Download Response Characteristics of Spinal Cord Dorsal Horn Neurons in

J Neurophysiol 92: 1391–1399, 2004;
10.1152/jn.00121.2004.
Response Characteristics of Spinal Cord Dorsal Horn Neurons in Chronic
Allodynic Rats After Spinal Cord Injury
Jing-Xia Hao,* Ron C. Kupers,* and Xiao-Jun Xu
Karolinska Institutet, Department of Laboratory Medicine, Division of Clinical Neurophysiology, Karolinska University HospitalHuddinge, S-141 86 Stockholm, Sweden
Submitted 23 February 2004; accepted in final form 24 April 2004
INTRODUCTION
Many patients with spinal cord injury (SCI) or spinal cord
disease suffer from chronic debilitating pain (Loeser 2000)
arising from spinal levels around the site of injury (Siddall et
al. 2002). This pain is often mixed with nociceptive, visceral,
and neuropathic components. The neuropathic component appears to be the most challenging for clinical pain management
(Siddall et al. 2002). SCI pain is difficult to treat and responds
poorly to opioids and ablative neurosurgical interventions
(Boivie 1994; Finnerup et al. 2002). As improved medical
management of acute CNS injury has greatly increased survival rates, chronic pain has emerged as a major clinical
problem in this population of patients (Casey 1991; Yezierski
and Burchiel 2002).
Several models of SCI have been developed in recent years
that allow the direct examination of neuronal mechanisms
* J.-X. Hao and R. C. Kupers contributed equally to this work.
Address for reprint requests and other correspondence: X.-J. Xu, Section of
Clinical Neurophysiology, Karolinska University Hospital-Huddinge, Karolinska
Institutet, S-141 86 Stockholm, Sweden (E-mail: [email protected]).
www.jn.org
mediating SCI-induced pain (Christensen et al. 1996; Scheifer
et al. 2002; Siddall et al. 1995; Vierck et al. 2000; Xu et al.
1992, 1994; Yezierski et al. 1998). We have established a rat
model of SCI pain after spinal ischemia that in several aspects
mimics chronic pain in man (Hao et al. 1996; WiesenfeldHallin et al. 1997; Xu et al. 1992, 1994). The main features of
the model are marked mechanical and cold allodynia and/or
hyperalgesia at skin areas rostral to the dermatome of the
injured spinal segments. Autotomy and/or excessive scratching
behaviors directed toward the caudal part of the body were also
observed and suggest the presence of spontaneous pain, itch,
and/or dysesthesia.
The present study was undertaken to explore the spinal
neuronal mechanisms underlying the sensory abnormalities
observed in this model. Thereto, we recorded and compared the
response characteristics of dorsal horn neurons in normal and
SCI rats displaying chronic allodynia-like behaviors. We examined the receptive fields, rates of spontaneous activity, and
the responses to noxious and innocuous mechanical and cold
stimulation of these neurons. Preliminary findings have been
communicated as an abstract (Kupers et al. 1997).
METHODS
Animals
Adult female Sprague-Dawley rats (B & K Universal, Stockholm,
Sweden) were used. Animals were housed on a 12 h light/dark cycle.
Food and water were available ad libitum. Experiments were carried
out on normal control rats and on rats that had been submitted to SCI.
All experiments were approved by the local research ethics committee
and were in accordance with the ethical guidelines of the International
Association for the Study of Pain (IASP).
Induction of ischemic SCI
A photochemically induced ischemic SCI was produced as described previously (Hao et al. 1996; Xu et al. 1992). In brief, rats
weighing 200 –220 g were anesthetized with chloral hydrate (300
mg/kg ip), and one jugular vein was cannulated. A midline incision
was made on the skin overlying vertebral segments T12–L1. The
animals were positioned beneath an argon laser beam and irradiated
for 10 min with the beam directed toward vertebral segment T12 or
T13 (spinal segments L3–5). Immediately prior to and 5 min after the
start of the irradiation, erythrosin B (Red N°3, Aldrich-Chemie,
Steinheim, Germany) dissolved in 0.9% saline was injected intravenously at a dose of 32.5 mg/kg. A tunable argon ion laser (Innova
model 70, Coherent Laser Product Division, Palo Alto, CA) operating
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
0022-3077/04 $5.00 Copyright © 2004 The American Physiological Society
1391
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
Hao, Jing-Xia, Ron C. Kupers, and Xiao-Jun Xu. Response characteristics of spinal cord dorsal horn neurons in chronic allodynic rats
after spinal cord injury. J Neurophysiol 92: 1391–1399, 2004;
10.1152/jn.00121.2004. The physiological mechanisms of chronic
pain in patients with spinal cord injury (SCI) are poorly understood.
In the present study, we explored response characteristics of dorsal
horn neurons of spinally injured rats exhibiting chronic pain (pain-like
response to innocuous mechanical and cold stimulation). Several
abnormalities were found in the distribution and response characteristics of dorsal horn neurons in chronic allodynic rats. First, 17% of
the recorded neurons (vs. 0% in control animals) had no receptive
field. Most of these units were located at or close to the lesioned spinal
segment, and they discharged spontaneously at high frequencies.
Allodynic rats also showed a significant decrease in the proportion of
low-threshold (LT) neurons and an increase in the proportion of wide
dynamic range (WDR) neurons. The rate of spontaneous activity of
high-threshold (HT) neurons was significantly higher in allodynic
compared with control rats. Moreover, HT neurons in allodynic
animals showed increased neuronal responses to mechanical stimulation. WDR neurons responded with higher discharge rates to innocuous von Frey hair stimulation in allodynic compared with control
rats. The percentage of WDR and HT neurons showing afterdischarges to noxious pinch was also significantly increased in the
allodynic rats. The proportion of WDR and HT neurons responding to
innocuous cold stimulation respectively increased from 53 and 25% in
control rats to 91 and 75% in allodynic animals. These results suggest
that the chronic pain-like behaviors in spinally injured rats may be
generated and maintained by abnormalities in dorsal horn neurons.
1392
J.-X. HAO, R. C. KUPERS, AND X.-J. XU
at 514 nm was used. The average beam output power was 160 mW.
The beam covers the entire width of the vertebra and the length is
⬃1–2 mm. The penetration of the laser for the 10-min duration of
irradiation is confined to dorsal spinal cord as estimated from the
resulting injury. During irradiation, the animal’s temperature was
maintained between 37 and 38°C. The wound was then closed in
layers and the rats were returned to their home cages. The bladder was
emptied manually for ⬃1 wk until normal function was regained.
set at 50% above the noise level) connected to a loudspeaker and fed
into a Macintosh computer through a 1401 interface (CED, Cambridge, UK) for data acquisition. Peristimulus time histograms (bin
width: 1 s) were displayed on-line. Data were analyzed off-line with
Spike 2 software (CED). The pattern of ongoing activity was analyzed
from interspike interval histogram and by a Spike2 script “burst
analysis.”
Stimulation procedures
Selection of animals
Surgical preparation for recording
Rats were anesthetized with urethan (1.2 g/kg ip). A tracheotomy
was performed, and animals were artificially ventilated. The right
carotid artery was catheterized, and rats were paralyzed with pancuronium bromide (Pavulon, Organon, 2.5 mg/kg ip). The animals were
mounted into a stereotactic frame and a laminectomy was performed
to expose spinal segments T12–L5. The dura was opened, and the cord
was covered with paraffin oil. The vertebral column was stabilized by
means of vertebral and hip clamps. Body temperature was maintained
between 37 and 38°C by means of a feedback-controlled heating pad.
Blood pressure and electrocardiographs were monitored throughout
the experiment.
Single units were isolated from background activity and their
receptive fields (RF) to mechanical stimuli were mapped. The rate of
spontaneous activity and the neuron’s responsiveness to thermal and
mechanical stimulation were quantified. First, baseline spontaneous
firing rate was measured until it had reached a steady firing rate. Then
the neuron’s response to brushing the RF with a cotton swab for ⬃5
s was tested. This was followed by squeezing the skin in the center of
the RF with a serrated forceps (for ⬃5 s). When the spontaneous firing
rate had returned to its baseline level, an ascending series of five von
Frey hairs (0.7, 6.8, 24.1, 44.7, and 69.2 mN) was applied in the center
of the neuron’s RF . Stimulation was applied either dynamically at a
frequency of 1 Hz (5 applications over a 5-s period) or statically
whereby tonic pressure was applied for 5 s. Finally, ethyl chloride was
briefly (⬃1 s) sprayed onto the receptive field.
Analysis of neuronal responses and classification of the units
Evoked responses were calculated by subtracting the background
firing rate from the mean discharge frequency during stimulation. An
afterdischarge was considered to be present when the mean firing rate
of the cell stayed ⱖ30% above baseline firing rate after stopping the
stimulation. Afterdischarges were calculated for the first 5-s period
after stopping the stimulus and in subsequent 10-s epochs.
Cells were classified with respect to their responses to mechanical
stimulation (Willis and Coggeshall 1991). The same classification
scheme as described by Chung et al. (1986) was used. Cells responding best to brush stimuli were classified as low-threshold (LT) cells.
Cells responding maximally to pinch, but with a brush response
⬎10% of the maximal response, were classified as wide dynamic
range (WDR) cells. Neurons with the largest response during pinch
and with either no brush response or a brush response ⬍10% of the
maximal response, were considered as high-threshold (HT) cells.
Cells for which no peripheral receptive field could be detected were
classified as NoRF (no receptive field) cells.
Statistical analysis
The results are expressed as the means ⫾ SE. Statistical analysis
was done by means of a one-way ANOVA, followed by Fisher PLSD
test, unpaired Student’s t-test or the ␹2 test. Values of P ⬍ 0.05 were
considered as statistically significant.
Electrophysiological recording
Extracellular recordings were made using low-impedance (2– 4 m⍀
at 1 kHz; Frederick Haer), glass-coated, tungsten microelectrodes
which were inserted into the left spinal cord by means of a hydraulic
microdrive (David Kopf). In normal rats, recordings in the rostrocaudal direction ranged from spinal segments T12–L4. The exact recording level was verified by identifying the dorsal roots. In SCI rats, the
lesion site could easily be visualized. In these rats, recordings were
made from the edge of the lesion up to three spinal segments rostrally.
At each spinal segment, three recording tracts were explored in the
mediolateral direction. As search stimuli, we used gentle tapping,
stroking of the skin with a cotton tip, and light squeezing of the skin
with the finger tips or a serrated forceps.
Action potentials were amplified, filtered and displayed on an
oscilloscope. They were also led into a window discriminator (level
J Neurophysiol • VOL
RESULTS
Behavioral testing
Before starting the recording sessions, rats were again tested
for the presence of allodynia. The average vocalization threshold to von Frey hair stimulation for the nine control rats was
832.5 ⫾ 37.9 mN. The 11 allodynic rats used had an average
vocalization threshold of 7.3 ⫾ 0.8 mN (range: 0.7–9.2 mN).
Whereas none of the control rats exhibited abnormal reactions
to cold stimulation, the SCI rats showed marked pain-like
reactions to ethyl chloride spray. The allodynic body areas
covered the caudal back, trunk, or lower abdomen. Regions of
mechanical and cold allodynia were usually overlapping.
92 • SEPTEMBER 2004 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
Responses to mechanical (von Frey hairs) and cold (ethyl chloride
spray) stimuli were tested approximately once a week after irradiation.
Signs indicating the presence of chronic mechanical and cold allodynia had variable onset latencies (2 wk to ⱕ3 mo postirradiation).
Once developed, the allodynia-like behavior remained stable with no
signs of spontaneous recovery. The rats used in the present study were
irradiated 3–9 mo previously, and all exhibited robust allodynia-like
responses to mechanical stimuli for ⱖ2 mo prior to the electrophysiological investigations.
The presence of mechanical allodynia was determined by examining the vocalization threshold to mechanical stimulation applied by a
set of calibrated von Frey hairs (ranging from 0.04 to 0.2155 mN).
The rat was gently held by the experimenter, and a von Frey hair was
repeatedly pressed to the skin until bending. Stimulation frequency
was ⬃0.5 Hz and each hair was applied 5–10 times at different areas
on the flanks and back. For normal rats, the vocalization threshold was
usually between 715 and 930 mN. The SCI rats were considered to be
allodynic when they consistently exhibited vocalization thresholds
⬍70.5 mN. The presence of cold allodynia was tested with ethyl
chloride spray. The fur was shaved for cold test. Whereas normal
animals responded with a simple orientation reaction, allodynic rats
vocalized and showed escape behaviors after application of cold spray
to the allodynic area. When sprayed on the hands of normal human
volunteers, ethyl chloride produces a distinct but nonpainful sensation
of cold.
DORSAL HORN NEURONS IN RATS
General electrophysiological findings and recording depths
The data presented include recordings from a total of 100
units among which 53 were recorded in allodynic and 47 in
normal rats. Units with input from joints were excluded.
Because mechanical stimulation was used as a search stimulus,
all the neurons were mechanoresponsive except for those
without a receptive field (see following text). Average recording depths did not differ between allodynic (494 ⫾ 32 ␮m) and
control (438 ⫾ 30 ␮m) rats.
Receptive fields
Classification of neurons
The relative proportion of LT, WDR, and HT neurons
differed significantly between normal and allodynic rats (Table
1). The most striking differences were the higher proportion of
TABLE 1. Classification of neurons recorded in control and
allodynic rats
Recording
levels
Control
T12–13
L1–2
L3–4
Total
Allodynic
⫹3
⫹2
⫹0–1
Total
WDR and the smaller proportion of LT units in the allodynic
rats (Table 1). Unlike in normal rats, the distribution of
different types of neurons covaried with the rostrocaudal recording level in the allodynic rats (Table 1). This was primarily
because the neurons without RF were mostly found close to the
edge of the lesion. However, the HT neurons also tended to be
located in the same region, with seven of nine HT neurons
recorded one segment rostral to the lesion (Table 1).
Spontaneous activity (SA)
The majority of neurons in control rats had either no or a
background SA ⬍1 Hz (Fig. 1). There was no significant
difference in the levels of SA among the three types of neurons
(Fig. 1A). In the allodynic rats, the majority of neurons had a
SA ⬎1 Hz. The neurons without RF usually spontaneously
discharged at high frequencies. All but one had a SA ⬎5 Hz.
HT neurons in allodynic rats also exhibited higher level of SA
with three units showing a SA ⬎15 Hz. In allodynic rats,
average firing rates for HT or neurons without RF were
significantly higher than that of LT or WDR neurons (Fig. 1A).
Furthermore, SA rate of HT neurons was significantly higher in
allodynic compared with control rats. No differences in SA
were found for LT and WDR neurons between the two groups
(Fig. 1A).
When the neurons in the allodynic rats were classified with
respect to recording level, it appeared that there was an increase in SA in neurons situated close to the lesion site. As
shown in Fig. 1B, the number of units with a SA of ⬎5 Hz was
largest at recording sites close to the lesion. This is largely
attributable to the presence of neurons with no RF in this area.
In contrast, we found no rostrocaudal gradient in the level of
SA in control rats (Fig. 1B).
We also analyzed the spontaneous firing pattern of the NoRF
neurons. Only two of these neurons had a regular firing pattern,
whereas the remaining neurons fired in a burst-like manner.
Only in one neuron, the SA was of very short interspike
interval. Further analysis of patterns of background activity
using a Spike2 script burst analysis also confirmed the presence
of burst-like activity in the majority of the neurons without RF.
Responses to mechanical stimulation
LT or WDR neurons in control and
allodynic rats did not differ with respect to their responses to
brush (Fig. 2). Similarly, the brush responses of the WDR
neurons in the two groups also did not differ significantly from
each other (Fig. 2).
BRUSH STIMULATION.
LT
WDR
HT
No RF
Total
5
11
3
19
4
12
4
20
3
3
2
8
0
0
0
0
12
26
9
47
3
2
2
7
8
13
6
27
1
1
7
9
0
2
8
10
12
18
23
53
The numbers of low-threshold (LT), wide dynamic range (WDR), and
high-threshold (HT) neurons or neurons without receptive field (RF) (NoRF)
are shown for either total samples or different recording levels. The ␹2 test
indicated a significant overall difference between control and allodynic rats in
the distribution of LT, WDR, and HT neurons (␹2 ⫽ 6, 5, df ⫽ 2, P ⬍ 0.05).
There was also a significantly different distribution according to recording
level in allodynic (␹2 ⫽ 13.9, df ⫽ 6, P ⬍ 0.05) but not in normal (␹2 ⫽ 4.1,
df ⫽ 4, P ⬎ 0.05) rats.
J Neurophysiol • VOL
LT neurons in both groups responded
to von Frey hair stimulation especially when stimulation was
applied dynamically. Allodynic and control rats did not differ
with respect to their discharge patterns of LT neurons to von
Frey stimulation (data not shown). HT neurons in both groups
responded poorly to von Frey stimulation with only a few
neurons responding to the highest strength tested (69.2 mN).
In normal rats, von Frey stimulation of increasing strength
produced proportional increases in firing rate of WDR neurons
(Fig. 3B, 4). On the other hand, WDR neurons in allodynic rats
often failed to encode stimulus intensity (Fig. 3C, 4). These
neurons already responded maximally to the lowest intensity
without further increasing their firing rate at higher stimulus
VON FREY STIMULATION.
92 • SEPTEMBER 2004 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
The 47 neurons recorded in normal rats all had an identifiable cutaneous RF on the left hindpaw, hindlimb, or flank area.
The size of the RFs varied from very small (e.g., a single toe)
to large (e.g., several square centimeters on the flank). The
neurons recorded from lumbar segments had RFs mostly on the
paw, limb, or hip, whereas neurons recorded from low thoracic
segments usually had their RFs on the flank.
In contrast to the neurons recorded in control rats, 10 of 53
units in allodynic rats had no peripheral RFs. The mean
recording depth of these units was 482 ⫾ 79 ␮m (range:
50 – 853), and the majority of these neurons were clustered
around the edge of the injured spinal segment. They had a high
rate of spontaneous activity (see following text). The remaining 43 neurons had identifiable cutaneous RFs. As the injury
site was located at spinal segments L2– 4 and recordings were
made one to three segments rostral to the lesion, the explored
segments were in the high lumbar and low thoracic region.
Consequently, the receptive fields were most often located on
the hip and flank area with only few units with a RF on the hind
limb. No major changes were observed in terms of size and
continuity of RFs between chronic allodynic and control rats.
1393
1394
J.-X. HAO, R. C. KUPERS, AND X.-J. XU
FIG. 2. Average increases in discharge rates over
baseline activity in dorsal horn neurons during a 5-s
brush (䊐) or pinch (o) in control and allodynic rats. The
type and number of neurons are indicated under the
columns. No difference between control and allodynic
rats was seen for any of the 3 neuronal cell groups
(unpaired t-test).
J Neurophysiol • VOL
92 • SEPTEMBER 2004 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
FIG. 1. A: the average rate of spontaneous activity
(SA) recorded from control (䊐) and allodynic (o) rats,
classified by cell types. The data are expressed as
means ⫾ SE, and the number of neurons studied is
indicated under the columns. ANOVA indicated a significant difference in the rate of SA among different
types of neurons in allodynic [F(3,47) ⫽ 8.3, P ⬍ 0.001]
but not in normal [F(2,43) ⫽ 1.8, P ⬎ 0.05] rats. The
Fisher PLSD test was usd for post hoc analysis. *P ⬍
0.01 and ***P ⬍ 0.001 compared with either lowthreshold (LT) or wide dynamic range (WDR) neurons
in the allodynic rats. †P ⬍ 0.05 compared with highthreshold (HT) neurons in normal rats (unpaired t-test).
B: percentage of neurons exhibiting SA ⱕ1 Hz (䊐),
between 1 and 5 Hz (o) or ⬎5 Hz (■) as classified by the
rostrocaudal recording level. The number of neurons
included in each group is indicated under the columns.
The ␹2 test shows that in control rats the degree of SA
is not influenced by the recording location (␹2 ⫽ 1.4,
df ⫽ 4, P ⬎ 0.05). In contrast, SA varies significantly
according to recording location in allodynic rats (␹2 ⫽
9.9, df ⫽ 4, P ⬍ 0.05).
DORSAL HORN NEURONS IN RATS
1395
intensities (Fig. 3C). WDR neurons in allodynic rats had
significantly lower static von Frey hair thresholds compared
with control rats (Fig. 4). They also responded with higher
spike frequencies to von Frey stimulation than control rats,
although the difference did not reach statistical significance for
any of the stimulus intensities examined (Fig. 4). The difference between allodynic and control rats was most evident for
the lower stimulation intensities and tended to fade off at
higher intensities (Fig. 4).
PINCH STIMULATION. Control and allodynic rats did not differ
with respect to their responses of WDR and HT neurons to
pinch. (Fig. 2).
Table 2 shows the incidence of afterdischarges to brush and pinch stimulation for LT, WDR, and HT
neurons in normal and allodynic rats. Only one LT neuron in
the control groups showed an afterdischarge to pinch. Afterdischarges were more frequently observed in WDR and HT
neurons of control rats. The proportion of WDR and HT
neurons exhibiting afterdischarges to pinching was significantly higher in allodynic compared with control rats (Table 2).
AFTERDISCHARGES.
J Neurophysiol • VOL
In addition, three WDR neurons also showed afterdischarges to
innocuous brushing. Figure 5 shows examples of prolonged
afterdischarges of a WDR neuron to brushing and of a HT
neuron to pinching in allodynic rats.
Responses to cold stimulation
Because our cold stimulus (ethyl chloride spray) involves an
initial innocuous mechanical component, WDR or LT neurons
were classified as cold responsive only if their response to the
cold spray exceeded 5 s. In control rats, a substantial portion of
LT or WDR neurons were cold-responsive, whereas the majority of HT neurons were not (Table 3). The magnitude and
duration of the cold response varied across neurons.
The proportion of LT cells responding to cold spray was not
significantly altered in the allodynic rats (Table 3). In contrast,
the proportion of WDR or HT neurons responding to cold
spray was significantly higher in the allodynic rats (Table 3).
Figure 6 shows the average discharge frequencies of LT,
WDR, and HT cells in normal and allodynic rats after cold
stimulation. There is a shift toward higher discharge rates in
92 • SEPTEMBER 2004 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
FIG. 3. Illustration of typical response
patterns of dorsal horn neurons to mechanical stimulation with von Frey hairs applied
to the center of the receptive field for ⬃5 s.
The neurons illustrated are a LT (A) and a
WDR (B) neuron from a control rat and a
WDR neuron from an allodynic rat (C). The
force of the von Frey hairs and forms of
application, dynamic (d) or static (s), are
indicated. The bin width is 1 s. The LT
neuron (A) was sensitive to dynamic von
Frey hair application but did not encode
stimulus intensity. This neuron was recorded
from spinal segment L1 at a depth of 419 ␮m
and had a receptive field on the flank. The
WDR neuron in B reacted equally well to
both dynamic and static application (threshold: 0.7 g) and showed increased responses
to increasing levels of stimulus intensity.
This neuron was recorded from spinal segment T13 at a depth of 404 ␮m and had a
receptive field on the flank. The WDR neuron in C had a response threshold of 0.07 g
to both dynamic and static stimulation, and
the response peaked at 0.7 g. This neuron
was recorded from spinal segment L1, which
was one segment rostral to the lesion at a
depth of 153 ␮m. The receptive field of this
neuron was also on the flank area.
1396
J.-X. HAO, R. C. KUPERS, AND X.-J. XU
the allodynic animals for the WDR and HT neurons, although
the differences for most time periods are not statistically
significant because of the large variability (Fig. 6). The neuronal response to the cold stimulus was often very prolonged in
the allodynic rats.
FIG. 5. Peristimulus histograms showing afterdischarges of dorsal horn
neurons to mechanical stimulation (bin width: 1 s). A: a WDR neuron showing
prolonged afterdischarges to brush. This neuron was recorded from spinal
segment T13, 3 segments rostral to the lesion at a depth of 715 ␮m. The
receptive field is on the flank area. B: a HT neuron with a largely prolonged
response to pinch. This neuron was recorded from spinal segment L3, 1
segment rostral to the lesion, at a depth of 404 ␮m.
In normal rats, the reported proportion of different neuronal
types depends on factors such as the classification criteria used,
the type of preparations, anesthesia, etc. In the present study,
we used an intact rat preparation with urethan anesthesia, and
we recorded from neurons with response characteristics resembling those of LT, WDR, and HT neurons (Chung et al. 1986;
Willis and Coggeshall 1991). The relative proportion of these
neuronal types in our sample of normal rats is similar to that
reported in most previous studies in the field. In contrast, we
encountered considerably more WDR neurons in the allodynic
rats two to three segments above injury. In addition, we
recorded from neurons without identifiable receptive field.
Because the neurons in normal and allodynic rats were recorded from different animals, it is impossible to know the
original phenotype of these WDR neurons. We speculate that
the increased proportion of WDR neurons reflects either the
appearance of novel innocuous input to HT neurons or increased responses of LT neurons to noxious input. Both possibilities imply that there is an increased excitability in sensory
pathways in the spinal cord underlying the behavioral allodynia.
It is well documented that the different types of input to a
given neuron are not hard-wired and can exhibit considerable
plasticity under conditions such as persistent noxious stimulation, deafferentation, and disinhibition (Willis 1994; Woolf
2. Number of cells with afterdischarges after
mechanical stimulation
TABLE 3. Number of cells with cold responses in control
and allodynic rats
DISCUSSION
TABLE
WDR
Brush
Pinch
HT
Pinch
Control
Allodynic
0/18
5/18
3/25
12/25*
3/8
6/9*
*P ⬍ 0.05 compared to control with ␹2 test.
J Neurophysiol • VOL
LT
WDR
HT
Control
Allodynic
4/16
10/19
2/8
3/7
18/21*
7/8*
The ␹2 test indicates that the proportion of cold-responsive neurons is
significantly higher for WDR and HT neurons in the allodynic rats (*P ⬍
0.05).
92 • SEPTEMBER 2004 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
FIG. 4. Average increase in discharge rate over baseline SA of WDR
neurons during 5-s von Frey hair application in either dynamic (A) or static
mode (B) in control (F) and allodynic (■) rats. The data are expressed as
means ⫾ SE. Stimulation intensity is expressed on a log scale. C: response
thresholds of individual dorsal horn WDR neurons to von Frey stimulation
applied either in dynamic or static mode. Each neuron is represented by a circle
and the median value for each group is indicated by the height of the columns.
*P ⬍ 0.05 between normal and allodynic rats with Mann-Whitney U test.
Stimulation intensity is expressed on a log scale.
DORSAL HORN NEURONS IN RATS
1997). Such plasticity may be achieved by alterations in
neuronal responsiveness, activation of silent synapses, and
changes in anatomical connections between neurons (Wall and
McMahon 1994; ; Willis 1994; Woolf 1997; Woolf et al.
1992). These mechanisms may contribute to the changes in the
response characteristics of dorsal horn neurons in chronic
allodynic rats. For instance, we and others have shown that
dysfunction of the spinal inhibitory GABA and glycine system
may be involved in the generation of mechanical allodynia in
normal and SCI rats (Hao et al. 1992; Sorkin and Puig 1996;
Yaksh 1989). Moreover, SCI has been shown to be associated
with sprouting of afferent sensory fibers (Christensen and
Hulsebosch 1997).
In the allodynic sample, we encountered neurons without
identifiable peripheral RF. This is in line with results from
studies showing spontaneously active neurons without RF in
spinal cord dorsal horn or sensory thalamus after peripheral
nerve, dorsal root, or spinal cord injury (Hoheisel et al. 2003;
Laird and Bennett 1993; Lenz 1991; Lenz et al. 1994; Loeser
and Ward 1967; Sotgiu 1993). Although we do not know the
original phenotype of these neurons, it is unlikely that they
would have joint input as SA in the majority of NoRF neurons
is not regular. It is therefore more reasonable to assume that the
normal cutaneous afferent input (either direct or indirect) to
J Neurophysiol • VOL
these neurons has been interrupted by the spinal cord injury,
resulting in a lack of a RF. Moreover, because the neurons
without RF were recorded in the vicinity of the lesion, the input
to these neurons is likely to come from spinal segments at or
caudal to the lesion. Thus the original RFs of these neurons are
likely to be at skin areas rendered anesthetic by the injury; this
implies that abnormal spontaneous discharges of neurons that
project to higher centers may give rise to dysesthesia or pain
referred to the anesthetic skin area.
Dysesthetic pain felt caudal to the level of a spinal lesion
presents a major component of the pain syndrome in patients
with spinal cord injury, often defined as below-level pain
(Davidoff and Roth 1991; Finnerup et al. 2003; Siddall et al.
2002; Yezierski 1996). In our SCI rats, autotomy behavior
directed toward the hind paws and tail (Xu et al. 1992) and
excessive scratching leading to skin damage (J.-X. Hao, unpublished observation) also suggests that the animals suffer
chronic pain below the level of the lesion. Excessive scratching
has also been reported in other models of SCI (Hoheisel et al.
2003; Scheifer et al. 2002; Yezierski et al. 1998). It is thus
possible that some of the behaviors and underlying sensory
experiences may be caused by the abnormal discharges of the
neurons without receptive field. Lenz and colleagues (1994)
recorded from neurons in the principal sensory nucleus of the
thalamus in patients with spinal cord transection. They found
that the majority of neurons representing the leg (anesthetic
part of the body) had no RF. Interestingly, some neurons
recorded from this or adjacent area did have RFs that were
often located on the border of the anesthetic body part. Electrical microstimulation of these neurons produced sensations
that were referred to the anesthetic body parts (Lenz et al.
1994). These authors suggested that the presence of these
neurons may be the physiological basis for phantom sensation
and phantom pain (Lenz et al. 1994).
Allodynic rats had a larger proportion of neurons with high
spontaneous activity. These neurons were located near the edge
of the lesion. This is similar to the results of Hoheisel et al.
(2003), who also found that neurons close to the spinal contusion exhibited increased level of ongoing activity. This is likely
to be due to a clustering of HT neurons and neurons without
RFs in this area. Interspike interval analysis indicated that the
discharges are irregular and burst-like in the majority of these
SA neurons. In patients, ongoing pain is known to fluctuate,
which may be related to the presence of ongoing activity and
its irregular nature. In the allodynic rats, the rate of average
spontaneous activity was significantly higher in HT compared
with WDR or LT neurons. In addition, SA of HT neurons was
significantly higher in the allodynic compared with the normal
rats. Because HT neurons receive input from nociceptors and
are involved in pain signaling (Chung et al. 1986), the high rate
of SA in HT neurons may give rise to spontaneous pain
sensation at the level around sensory loss.
Systemic lidocaine suppresses SA in SCI rats (J.-X. Hao and
R. Kupers, unpublished observations), indicating the possible
involvement of sodium channels. The ectopic discharges generated in the neuroma and DRG cells after peripheral nerve
injury are also related to activation of sodium channels (Devor
1991; Devor et al. 1993), and it has been reported recently that
the sodium channel NaV1,3 was upregulated in the spinal cord
after spinal contusion, and antisense directed against this channel reduced hyperexcitability of dorsal horn neurons in spinally
92 • SEPTEMBER 2004 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
FIG. 6. Average increase in discharge rates over baseline SA in LT (A),
WDR (B), and HT (C) neurons after cold stimulation in control (䊐) and
allodynic (o) rats. The responses are shown for the time periods 0 –5, 6 –10,
11–20, and 21–30 s after cold application, and are expressed as means ⫾ SE.
**P ⬍ 0.01 compared with normal rats with nonpaired t-test.
1397
1398
J.-X. HAO, R. C. KUPERS, AND X.-J. XU
J Neurophysiol • VOL
allodynia. This suggests that the enhanced cold responses are
mediated by capsaicin-sensitive C-afferent input (Hao et al.
1996).
In conclusion, the present study documents a series of
abnormal electrophysiological properties of dorsal horn neurons in rats with chronic pain-related behaviors after SCI.
Based on the similarities between the neuronal and behavioral
abnormalities, it is suggested that the behavioral symptoms in
SCI rats may be explained by neuronal changes in the dorsal
horn around the level of injury. The neuronal response pattern
in SCI rats varied according to the segmental recording level,
indicating that they are caused by the spinal lesion and subsequent plasticity. Part of the abnormalities are most likely the
result of deafferentation in the zone immediately rostral to the
spinal lesion. The abnormalities observed in the segmental area
most rostral to the lesion may be sustained in part by the
abnormal input from the more caudal segments. Moreover, SCI
induces complex neurochemical changes in areas rostral to the
lesion, and this may also contribute to the neuronal abnormalities.
ACKNOWLEDGMENTS
We thank Prof. Z. Wiesenfeld-Hallin for correcting the manuscript and for
useful discussions.
Present address of R. C. Kupers: Center for Functionally Integrative Neuroscience(CFIN), Aarhus University, Noerrebrogade 44, DK-8000 Aarhus,
Denmark
GRANTS
The present study was supported by the Swedish Medical Research Council
(Project 04X-12168) and research funds of the Karolinska Institute. R. C.
Kupers was supported by a postdoctoral training and mobility grant (ERBCHBICT 941808) from the European Commission.
REFERENCES
Boivie J. Central pain. In: Textbook of Pain, edited by Wall PD and Melzack
R. London: Churchill Livingstone, 1994, p. 871– 892.
Bushnell MC, Duncan GH, Dubner R, and He L-F. Activity of trigeminothalamic neurons in medullary dorsal horn of awake monkeys trained in
a thermal discrimination task. J Neurophysiol 52: 170 –187, 1984.
Casey KL. Pain and Central Nervous System Disease: The Central Pain
Syndromes. New York: Raven, 1991.
Christensen MD, Everhart AW, Pickelman JT, and Hulsebosch CE.
Mechanical and thermal allodynia in chronic central pain following spinal
cord injury. Pain 68: 97–107, 1996.
Christensen MD and Hulsebosch CE. Spinal cord injury and anti-NGF
treatment results in changes in CGRP density and distribution in the dorsal
horn in the rat. Exp Neurol 147: 463– 475, 1997.
Chung J-M, Surmeier DJ, Lee KH, Sorkin LS, Honda CN, Tsong Y, and
Willis WD. Classification of primate spinothalamic and somatosensory
thalamic neurons based on cluster analysis. J Neurophysiol 56: 308 –327,
1986.
Davidoff G and Roth EJ. Clinical characteristics of central(dysesthetic) pain
in spinal cord injury patients, In: Pain and Central Nervous System Disease:
The Central Pain Syndromes, edited by Casey KL. New York: Raven, 1991,
p. 77– 84.
Devor M. Neuropathic pain and injured nerve: peripheral mechanisms. Br Med
Bull 47: 619 – 630, 1991.
Devor M, Govrin-Lippman R, and Angelides K. Na⫹ channel immunolocalization in peripheral mammalian axons and changes following nerve
injury and neuroma formation. J Neurosci 13: 1976 –1992, 1993.
Drew GM, Siddall PJ, and Duggan AW. Responses of spinal neurons to
cutaneous and dorsal root stimuli in rats with mechanical allodynia after
contusive spinal cord injury. Brain Res 893: 59 – 69, 2001.
Dubner R, Kenshalo DR Jr, Maixner W, Bushnell MC, and Oliveras J-L.
The correlation of monkey medullary dorsal horn neuronal activity and the
perceived intensity of noxious heat stimuli. J Neurophysiol 62: 450 – 457,
1989.
92 • SEPTEMBER 2004 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
injured rats (Hains et al. 2003). Clinically, drugs with sodium
channel blocking properties are to some extent useful in treating SCI pain (Finnerup et al. 2002).
Several abnormalities in response to mechanical stimulation
were observed in the allodynic rats, including increased responses to pinch in HT neurons, increased afterdischarges to
mechanical stimulation, and reduction of response threshold to
von Frey hairs for WDR neurons. These data suggest that the
responses of dorsal horn neurons to mechanical input are
enhanced in allodynic rats, which may be reflected in the
behavioral changes. Significantly increased response levels of
dorsal horn neurons to mechanical stimulation have been
reported after neurochemical or contusion injury (Drew et al.
2001; Yezierski and Park 1993), while Hoheisel and colleagues
(2003) found a nonsignificant trend toward lower von-Frey hair
thresholds in LT mechanosensitive neurons after SCI. WDR
neurons have been suggested to play a key role in intensity
coding of noxious stimulation (Bushnell et al. 1984; Dubner et
al. 1989). Thus a reduction in their response threshold and an
increase in their response to low-intensity mechanical stimulation may underlie the mechanical allodynia.
Prolonged afterdischarges of dorsal horn neurons to mechanical stimulation has been shown in rats after partial sciatic
nerve injury (Laird and Bennett 1993; Palecek et al. 1992;
Sotgiu et al. 1995) or spinal cord injury (Drew et al. 2001).
Similar phenomena have been encountered in the present
study. These prolonged discharges, together with an increased
response to pinch stimulation, may account for the exaggerated
behavioral reactions of rats to mechanical stimulation. Although not systematically studied, the phenomenon of mechanical hyperalgesia also exists in SCI rats. Also the painful
behavioral response pattern to light mechanical stimulation in
allodynic rats is more intense than the response pattern to
strong mechanical stimulation in normal rats.
Enhanced responsiveness to cold is one of the hallmarks of
our SCI pain model. In the present study, we used ethyl
chloride spray as cold stimulus. In normal rats, a substantial
portion of LT and WDR neurons react to this type of cold
stimulation, possibly because of the mechanical component of
the stimulus. To minimize the potential confounding effect of
mechanical activation produced by the cold spray, we limited
our data analysis to those cells showing a response duration
exceeding 5 s. The high percentage of cold-responsive neurons
encountered compared with earlier studies (Laird and Bennett
1993; Pertovaara et al. 1997) may be due to the fact that we
used a strong cold stimulus applied over a relatively large skin
area.
A higher percentage of WDR and LT neurons in the allodynic rats reacted to cold stimulation. In addition, HT neurons,
which normally react poorly to cold, showed clear cold responses. It is unlikely that the presence of SCI influenced the
properties of the skin receptors. Our results therefore suggest
that the normal input from cold receptors to HT neurons is
inhibited under normal conditions and that the reduction of this
inhibition by SCI may give rise to increased cold responses.
The increased incidence and magnitude of cold responses of
HT neurons may be primarily responsible for the cold allodynia or hyperalgesia observed in this model. Interestingly, this
hypothesis fits well with results from our previous study
showing that treatment with resinirefatoxin, an ultrapotent
capsaicin analogue, totally abolishes cold but not mechanical
DORSAL HORN NEURONS IN RATS
J Neurophysiol • VOL
nisms, Management, edited by Yezierski RP and Burchiel KJ. Seattle: IASP
Press, 2002, p. 9 –24.
Sorkin LS and Puig S. Neuronal model of tactile allodynia produced by spinal
strychnine: effects of excitatory amino acid receptor antagonists and a
mu-opiate receptor agonist. Pain 68: 283–292, 1996.
Sotgiu M-L. Descending influence on dorsal horn neuronal hyperactivity in a
rat model of neuropathic pain. Neuroreport 4: 21–24, 1993.
Sotgiu, M-L, Biella G, and Riva L. Poststimulus afterdischarges of spinal
WOR and NS units in rats with chronic nerve constriction. Neuro Report 6:
1021–1024, 1995.
Vierck CJ Jr, Siddall P, and Yezierski RP. Pain following spinal cord injury:
animal models and mechanistic studies. Pain 89: 1–5, 2000.
Wall PD and McMahon SB. Long range afferents in rat spinal cord. III.
Failure of impulse transmission in axons and relief of the failure after
rhizotomy of dorsal roots. Philos Trans R Soc Lond B Biol Sci 343:
211–223, 1994.
Wiesenfeld-Hallin Z, Hao J-X, and Xu X-J. Mechanisms of central pain. In:
Proc. 8th World Congress on Pain, Prog. Pain Res. Manag, edited by
Jensen, TS, Turner JA, and Wiesenfeld-Hallin Z. Seattle: IASP Press, 1997,
vol. 8, p. 575–590.
Willis WD. Central plastic responses to pain. In: Proc. 7th World Congress on
Pain, Prog. Pain Res. Manag, edited by Gebhart GF, Hammond DL, and
Jensen TS. Seattle: IASP Press, 1994, vol. 2, p. 575–590.
Willis WD and Coggeshall RE. Sensory Mechanisms in the Spinal Cord. New
York: Plenum, 1991.
Woolf CJ. Molecular signals responsible for the reorganization of the synaptic
circuitry of the dorsal horn after peripheral nerve injury: the mechanisms of
tactile allodynia. In: Molecular Neurobiology of Pain, Proc. Pain Res.
Manag, edited by Borsook D. Seattle: IASP Press, 1997, vol. 9, p. 171–200.
Woolf CJ, Shortland P, and Coggeshall RE. Peripheral nerve injury triggers
central sprouting of myelinated afferents. Nature 355: 75–77, 1992.
Xu X-J, Hao J-X, Aldskogius H, Seiger Å, and Wiesenfeld-Hallin Z.
Chronic pain-related syndrome in rats after ischemic spinal cord lesion: a
possible animal model for pain in patients with spinal cord injury. Pain 48:
279 –290, 1992.
Xu X-J, Hao J-X, Seiger Å, Hughes J, Hökfelt T, and Wiesenfeld-Hallin Z.
Chronic pain-related behaviours in spinally injured rats: evidence for functional alterations of the endogenous cholecystokinin and opioid systems.
Pain 56: 271–277, 1994.
Yaksh TL. Behavioral and autonomic correlates of the tactile evoked allodynia produced by spinal glycine inhibition: effect of modulatory receptor
systems and excitatory amino acid antagonists. Pain 37: 111–123, 1989.
Yezierski RP. Pain following spinal cord injury: the clinical problem and
experimental studies. Pain 68: 185–199, 1996.
Yezierski RP and Burchiel KJ. Spinal Cord Injury Pain: Assessment,
Mechanisms, Managements. Prog. Pain Res. Manage. Seattle: IASP Press,
2002, vol. 23.
Yezierski RP, Liu S, Ruenes GL, Kajander KJ, and Brewer KL. Excitotoxic spinal cord injury: behavioral and morphological characteristics of a
central pain model. Pain 75: 141–155, 1998.
Yezierski RP and Park S-H. The mechanosensitivity of spinal sensory
neurons following intraspinal injections of quisqualic acid in the rat. Neurosci Lett 157: 115–119, 1993.
92 • SEPTEMBER 2004 •
www.jn.org
Downloaded from http://jn.physiology.org/ by 10.220.32.247 on June 18, 2017
Finnerup NB, Johannesen IL, Fuglsang-Frederiksen A, Bach FW, and
Jensen TS. Sensory function in spinal cord injury patients with and without
central pain. Brain 126: 57–70, 2003.
Finnerup NB, Johannesen IL, Sindrup SH, Bach FW, and Jensen TS.
Pharmacological treatment of spinal cord injury pain. In: Spinal Cord Injury
Pain: Assessment, Mechanisms, Management, edited by Yezierski RP and
Burchiel KJ. Seattle: IASP Press, 2002, p. 341–352.
Hains BC, Klein JP, Saab CY, Graner MJ, Black JA, and Waxman SG.
Upregulation of sodium channel Nav1,3 and functional involvement in
neuronal hyperexcitability associated with central neuropathic pain after
spinal cord injury. J Neurosci 23: 8881– 8892, 2003.
Hao J-X, Xu X-J, Yu Y-X, Seiger Å, and Wiesenfeld-Hallin Z. Baclofen
reverses the hypersensitivity of dorsal horn wide dynamic range neurons to
low threshold mechanical stimuli after transient spinal cord ischemia:
implication for a tonic GABAergic inhibitory control upon myelinated fiber
input. J Neurophysiol 68: 392–396, 1992.
Hao J-X, Yu W, Xu X-J, and Wiesenfeld-Hallin Z. Capsaicin-sensitive
afferents mediate chronic cold, but not mechanical, allodynia-like behavior
in spinally injured rats. Brain Res 196: 116 –118, 1996.
Hoheisel U, Scheifer C, Trudrung P, Unger T, and Mense S. Pathophysiological activity in rat dorsal horn neurons in segments rostral to a chronic
spinal cord injury. Brain Res 974: 134 –145, 2003.
Kupers RC, Hao J-X, Xu X-J, Wiesenfeld-Hallin Z, and Persson JKE.
Abnormal responses of dorsal horn neurons in rats with chronic allodynia
produced by ischemic spinal cord injury. Soc Neurosci Abstr 23: 438, 1997.
Laird JM and Bennett GJ. An electrophysiological study of dorsal horn
neurons in the spinal cord of rats with an experimental peripheral neuropathy, J Neurophysiol 69: 2072–2085, 1993.
Lenz FA. The thalamus and central pain syndromes: human and animal
studies, In: Pain and Central Nervous System Disease: The Central Pain
Syndromes, edited by Casey KL. New York: Raven, 1991, p. 171–182.
Lenz FA, Kwan HC, Martin R, Tasker R, Richardson RT, and Dostrovsky
JO. Characteristics of somatotopic organization and spontaneous neuronal
activity in the region of the thalamic principal sensory nucleus in patients
with spinal cord transection. J Neurophysiol 72: 1570 –1587, 1994.
Loeser JD. Generalized pain syndromes: pain in spinal cord isjury patients. In:
Bonica’s Management of Pain, edited by Loeser JD. (3rd ed.). Philadelphia:
Lippincott, 2000, p. 613– 620.
Loeser JD and Ward AA Jr. Some effects of deafferentation on neurons of
the cat spinal cord. Arch Neurol 17: 629 – 636, 1967.
Palecek J, Pelecekova V, Dougherty PM, Carlton SM, and Willis WD.
Responses of spinothalamic tract cells to mechanical and thermal stimulation of skin in rats with an experimental peripheral neuropathy. J Neurophysiol 67: 1562–1573, 1992.
Pertovaara A, Kontinen VK, and Kalso EA. Chronic spinal nerve ligation
induces changes in response characteristics of nociceptive spinal dorsal horn
neurons and in their descending regulation originating in the periaqueductal
gray in the rat. Exp Neurol 127: 428 – 436, 1997.
Scheifer C, Hoheisel U, Trudrung P, Unger T, and Mense S. Rats with
chronic spinal cord transection as a possible model for the at-level pain of
paraplegic patients. Neurosci. Lett 323: 117–120, 2002.
Siddall PJ, Xu CL, and Cousins MJ. Allodynia following traumatic spinal
cord injury in the rat. Neuroreport 6: 1241–1244, 1995.
Siddall PJ, Yezierski RP, and Loeser JD. Taxonomy and epidemiology of
spinal cord injury pain. In: Spinal Cord Injury Pain: Assessment, Mecha-
1399