European Journal of Neuroscience, Vol. 4 , p p . 183 - I88

0 1992 European Neuroscience Association

Ca2+-MediatedPlateau Potentials in a Subpopulation of lnterneurons in the Ventral Horn of the Turtle Spinal Cord J. Hounsgaard and 0. Kjaerulff Institute of Neurophysiology, Panum Institute, Blegdamsvej 3C, University of Copenhagen, DK-2200 Copenhagen N., Denmark

Key words: spinal networks, slice preparation

Abstract The response properties of interneurons in the ventral horn were studied in transverse slices of segments D8 to S2 from the turtle spinal cord, using the current clamp technique. In about half of the neurons the response properties were dominated by their ability to generate plateau potentials. In these cells the plateau potential could account for delayed onset of spiking and a phase of increasing spike frequency during depolarizing current pulses and for a depolarizing afterpotential following the stimulus. The cells usually received monosynaptic and polysynaptic input from the ipsilateral dorsal root and occasionally from the contralateral root. The plateau potential was insensitive to tetrodotoxin but blocked by nifedipine and by replacing Ca2+ with Co2+ in the medium. It is concluded that the response properties of neurons in the ventral horn outside the motor nucleus have differentiated response properties that may well contribute to spinal motor function. Introduction

In vertebrates the intrinsic functional capabilities of the spinal motor system have been thought to reside predominantly with the wiring pattern among excitatory and inhibitory ventral horn neurons (Baldissera er a l . , 1981; Grillner, 1981). At least for motoneurons, however, it is increasingly clear that voltage-sensitive ion channels in the somatic and dendritic membrane also contribute to the formation of spike patterns (Barrett and Barrett, 1976; Schwindt and Crill, 1984; Hounsgaard et a l . , 1988a) and constitute a potential target for synaptic modulation (Hounsgaard et al., 1988b; Hounsgaard and Kiehn, 1989; Kiehn, 1991). It seems of interest to explore in what form and to what extent postsynaptic response properties contribute to synaptic processing in other cell types in the spinal motor system. First, such properties are known to be of extreme importance in central pattern generators in invertebrates (Harris-Warrick, 1989; Getting, 1989). Secondly, active response properties have been demonstrated in neurons throughout the vertebrate brain (Jahnsen, 1986) and are thought to be of specific functional importance (LlinBs, 1988; Midtgaard and Hounsgaard, 1989). In the present paper we identify a subset of neurons with complex response properties in the proximal part of the ventral horn in the lumbar spinal cord of the turtle. These properties include delayed onset of spiking and initially accelerating firing frequency during depolarizing current pulses, as well as afterdischarge and overt bistability. We find that a nifedipine-sensitive Ca2+-mediated plateau potential is the Correspondence to: J . Hounsgaard, as above Received 9 April 1991, revised 30 May 1991, accepted 10 October 1991

main factor underlying these properties. The present findings suggest that active postsynaptic response properties, in addition to their role in motoneurons, contribute to the function of spinal networks in general. The results have been presented in an abstract (Kjaerulff and Hounsgaard, 1991). Materials and methods

The procedure to obtain and use transverse slice preparations from the turtle spinal cord has been described in detail elsewhere (Hounsgaard et a f . ,1988a; Hounsgaard and Nicholson, 1990). Turtles (Pseudemys scripra and Chrysemys picta, carapace length 10-20 cm) were anaesthetized with 50 mg pentobarbitone injected intraperitoneally, and perfused with 0.5 1 normal medium transcardially, followed by isolation of segments Dg-S2 of the lumbar enlargement. Slices cut transversely to a thickness of - 3 mm were transferred to small vials, one for each segment. During dissection and at all later stages, the preparations were kept in medium of the following composition (mM): NaCI, 120; KCl, 5; NaHC03, 15; CaC12, 3; MgCI2, 3; glucose, 20; this solution was saturated with carbogen containing 2 % C02 and 98 % 02. For an experiment a slice was mounted on end in the recording chamber and superfused with medium at room temperature. The funiculi and the dorsal and ventral horns were clearly visible, and, aided by

184 Plateau potentials in turtle interneurons

I0.5nA

a stereomicroscope, the recording electrode, mounted vertically, could be inserted in the recording area. In some experiments dorsal roots were stimulated using suction electrodes. Intracellular recordings, command voltages for current injection through a bridge circuit and sweep trigger signal were stored on FM tape for off-line analysis. Laser printer hard copies of relevant sweeps were obtained after digitization and formatting on a computer. Cells to be stained were injected iontophoretically with biocytin during intracellular recording. Electrodes contained 2 % biocytin in 1 M K-acetate. All cells in good condition after being injected with positive current for at least 5 min were reliably stained using the histological procedures described by others (Horikawa and Armstrong, 1988).

Results Electrophysiology

FIG. 1. Evoked spike patterns in non-plateau and plateau-generating cells of the ventral horn. (A) In non-plateau cells a depolarizing current pulse resulted in declining spike frequency during and after hyperpolarization following the pulse. (B, C) Response patterns of plateau-generating cells. The high threshold plateau in B is activated late in the spike train and gives rise to a phase of increasing spike frequency. In C the low threshold plateau is almost fully activated prior to onset of spike train. Recordings from three cells.

Recordings were obtained from the major part of areas V-VI and VII-VIII (Kusuma e t a / . , 1979). The recording area had a dorsal border at the level of the central canal and extended halfway into the ventral horn. The database consists of stable intracellular recordings from neurons with the ability to maintain firing throughout a depolarizing square current pulse lasting 3 s. Based on the firing pattern in response to such current pulses, the 171 neurons making the total database were divided into two groups. In group 1 neurons ( n = 82) the spike frequency declined from an early peak to a lower steady-state value, as shown in Figure 1A. The neurons of group 2 (n = 89) displayed response properties dominated by their ability to generate plateau potentials. The relative threshold for action potentials and plateau potentials varied considerably among the neurons in this group, as illustrated in Figure 1B and C. In some cells the threshold for the plateau potential was higher than the spike threshold, evidenced by the late onset of accelerating spike frequency during an evoked spike train (Fig. 1B). In other cells a plateau was recruited prior to the onset of firing (Fig. IC). There was no significant difference in spike amplitude between neurons of groups 1 and 2 (Student’s I-test, P > 10%). as shown in Table 1. This table also demonstrates that no gross differences were found when comparing spike duration, membrane potential and input resistance. In the remaining part of the paper we focus on the neurons with a plateau potential (group 2 neurons). Response patterns Characteristic response patterns for plateau-generating cells are illustrated in more detail in Figure 2. The response to current pulses

TABLE1. Electrophysiological parameters Total sample Groupb

High spike amplitudea

Number of

Spike amplitude‘ (mv)

neurons ~

1

2

82 89

~

Group

Spike durationd (ms)

Membrane potential (mV)

Input resistance (MfO

1

0.92b0.44 (n = 18) 0.84*0.33 In = 20)

57*7 (n = 7) 63*9 (n = 12)

102*53 (n 130*49 (n

~

65*1l 63*13

2

Values for spike amplitude, spike duration, membrane potential and input resistance are mean f SD. aValues measured from neurons with a spike amplitude equal to or higher than the 22% percentile (= 70 mV in both groups). bSee text. ‘Measured between the lower level of the fast spike afterhyperpolarization and the peak. dMeasured halfway between the threshold and the peak.

= =

11) 13)

Plateau potentials in turtle interneurons 185 of increasing amplitude is shown in Figure 2A. Well below threshold for action potentials, a steady level of depolarization was attained during depolarizing current pulses (lower sweep). Above threshold a depolarizing voltage trajectory is clearly seen before the first spike. This prepotential develops over several seconds at stimulus intensities near threshold, and increasingly rapidly with higher stimulus intensities. In the experiment illustrated in Figure 2B the net current applied during the stimulus was kept constant with the holding current at three different levels. During the stimulus it is apparent that the spike train is delayed and the phase of acceleration shortened with increasing depolarizing holding current. The most dramatic effect of holding current is on events immediately after the stimulus. In the records shown, these after-effects shift

C

--

A

- I0.5nA 1s

from a brief afterhyperpolarization in the lower trace to a transient afterdepolarization in the middle trace, and finally to sustained firing in the upper trace, as a consequence of increasing the depolarizing holding current. At holding currents that allowed sustained firing after a stimulus the resting membrane potential could be reinstated by a hyperpolarizing current pulse, as shown in Figure 2C. In this sense the cells were bistable. Synaptic input after doisal root stimulation

In the majority of cells tested synaptic responses could be evoked by stimulating the ipsilateral dorsal root. The most common response was a presumed monosynaptic excitatory potential graded with stimulus

1.5s

-

I0.5nA

3s

FIG. 2. Spike pattern in plateau-generating cell in response to change in stimulus current (A) and holding current (B). (C) illustrates bistability in response to sequential depolarizing and hyperpolarizing current pulses at depolarizing holding current. All records from the same cell. In A a holding current of -0.2 nA was applied throughout. The upper sweep in B and the sweep in C were obtained without holding current.

186 Plateau potentials in turtle interneurons

A

A

1 OOms

B

A 1 OOms

I0.5nA

A

3s FIG. 3. Synaptic responses evoked in plateau-generating cells by ipsilateral stimulation of dorsal root. (A) Monosynaptic response graded in amplitude by stimulus intensity. (B)Polysynaptic response graded in amplitude, duration and latency by stimulus intensity. (C) Sustained activity induced by depolarizing current pulse (upper sweep) and by polysynaptic excitation in response to a dorsal root stimulus. B and C show recordings from the same cell with a holding current of +O. 15 nA.

intensity, as shown in Figure 3A. This response had a short latency without 'jitter'. Often, however, additional (probably polysynaptic) components, graded in latency and amplitude with stimulus intensity, were also evoked (Fig. 3B). In such cases synaptic activity could persist for several hundred ms after a single stimulus, and sometimes even replace injected current as the stimulus for sustained firing in cells in the bistable mode (Fig. 3C).

Ca'+ -mediated plateau potentials In group 2 neurons, application of tetrodotoxin (1 pM) uncovered a plateau potential (n = 16). This potential developed during the stimulus with a time course similar to that of the subthreshold prepotential and the phase of accelerating spike frequency in normal medium (Fig. 4A, B). Note that the transient depolarizing afterpotential also persisted in the presence of tetrodotoxin. Application of 1 pM nifedipine (n = 2) or replacement of Ca2' with 3 mM Co2+ (n = 6)greatly reduced the plateau potential during the pulse as well as the afterdepolarization. This is shown for nifedipine in Figure 4C. In one additional experiment the phase of acceleration and the afterdepolarization seen in normal medium were abolished after application of 10 1M nifedipine, and a similar result was obtained in an experiment using 3 mM Co2+. Morphology Three neurons without a plateau and 11 plateau-generating cells were stained with biocytin-coupled horseradish peroxidase. Two examples, one from each group, are shown in Figure 5 . When comparing the

-

-

I0.5nA 2s

FIG. 4. Ionic mechanisms of plateau potential. (A) Control sweep in normal medium. (B) Persistent plateau after application of tetrodotoxin ( 1 pM). (C) Plateau reduced by nifedipine (10 pM).

Plateau potentials in turtle interneurons 187

Fic. 5 . Micrographs of two biocytin-labelled ventral horn neurons. (A) Neuron without plateau potentials. The cell body and the proximal dendrites are clearly seen whereas most of the axon is not in focus. In the insert a detail from the same cell is shown at slightly higher magnification. The axon is in focus and is seen to project contralaterally. (B)Plateau-generating neuron. Contralateral projection is evident as in A (arrowheads). Apart from the lack of a dorsally extending dendrite in B, the neurons in A and B are similar. Scale = 100 p n in A and B.

188 Plateau potentials in turtle interneurons cells, the first point to be noted is the similarity in dendritic ramification and distribution as well as in the location of the soma. Secondly, contralateral projection is seen in both neurons, and was also evident in another plateau-generating neuron (not shown). In seven neurons with plateau potentials and in one without, contralateral projection could be inferred from the course of the axon, although it was cut in or close to the anterior white commissure before actually crossing the midline. To summarize, Figure 5 illustrates our general impression that it was not possible to use morphological criteria to distinguish the plateau-generating neurons from those not exhibiting this ability.

Discussion Based on their response properties, we have defined a subpopulation of interneurons in the ventral horn of the lumbar spinal cord in the turtle. The cells are distinguished by the delayed onset of spiking and the initial acceleration of the spike frequency during a depolarizing current pulse, and by the presence of an afterdischarge and bistability for a certain range of holding current. These properties are most simply explained by the presence of a nifedipine-sensitive, Ca2+-mediated plateau potential. The response properties of the neurons under study and the underlying mechanisms are similar to those induced by serotonin in spinal motoneurons (Hounsgaard and Kiehn, 1989). It is of interest, however, that plateau potentials in interneurons were present without addition of promoting transmitters and K channel blockers, and were unaffected by the presence of tetrodotoxin. This suggests that the ability to generate plateau potentials may be a ground state for the cells studied here rather than a latent property, as in motoneurons (Hounsgaard and Kiehn, 1989; Hounsgaard and Nicholson, 1990). It is important to emphasize that our experiments are unable to tell whether modulatory transmitters exist that can shift cells between the plateau and the non-plateau-generating categories. In a pilot study, however, we were not able to convert non-plateau-generating cells to plateau-generating cells by application of apamin (n = 2), serotonin (n = 2) or substance P (n = 1) (unpublished observations). In general terms, plateau potentials facilitate shifts between states with high and low spike activity. A priori, neurons with these properties therefore seem useful in a network engaged in phasic activity. Interestingly, the majority of neurons in laminae 7 and 8 in the lumbar spinal cord of the cat are phasically active during scratching (Berkinblit et al., 1978). Gradual activation of plateau potentials may also offer an alternative explanation for the ‘windup’ phenomenon recently investigated in the neonatal rat (Thompson et al., 1990). In addition, postsynaptic plateau potentials remain a possible mechanism for the sustained discharge of sensory neurons in the turtle spinal cord after a single peripheral stimulus (Currie and Stein, 1990). At present, however, the functional significance of plateau potentials in the cells studied here cannot be assessed, since afferent and efferent connections are not yet known. Information on connectivity is also crucial to determine whether the cells with plateau potentials belong to a single functional group. Unfortunately an anatomical analysis to decide the issue is hampered by the lack of a clear morphological distinction between plateau-generating and non-plateau-generating cells. +

Acknowledgements This work was supported by Lundbeck Fonden, Alfred Benzons Fond and by the Danish Medical Research Council. We wish to thank Morten Msller and Jens Damsgaard Mikkelsen for providing the micrographs.

References Baldissera, F., Hultborn, H. and Illert, M. (1981) Integration in spinal neuronal systems. In Brooks, V. B. (ed.), Handbook of Physiology, Section I , Vol. 11, Part I , Motor Conrrol. American Physiological Society, Bethesda, MD, pp. 509-595. Barrett, E. F. and Barrett, J. N. (1976) Separation of two voltage-sensitive potassium currents, and demonstration of a tetrodotoxin-resistant calcium current in frog motoneurones. J. Physiol., 225, 737-774. Berkinblit, M. B., Deliagina, T. G., Feldman, A. G., Gelfand, I. M. and Orlovsky, G. N. (1978) Generation of scratching. I. Activity of spinal interneurons during scratching. J. Neurophysiol., 41, 1040- 1057. Currie, S. N. and Stein, P. S. G. (1990) Cutaneous stimulation evokes long-lasting excitation of spinal interneurons in the turtle. J. Neurophysiol., 64, 1134- 1148. Getting, P. A. (1989) Emerging principles governing the operation of neural networks. Annu. Rev. Neurosci., 12, 185-204. Grillner, S. (1981) Control of locomotion in bipeds, tetrapods, and fish. In Brooks, V. B. (ed.),Handbook of Physiology, Vol. 11. American Physiological Society, pp. 1177- 1236. Harris-Warrick, R. M. and Johnson, B. R. (1989) Motor pattern networks: flexible foundations for rhythmic pattern production. In Carew. T. J. and Kelly, D. B. (eds), Perspectives in Neural Systems and Behavior. Alan R. Liss Inc., New York, pp. 51-71. Horikawa, K. and Armstrong, W. E. (1988) A versatile means of intracellular labeling: injection of biocytin and its detection with avidin conjugates. J. Neurosci. Methods. 25, 1 - 11. Hounsgaard, J. and Kiehn, 0. (1989) Serotonin-induced bistability of turtle motoneurones caused by a nifedipine-sensitive calcium plateau potential. J. Physiol. (Lond.).414, 265-282. Hounsgaard, J . and Nicholson, C. (1990) The isolated turtle brain and the physiology of neuronal circuits. In Jahnsen, H. (ed.), Preparations of Vertebrate Central Nervous System In Virro. John Wiley and Sons Ltd, pp. 155-181. Hounsgaard, J., Kiehn, 0. and Mintz, I. (1988a) Response properties of motoneurones in a slice preparation of the turtle spinal cord. J. Physiol. (Lond.), 398, 575-589. Hounsgaard, J.. Hultborn, H., Jespersen, B. and Kiehn, 0. (1988b) Bistability of a-motoneurones in decerebrate and acute spinal cats after I.V. injection of 5-hydroxytryptophan. J. Physiol. (Lond.), 405, 345 -367. Jahnsen, H. (1986) Responses of neurons in isolated preparations of the mammalian central nervous system. Prog. Neurobiol., 27, 351 -372. Kiehn, 0. (1991) Plateau potentials and active integration in the ‘final common pathway’ for motor behaviour. Trends Neurosci., 14, 68-73. Kjaerulff, 0. and Hounsgaard, J. (1991) Neurones in lamina 7 and 8 of the turtle spinal cord: Ca+ ‘mediated plateau potentials. Eur. J . Neurosci., 3 (Suppl.), p. 204. Kusuma, A,, Ten Donkelaar, J. and Nieuwenhuys, R. (1979) In Northcutt. R. G. and Ulinsky, U. (eds), Bio/ogy of the Reptilia. Vol. 10. Academic Press, New York. pp. 59-109. Lliniis, R. R. (1988) The intrinsic electrophysiological properties of mammalian neurons: insights into central nervous system function. Science, 242, 1654- 1668. Midtgaard, J. and Hounsgaard, J. (1989) Nerve cells as source of time scale and processing density in brain function. Inr. J. Neural Syst., 1. 89-94. Schwindt, P. C. and Crill, W. E. (1984) Membrane properties of cat spinal motoneurones. In Davidoff, R. A. (ed.),Handbook of the Spinal Cord. Marcel Dekker Inc., New York, Basel, pp. 199-242. Thompson, S. W. N., King, A. E. and Woolf, C. J. (1990) Activity-dependent changes in rat ventral horn neurons in vitro ; summation of prolonged afferent evoked postsynaptic depolarizations produce a D-2-amino-5-phosphonovaleric acid sensitive windup. Eur. J. Neurosci., 2, 638-649.