biphasic effects of substance P on respiratory activity and respiration-related neurons in ventrolateral medulla in the neonatal rat brainstem in vitro, Acta Physiol. Scand. 174 (2002) 67–84. [9] I.C. Solomon, Excitation of phrenic and sympathetic output during acute hypoxia: combination of medullary oxygen detectors, Resp. Physiol. 121 (2000) 1001–1117. [10] W.M. St.-John, J.F.R. Paton, Characterization of eupnea, apneusis and gasping in a perfused rat preparation, Resp. Physiol. 123 (2000) 201 –213.
Two types of rhythm in the respiratory network output in the isolated ventrolateral medulla in the neonatal rats Y.N. Shvareva,b,*, H. Lagercrantza, Y. Yamamotoa,c a
Department of Woman and Child Health, Q2: 07, Neonatal Research Unit, Astrid Lindgren Children’s Hospital, Karolinska Institutet, SE-171 76 Stockholm, Sweden b Behavioral Phenogenetics Laboratory, Institute of Cytology and Genetics, Siberian Division of the Russian Academy of Sciences, Novosibirsk 630090, Russia c Department of Anesthesiology and Intensive Care, Karolinska Hospital, SE-171 76 Stockholm, Sweden Received 12 March 2003; received in revised form 30 April 2003; accepted 27 May 2003
Abstract Effects of substance P and extracellular [Kþ]o on respiratory motor activity in the ventrolateral medulla in neonatal rat (0– 4 days old) brainstem-spinal cord preparation were studied. In addition to fictive eupneic rhythm (8 – 13 bursts/minute), the respiratory motor output was composed of biphasic bursts which might underlie the sigh pattern in vivo. These bursts had considerably lower frequency (0.15– 0.86 bursts/minute) and appeared when inspiratory neurons generated augmented biphasic discharges. The two rhythms were differently affected when the respiratory network excitability was increased by substance P or decreased by lowering external [Kþ]o, the effects on biphasic burst frequency being considerably greater. The augmented bursts could suppress inspiratory, but not pre-inspiratory neuron discharge, suggesting that pre-inspiratory neurons formed a supplementary rhythmic network which was not directly affected by biphasic burst generation. q 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Substance P; Respiration-related neurons; Respiration; [Kþ]o; Rhythm generation in vitro; Neonatal rat brainstem
During the last decades, the ventrolateral medulla (VLM), mainly its particular area called the pre-Bo¨tzinger complex (preBo¨tC), was suggested to be the crucial site for generation of respiratory rhythm in mammals [1,7]. However, the correspondence of the uniform rhythmic bursts of in vitro preparations to multiple rhythms underlying normal breathing in vivo has been controversial [1,10]. Recently, evidence was presented in slice preparation in mice that the respiratory network could generate multiple respiratory patterns. In population activity recorded in the preBo¨tzC and respiration-related neurons the patterns considered as fictive eupneic rhythm, sigh and gasp were demonstrated [3]. In this study we have found that respiratory output in the neonatal rat brainstem preparation in addition to eupneic rhythm, consisted of biphasic bursts which might underlie the sigh pattern in vivo. Both rhythms were differently affected when the respiratory network excitability was decreased by lower external [Kþ]o or increased by substance *
Corresponding author. Tel.: þ 46-8-517-77359; fax: þ46-8-517-77353. E-mail address: [email protected] (Y.N. Shvarev).
P (SP) which was earlier shown to activate respiratory system [1,2,8]. Experiments were performed on the brainstem-spinal cord preparation of newborn (0 – 4 days old) Wistar rat pup. The brainstem was dissected under deep ether anesthesia by cutting between the VIth cranial nerve roots and the lower border of the trapezoid body rostrally, and between 6th and 7th spinal nerves caudally. The isolated preparation was continuously perfused at the rate of 4.5 ml min21 in a 2-ml chamber with a solution containing (mM): 124 NaCl, 5 KCl, 1.2 KH2PO4, 2.4 CaCl2, 1.3 MgSO4, 26 NaHCO3, 30 glucose; equilibrated with 95% O2 and 5% CO2 at 288C. SP (Sigma) was applied by superfusion for 6 – 10 min in concentrations of 10 nM, 50 nM and 1 mM; pH was set at 7.4. In another set of experiments, the control perfusate with 6.2 mM [Kþ]o was replaced by 3.2 mM [Kþ]o for 30– 40 min. Respiratory activity was recorded from C4 ventral roots. Whole-cell records from respiratory neurons classified according to Onimaru and Homma [5] were obtained using the modified ‘blind’ patch-clamp technique, as described in detail previously [8]. The membrane potential was measured in a current-clamp mode using a CEZ-3100
0304-3940/03/$ - see front matter q 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0304-3940(03)00645-1
54
Y.N. Shvarev et al. / Neuroscience Letters 347 (2003) 53–56
voltage clamp amplifier (Nihon Hohden, Tokyo, Japan). Respiration-related neurons were found 50– 450 mm from the ventral surface, in the area between caudal and rostral part of the VLM [1]. The burst frequency was calculated as the number of C4 or respiratory neuron bursts per minute. Coefficient of variation was calculated over 9 min interval for eupneic rhythm and over five to seven cycles for type II bursts. Average data are given as mean ^ SEM. Significance level was set at P , 0:05. Two types of inspiratory pattern were observed under normal experimental conditions in C4 activity and in respiration-related neuron discharges in the VLM. The first pattern considered as in vitro eupneic rhythm [1] was represented by regular bursts with typical decrementing or bell-shaped pattern (Fig. 1) and frequency varying in a range of eight to thirteen bursts per minute (mean 10.8 ^ 2.3, coefficient of variation 0.026 ^ 0.002, n ¼ 24). In filtered intracellular recordings from inspiratory neurons (n ¼ 12) the eupneic pattern was represented by bell-shaped inspiratory drive potentials (Figs. 1D – F). The second pattern has not been described earlier in the brainstem preparation. In C4 activity it was characterized by biphasic shape including the first phase similar to eupneic burst and the second phase when the signal varied in amplitude. During this type of bursts the inspiratory drive potentials had a biphasic augmented shape, the later component having considerably bigger amplitude. In preinspiratory neurons (n ¼ 7) the excitatory drive potential during biphasic burst generation was not affected (Fig. 1H); inhibitory potentials in both pre-inspiratory and expiratory neurons (n ¼ 5) were clearly bigger (Figs. 1H,G). This type of respiratory activity was also regular, however the frequency was 10 – 30 times lower (Fig. 1I). The interburst interval varied in a range of 70 – 400 s corresponding to frequency of 0.15 – 0.86 bursts per minute with the coefficient of variation of 0.11 ^ 0.001 (n ¼ 11). The particular characteristic of the biphasic activity was its influence on the baseline burst periodicity. Visual data inspection distinguished four groups of observations in changing eupneic timing during biphasic burst generation (Fig. 1A). In order to analyze these observations we created scatter plots of the changes in eupneic burst cycle length around the biphasic bursts (Fig. 1B). Data seemed to group into four distinct clusters. We used cluster analysis to determine, whether these four groups were really separated or formed a diffuse single group. Cluster analysis showed that the eupneic cycle timing around biphasic bursts clustered into the discrete states (Figs. 1A,B; K-mean clustering, ANOVA, P , 0:0001). Cluster ‘b’ was formed by observations where the eupneic periods after the biphasic bursts, though being gradually delayed, were in a range of baseline period. At the same time other clusters consisted of observations where interburst periods following (‘a’), preceding (‘d’) or both following and preceding the biphasic
Fig. 1. Two types of respiratory rhythm in C4 roots and in respirationrelated neurons. (A) Here and throughout, asterisk labels the biphasic burst. In (a), traces represent the real and the integrated C4 activity. (b–d), only the integrated C4 data are depicted. T0 is the baseline eupneic cycle period. T1 is the duration of the eupneic cycle measured from the onset of the preceding eupneic burst until the onset of the biphasic burst. T2 is the time from the second peak of the biphasic burst to the onset of subsequent eupneic burst. T3 is the time from the onset of biphasic burst to the second peak of the same burst. (B, C) Scatter plots of the changes in the burst cycle length around the biphasic burst, represented as T2/T1 versus T1/T0 (B) and T2/T1 versus T3/T1 (C); T0 here is mean of two baseline eupneic cycles. Note several significantly separate groups of observations (K-mean clustering, ANOVA, P , 0:0001); the circles demonstrate the independent clusters. Fourteen records with three to seven biphasic bursts per record were pooled together giving 75 observations. ‘a’ and ‘d’ clusters in (C) represent the same observations as in (B), whereas the middle cluster in (C) is composed of the observations forming ‘b’ and ‘c’ clusters in (B). (D, H) Examples of C4 bursts and simultaneously recorded discharge patterns of inspiratory I (D), inspiratory II (E), inspiratory III (F), expiratory (G) and pre-inspiratory (H) neurons. The upper trace represents integrated C4 activity, the middle and lowest traces show membrane potential trajectories and their 2 Hz-low-pass filtered records, respectively. Note the presence of two distinct discharge patterns clearly visible in all pictures in integrated C4 activity and in filtered intracellular recordings. Inspiratory neurons discharge action potentials during inspiratory phrenic (C4) activity. Type I neurons receive excitatory postsynaptic potentials (EPSPs) prior to the onset and after the termination of inspiratory nerve bursts probably from pre-inspiratory neurons. Type II inspiratory neurons show EPSPs only during the inspiratory phase. Type III neurons are hyperpolarized by synchronized inhibitory postsynaptic potentials during both pre- and postinspiratory phase. Pre-inspiratory neurons are characterized by pre- and post-inspiratory excitation and inspiratory-related inhibition. Expiratory neurons discharge action potentials between the inspiratory phases, and are inhibited during the inspiratory phase. (I) Example of long-term whole-cell record from a type II inspiratory neuron.
bursts (‘c’) were divisible by two baseline periods. This means that one eupneic cycle was skipped. Biphasic bursts resemble the pattern recently described in rhythmic slice preparations in mice as in vitro analogue of
Y.N. Shvarev et al. / Neuroscience Letters 347 (2003) 53–56
sigh [3]. Authors reported that subsequent to sigh eupneic burst was gradually delayed by 1.5 –3 cycles and linearly depended on the phase of its occurrence within the eupneic cycle. We also obtained a significant correlation between the changes in baseline cycle around the biphasic burst and the phase of the second peak occurrence (T3/T1) (Fig. 1C, r ¼ 0:69, P , 0:0001, the regression line not shown). However, similar to Fig. 1B, the data grouped into three clusters (K-mean clustering, ANOVA, P , 0:0001) which again revealed existence of discrete cycle delays. Hence, the strong correlation seems to be defined by the shifting of the cluster centers. Thus, biphasic burst generation either slightly and gradually delays the following regular rhythmic burst, probably due to prolonged refractory period as consequence of augmented discharge or suppresses both preceding and following eupneic cycles. It is interesting that when eupneic cycles before and after biphasic bursts were considerably longer (‘a’, ‘d’ and ‘c’ clusters), six of seven pre-inspiratory neurons continued to discharge at the normal rate. However, four of five inspiratory III and four of five expiratory neurons received phasic inhibitory potentials (Fig. 2), which suggests that these neurons receive inhibitory synaptic inputs from preinspiratory neurons [5]. SP induced a biphasic effect on eupneic respiratory frequency: a pronounced decline (phase P1) followed by a weaker increase in burst rate (phase P2) (Fig. 3A). The transient frequency decrease coincided with increased excitability of respiration-related and tonic neurons and increased tonic discharges of motor C4 output, as described in detail earlier [8]. Frequency decrease was probably defined by phasic dissynchronization within respiratory network during mass activation of SP-sensitive neuronal groups in the VLM, while increase was associated with sustained membrane depolarization. At the same time, the biphasic burst frequency was monotonously and sharply increased, similar to changes in sigh frequency observed by Lieske et al. [3]. In three of eight cases normal respiratory rhythm was completely suppressed when SP of 1 mM was applied, so that respiration activity was represented exclusively by biphasic patterns (Fig. 3B). Frequency changes in both types of bursts depended on SP concentration (P , 0:001, one-way ANOVA), but the range of
Fig. 2. (A –C) Examples of activity of the respiration-related neurons and delay in respiratory cycle after biphasic burst with skipped eupneic cycle. From up to down, the traces represent the real and integrated C4 activity, membrane potential trajectories and their 2 Hz-low-pass filtered records, respectively. Arrows indicate the phasic inhibitory (A,C) and excitatory (B) drives at control frequency.
55
Fig. 3. Effects of SP on eupneic and biphasic burst frequency. (A) An example of changes in the C4 activity (upper panel) and in the membrane potential trajectory and discharge pattern (middle panel) in an inspiratory I neuron. The lowest panel represents the same intracellular record after 2 Hz low-pass filtration. ‘P1’ and ‘P2’ denote two phases in eupneic-like burst frequency changes, indicating a decrease in burst rate (P1) followed by its increase (P2). (B) In the inspiratory II neuron, example of switching from mixed respiratory rhythm to the rhythm represented exclusively by biphasic pattern, from up to down: C4 activity, integrated C4 activity and membrane potential trajectory.
changes was considerably bigger in biphasic burst (650%) as compared to eupneic pattern (42%) (P , 0:001, two-way ANOVA) (Fig. 4A). Decrease in external [Kþ]o from control level of 6.2mM to 3.2 mM produced significant frequency decrease in both types of rhythms (Fig. 4B). However, frequency changes in biphasic bursts were three times bigger than in eupneic ones (to 36 ^ 9% of control versus 81 ^ 3%, P , 0:001, t-test). It is interesting that during skipped cycle all inspiratory neurons ceased to charge, while pre-inspiratory ones were firing at the regular rate. These observations and the lack of biphasic burst influence on excitatory drive potential suggest that pre-inspiratory neurons compose a supplemen-
Fig. 4. The differential effects of SP application (A) and external [Kþ]o decrease (B) on eupneic and biphasic burst frequency. (A) During P1 phase (white columns), the eupneic burst frequency dose-dependently decreased to 81 ^ 9.4, 84 ^ 6, 42 ^ 5.8%, while during P2 phase (grey columns) it increased to 110 ^ 6.4, 119 ^ 4, 119 ^ 3%, when SP of 10 nM (n ¼ 9), 50 nM (n ¼ 7) and 1 mM (n ¼ 8) was applied (P , 0:0001, two-way ANOVA). At the same time, biphasic burst frequency (black columns) dose-dependently increased to 269.6 ^ 51, 356.8 ^ 33 and 651 ^ 76%, correspondingly (P , 0:0001, one-way ANOVA). (B) 3.2 mM [Kþ]o (n ¼ 8) induced significant decrease in eupneic rhythm frequency, to 81 ^ 3% of control and in biphasic rhythm frequency, to 36 ^ 9% (P , 0:001, t-test).
56
Y.N. Shvarev et al. / Neuroscience Letters 347 (2003) 53–56
tary network which is not directly affected by biphasic burst generation. Our results are consistent with the report that respiratory output in the brainstem preparation depends on interaction of two coupled rhythmically active networks, one composed of rhythmic PreBo¨ tC inspiratory and another of pre-inspiratory neurons [4]. Pre-inspiratory neurons were shown to supply periodic drive to the inspiratory network, and in a case of transmission failure skipped respiratory cycles were observed. Thus, we found that the eupneic and biphasic burst patterns have different frequencies and sensitivity to changes in respiratory network excitability. Our results suggest also that within the rhythmic inspiratory network in the neonatal brainstem, the biphasic rhythm generation interferes with eupneic rhythmogenesis, leading to transient uncoupling of two networks responsible for rhythm generation with a time window within two baseline cycles. The underlying ionic basis for these changes is not known, nor whether two rhythms are generated by the same neuronal network. These questions need further investigation. Biphasic bursts behavior resembles the sigh pattern in intact mammals represented by periodic biphasic augmented burst occurring at low frequencies, followed by brief apnea [6]. This suggestion is supported by evidence that local perturbation in the pre-Bo¨ tC alone was sufficient to change the respiratory output pattern inducing sigh and gasp patterns in normoxic deafferented animals [9]. In our experiments, biphasic bursts in the respiratory output had no stable augmented shape, and the eupneic cycle delay around them tended to be transposed into several discrete states. However, this pattern, as with eupneic bursts, seems to result from restricted experimental conditions when afferent inputs were removed [1,7]. In conclusion, the respiratory motor output in the neonatal brainstem preparation in addition to eupneic rhythm was composed of biphasic bursts which might underlie the sigh pattern in vivo. These biphasic bursts had considerably lower frequency and appeared when inspiratory neurons generated augmented biphasic discharges. Two rhythms were differently affected when the respiratory network excitability was changed, the effects on biphasic burst frequency being considerably greater. The augmented
bursts could suppress inspiratory, but not pre-inspiratory neuron discharges, suggesting that pre-inspiratory neurons formed a supplementary rhythmic network which was not directly affected by biphasic burst generation.
Acknowledgements We thank Drs Julia Skrinskaya, David Parker and Zoltan Nagy for help with the manuscript. This study was supported by grants from Sa¨ llskapet Barnava˚ rd Foundation (1999), Stiftelsen Frimurare Orden (2001) and Swedish Medical Research Council (MFR, 19X-05234-21DK).
References [1] K. Ballanyi, H. Onimaru, I. Homma, Respiratory network function in the isolated brainstem-spinal cord of newborn rats, Prog. Neurobiol. 59 (1999) 583–634. [2] P.A. Gray, J.C. Rekling, M. Bocchiaro, J.L. Feldman, Modulation of respiratory frequency by peptidergic input to rhythmogenic neurons in the pre-Bo¨ tzinger complex, Science 826 (1999) 1566–1568. [3] S.P. Lieske, M. Thoby-Brisson, P. Telgkamp, J.M. Ramirez, Reconfiguration of the neural network controlling multiple breathing pattern: eupnea, sighs and gasps, Nat. Neurosci. 3 (2000) 600–607. [4] N.M. Mellen, W.A. Janczewski, C.M. Bocchiaro, J.L. Feldman, Opioid-induced quantal slowing reveals dual networks for respiratory rhythm generation, Neuron 37 (2003) 821–826. [5] H. Onimaru, I. Homma, Whole cell recordings from respiratory neurons in the medulla of brainstem-spinal cord preparations isolated from newborn rats, Pflu¨ gers Arch. 420 (1992) 399– 406. [6] J. Orem, R.H. Trotter, Medullary respiratory neuronal activity during augmented breaths in intact unanesthetized cats, J. Appl. Physiol. 74 (1993) 761 –769. [7] W. Richter, K.M. Spyer, Studying rhythmogenesis of breathing: comparison of in vivo and in vitro models, Trends Neurosci. 24 (2001) 464 –472. [8] Y.N. Shvarev, H. Lagercrantz, Y. Yamamoto, Biphasic effects of substance P on respiratory activity and respiration-related neurons in ventrolateral medulla in the neonatal rat brainstem in vitro, Acta Physiol. Scand. 174 (2002) 67–84. [9] I.C. Solomon, Excitation of phrenic and sympathetic output during acute hypoxia: combination of medullary oxygen detectors, Resp. Physiol. 121 (2000) 1001–1117. [10] W.M. St.-John, J.F.R. Paton, Characterization of eupnea, apneusis and gasping in a perfused rat preparation, Resp. Physiol. 123 (2000) 201 –213.