Intracellular and extracellular electrophysiology of nigral dopaminergic neurons—2. Action potential generating mechanisms and morphological correlates

NeuroscienceVol. 10, No. 2, pp. 317-331, 1983 Printed in Great Britain

0306-4522/83$3.00+ 0.00 Pergamon Press Ltd IBRO

INTRACELLULAR AND EXTRACELLULAR ELECTROPHYSIOLOGY OF NIGRAL DOPAMINERGIC NEURONS-2. ACTION POTENTIAL GENERATING MECHANISMS AND MORPHOLOGICAL CORRELATES A. A. GRACE and B. S. BUNNEY Departments of Psychiatry and Pharmacology, Yale University School of Medicine, New Haven, CT 06510, U.S.A. Abstract-Intracellular recordings from identified nigral dopamine neurons in the rat revealed that their potentials are composed of four components: (1) a slow depolarization, (2) an initial segment spike, (3) a somatodendritic spike, and (4) an afterhyperpolarization. By combining intracellular and extracellular recording techniques with anatomical studies using intracellular injections of Lucifer yellow, an attempt was made to localize each of these potentials to various neuronal compartments. Lucifer yellow injections demonstrated that the dopamine neurons recorded have a pyramidal or polygonal shaped soma, 12-30 pm in diameter, with 34 thick major dendrites which extend lO-50pm from the soma before bifurcating. The axon appears to rise from a major dendrite 15-30pm from the soma. Based on this anatomical configuration, results from the electrophysiological studies suggest that: (1) the slow depolarization is a pacemaker-like conductance most likely localized to the somatic region, (2) the initial segment spike is a low-threshold spike probably located at the axon hillock, (3) the somatodendritic spikes are long duration spikes that rapidly inactivate with depolarization, have a high threshold, and are localized to the dendritic regions. The action potential is then terminated by a long duration afterhyperpolarization. Our data further suggest that spike generation may be initiated by a slow depolarization at the soma triggering a spike in the low-threshold axon hillock which then spreads across the already-depolarized soma to trigger the dendritic spike. Based on the above findings, dopamine neurons can be compartmentalized electrophysiologically and morphologically into subcomponents, each associated with spikes and specific ionic currents. The high threshold dendritic component of the action potential demonstrates rapid inactivation with depolarization, and thus occurs over a rather narrow range of membrane polarization. This limited range of action potential generation may be important in control of dendritic dopamine release and/or modulation of electrical coupling between dopaminergic neurons.

Action potentials of mammalian neurons are known to be composed of several subcomponents corresponding to the multiple excitable membrane regions of the neuron’s surface. The excitatory and inhibitory inputs to the neuron summate to trigger a spike at the membrane area with the lowest threshold, most commonly the axon hillock/initial segment (IS)” region.2~3~‘2~‘3~‘4~‘5 This spike excites the spike generating regions with higher thresholds-the axon followed by the somatic/dendritic (SD)” area.13 However, this sequence of events does not always occur with maximum reliability, i.e. there may be a low “safety for transmission between spike generating factor”” regions. Thus, a delay may be observed between the IS and the SD spike.’ This can be observed intracellularly as an inflection in the rising phase of an action potential and extracellularly as a notch in the positive phase of a spike.‘2*22.24 Please send reprint requests to: Dr. B. S. Bunney, Neuropsychopharmacology Research Unit, Yale University School of Medicine, P.O. Box 3333, New Haven, CT 06510, U.S.A. Abbreviations: DA, dopaminergic; IS, initial segment; SD, somatodendritic; SN, substantia nigra. 317

Dopaminergic (DA) neurons are known to fire action potentials which often have a prominent notch in the positive phase of the extracellular and intracellular spikes,‘0*25~26~3’ suggesting that they also may have a low safety factor for SD spike invasion.55 Indeed, during antidromic activation of DA neurons from the caudate nucleus, a failure of the SD spike is the most commonly observed phenomenon.25.26,3’ In

this study, we used intracellular and extracellular recording techniques to examine, in detail, nigral DA neuron IS and SD spike generation. These techniques were also used to investigate the origin of the slow depolarizations found to precede most DA neuron action potentials. EXPERIMENTAL PROCEDURES The intracellular and extracellular recording techniques were essentially the same as those described in the first paper of this series.26 In one set of experiments (n = 12) Lucifer yellow,63 3% in 2 M LiCl, was injected into DA cells at the end of the recording period by a constant 2nA hyperpolarizing current. The animals were then perfused with 10% formalin in saline, the brain removed, and 50 pm sections cut through the substantia nigra. Sections were examined with a Zeiss fluorescence microscope, and recon-

318

A. A. Grace and B. S. Bunney

structions of photographs of dye-injected neurons were made. This technique allowed us to outline DA cell processes in a Golgi-like manner for morphological studies.” IS-SD delays were calculated by measuring the time between peaks of the differentiated intracellular records, Differentiation was performed by a PDP I l/34 on waveforms obtained with a Nicolet Explorer III digital oscilloscope by taking successive differences between consecutive points. Signal averaging of prespike slow depolarizations was accomplished using a Nicolet 4094 digital oscilloscope in conjunction with the Nicolet sweep averaging program. A stable trigger was obtained by using a window discriminator output as an external trigger source with input from a parallel signal filtered for low frequencies.

RESULTS

Morphology Dopaminergic

neurons

injected

with Lucifer

yellow

(n = 12, Fig. 1) were generally

12-30pm in diameter, depending upon the axis of measurement of their pyramidal or polygonal cell bodies. The cells had 3-6 thick (4-8 pm diameter) major dendrites which extended l&50 pm before bifurcating. These dendrites could occasionally be followed over 200 pm as they coursed ventrally and were often found to form tufted endings. Although the axon cannot be conclusively identified without electron microscopy,5’ the most likely candidate appears to be a thin [approx. 1 pm) constant diameter process which was consistently seen to arise from a major dendrite approximately 15-30 pm from the cell body. Dye coupling between DA cells was also frequently observed and will be described in the next paper.27 Antidromically

activated spikes

Antidromic activation of extracellularly recorded DA cells from their terminal fields in the caudate nucleus (n = 40) resulted either in a full amplitude spike (Fig. 2A) or in a spike of smaller amplitude than the action potential (Fig. 2B). Both the full amplitude spike and the small spikes could be consistently collided with a spontaneously occurring action potential (Fig. 2C). Intracellular recording (Figs. 2D-F) revealed that this truncated spike could follow high frequency (50 Hz or higher) antidromic activation with a constant latency (n = 25, Fig. 2E) and collide with a full amplitude directly elicited action potential (n = 40, Fig. 2F) thereby fulfilling all of the criteria for antidromic activation.43 The possibility that this small spike may be a full action potential which was greatly decreased in amplitude (e.g. due to a concomitant IPSP” shunting the current out of the cell soma”) was also examined. Depolarizing current injection during antidromic activation caused the truncated spike to become a full amplitude action potential (n = IO, Fig. 3). suggesting that the small potential is probably the axon hillock or initial segment (IS) spike as described in other preparations (see Discussion). The IS spike was observed to be embedded in the rising phase of the action potential (Fig. 3E). The converse was also observed. Thus,

when antidromic activation resulted in a full action potential, hyperpolarizing current injection could block invasion of the soma, resulting in the lower amplitude spike only (n = 7, Fig. 4). In both cases, the antidromically elicited potentials could be collided with directly elicited somatic spikes. Directly elicited spikes The lower threshold of the IS portion of a neuron could allow the IS spike to be initiated independently by direct intracellular current injection, provided the SD spike threshold is high enough. Intracellular current injection into slowly firing or non-firing DA neurons often resulted in elicitation of the IS spike alone (n = 15, Fig. 5B). Higher levels of injected current could elicit the full IS-SD spike (Fig. 5C), thus demonstrating the threshold differences between these two spike components. Further depolarization would often result in one or more action potentials followed by a single IS spike before firing accommodation occurred. Both the IS spike and the full ISSD spike triggered the axon to fire, since both the directly elicited IS spike and the directly elicited IS-SD spike could be collided with an antidromically elicited IS spike (Figs 5B,C). In a DA cell firing spontaneously, the IS spike could rarely be elicited independently by direct current injection, possibly because the DA cell was already depolarized sufficiently to allow the IS spike to trigger the SD regions. Thus, current injection into a DA cell near threshold for spontaneous spike generation always led to a full IS-SD spike. However, the IS spike could still be antidromically activated without triggering the SD spike (n = 11, Figs SD,E,F and G). Spontaneous

occurrence of’ the initial segment

.spike

During intracellular recording, high levels of depolarizing current injection caused the SD regions of the neuron to inactivate, and therefore the IS spike was observed independently (n = 6). This observation was probably not an artifact of increased SD spike lability due to cell damage from microelectrode penetration. since during extracellular recording the IS spike could also be observed independently in a very depolarized, fast firing DA cell. Thus, following haloperidol pretreatment (0.1 mg/kg i.v., to increase cell firing rate and burst firing frequency), extracellular recording from a fast firing, bursting DA cell revealed progressive inactivation of the SD spike during bursts which occasionally resulted in the independent occurrence of a presumed IS spike as the last spike in a burst (n = 4, Fig. 6). The fact that this spike was positive-going (indicating an outward current) ruled out the possibility that this was a consequence of recording extracellularly directly apposed to the IS region. Somatodendritic

(SD) spike

Hyperpolarization of the soma delayed and then prevented invasion of the soma by the IS spike (n = 8.

Fig. 1. Drawing of Lucifer yellow injected dopaminergic neuron. The injected neuron is shown in the inset; the medially-directed putative axon is marked with an arrow. Two lightly labeled neighboring neurons were not included in this composite. Dendrites emerge from the neuron’s poles and bifurcate approximately 1530 pm from the center of the soma. Three major dendrites course 1W200 pm ventrally into the zona reticulata, and end in tufts. Two horizontally projecting dendrites were also observed, with the axon eminating from one of these. Vertical arrow points dorsally; horizontal arrow points medially.

30 pm

Nigral dopamine neurons: Electrical characterization

321

Fig. 2. Extracellular and intracellular recordings of antidromically activated dopamine neurons. Stimulation of the caudate nucleus (solid arrows) during extracellular recording produced either a full IS-SD spike (A) or an IS spike (B), often in the same preparation. Both spikes could be collided (C, open arrow) with a spontaneously occurring spike. Stimulation of the caudate nucleus (solid arrows) during intracellular recording yielded a similar phenomenon. Although a full spike could be observed during antidromic activation (D), the more usual circumstance was an IS spike which occasionally triggered the full IS-SD spike (E). Again, both action potentials could be collided (open arrow in F) with a directly elicited action potential (0.2 nA current injection at bar). In this and all subsequent Figures containing extracellular records positive deflection is in the upwards direction.

4). However, in a spontaneously firing DA neuron intracellular depolarizing current injection was also observed to increase the IS-SD delay (Fig. 7A). Indeed, at depolarizations greater than -40 mV the SD spike failed altogether, leaving only an IS spike. The comparatively greater lability of the SD spike can be observed by plotting the change in IS and SD spike amplitude at various membrane potentials (Fig. 7B). Thus, the SD spike demonstrates large decreases in amplitude with depolarization, whereas the IS spike varies little in amplitude under the same conditions. By differentiating the intracellular recordings with respect to time, the changes in IS-SD separation with depolarization may be more clearly demonstrated (Fig. 8). Since the time differentials of intracellular recordings have been described to resemble extracellular recordings,23@’ we have compared the wave patterns of differentiated intraceullar recordings with action potentials observed during extracellular recording. As extracellularly recorded Fig.

action potentials could be observed which closely resembled the differentiated intracellular records (Fig. 8), this increased IS-SD separation was probably not due to artifacts created by the current injection. Those extracellular recordings obtained to correspond to the most depolarized intracellular recordings were taken during late segments of a spontaneous burst. Slow depolarization preceding action potentials

During intracellular recording we have reported26 that DA neurons exhibit a slow ramp depolarization which precedes the action potential. This slow depolarization usually begins 40 to 120 ms (average 60 f 25 ms, n = 50, mean &-S.D.) prior to the firing of an action potential and attains an amplitude of 7 to 20 mV (average 13 f 3 mV, n = 50, mean f SD). The possible site of origin of this current may be inferred by using extracellular recording if a slow

322

A. A. Grace and 8. S. Bunney

D

’

Fig. 3. During intracellular recording. antidromic activation of dopaminergic neurons from the caudate (soiid arrows) in most cases resulted in an IS (initial segment) spike only (A), which could be reliably collided (B, open arrow) with a directly elicited action potential (B, 0.3 nA current injection at bar). Injection of increasing levels of depolarizing current (at bars) could transform the antidromically activated IS spike (C, 0.2 nA) into the full IS-SD action potential (D, 0.28 nA). (E) Superimposed tracings of antidromically activated DA cell with various levels of depolarizing current injection (O-O.3 nA at bar). The IS spike can be observed giving rise to the inflection in the rising phase of the action potential (arrowhead). A

C

--!F---D

Fig. 4. During intraceIluiar recording, stimu~tion of the caudate nucleus (solid arrows) occasionally resulted in an antidromically activated full initial se~ent-somatodcndritic spike (A). This action potential collided (B, open arrow) with directly elicited action potentials (B, 0.3 nA current injection at bar, 2 repetitions). Injection of increasing levels of hyperpolarizing current (0. I5 nA and 0.25 nA at bar, C and D, respectively) first increased the IS-SD break (arrowhead, C) and then blocked the SD spike altogether (D). Two repetitions of threshold hyperpolarizing current injection (0.2 nA, E) revealed the IS spike giving rise to the IS-SD break (arrowhead) in the full action potential.

323

Nigral dopamine neurons: Electrical characterization

C

Fig. 5. Antidromic activation and collision of initial segment (IS) spike with directly elicited IS and IS-somatodendritic spikes during intracellular recording from DA neurons. (A) Caudate stimulation (arrow) resulted in an IS spike. Subthreshold current injection (0.4nA bar) did not interfere with this activation. (B) A higher level of current injection (0.45 nA) resulted in a directly elicited IS spike which collided with the antidromically driven IS spike (small arrow). (C) Still higher levels of injected (0.5 nA) current resulted in a full action potential, which also collided with the antidromically driven IS spike (small

arrow). This occurrence (A-C) was more common in spontaneously hyperpolarized, non-firing neurons. If the DA cell is closer to threshold, however, a different result is noted (D-G). (D) Again, subthreshold levels of current injection (0.25 nA at bar) did not interfere with caudate nucleus stimulation (at arrow) driving an IS spike. Increasing levels of current injection (0.35 nA in E and 0.4 nA in F, respectively) did not result in an IS spike (E), but instead resulted in a full IS-SD action potential which collided with the antidromically activated IS spike (F, small arrow). Repetition of the current in (F, 0.4 nA) occasionally resulted in direct spike failure and consequent lack of collision (G), but, again, did not result in a directly elicited IS spike. Thus, when the DA cell is close to firing threshold (D-G), the IS spike cannot be elicited directly without triggering the SD spike.

potential shift with a time course similar to this slow depolarization is observable. For example, if extracellular recording from the soma reveals a current that is in the negative direction, then the inward current responsible for it must be generated locally at the soma. To investigate the possible origin of this current it was necessary to use unfiltered DC recordings. Action potentials were captured on a Nicolet digital oscilloscope adjusted to obtain IOOms of prespike baseline data. A small but consistent negative deflection was present when the sweeps were averaged 50-400 times (n = 16), and attained an amplitude of 2@4OpV. These slow currents began approximately 60 to 130 ms prior to spike initiation. Although this negative deflection is small in amplitude and necessitates signal averaging for its obser-

vation, it is nevertheless suggestive of a slow inward current at the neuron’s soma. Comparison cellularly

of spontaneous potentials

observed intra-

A total of five different potentials were observed to occur spontaneously during intracellular recording and are compared in Table 1. Three of these, the IS spike, the SD spike, and the slow depolarization, are discussed in this paper and the electrical coupling potential is described in the next paper. The afterhyperpolarization is associated only with the SD portion of the spike, and varies widely in amplitude and duration. The decay of this afterhyperpolarization precedes the slow depolarization of the next

324

A. A. Grace and B. S. Bunney

I

5

mSec

1

Fig. 6. Typical inactivation of somatodendritic spikes recorded extracellularly from a bursting dopaminSD portion of the spike demonstrates a decrease in in the IS spike amplitude. In (E), the SD spike has

ergic cell. (A-D) As the burst progresses, the amplitude, with comparatively little variation

completely failed, leaving only the IS spike.

spike. Spontaneously occurring EPSPs were not observed in the dopamine cells studied. DISCUSSION

In this study we have investigated the properties of the DA neuron axon hillock and somato-dendritic spike generating sites in order to characterize the regions of DA neurons capable of spike generation. An attempt will be made to correlate the electrophysiological data obtained with the observed niorphological features of DA cells. Morphology

Intracellular injection of the dye Lucifer yellow into identified DA neurons permitted us to make light microscope observations of their cellular morphology. The DA cells studied gave rise to 3-6 dendrites which bifurcate and course ventrally, as described in Golgi studies.29*35~56*58 One process was tentatively identified as the axon due to its small, constant diameter, its unbranched nature and its medial course (the characteristic course of DA cell

axons in the SN).42 This putative axon, when it could be distinguished, was always observed to arise from a major dendrite instead of the soma (n = S), as also reported for nigral compacta neurons using Golgi studies’j and intracellular horseradish peroxidase injection.j4 Such an arrangement would place the axon hillock spike generating region at a relatively long electrotonic distance from the soma. Our electrophysiological data support this hypothesis (see below). Ident$cation spikes

and characterization

of’ initial segment

Antidromic activation. Stimulation of the caudate nucleus during extracellular recordings from DA neurons25.26,3’rarely results in antidromic activation of a full action potential. The most commonly observed occurrence is activation of a smaller amplitude spike. This small spike fulfills all the criteria of antidromic activation.43 Even if a full action potential is activated antidromically, the SD spike cannot follow high frequency stimulation and is reduced to this smaller spike. Similar results were also obtained

325

Nigral dopamine neurons: Electrical characterization I

IS-SD

50

B

DELAY

0. \\ %\ I 50

45

d\

SD Spaks

~\

\%

MEMBRANE

POTENTIAL

40

r i - 0.95

I\

I25

rlop?

= - I .32

/

(mV)

I5 IlOP.

IO

%\

Sp,k*

\\

, =

0.59

:

0 09


5

‘% \\

I

1

I

-65

-bO

-55

MEMBRANE

I -50

/

I -45

-40

POTENTIAL

‘, -35

\. .A.

1

-30

(mV)

Fig. 7. (A) Dependence of the initial segment-somatodendritic spike (IS-SD) delay on the membrane potential of dopaminergic cells. As the membrane potential is depolarized by constant current injection, the delay between the IS and the SD portions of the action potential increases. All action potentials were spontaneous. Asterisk marks the level of depolarization at which total SD spike inactivation occurs. (B) Dependence of IS and SD spike amplitude on DA neuron membrane potential. Measurements were made on spontaneously occurring action potentials, with the membrane potential varied by constant current injection. Although the IS spike amplitude (filled circles) was poorly correlated and showed little dependence on membrane potential, the SD spike (open circles) rapidly decreased amplitude with depolarization. Total SD failure occurred at the more depolarized levels.

during intracellular recording from these cells.25,26 Thus, since this small spike can be antidromically activated from the caudate and also can give rise to full amplitude action potentials, we have proposed it to be an active potential generated at the axon hillock/initial segment region firing independently of the SD spike. This conclusion was supported by our evidence obtained during intracellular recording: (1) The antidromically activated small spikes could be collided with full action potentials: (2) The small spike could be activated directly by low levels of current injection-higher current levels triggered full IS-SD spikes: (3) Directly elicited small spikes could be collided with similar antidromically activated spikes:24 (4) The antidromically activated small spikes could trigger action potentials when a depolarizing current was injected concurrently (as in 5, 14, l&67), and conversely: (5) SD invasion during antidromic activation could be blocked by hyperpolarizing current injection (as in References 12, 14, 20, 38, 67). The last two experiments rule out the possibility that this potential is a truncated IS-SD spike resulting from the concurrence of a normal amplitude action potential with an increased conductance synapse (e.g. IPSF) or an injured SD spike generating site. However, the concurrent IPSP often observed during antidromic activation may account for the variable size of the small spike. This small spike is also not of synaptic origin, since it fulfills all of the criteria of Van Essen6’ and Yau” for nonsynaptic

potentials: (I) steep rise and decay; (2) amplitude greater than IO mv; (3) same antidromic latency as IS-SD spike; (4) constant amplitude with high frequency antidromic activation; (5) collision with full IS-SD action potential; and (6) blockade of SD spike with hyperpolarization. Directly elicited initial segment spikes. The IS region has been described as the lowest threshold and therefore the site of region of neurons, 2,3~‘2~‘3~‘4~‘9*24 action potential generation.3,6.‘5 This lower current threshold is probably due to its smaller diameter and thus increased current density compared to SD spike generating regions. 8~‘6~6’In spontaneously hyperpolarized, non-firing DA neurons, the IS spike can be elicited by direct current injection without triggering an SD spike. IS spikes cannot be elicited directly independent of the SD spike in fast firing DA neurons. Although the direct triggering of IS spikes alone firing of is rare in some systems, ‘2.47the independent IS spikes has been described as common in caudate nucleus cells,48 where synapses on the IS neuronal region have been described.36 Safety factor for initial segment-somatodendritic invasion

spike

The IS-SD delay observed during antidromic activation is related to the difficulty that a spike arising in the IS region of a neuron has in triggering a spike in the SD portion. This reliability of information transfer is referred to as the “safety factor” for spike

326

A. A. Grace and B. S. Bunney INTRACELLULAR

-45mV

EXTRACELLULAR

DIFFERENTIATED

----JL

-5lmV

Fig. 8. Variation of the initial segment and somatodendritic portions of dopaminergic cell action potentials with the level of depolarization. During intracellular recording (left column), depolarizing current injection resulted in a gradual increase in the IS-SD latency in addition to an inactivation of the SD spike, which failed at the most depolarized level (- 33 mV). As seen in the differentiated intracellular recordings (which are known to resemble the equivalent extracellular recordings of these potentials, see text), the progressive inactivation is refiected in an increased relative prominence of the notch and loss in amplitude of the action potential. The right column shows extracellular recordings taken from various DA neurons chosen to match the waveforms of the differentiated intracellular recordings. The close similarity between the differential recordings and extracellularly obtained tracing demonstrate that such variations in extracellular spike shape occur spontaneously and are not due to distortions induced by intracellular current injection. Extracellular spikes corresponding to the most depolarized intracellular recordings were taken during late portions of spontaneous bursting activity. Table I. Electrical properties of dopaminergic neurons

.~~..-._____ Potential Action potential (SD spike) Slow depolarizing potential Initial segment (IS; small spike)

Time to peak (ms)

Total (ms)

55-75

0.75-1.7

1.64.7

Active

Nature

7-20

40-120

60 k 25

Active (?)

9-25

0.2-0.8

0.5-1.7

Active

-

1-6

Passive f?)

0.5-2.0

2-6

Active

(-3)-(-0.5) Electrical coupling potential (fast uotentiaW

Duration

Amplitude (mV)

2-15

~Antidromic-elicited by stimulation in the caudate nucleus. *Direc+licited by intracellular current injection.

Occurrence Spontaneous antidromic? direct* Spontaneous Spontaneous antidromic direct Only after SD spikes Spontaneous antidromic

Nigral

dopamine

neurons:

Electrical

characterization

327

of the electrode from the cell (i.e. no inversion was invasion.” Dopaminergic cells would appear to have noted either when the electrode was withdrawn until a low safety factor for IS-SD invasion for the the cell action potential was no longer visible or when following reasons: (1) a relatively long IS-SD delay; the electrode approached the cell to the point of (2) low reliability of antidromic activation in eliciting puncture); (2) whether or not continued vertical and SD spike; (3) rapid inactivation of SD spikes with displacement of the electrode punctured the cell or depolarization or during bursts; (4) triggering of IS passed by the soma without injuring it; and (3) with spikes alone with low levels of depolarizing current the electrode used [i.e. either a low-impedance extracinjection; and (5) tentative morphological evidence ellular electrode (8 mQ) or a high impedance intrafor relatively large distances between putative IS and cellular electrode (over 30 mn)]. Lastly, this biphasic SD spike generating regions. extracellular spike was always observed prior to The IS-SD delay during spontaneous firing of DA somatic penetration with the intraceullular electrode. neurons averages about 0.35 ms, as compared to As expected, the IS and SD spikes recorded intra0.05-0.4 ms in motoneurons’ and 0.15 ms in cerebellar Purkinje cells.‘8 Hyperpolarization during anticellularly both lead to membrane depolarization, thus suggesting net inward current flow at the spike generdromic activation of full IS-SD action potentials in ating sites. However, in the extracellularly recorded DA neurons delayed and finally blocked the SD spike, as reported in other systems.2~‘2~13~14~20~2’~24~3x~44 action potential observed just prior to penetration of However, in DA neurons depolarization also inthe soma of the DA cell, the IS and the SD spike are both positive-going. This is indicative of an outward creased the IS-SD latency, an effect opposite to that current, meaning that the soma is a sink, and not a observed previously in other systems5.‘4.30 Ecclesi6 source, of either the IS or the SD spike. Thus, both reported that depolarization can have a biphasic of these spikes would appear to be generated distally effect on IS-SD latency: first decreasing and then from the electrode placed near the soma. This distal increasing latency with increasing depolarization levlocation of the IS spike would account for its relaels. A possible reason for the increased IS-SD latency tively small size recorded in the soma (about 15 mV, in DA cells with depolarization could be that the SD as compared to 3&40 mV in the spinal motogenerating area of the cell is partially inactivated, neuron*.20) and low safety factor for SD spike inithus leading to slower SD activation.65 Alternately, tiation. Morphological studies (References 35, 54 and the site of SD spike generation could move to more electrotonically distal areas, resulting in a decreased Fig. 1) provide an anatomical correlation with these amplitude and increased duration of the SD spike. electrophysiological findings-axons of DA cells apThus, intracellular recording revealed that even durparently arise from a dendrite rather than directly from the soma. If this is the case, any synapses on the ing relatively low levels of depolarization the SD dendrite between the axon hillock and the soma spike rapidly decreases in amplitude, leaving only an would profoundly affect the cell’s firing IS spike at more depolarized levels (Figs 7A, B; 8). In thresho1d”,43.52,62 and might also account for the addition, during partial depolarization the SD spike variability in the size of the IS spike observed in the decreases in amplitude, whereas the IS spike amplisoma during antidromic activation. Possible support tude is little affected. Taken together, these data for this speculation comes from the finding that indicate that the SD spike of DA neurons is very labile compared to the IS spike. Inactivation of DA degenerating boutons have been noted on primary nigral dendrites after caudate lesions.4 Such an ananeuron spike generation can also result from pharmacological manipulations of the dopaminergic system. tomical arrangement is not without precedent as in the Dopaminergic neurons have been reported to enter a axon hillock synapses have been reported mammalian nervous system (e.g. in cerebellum,32 state of apparent depolarization block in response to hippocampus, caudate36 and cortex34*69). repeated haloperidol administration,’ acute kainic As noted previously, the SD spike generation site acid injection into the caudate nucleus,’ and cholecystokinin iontophoresis.60 must also be located distal to the soma. The higher threshold for SD spike generation by direct current Sites of current Jlow in the dopaminergic cell injection, in addition to the failure of comparatively large IS spikes recorded in the soma to trigger an SD The sites of active current flow across a cell’s spike, would also argue for a distally located SD membrane may be estimated by comparing intraspike generation site. One possible location for this cellularly and extracellularly recorded action potentia1s.22,33.45Although the intracellularly and extraSD spike generating site would be the convergence point of nigral cell branching dendrites reported in cellularly recorded action potentials were not, in this case, obtained simultaneously, we feel justified in compacta neurons in Golgi studies.58 The glial wrapping observed surrounding compacta cell primary making this comparison as both types of action dendrites29v56 is reminiscent of a similar morphology potentials have the same subcomponents and these subcomponents have equivalent time courses and in apical dendrites of olfactory bulb mitral cells, activation characteristics. Furthermore, the direction where they are believed to aid in active potential and time course of the extracellularly observed spike conduction (see Reference 59 for discussion). Such a subcomponents do not change: (1) with the distance distal location may account for DA neurons having

328

A. A. Grace and B. S. Bunney

POTENTIALS _EXTRACELLULAR Et ECTRODE

EXTRACELLULAR

INTRACELLULAR

B

3

PROXtNAL\

DISTAL

DENDRllCS

DEMORITES

@--k& r’

. .._. ,,-.

8.

2 mSec

Fig. 9. A hypothetical model of spike generation in dopaminergic neurons is shown in this compartmental representation. The sites of potential generation in the DA neuron can be estimated by comparison of intracellular and extracellular recordings. By using a compartmental model (B) to represent the functional divisions of the DA neuron (A), the sites of current flow can be compared to the recorded potential changes. The flow of positive ionic currents is shown by the arrows (solid arrows = current sources; dashed arrows = current sinks), and the corresponding intracellular and extracellular records outlined are the resultant potential changes, assuming extracellular recordings are from the soma. The portion of the intracellular and extracellular potential changes corresponding to the arrows shown are darkened. Inward current always results in a positive deflection measured by the intracellular electrode. Potential changes at the extracellular electrode depend on whether current is directed away from the extracellular compartment (negative deflection) or toward the extracellular compartment (positive deflection) at the electrode location, i.e. the soma. (Cl) The slow inward current at the soma depolarizes the cell; the somatic location is established by the negative deflection at the extracellular electrode. This slow depola~zation brings the lowest threshold region of the cell, the IS, to threshold and results in a potential, the IS spike (C2). The positive deflection in the extracellular electrode indicates that the soma is the current sink for the inward current generated distally; presumably at the initial segment (inward arrows). This IS spike travels across the soma [already depolarized by the slow inward current from (Cl)] to trigger the dendritic action potential (C3). The positive deflection at the extracellular electrode indicates that the soma is a sink for this current also; thus, the current source is located distally to the soma-presumably in the dendrites.

Nigral dopamine neurons: Electrical characterization a comparatively small action potential amplitude and long duration (typically 55-65 mV) (average = 2.75 ms). 26 Both properties could be due to electrotonic degradation of the SD spike during conduction from a distal site or sites to the soma. Prior to firing action potentials, DA neurons were observed to have a long (approximately 60 ms) slow depolarization leading to spike initiation.26 We have found that, during extracellular recording (presumably from the soma), a slow negative potential shift occurs prior to spike firing, which has a similar time course as the intracellularly observed slow depolarization. Although this potential is small in amplitude and requires averaging to be observed, the data tentatively suggest that the slow depolarization occurs by way of an inward current generated at the somatic region. In Aplysiu a slow depolarization leading to bursting has been localized to the somatic region of neurons.’ In summary, our data suggest that the spontaneous generation of action potentials in DA neurons occurs as follows: a slow inward current generated at the soma brings the axon hillock region near threshold. Either this depolarization or a summation of this depolarization with an electrical coupling potentia12’ causes the axon hillock to fire. This IS spike then induces a regenerative action potential which travels down the axon, as well as spreads over the depolarized soma and triggers a dendritic (SD) spike (Fig. 9). This model would explain the typically observed lack of antidromic activation of the SD spike: since the antidromically driven IS spike would not be expected to occur at the peak of a slow depolarization at the soma (which precedes spontaneously occurring spikes), the IS spike alone would not be of sufficient amplitude to cause this somatic depolarization and

329

bring the dendrites to spike threshold. The concomitant orthodromically driven IPSP observed prior to the antidromic spike25,26would also tend to inactivate any ongoing slow depolarizations. Waxmana has proposed that neurons are compartmentalized into regions which sequentially transform information. Such a proposal may be particularly relevant for DA neurons. Drawing from the present results and models based on data collected in other systems, a hypothesis for DA cell compartmentalization can be developed: (1) dendritesreceive synaptic input,37*46,53 transmit and receive electrical coupling potentials (lateral excitation),27 and release DA (lateral inhibition);28 (2) soma-generates slow inward currents leading to spiking (Fig. 15);26(3) axon hillock-integrates excitatory and inhibitory inputsI and generates propagating action potentials; and (4) terminals-releases DA in impulse-dependent manner under the regulation of local factors [e.g. autoreceptors;4’ tyrosine hydroxylase activation (see References 50, 57) etc.]. Because all of these regions are also areas which probably possess low conduction safety factors, *J~*~’ fine control of their interrelations is possible. 49Thus, even low levels of inhibitory input to the regions with a low safety factor may easily block invasion of these areas by spikes. authors would like to thank G. Shepherd and M. Nowycky for suggestions on the manu-

Acknowledgements-The

script; R. Llinas for advice on localization of cellular currents; D. Stagg for computer programming; L. Meltzer for lending. us a few cells; W. Stewart for nraciouslv supplying Lucifer yellow; R. Jackson, S. LaFlamme and C.-L. Pun for technical assistance and L. Williams and D. Bennett for their patience and excellent typing. Supported by USPHS grants MH28849 and MH25642 and by the State of Connecticut.

REFERENCES Alving B. 0. (1968) Spontaneous activity in isolated somata of Aplysia pacemaker neurons. J. gen. Physiol. 51, 29-45. Araki T. and Otani T. (1955) Response of single motoneurons to direct stimulation in toad’s spinal cord. J. Neurophysiol.

18, 412-485.

Araki T. and Terzuolo C. A. (1962) Membrane currents in spinal motoneurons associated with the action potential and synaptic activity. J. Neurophysiol. 25, 112-789. Bak I. J., Choi W. B., Hassler R., Usunoff K. G. and Walgner A. (1975) Fine structural synaptic organization of the corpus striatum and substantia nigra in rat and cat. In Advances in Neurology, Vol. 9, Dopaminergic Mechanisms (eds Caine D. B., Chase T. N. and Barbeau A.) pp. 25-41. Raven Press, New York. Baylor D. A. and Nicholls J. G. (1969) After-effects of nerve impulses on signaling in the central nervous system of the leech. J. Physiol., Lond. 203, 571-589. Bishop P. O., Burke W. and Davis R. (1962) The interpretation of the extracellular response of single lateral geniculate cells. J. Physiol., Lond. 162, 451472. Braszko J. J., Bannon M. J., Bunney B. S. and Roth R. H. (1981) Intrastriatal kainic acid: acute effects on electrophysiological and biochemical measures of nigrostriatal dopaminergic activity. J. Pharmac. exp. Thdr. 216, 289-293.

8. Brock L. G., Coombs J. S. and Eccles J. C. (1953) Intracellular recording from antidromically activated motoneurons. J. Physiol.,

Lond. 122, 429, 461.

9. Bunney B. S. & Grace A. A. (1978) Acute and chronic haloperidol treatment: comparison of effects on nigral dopaminergic cell activity. Life Sci. 23, 1715-1728. 10. Bunney B. S., Walters J. R., Roth R. H. and Aghajanian G. K. (1973) Dopaminergic neurons: effect of antipsychotic drugs and amphetamine on single cell activity. J. Pharmac. exp. ThPr. 185, 56571. 11. Conradi S. (1969) Observations on the ultrastructure of the axon hillock and initial axon segment of lumbosacral motoneurons in the cat. Acta physiol. stand. Suppl. 332, 6584.

330

A. A. Grace and B. S. Bunney

12. Coombs J. S., Curtis D. R. and Eccles J. C. (1957) The interpretation of spike potentials of motoneurons. J. Physiol., Lond. 139, 198-231. 13. Coombs J. S., Curtis D. R. and Eccles J. C. (1957) The generation of impulses in motoneurons. J. Physiol., Lond. 139, 232-249. 14. Coombs J. S., Eccles J. C. and Fatt P. (1955) The electrical properties of the motoneuron membrane. J. Physiol., Lond. 130, 291-325. 15. Curtis D. R., Eccles J. C. and Lundberg A. (1958) Intracellular recording from cells in Clarke’s column. Acra physiol. stand. 43, 303-314. 16 Eccles J. C. (1955) The central action of antidromic impulses in motor nerve fibers. eflugers Arch. ges. Physiol. 260, 385415. 17 Eccles J. C. (1957) The Physiology of Nerve Cells. Johns Hopkins Press, Baltimore. 18 Eccles J. C., Faber D. S. and Taborikova H. (1971) The action of a parallel fiber volley on the antidromic invasion of Purkinje cells of cat cerebellum. 19 Fatt P. (1957) Sequence of events in synaptic activation of a motoneuron. J. Neurophysiol. 20, 61-80. 20 Frank K. and Fuortes M. G. F. (1955) Potentials recorded from the spinal cord with microelectrodes. J. Physiol., Land. 130, 625-654. 21 Frank K. and Fuortes M. G. F. (1956) Stimulation of spinal motoneurons with intracellular electrodes. J. Ph_wiol., Lond. 134, 451470. 22 Freygang W. H. (1958) An analysis of extracellular potentials from single neurons in the lateral geniculate nucleus of the cat. J. gen. Physiol. 41, 543-564. 23. Freygang W. H. and Frank K. (1959) Extracellular potentials from single spinal motoneurons. J. gen. Physiol. 42, 749-760. 24. Fuortes M. G. F., Frank K. and Becker M. C. (1957) Steps in the production of motoneuron spikes. J. gen. Physiol. 40, 735-752. 25. Grace A. A. and Bunney B. S. (1980) Nigral dopamine neurons: intracellular recording and identification with L-DOPA injection and histofluorescence. Science N. Y. 210, 654656. and extracellular electrophysiology of nigral dopaminergic 26 Grace A. A. and Bunney B. S. (1983) Intracellular neurons-l. Identification and characterization. Neuroscience 10, 301-315. and extracellular electrophysiology of nigral dopaminergic 27. Grace A. A. and Bunney B. S. (1983) Intracellular neurons-3. Evidence for electrotonic coupling. Neuroscience 10, 333-348. 28. Groves P. M., Wilson C. J., Young S. J. and Rebec G. V. (1975) Self-inhibition by dopaminergic neurons. Science. N.Y. 190, 522-529. 29. Gulley R. L. and Wood R. L. (1971) The fine structure of the neurons in the rat substantia nigra. Tissue and Cell 3, 675-690. 30. Gustafsson B. and Lipski J. (1980) Effect of membrane polarization and synaptic activity on the timing of antidromic invasion. Bruin Res. 181, 61-74. 31. Guyenet P. G. and Aghajanian G. K. (1978) Antidromic identification of dopaminergic and other output neurons of the rat substantia nigra. Bruin Res. 150, 69-84. 32. Hamori J. (1981) Synaptic input to the axon hillock and initial segment of inhibitory interneurons in the cerebellar cortex of rat. Cell. Tissue Res. 217, 553-562. 33. Hubbard J. J., Llinas R. and Quastel D. M. J. (1969) Electrophysiologicol Analysis of Synaptic Transmission, pp. l-372. Williams & Wilkins, Baltimore. 34. Jones E. G. and Powell T. P. S. (1969) Synapses on the axon hillocks and initial segments of pyramidal cell axons in the cerebral cortex. J. Cell Sci. 5, 495-507. 35. Juraska .I. M., Wilson C. J. and Groves P. M. (1977) The substantia nigra of the rat: a Golgi study. J. camp. Neural. 172, 585-600. of the caudate nucleus. Phil. Trans. R. Sot. Ser. 36. Kemp J. M. and Powell T. P. S. (1971) The synaptic organization E 262, 403412. 37. Kim J. S., Bak I. J., Hassler R. and Okada Y. (1971) y-Aminobutyric acid (GABA) in the extrapyramidal motor system-II. Some evidence for the existence of a type of GABA-rich strio-nigral neuron. Expl. Brain Res. 14, 95, 104. properties of lateral reticular 38. Kitai S. J., Kennedy D. J., DeFrance J. F. and Hatada K. (1974) Electrophysiological nucleus cells--I. Antidromic activation. Expl. Brain Res. 21, 403418. 39. Kosaka T. (1980) The axon initial segment as a synaptic site: ultrastructure and synaptology of the initial segment of the pyramidal cell in the rat hippocampus (CA3 region). J. Neurocyfol. 9, 861-882. C. (1955) Synaptic inhibition in an isolated nerve cell. J. gen. Physiol. 39, 155-184. 40. Kuffler S. W. and Eyzaguirre 41. Langer S. Z. (1981) Presynaptic regulation of the release of catecholamines. Pharmac. Rev. 32, 337-362. of catecholamine neurons in the rat central nervous system. In 42. Lindvall 0. and Bjijrklund A. (1978) Organization Handbook qf Psychopharmacology (eds Iversen S. D. and Snyder S. H.) Vol. 9, pp. 139-231. Plenum, New York. 43. Lipski J. (1981) Antidromic activation of neurones as an analytic tool in the study of the central nervous system. J. Neurosci. Meth. 4, l-32. coupling between neurons in cat inferior olive. J. Neurophysiol. 44. Llinas R., Baker B. and Sotelo C. (1974) Electrotonic 37, 56C-57 I. 45. Llinas R. and Yarom Y. (1981) Properties and distribution of ionic conductances generating electroresponsiveness of mammalian inferior ohvary neurones in vitro. J. Physiol., Lond. 315, 569-584. 46. McGeer P. L., Fibiger H. C., Maler L., Hattori T. and McGeer E. G. (1974) Evidence for descending pallido-nigral GABA-containing neurons. In Advances in Neurology Second Canadian-American Conference on Parkinson’s Disease (eds McDowell F. and Barbeau A.) Vol. 5, pp. 153-160. Raven Press, New York. in the in situ spinal 47. Magherini P. C., Precht W. and Schwindt P. C. (1976) Electrical properties of frog motoneurons cord. J. Neurophysiol. 39, 459-473. 48. Matsuda Y. and Jinnai K. (1980) Afferent inputs to caudate output neurons as monitored by antidromic spike invasion of the cell soma. Bruin Res. 188, 560-565. in a lobster motor 49. Nagy F., Dickinson P. S. and Moulins M. (1981) Rhythmical synaptic control of axonal conduction neuron. J. Neurophysiol. 45, 1 lO9- 1124.

Nigral

dopamine

neurons:

Electrical

characterization

331

50. Nowycky M. C. and Roth R. H. (1978) Dopaminergic neurons: role of presynaptic receptors in the regulation of transmitter biosynthesis. Prog. Neuropsychopharmac. 2, 139-I 58. 51. Palay S. F., Sotelo C., Peters A. and Orkland P. M. (1968) The axon hillock and the initial segment. J. Cell Biol. 38, 193-201. 52. Peters A., Proskauer C. C. and Kaiserman-Abramof I. R. (1968) The small pyramidal neuron of the rat cerebral cortex. The axon hillock and initial segment. J. Cell Biol. 39, 604-619. 53. Precht W. and Yoshida M. (1971) Blockage of caudate-evoked inhibition of neurons in the substantia nigra by picrotoxin. Brain Res. 32, 229-233. 54. Preston R. J., McCrea R. A., Chang H. T. and Kitai S. T. (1981) Anatomy and physiology of substantia nigra and retrorubal neurons studied by extra- and intracellular recording and by horseradish peroxidase labeling. Neuroscience 6, 331-344. 55. Renshaw B. (1942) Effects of presynaptic volleys on spread of impulses over the soma of the motoneuron. J. Neurophysiol. 5, 235-243. 56. Rinvik E. and Grofova I. (1970) Observations on the fine structure of the substantia nigra in the cat. Expl. Brain Res. 11, 229-248. 57. Roth R. H. (1979) Dopamine autoreceptors: pharmacology, function and comparison with postsynaptic dopamine receptors. Comm. Psychopharmac. 3, 429445. 58. Schwyn R. C. and Fox C. A. (1974) The primate substantia nigra: a Golgi and electron microscopic study. J. Hirnforsch. 16, 95-126. 59. Shepherd G. M. (1979) The Synaptic Organization of the Brain. Second edition. Oxford University Press, New York. 60. Skirboll L. R., Grace A. A., Hommer D. W., Rehfeld J., Goldstein M., Hokfelt T. and Bunney B. S. (1981) Peptide-monoamine coexistence; studies of the actions of cholecystokinin-like peptide on the electrical activity of midbrain dopamine neurons. Neuroscience 6, 21 I I-2124. 61, Smith D. 0. (1980) Morphological aspects of the safety factor for action potential propagation at axon branch points in the crayfish. J. Physiol., Lond. 301, 361-369. 62. Steriade M., Apostal V. and Oakson G. (1971) Control of unitary activities in cerebellothalamic pathway during wakefulness and synchronized sleep. J. Neurophysiol. 34, 389413. 63. Stewart W. W. (1978) Functional connections between cells as revealed by dye-coupling with a highly fluorescent naphthahmide tracer. Cell 14, 741-759. 64. Takato M. and Goldring S. (1979) Intracellular marking with Lucifer Yellow CH and horseradish peroxidase of cells electrophysiologically characterized as gha in the cerebral cortex of the cat. J. camp. Neurol. 186, 173-188. 65. Taut L. (1962) Site of origin and propagation of spikes in the giant neuron of Aplysiu. J. gen. Physiol. 45, 1077-1097. 66. Terzuolo C. A. and Araki T. (1961) An analysis of intra- versus extracellular potential changes associated with activity of single spinal motoneurons. Ann. N.Y. Acad. Sci. 94, 547-558. 67. Van Essen D. C. (1973) The contribution of membrane hyperpolarization to adaptation and conduction block in sensory neurones of the leech. J. Physiol., Lond. 230, 509-534. 68. Waxman S. G. (1972) Regional differentiation of the axon: a review with special reference to the concept of the multiplex neuron. Brain Res. 47, 269-288. 69. Westrum L. E. (1970)Observations on initial segments of axons in the prepyriform cortex of the rat. J. camp. Neurol. 139, 337-356. 70. Yau K.-W. (1976) Receptive fields, geometry and conduction block of sensory neurones in the central nervous system of the leech. J. Physiol., Lond. 263, 513-538. (Accepted 9 February 1983)

NSC 10/2-t