Neurons with distinctive firing patterns, morphology and distribution in laminae V–VII of the...
-
Upload
peter-szucs -
Category
Documents
-
view
213 -
download
1
Transcript of Neurons with distinctive firing patterns, morphology and distribution in laminae V–VII of the...
Neurons with distinctive firing patterns, morphology anddistribution in laminae V–VII of the neonatal rat lumbarspinal cord
Peter Szucs, Francis Odeh, Karolina Szokol and Miklos AntalDepartment of Anatomy, Histology and Embryology, Faculty of Medicine, Medical & Health Science Centre, University of Debrecen,Debrecen, H-4012 Hungary
Keywords: biocytin labelling, slice preparation, spinal interneurons, whole cell patch clamp
Abstract
It is generally accepted that neurons in the ventral spinal grey matter, a substantial proportion of which can be regarded as constituentsof the spinal motor apparatus, receive and integrate synaptic inputs arising from various peripheral, spinal and supraspinal sources.Thus, a profound knowledge concerning the integrative properties of interneurons in the spinal ventral grey matter appears to beessential for a fair understanding of operational principles of spinal motor neural assemblies. Using the whole cell patch clampconfiguration in a correlative physiological and morphological experimental approach, here we demonstrate that the intrinsic membraneproperties of neurons vary widely in laminae V–VII of the ventral grey matter of the neonatal rat lumbar spinal cord. Based on their firingpatterns in response to depolarizing current steps, we have classified the recorded neurons into four categories: ‘phasic’, ‘repetitive’,‘single’ and ‘slow’. Neurons with firing properties characteristic of the ‘phasic’, ‘repetitive’ and ‘single’ cells have previously beenreported also in the superficial and deep spinal dorsal horn, but this is the first account in the literature in which ‘slow’ neurons have beenrecovered and described in the spinal cord. The physiological heterogeneity in conjunction with the morphological correlation anddistribution of neurons argues that different components of motor neural assemblies in the spinal ventral grey matter possess differentsignal processing characteristics.
Introduction
There is general agreement that neural assemblies that are responsible
for motor rhythm and pattern generation, and thus for the initiation of
rhythmic alternating motor behaviours like locomotion are confined to
the ventral aspects of the spinal grey matter (Szekely, 1963; Pratt &
Jordan, 1987; Schmidt et al., 1988; Pearson, 1993; Kjaerulff & Kiehn,
1996). It is also established that spinal motor neural circuits constitute
a functionally heterogeneous population of neurons. It has been shown
that some of these cells contribute to networks generating pacemaker
like activity (Ho & O’Donovan, 1993; Kjaerulff & Kiehn, 1996), while
others retain mechanisms that render a well-balanced spatial and/or
temporal integration of synaptic inputs arising from various sources
possible (Harrison & Jankowska, 1985; Morisset & Nagy, 1998;
Garraway & Hochman, 2001; Nakatsuka et al., 2002). For instance,
in addition to a strong drive from the spinal motor rhythm and pattern
generating circuits, last-order premotor interneurons, neurons that
conduct nerve signals to motoneurons, may also receive sensory inputs
and commands from various supraspinal motor centres (Moschovakis
et al., 1991; Jankowska, 1992; Degtyarenko et al., 1996, 1998). Thus,
they have crucial importance in the integration of activities generated
by the spinal motor apparatus, sensory information and nerve impulses
descending from voluntary and nonvoluntary higher motor centres.
Although the sources of synaptic inputs that are received by other
elements of the spinal motor neural assembly may vary in a wide range,
it is quite possible that similar to last-order premotor interneurons,
most, if not all, of the neurons in this circuitry receive multiple inputs.
The integration of various synaptic inputs on the somato-dendritic
membrane depends on several factors including the intrinsic passive
and active membrane properties as well as the morphology of the
neurons. Therefore extensive morphological studies in the ventral
aspect of the spinal grey matter has been performed (Kjaerulff et al.,
1994; Moschovakis et al., 1992; Puskar & Antal, 1997). The input
properties and the participation of some ventral horn neurons in various
reflex pathways have also been characterized (Brink et al., 1983;
Edgely & Jankowska, 1987; Wheatley et al., 1994; Jonas et al., 1998).
However, our present comprehension concerning the active and pas-
sive membrane properties of these neurons is limited, although a
profound knowledge of this matter appears to be a prerequisite for
a fair understanding of fundamental operational characteristics of
spinal neural assemblies in laminae V–VII.
Accordingly, in the experiments presented here we investigated the
firing patterns of neurons in response to suprathreshold depolarizing
current pulses in laminae V–VII of the neonatal rat lumbar spinal cord.
We have also made an attempt to find some correlation among the
physiological and morphological properties as well as the location of
the investigated neurons. Preliminary observations from this study
have been reported in abstract form (Szucs et al., 2002).
European Journal of Neuroscience, Vol. 17, pp. 537–544, 2003 � Federation of European Neuroscience Societies
doi:10.1046/j.1460-9568.2003.02484.x
Correspondence: Dr Peter Szucs, as above.
E-mail: [email protected]
Received 12 August 2002, revised 26 November 2002, accepted 2 December 2002
Materials and methods
Preparation of spinal cord slices
Experiments were carried out on new-born and young (P0–8) Wistar–
Kyoto rats (Godollo, Hungary). All animal study protocols were
approved by the Animal Care and Protection Committee at the
University of Debrecen, Hungary, and were carried out in accordance
with the European Communities Council Directives. Under deep
isofluorane anaesthesia, the animals were decapitated and the lumbar
spinal cord was dissected in an ice cold artificial cerebrospinal fluid
(ACSF, pH 7.4) that contained (in mM): NaCl, 130; NaHCO3, 24; KCl,
3.5; NaH2PO4, 1.25; CaCl2, 1; MgSO4, 3; and glucose, 10, and was
saturated with 95% O2 and 5% CO2. After removing the meninges,
blocks of the lumbar spinal cord were embedded into agar and
sectioned at 300–400 mm on a Vibratome. Slices were incubated in
ACSF at room temperature for at least 1 h prior to recording.
Electrophysiological recordings
Slices were transferred into a recording chamber, which was constantly
perfused with oxygenated ACSF. Neurons located in laminae V–VII
were visually identified with a Zeiss Axioskop FS microscope
equipped with a �40 water immersion objective, DIC filter and an
infrared CCD camera system (Hamamatsu, Japan). Conventional
whole-cell patch clamp recordings were performed in current clamp
mode (analogous to bridge mode of conventional microelectrode
amplifiers) using an Axopatch 1D amplifier (Axon Instruments, Union
City, CA, USA) and pipettes with a resistance of 4–7 MO. The
electrode filling solution contained (in mM): K-gluconate, 126;
KCl, 4; ATP, 4; GTP, 0.3; Phosphocreatine, 10; Hepes, 10; and
0.5% biocytin (Sigma, St. Louis, MO, USA) at pH 7.2. To prevent
spontaneous firing, which is not unusual in slice preparations due to
either electrode penetration or disruption of neuronal circuitry during
slicing, the membrane potential was kept constant in the range between
�60 mV and �65 mV throughout the experiment with a tonic hyper-
polarizing DC current injection in the range of 0–100 pA, a commonly
used method (Maccaferri et al., 2000). If the DC current needed to
keep the cells stable increased above the given range during the
experiment, the recording was discarded from the study. Firing
patterns were obtained by applying 800 ms long incrementing current
steps ranging from �30 pA to 100 pA. Capacitance and series resis-
tance were not compensated. All recordings were performed at room
temperature.
Data acquisition and analysis
Data were recorded on an IBM PC, filtered at 5 kHz, digitized with a
Digidata 1320 A A/D board and analyzed by using pClamp (Axon
Instruments, Union City, CA, USA), Origin (Microcal Software,
Northampton, MA, USA) and Whole Cell Program and Electrophy-
siology Data Recorder (courtesy of Dr J. Dempster, University of
Strathclyde, UK) software packages.
From firing patterns evoked by the application of suprathreshold
current steps (100 pA, 800 ms) the following parameters were ana-
lyzed: spike frequency, interspike intervals (ISI), ratio between the
amplitudes of the first and last spikes.
Membrane time constant (t) was determined by fitting one decaying
exponential function to the cells’ voltage response to small negative
current steps (�10 pA, 800 ms). Current–voltage (I–V) curves were
constructed by plotting the amplitude of the membrane potential
change against the amplitude of the applied subthreshold current step.
Action potential and afterhyperpolarization (AHP) amplitudes were
measured from action potential threshold. The half-width of the action
potentials was determined at half peak amplitude measured from
threshold. Action potential threshold and the shape of the AHP was
evaluated at the lowest current step that generated action potentials
from resting membrane potential (Ruscheweyh & Sandkuhler, 2002).
Histological processing and morphological evaluation
Following the recording session, the slices were placed between two
Millipore filters (Millipore Corporation, Bedford, MA, USA) to avoid
deformation and transferred into a fixative containing 4% paraformal-
dehyde, 1.25% glutaraldehyde and 0.2% picric acid in 0.1 M phosphate
buffer (pH 7.4) for 1–4 days. Slices were sectioned at 60mm on a
Vibratome. To visualize biocytin that diffused from the patch pipette
into the recorded neuron, sections were treated according to the avidin-
biotinylated horseradish peroxidase method (Extravidin, diluted
1 : 1000, Vector Laboratories, Burlingame, CA, USA) and the histo-
chemical reaction was completed with a diaminobenzidine (Sigma, St.
Louis, MO, USA) chromogen reaction. Sections were counterstained
with toluidine blue, dehydrated and mounted with DPX (Fluka, Buchs,
Switzerland).
The somata as well as the dendritic and axonal arbors of the recorded
and labelled neurons were recovered and three dimensionally recon-
structed from serial sections using a computer aided NEUROLUCIDA
reconstruction system (Microbrightfield Inc., Villiston, VT, USA)
installed onto a Leitz Laborlux microscope. To define the laminar
location of the recorded and labelled cells in the spinal grey matter we
have applied the segmentation scheme of Kjaerulff et al., (1994) that
they constructed specifically for the neonatal rat lumbar spinal cord.
The maximum cross-sectional area as well as the minimum and
maximum diameters of the somata were measured. The ratio of the
maximum and minimum diameters (aspect ratio) was calculated.
Dendrograms of the labelled neurons were constructed, and the
numbers of stem dendrites and branch points along the dendritic trees
in relation to the distance from the cell body were counted. No variance
analysis of the dendritic and axonal distribution was performed inside
or between the different cell groups.
All curve fits were performed using the built-in iterative Levenberg–
Marquadt algorithm in ClampFit (Axon Instruments, Union City, CA,
USA). All numerical data are presented as mean� SEM. Statistical
significance was assessed using Student’s t-test.
Results
Seventy neurons were recorded. In certain cases more than one cell
was recorded in the same slice. Both the electrophysiological record-
ings and the biocytin labelling were successful in 47 cases. Conse-
quently, data presented in this study are based on the results obtained
from these 47 cells. The average resting membrane potential of the
recorded cells were �51.71� 1.52 mV. On the basis of their firing
patterns evoked by suprathreshold depolarizing current pulses the
recorded neurons were divided into four groups that we classified
as: (i) ‘phasic’ neurons (10 cells); (ii) ‘repetitive’ neurons (15 cells);
(iii) ‘single’ neurons (9 cells); and (iv) ‘slow’ neurons (8 cells). Five
neurons could not be classified according to this scheme; their firing
patterns showed similarities partly to the ‘phasic’ partly to the ‘repe-
titive’ neurons.
Firing patterns
‘Phasic’ neurons discharged a continuous train of action potentials
during suprathreshold depolarizing current pulses (Fig. 1A). In addi-
tion, they showed a marked spike accommodation including a pro-
gressive increase in interspike intervals and attenuation of spike
amplitude during the 800 ms long depolarizing current step (Fig. 1B).
Short duration monophasic AHPs were also observed after the spikes
538 P. Szucs et al.
� 2003 Federation of European Neuroscience Societies, European Journal of Neuroscience, 17, 537–544
(Fig. 1C). The average ratios between the first and last spike ampli-
tude and interspike interval were 5.56� 0.72 and 0.44� 0.03, respec-
tively.
Similar to ‘phasic’ cells, ‘repetitive’ neurons also discharged a
continuous train of action potentials during suprathreshold depolariz-
ing current pulses (Fig. 1A). In contrast to the ‘phasic’ cells, however,
we did not observe any obvious sign of spike accommodation or
attenuation (Fig. 1B), but spikes were always followed by a marked
slow monophasic AHP (Fig. 1C).
In contrast to all of the other cell types, ‘single’ neurons fired a
single (in some cases two) action potential upon suprathreshold
depolarization (Fig. 1A). Consecutive spikes could be evoked only
after repolarization to resting membrane potential. In these cells only a
small AHP could be detected after the single spike (Fig. 1C).
Similar to the ‘repetitive’ cells, ‘slow’ neurons also discharged a
continuous train of action potentials during suprathreshold depolar-
ization (Fig. 1A). There was no obvious sign of spike accommodation
or attenuation, and spikes were always followed by a long-lasting
Fig. 1 Firing patterns of recorded neurons in laminae V–VII. (Column A) Traces show the voltage response to the current steps indicated below. The small inserts(Column B) demonstrate the relationship between the consecutive action potential amplitudes. ‘Phasic’ neurons showed a marked accommodation of both spikeamplitude and interspike intervals. ‘Repetitive’ neurons maintained their spike amplitudes and frequency during the suprathreshold depolarization induced dischargeperiod. ‘Single’ neurons fired only one action potential. ‘Slow’ neurons showed a low spike frequency but no obvious sign of spike accommodation or attenuation.(Column C) Enlarged initial spikes of a neuron from each firing pattern group. Arrowheads indicate the AHPs following the spikes.
� 2003 Federation of European Neuroscience Societies, European Journal of Neuroscience, 17, 537–544
Interneurons in laminae V–VII of the spinal cord 539
monophasic AHP (Fig. 1B and C). However, the average spike fre-
quency of ‘slow’ neurons (8.9� 0.6Hz) was significantly lower than
that of the ‘repetitive’ cells (21.83� 1.5Hz) (P< 0.001).
Passive and active membrane parameters
Although the resting membrane potential and input time constant
tended to be smaller in ‘slow’ and ‘single’ neurons than in ‘phasic’ and
‘repetitive’ cells, statistical analysis revealed that the values of resting
membrane potentials, input time constants, action potential overshoots
and half-widths did not differ significantly among the neurons with
different firing patterns. The smallest average input time constant was
observed in the ‘single’ group (26.6� 8.82 ms). Also, ‘phasic’ and
‘single’ cells showed a tendency for wider action potentials than that of
the other two firing pattern groups, although a statistically significant
difference could not be established (Table 1).
However, certain parameters made the ‘single’ and ‘slow’ cells
unique and obviously different from the ‘phasic’ and ‘repetitive’
neurons. The amplitudes of action potentials were similar in ‘phasic’,
‘repetitive’ and ‘slow’ neurons. ‘Single’ cells presented action poten-
tials with the smallest amplitude (54.44� 4.14 mV) and overshoot
(16.84� 6.16 mV) but with the highest threshold (�40.94� 5.95 mV)
among all types of neurons. Action potential thresholds were similar in
the ‘phasic’ and ‘repetitive’ group, while ‘slow’ neurons had the lowest
threshold values (�45.2� 1.47). Similarly to the ‘phasic’ and ‘repe-
titive’ cells, ‘slow’ neurons presented a strong afterhyperpolarization,
while the average amplitude of AHP was significantly smaller in the
‘single’ cell population (Table 1). The current–voltage curves appeared
to be flat in ‘slow’ and ‘single’ neurons, while they were steeper in
‘repetitive’ and steepest in the ‘phasic’ cells (Fig. 2), indicating that the
input resistance of ‘slow’ and ‘single’ cells is lower than that of
neurons in the ‘repetitive’ and especially in the ‘phasic’ groups. Indeed
‘slow’ cells presented the smallest input resistance among the recorded
neurons (Table 1).
Laminar distribution and morphometric parameters
Of the 47 reconstructed neurons, 13 were located within laminae V–VI
and 34 in lamina VII. Similarly to membrane parameters, the dis-
tribution and morphology of ‘phasic’ and ‘repetitive’ neurons were
similar to each other, while ‘single’ and ‘slow’ cells showed unique
characteristics.
Most of the ‘phasic’ (8 out of 10) and ‘repetitive’ (13 out of 15)
neurons presented multipolar perikarya with 4–6 primary dendrites,
and were distributed equally in laminae V–VII (Fig. 3). Of the nine
‘single’ neurons seven were recovered in the lateral aspect of laminae
V–VI, and only two were located in lamina VII. In these seven neurons
the small multipolar perikarya gave rise to 3–5 stem dendrites, and
presented an extensive but poorly arborizing dendritic tree with a
preferred orientation in the dorso-ventral direction (Fig. 3). In five cells
out of the seven, the axon was also labelled. After arising from the cell
body the axons turned ventrally and terminated in the ventral grey
matter. From the neurons investigated in this study, ‘slow’ neurons
presented the largest and most elongated perikarya. Six out of the eight
recovered ‘slow’ neurons had 6–8 primary dendrites, were arranged in
two groups that arose from the opposite poles of the elongated somata
Fig. 2 Current–voltage relations of neurons in laminae V–VII. The graph showsthe amplitude of the membrane voltage change plotted against the appliedcurrent step. All numerical data are presented as mean� SEM.
Table 1. Passive and active membrane parameters of neurons in laminae V–VII
Neuron type
Phasic(n¼ 9)
Repetitive(n¼ 14)
Slow(n¼ 7)
Single(n¼ 8)
Significance(Student’s t-test)
Resting membrane potential (mV) �51� 4.2 �53.6� 2.54 �49� 2.79 �47� 3.97 NSInput time constant (ms) 50.83� 7.74 41.81� 5.4 32.36� 4.42 26.6� 8.82 NSInput resistance (MW) 994� 197 822� 151 304� 59 643� 166 Phasic vs. slow, P< 0.01;
repetitive vs. slow, P< 0.01;phasic vs. single, P< 0.05.
Action potential threshold (mV) �43.8� 2.05 �43.7� 3.17 �45.2� 1.47 �40.94� 5.95 NSAction potential amplitude (mV) 69� 4.48 65.79� 2.57 69.03� 2.3 54.44� 4.14 Phasic vs. single, P< 0.05;
repetitive vs. single, P< 0.05;slow vs. single, P< 0.01.
Action potential overshoot (mV) 25.15� 5.28 23.97� 3.95 23.42� 4.06 16.84� 6.16 NSAction potential half-width (ms) 1.41� 0.17 1.08� 0.08 1.1� 0.07 1.52� 0.2 NSAHP amplitude (mV) 13.26� 2.49 16.98� 2.3 14.79� 2.16 8.91� 1.35 Repetitive vs. single, P< 0.01;
slow vs. single, P< 0.05.
Average spike frequency (Hz) NA 21.83� 1.5 8.9� 0.6 NA Repetitive vs. slow, P< 0.001.
AHP, afterhyperpolarization; NA, not applicable; NS, no significant differences.
� 2003 Federation of European Neuroscience Societies, European Journal of Neuroscience, 17, 537–544
540 P. Szucs et al.
and formed a richly arborizing bushy dendritic arbor. In contrast to the
other cells, ‘slow’ neurons were confined to the ventromedial part of
lamina VII (Fig. 3). In the case of seven neurons the axon was also
labelled. The labelled axon arose from the cell body and extended
either towards the midline or turned laterally. Of the five medially
orientated axons, four crossed the midline in the anterior commissure
and terminated in the contralateral ventral grey matter. The laterally
orientated axons targeted the lateral motor column.
Measuring the cross-sectional areas as well as the largest and
smallest diameters of stained perikarya it turned out that the shape
and size of cell bodies of ‘phasic’, ‘repetitive’ and ‘single’ neurons
were very similar (Table 2). According to these parameters, however,
the perikarya of ‘slow’ neurons turned out to be different from the
others. They presented a significantly larger and apparently more
elongated perikarya than neurons in the other groups (Table 2).
Analysis of branch point distribution histograms obtained from
dendrograms of neurons with different firing patterns showed that
in this regard the dendritic arborization pattern of ‘phasic’, ‘repetitive’
and ‘single’ neurons were quite similar to each other. However, ‘slow’
neurons presented dendritic structures that were substantially different
from the dendritic arbores of cells in the other groups. The long,
slender dendrites of ‘phasic’, ‘repetitive’ and ‘single’ neurons
Fig. 3 Photomicrograph and Neurolucida reconstruction of representative samples of ‘phasic’, ‘repetitive’, ‘single’ and ‘slow’ neurons. Cell bodies and dendrites areshown in black, while the axons are grey. Dots on the schematic representations of the neonatal spinal cord slice adopted from Kjaerullf et al. (1994) indicate the locationof successfully recorded and labelled neurons in each group. On the drawings, the borders between the grey and white matter are drawn with dashed lines, whereascontinuous grey lines represent the borders of Rexed laminae. Roman numbers identify Rexed laminae V–VI and VII. D, dorsal; M, medial. Scale bars, 100mm.
� 2003 Federation of European Neuroscience Societies, European Journal of Neuroscience, 17, 537–544
Interneurons in laminae V–VII of the spinal cord 541
branched infrequently, in a way that most of the branching points were
scattered within a distance of 100–150 mm from the soma (Fig. 4). In
the ‘phasic’ and ‘single’ groups, one or two primary dendrites occa-
sionally extended away from the cell body in a way that they gave
branches only in a distance of 300–600mm from the soma. This distant
accumulation of branching points presented a second smaller hump on
the branch point distribution histograms (Fig. 4). In contrast to this,
‘slow’ neurons formed a richly arborizing bushy dendritic arbor in
which the dendrites continuously branched as they receded away from
the cell body (Fig. 4).
Discussion
In this study we have documented that intrinsic membrane properties
as well as the morphology of neurons vary widely in the ventral
grey matter of the neonatal rat lumbar spinal cord. Based on their
firing patterns in response to depolarizing current steps, we have
classified the recorded neurons into four categories, and distinguished
‘phasic’, ‘repetitive’, ‘single’ and ‘slow’ neurons. Neurons with firing
properties characteristic of the ‘phasic’, ‘repetitive’ and ‘single’ cells
have previously been reported in the superficial and deep dorsal
horn of the spinal cord (Hochman et al., 1997; Prescott & Koninck,
2002) as well as among cultured spinal neurons (Jo et al., 1998).
However, this is the first account in the literature in which ‘slow’
neurons have been recovered and described in the spinal cord. Al-
though the input resistance values of the neurons in our sample were
slightly different from the values described in the above studies, other
authors report similar data especially in young animals (Hochman
et al., 1997).
We have frequently observed spontaneous excitatory and inhibitory
postsynaptic potentials throughout the recordings. Although this
spontaneous activity in the network might modify the discharge
patterns of the recorded neurons, it is likely that this external influence
had little, if any effect on our results, as ‘phasic, ‘repetitive’ and
‘single’ neurons have previously been identified also among cultured
spinal neuron (Jo et al., 1998), where this network effect can obviously
be excluded. Thus we assume that the different discharge patterns of
neurons that we have demonstrated here are primarily determined by
the intrinsic membrane properties of the recorded neurons.
‘Phasic’ and ‘repetitive’ neurons
The separation of ‘phasic’ and ‘repetitive’ neurons into two distinct
categories might be artificial. They may only represent the two
extremes of one heterogeneous cell population. The fact that in our
sample there were five neurons that could not be classified according to
this scheme, as their firing patterns showed similarities partly to the
‘phasic’ partly to the ‘repetitive’ neurons argues in favour of this
notion. All of the ‘phasic’, ‘repetitive’ and unclassified neurons show a
continuous high frequency spiking to suprathreshold stimulation,
where spikes are followed by AHPs with similar amplitude. This
AHP is likely to be responsible for the spike frequency accommodation
in the case of ‘phasic’ cells. The mechanism behind the decreasing
spike amplitude in ‘phasic’ neurons can be either deactivation of
sodium channels and/or a raised spike threshold level. One could also
argue that transient electrode blockage during the depolarizing current
step could result in a similar spike amplitude decrease. However, the
random occurrence/development of such electrode blockage during
the course of the experiment would also change the firing pattern of
neurons in all the other groups; this was not observed in our experi-
ments. Furthermore, this typical firing pattern was also reported earlier
by others in cells with membrane properties similar to the neurons in
our ‘phasic’ group (Hochman et al., 1997; Jo et al., 1998).
Paradoxically neurons with a characteristic ‘repetitive’ discharge
pattern, present slightly larger and markedly longer AHPs, without
obvious signs of spike accommodation. It is possible that the more
developed AHP may allow a more effective de-inactivation of sodium
channels, and therefore ensures the uniform amplitudes and threshold
levels of spikes in case of the ‘repetitive’ cells (Thomson et al., 1989).
Such a mechanism requires further investigation.
Neurons with the ‘phasic’ and ‘repetitive’ properties are scattered all
over the entire cross-sectional area of the spinal grey matter. In
addition to laminae V–VII, where we located them in the present
study, neurons with identical firing patterns were identified also in the
superficial and deep spinal dorsal horn (Hochman et al., 1997; Prescott
& Koninck, 2002). Studying the physiological properties of these
neurons in lamina I, Prescott & Koninck, 2002) concluded that
‘repetitive’ (tonic in their classification scheme) and ‘phasic’ cells
respond in a graded fashion over a wide range of stimulus intensities
and their slow synaptic events and slow membrane time constant
promote temporal summation or integration of synaptic inputs.
‘Single’ neurons
‘Single’ neurons responded with only one, occasionally with two,
action potential at the onset of suprathreshold current steps. This may
indicate that these neurons are incapable of encoding stimulus inten-
sity through firing frequency but could follow high frequency stimula-
tion that may arise from various sources including burst firing or
rhythmically active neurons (Thomson et al., 1989). High frequency
stimulation of a neuron, however, can be accomplished only in those
cases when synaptic inputs generate synaptic potentials on various
areas of the somatodendritic membrane more or less at the same time.
Thus, ‘single’ neurons require a substantial coincident input from one
or different sources to allow the spatial summation of postsynaptic
potentials necessary for driving the neuron to fire. Once they are
activated they become incapable of generating any consecutive spikes
to subsequent incoming signals. To regain their responsiveness they
Table 2. Morphometric parameters of recorded and labelled neurons in laminae V–VII
Neuron type
Phasic(n¼ 10)
Repetitive(n¼ 15)
Slow(n¼ 8)
Single(n¼ 9)
Significance(Student’s t-test)
Average maximum cross-sectional area (mm2) 209� 19.69 268.55� 30.72 376.49� 37.61 209.32� 26.34 Phasic vs. slow, P< 0.01;repetitive vs. slow, P< 0.05;single vs. slow, P< 0.01.
Maximum diameter (mm2) 24.57� 2.63 27.42� 2.45 35.8� 3.91 21.26� 1.4 NSMinimum diameter (mm2) 13.38� 0.77 15.85� 1.09 17.47� 0.84 14.08� 0.79 NSAspect ratio 1.84� 0.16 1.81� 0.22 2.11� 0.3 1.51� 0.07 NS
NS, no significant differences.
� 2003 Federation of European Neuroscience Societies, European Journal of Neuroscience, 17, 537–544
542 P. Szucs et al.
need to be repolarized to resting membrane potential. Thus they can
also be regarded as onset detectors (Ruscheweyh & Sandkuhler, 2002).
In our sample most ‘single’ neurons were distributed in lamina
V–VI. Neurons with similar firing patterns were reported also in the
superficial and deep dorsal horn (Hochman et al., 1997; Prescott &
Koninck, 2002), where various types of primary afferents terminate
(Fyffe, 1984). These observations which show an overlapping between
the distribution of primary afferent terminals and ‘single’ neurons
Fig. 4 Dendritic branching patterns of neurons in laminae V–VII. (Column A) Averaged branching point distribution histograms of neurons in the four firing patterngroups. Number of dendritic branching points counted in 50 mm bins are plotted against the distance from the soma. (Columns B and C) Dendrograms constructedfrom three-dimensional reconstruction showing the length and branching pattern of dendrites originating from the cell body and Neurolucida reconstruction’s of arepresentative cell from each firing pattern group. Neurons in the ‘phasic’, ‘repetitive’ and ‘single’ groups had dendrites branching mainly within 100–150mmdistance from the soma. Cells in the ‘slow’ group presented dendrites that were continuously branching as they receded away from the cell body. Occasionally in the‘phasic’ and ‘repetitive’ groups a second distant accumulation of branching points occurred, which is represented on the histograms as a second small hump between300 and 600mm. Scale bars, 100mm. Numerical data in the histograms are presented as mean� SEM.
� 2003 Federation of European Neuroscience Societies, European Journal of Neuroscience, 17, 537–544
Interneurons in laminae V–VII of the spinal cord 543
suggest, that ‘single’ neurons among many other possible sources may
receive innervation also from primary sensory neurons. This notion is
reinforced by the finding that Ia inhibitory interneurons, that are known
to be heavily innervated by primary afferents (Jankowska, 1992), have
also been classified as ‘single’-spiking neurons in the adult cat spinal
cord (Jankowska, 2001).
‘Slow’ neurons
‘Slow’ neurons showed unique physiological and morphological
characteristics that made them substantially different from the other
cell types. During suprathreshold depolarization, they fired a contin-
uous and low frequency train of action potentials without any obvious
sign of spike accommodation or attenuation. In addition, this was the
only cell type that showed ongoing spontaneous firing (not demon-
strated here) at resting membrane potential, when there was no
hyperpolarizing current injected. The spontaneous firing together with
the strong and elongated afterhyperpolarization that always followed
the action potentials resembled a pacemaker like activity. It also
appears to be important that all ‘slow’ neurons recorded in this study
were confined to the ventromedial grey matter of the spinal cord, where
Ole Kiehn and his collaborators observed rhythmically active neurons
in the neonatal rat spinal cord (personal communication). Ablation
studies carried out on embryonic and neonatal chicken and rats
unequivocally demonstrated that this area of the spinal grey matter
is essential in the generation of rhythmic motor behaviours like
locomotion (Ho & O’Donovan, 1993; Kjaerulff & Kiehn, 1996; Tresch
& Kiehn, 1999). Taking all of this together, we assume that ‘slow’
neurons may represent essential constituents of rhythm and/or pattern
generating spinal motor neural assemblies that are responsible for the
generation of flexor-extensor and left-right alternation. The findings
that the axons of some ‘slow’ neurons project to the ipsilateral motor
column, whereas others cross the midline and terminate in the con-
tralateral grey matter appear to reinforce this notion.
Acknowledgements
This was work supported by the Hungarian National Science Fund (OTKA T032075) and the Ministry of Welfare of Hungary (ETT 04–032/2000). Theauthors would also like to thank SUPERTECH (Hungary), for all their technicalsupport and Dr G. Veress for his help with the figure preparation.
Abbreviations
ACSF, artificial cerebrospinal fluid; AHP, afterhyperpolarization; DIC,differential interference contrast; ISI, interspike intervals.
References
Brink, E., Harrison, P.J., Jankowska, E., Mccrea, D.A. & Skoog, B. (1983) Post-synaptic potentials in a population of motoneurones following activity ofsingle interneurones in the cat. J. Physiol. (Lond.), 343, 341–359.
Degtyarenko, A.M., Simon, E.S. & Burke, R.E. (1998) Locomotor modulationof disynaptic EPSPs from the mesencephalic locomotor region in catmotoneurons. J. Neurophysiol., 80, 3284–3296.
Degtyarenko, A.M., Simon, E.S. & Burke, R.E. (1996) Differential modulationof disynaptic cutaneous inhibition and excitation in ankle flexor motoneuronsduring fictive locomotion. J. Neurophysiol., 76, 2972–2985.
Edgely, S.A. & Jankowska, E. (1987) An interneuronal relay for group I and IImuscle afferents in the midlumbar segments of the cat spinal cord. J. Physiol.(Lond.), 389, 647–674.
Fyffe, R.E.W. (1984) Afferent fibres. In Davidoff, R.A., (Ed), Handbook ofthe Spinal Cord. Marcel Dekker Inc., New York, pp. 79–136.
Garraway, S.M. & Hochman, S. (2001) Modulatory actions of serotonin,norepinephrine, dopamine, and acetylcholine in spinal cord deep dorsal hornneurons. J. Neurophysiol., 86, 2183–2194.
Harrison, P.J. & Jankowska, E. (1985) Sources of input to interneuronesmediating group I non-reciprocal inhibition of motoneurones in the cat.J. Physiol. (Lond.), 361, 379–401.
Ho, S. & O’Donovan, M.J. (1993) Regionalization and intersegmental coordi-nation of rhythm-generating networks in the spinal cord of the chick embryo.J. Neurosci., 13, 1354–1371.
Hochman, S., Garraway, S.M. & Pockett, S. (1997) Membrane properties ofdeep dorsal horn neurons from neonatal rat spinal cord in vitro. Brain Res.,767, 214–219.
Jankowska, E. (1992) Interneuronal relay in spinal pathways from propriocep-tors. Prog. Neurobiol., 38, 335–378.
Jankowska, E. (2001) Spinal interneuronal systems: identification, multifunc-tional character and reconfigurations in mammals. J. Physiol. (Lond.), 533,31–40.
Jo, Y.H., Stoeckel, M.E. & Schlichter, R. (1998) Electrophysiological propertiesof cultured neonatal rat dorsal horn neurons containing GABA and met-enkephalin-like immunoreactivity. J. Neurophysiol., 79, 1583–1586.
Jonas, P., Bischofberger, J. & Sandkuhler, J. (1998) Corelease of two fastneurotransmitters at a central synapse. Science, 281, 419–424.
Kjaerulff, O., Barajon, I. & Kiehn, O. (1994) Sulphorhodamine-labelled cells inthe neonatal rat spinal cord following chemically induced locomotor activityin vitro. J. Physiol. (Lond.), 478, 265–273.
Kjaerulff, O. & Kiehn, O. (1996) Distribution of networks generating andcoordinating locomotor activity in the neonatal rat spinal cord in vitro: alesion study. J. Neurosci., 16, 5777–5794.
Maccaferri, G., Roberts, J.D., Szucs, P., Cottingham, C.A. & Somogyi, P. (2000)Cell surface domain specific postsynaptic currents evoked by identifiedGABAergic neurones in rat hippocampus in vitro. J. Physiol. (Lond.),524, 91–116.
Morisset, V. & Nagy, F. (1998) Nociceptive integration in the rat spinal cord:role of non-linear membrane properties of deep dorsal hornneurons. Eur.J. Neurosci., 10, 3642–3652.
Moschovakis, A.K., Sholomenko, G.N. & Burke, R.E. (1991) Differentialcontrol of short latency cutaneous excitation in cat FDL motoneurons duringfictive locomotion. Exp. Brain Res., 83, 489–501.
Moschovakis, A.K., Solodkin, M. & Burke, R.E. (1992) Anatomical andphysiological study of interneurons in an oligosynaptic cutaneous reflexpathway in the cat hindlimb. Brain Res., 586, 311–318.
Nakatsuka, T., Furue, H., Yoshimura, M. & Gu, J.G. (2002) Activation of centralterminal vanilloid receptor-1 receptors and alpha beta-methylene-ATP-sen-sitive P2X receptors reveals a converged synaptic activity onto the deepdorsal horn neurons of the spinal cord. J. Neurosci., 22, 1228–1237.
Pearson, K.G. (1993) Common principles of motor control in vertebrates andinvertebrates. Annu. Rev. Neurosci., 16, 265–297.
Pratt, C.A. & Jordan, L.M. (1987) Ia inhibitory interneurons and Renshaw cellsas contributors to the spinal mechanisms of fictive locomotion. J. Neuro-physiol., 57, 56–71.
Prescott, S.A. & Koninck, Y.D. (2002) Four cell types with distinctive mem-brane properties and morphologies in lamina I of the spinal dorsal horn of theadult rat. J. Physiol. (Lond.), 539, 817–836.
Puskar, Z. & Antal, M. (1997) Localization of last-order premotor interneuronsin the lumbar spinal cord of rats. J. Comp. Neurol., 389, 377–389.
Ruscheweyh, R. & Sandkuhler, J. (2002) Lamina-specific membrane anddischarge properties of rat spinal dorsal horn neurones in vitro. J. Physiol.(Lond.), 541, 231–244.
Schmidt, B.J., Meyers, D.E., Fleshman, J.W., Tokuriki, M. & Burke, R.E.(1988) Phasic modulation of short latency cutaneous excitation in flexordigitorum longus motoneurons during fictive locomotion. Exp. Brain Res.,71, 568–578.
Szekely, G. (1963) Functional specificity of spinal cord segments in the controlof limb movements. J. Embryol. Exp. Morphol., 11, 431–444.
Szucs, P., Odeh, F., Szokol, K. & Antal, M. (2002) Electrophysiological andMorphological characteristics of neurons in the intermediate gray matter(laminae V–VII) of the neonatal rat lumbar spinal cord in vitro. [abstract].FENS, 3, 185.
Thomson, A.M., West, D.C. & Headley, P.M. (1989) Membrane characteristicsand synaptic responsiveness of superficial dorsal horn neurons in a slicepreparation of adult rat spinal cord. Eur. J. Neurosci., 1, 479–488.
Tresch, M.C. & Kiehn, O. (1999) Coding of locomotor phase in populations ofneurons in rostral and caudal segments of the neonatal rat lumbar spinal cord.J. Neurophysiol., 82, 3563–3574.
Wheatley, M., Jovanovic, K., Stein, R.B. & Lawson, V. (1994) The activityof interneurons during locomotion in the in vitro necturus spinal cord.J. Neurophysiol., 71, 2025–2032.
� 2003 Federation of European Neuroscience Societies, European Journal of Neuroscience, 17, 537–544
544 P. Szucs et al.