www.elsevier.com/locate/pain

Pain 135 (2008) 300–310

Spread of excitation across modality borders in spinal dorsalhorn of neuropathic rats

Doris Schoffnegger, Ruth Ruscheweyh 1, Jurgen Sandkuhler *

Department for Neurophysiology, Center for Brain Research, Medical University Vienna, Spitalgasse 4, 1090 Vienna, Austria

Received 5 July 2007; received in revised form 13 December 2007; accepted 20 December 2007

Abstract

Under physiological conditions, nociceptive information is mainly processed in superficial laminae of the spinal dorsal horn,whereas non-nociceptive information is processed in deeper laminae. Neuropathic pain patients often suffer from touch-evoked pain(allodynia), suggesting that modality borders are disrupted in their nervous system. We studied whether excitation evoked in deepdorsal horn neurons either via stimulation of primary afferent Ab-fibres, by direct electrical stimulation or via glutamate microin-jection leads to activation of neurons in the superficial dorsal horn. We used Ca2+-imaging in transversal spinal cord slices of neu-ropathic and control animals to monitor spread of excitation from the deep to the superficial spinal dorsal horn. In neuropathic butnot control animals, a spread of excitation occurred from the deep to the superficial dorsal horn. The spread of excitation was syn-aptically mediated as it was blocked by the AMPA receptor antagonist CNQX. In contrast, block of NMDA receptors was inef-fective. In control animals, the violation of modality borders could be reproduced by bath application of GABAA and glycinereceptor antagonists. Furthermore, we could show that neuropathic animals were more prone to synchronous network activity thancontrol animals. Thus, following peripheral nerve injury, excitation generated in dorsal horn areas which process non-nociceptiveinformation can invade superficial dorsal horn areas which normally receive nociceptive input. This may be a spinal mechanism oftouch-evoked pain.� 2007 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.

Keywords: Peripheral nerve injury; Allodynia; Spinal cord; Modality borders; Ca2+-imaging; Ab-fibre; Pain

1. Introduction

Neuropathic pain may be due to an injury or a dys-function of the nervous system. One of its main symp-toms is allodynia, i.e. pain elicited by innocuous stimuli.Some types of allodynia are mediated by activation oflow-threshold, myelinated Ab-fibres, which conveytouch-related information from the periphery to thespinal cord [16,35]. Thus, information conveyed in Ab-fibres must gain access to and excite nociceptive neurons.

0304-3959/$34.00 � 2007 International Association for the Study of Pain. P

doi:10.1016/j.pain.2007.12.016

* Corresponding author. Tel.: +43 1 4277 62835; fax: +43 1 427762865.

E-mail address: juergen.sandkuehler@meduniwien.ac.at (J. Sand-kuhler).

1 Present address: Department for Neurology, University of Muen-ster, Muenster, Germany

Under physiological conditions, sensory modalitiesare associated to dorsal horn lamination. Fine primaryafferent Ad- and C-fibres, including most nociceptors,terminate in the superficial dorsal horn (laminae I/II).In contrast, thick Ab-fibres, that transmit non-nocicep-tive information, terminate in deeper laminae (III/IV)[48]. Consistent with this pattern of afferent termination,dorsal horn second order neurons in the superficial lam-inae mainly receive nociceptive input while neurons indeeper dorsal horn laminae mostly convey non-nocicep-tive (lamina III) or converged input (laminae IV–VI)[48]. Under physiological conditions, information fromdifferent modalities are kept separate to prevent cross-talk of sensory channels leading to phenomena such asallodynia. Mechanisms preventing sensory cross-talkare presently not known.

ublished by Elsevier B.V. All rights reserved.

D. Schoffnegger et al. / Pain 135 (2008) 300–310 301

In neuropathic animals, c-Fos expression, an indica-tion for neuronal activity, is increased in superficial dor-sal horn neurons following repeated touch stimuli. Thissuggests that mechanosensitive fibres play a major rolein allodynia by activating nociceptive specific laminaeI/II neurons [3]. Furthermore, in animal models of neu-ropathic pain (sciatic nerve transection, chronic constric-tion injury), Ab-fibre-mediated input to the nociceptivesuperficial dorsal horn increases substantially[27,28,34]. Controversy exists regarding mechanisms ofenhanced Ab-fibre input to nociceptive specific neuronsin laminae I/II of spinal cord in neuropathy.

According to one theory peripheral nerve injury trig-gers Ab-fibres to sprout from their normal terminationsite in the deep dorsal horn into lamina II and establishsynapse-like structures with lamina II neurons[26,50,51]. Another possibility how Ab-fibre-mediatedinformation reaches the superficial dorsal horn, is theopening of pre-existing polysynaptic pathways in thespinal dorsal horn. Sprouting and the opening of pre-existing pathways are not mutually exclusive.

Reduced inhibition in spinal dorsal horn is believedto contribute to neuropathic pain [2,32,45], but it isnot known whether disinhibition is sufficient for anAb-fibre-mediated input to the superficial dorsal horn.

It has been suggested that epileptic activity of noci-ceptive spinal dorsal horn neurons may constitute a pos-sible mechanism underlying lancinating neuropathicpain [38], but it is not known whether spinal dorsal hornneurons of neuropathic animals are prone to epilepticactivity spreading over modality borders.

Here, we used Ca2+-imaging in spinal cord slices of 6weeks old rats after a spared nerve injury (SNI) [9] toinvestigate whether Ab-fibre-mediated input may gainaccess from lamina III to the superficial dorsal horn inthe SNI model. We tested if, in the SNI model, excita-tion spreads from the deep to the superficial dorsal hornand which types of membrane receptors in spinal cordare involved.

2. Methods

2.1. Animals

All procedures used were in accordance with EuropeanCommunities Council directives (86/609/EEC) and wereapproved by the Austrian Federal Ministry for Education, Sci-ence,and Culture. Male Sprague–Dawley rats were used forthis study and were housed under a 12 h light/dark cycle withfree access to water and food.

2.2. Spared nerve injury (SNI) model

Rats (30–32 days old) were deeply anaesthetized with isoflu-rane (1.5–2%) and the sciatic nerve and its three terminalbranches were exposed after incision of the skin and the bicepsfemoris muscle at mid-thigh level [9]. The nerves were freed of

adhering tissue, the tibial, and common peroneal nerves weretightly ligated with 6/0 silk, and 2–3 mm of the nerves distalto the ligation were removed. Care was taken not to touchor stretch the sural nerve. The incision was then closed intwo layers. Sham animals whose sciatic nerve was exposedwithout ligating and sectioning its branches and naıve animalswere used as controls.

2.3. Behavioural tests

Behavioural tests were performed on SNI-operated andsham-operated animals between 08:00 am and 06:00 pm.Assessments were made two times prior to and on days 1, 3,5, 7, and 9 after the surgical procedure. Animals were placedon an elevated wire grid and allowed to habituate to the testingenvironment for approximately 15 min. Mechanical thresholdswere assessed with a series of calibrated von Frey monofila-ments with incremental stiffness (Stoelting, Wood Dale,USA). The von Frey test [6] was performed within the innerva-tion area of the sural nerve on the lateral plantar surface of thehindpaws [9]. The 50% probability withdrawal threshold wasdetermined using the up and down method of Dixon [10]. Ifthe animal responded to perpendicular touching of the footpadwith a von Frey hair for 3 s by brisk withdrawal of the respec-tive hindpaw, the response was considered positive. In theabsence of a paw withdrawal response, the next thicker hair(corresponding to a stronger stimulus) was tested. In case ofa positive response, the next weaker stimulus was applied.The procedure was repeated until six responses (either positiveor negative) were recorded. The 50% withdrawal threshold,which indicated the force of von Frey hair at which the animalreacted in 50% of the presentations, was calculated using themethod of Chaplan et al. [6].

2.4. Preparation of spinal cord slices

Nine to eleven days after the operation, SNI-operated andsham-operated rats were killed by decapitation under deepether anaesthesia, a laminectomy was performed, and the lum-bar spinal cord was removed into ice-cold incubation solutionconsisting of (mM): NaCl 95, KCl 1.8, KH2PO4 1.2, CaCl20.5, MgSO4 7, NaHCO3 26, glucose 15, sucrose 50, oxygenatedwith 95% O2, 5% CO2; pH 7.4, measured osmolarity 310–320 mosmol l�1. The dura mater and all ventral and dorsalroots were removed. About 500–600 lm thick transversal slicesfrom lumbar segments L4 to L6 were cut on a microslicer(DTK-1000, Dosaka, Kyoto, Japan). The slices were storedin incubation solution at room temperature (20–24 �C). Forexperiments in which spinal dorsal horn neurons were stimu-lated via the dorsal root, dorsal L4–L6 roots ipsilateral to theoperation site were retained, and 600 lm thick slices with anattached dorsal root were cut as described below. Slices fromnaıve animals (39–42 days old) were prepared in the same way.

2.5. Fluorometric Ca2+-measurements

For Ca2+-measurements, slices were incubated for 1½ h atroom temperature in incubation solution containing fura-2-acetoxymethyl ester (fura-2 AM, 10 lM) and pluronic acid(0.02%), followed by a P30 min washout period. A single,

302 D. Schoffnegger et al. / Pain 135 (2008) 300–310

fura-2-loaded slice was then transferred to the recording cham-ber where it was superfused with oxygenated recording solu-tion at 3 ml min�1. Recording solution was identical toincubation solution except for (mM): NaCl 127, CaCl2 2.4,MgSO4 1.3, and sucrose 0. Experiments were performed at25 �C. Slices were checked for vitality under an upright micro-scope (Olympus, BX50WI, Tokyo, Japan) equipped withDodt-infrared optics [11]. Determination of the number ofcells that take up fura-2 AM in comparison to the total num-ber of neurons seen in the imaging window revealed that75 ± 3% (n = 10 slices) of all neurons in the imaging windowtake up the calcium dye. In all experiments, Ca2+-imagingwas performed in the superficial dorsal horn (laminae I andII, boxes in Fig. 3A). Slices were illuminated with a monochro-mator and images (220 � 160 lm2, around one fourth of themediolateral expansion of the superficial dorsal horn) weretaken with a cooled CCD camera (both TILL Photonics, Gra-felfing, Germany). Paired exposures to 340 and 380 nm wereobtained at a frame rate of 2 Hz, and the ratio (F340/F380)was calculated to assess relative changes in Ca2+-concentrationwithout conversion to absolute concentration values (TILLVision software package). Changes in Ca2+-concentrationwere evaluated in single cells by placing a small region of inter-est over the cell body. All neuronal cell bodies in the imagingwindow that were loaded with fura-2 (as determined in the380 nm epifluorescence image) were evaluated. A neuron wasconsidered to be excited by electrical or chemical stimulation(see below) when its Ca2+-trace exhibited a peak within 1 safter the stimulation with an amplitude exceeding ten timesthe standard deviation of the baseline noise, measured in a4–5 s segment in which no spontaneous activity of the neuronswas present. A slice was considered to show spontaneous syn-chronous activity when, within a period of 5 min, there was atleast one episode where P5 neurons in the imaging windowexhibited synchronous (within 500 ms) Ca2+-transients in theabsence of any stimulation.

2.6. Electrical and chemical stimulation in the slice

To electrically stimulate neurons in the deep dorsal horn ofthe spinal cord, a unipolar glass stimulation electrode wasplaced 350–400 lm ventral from the dorsal white matter bor-der into the termination area of the injured tibial or the intactsural nerve (pipette tips in Fig. 3A) [31,49]. The somatotopicareas of the single branches of the N. suralis have been pub-lished [31,49]. We used these papers as a guideline for the loca-tion of the somatotopic area of the tibial and the sural nerve.Therefore, we measured the mediolateral expansion of thespinal dorsal horn and divided it into four equal parts(Fig. 3A). Stimulation in the somatotopic area of the tibialnerve was applied in the most medial quarter of the deep dor-sal horn whereas stimulation of the somatotopic area of thesural nerve took place in the third quarter of the deep dorsalhorn (see also Fig. 3A). The pipettes were pulled from borosil-icate glass (Hilgenberg, Maisfeld, Germany) with a flamingmicropipette puller (Model P-97, Sutter Instrument, Novato,CA, USA), filled with recording solution and exhibited tipresistances of 2–3 MX. Stimulation was applied via a constantcurrent stimulator (WPI, Sarasota, Florida, USA) and deliv-ered in trains of 20 pulses at 20 Hz, at 0.35 mA intensity and

0.1 ms pulse width. For chemical stimulation, glutamate(1 mM in recording solution) was applied via a picospritzer(Picospritzer II, Parker Instrumentation, Chicago, USA) tothe same site by pressure ejection from a glass pipette of thesame type (2–3 lm tip diameter, 8 pulses, 20 ms, 20 Hz,8 psi). For electrical and chemical stimulation, in a series ofpilot experiments, stimulation intensity was carefully adjustedin control animals to limit spread of excitation to the directvicinity of the stimulation pipette. Around 5–8 cells within aradius of 60 lm around the stimulation pipette showed a cal-cium transient in response to the stimulation in the deep dorsalhorn.

2.7. Primary afferent stimulation and compound action potential

recording

When preparing the spinal cord for cutting of slices withappending dorsal root, the proximal half of the dorsal rootwas divided into a cranial and a caudal branch. The caudalbranch was cut at the entrance to the spinal cord. A 600 lmthick slice was then made in continuity with the cranial dorsalroot branch (Fig. 2A). The distal dorsal root end was insertedinto a suction electrode for stimulation by a constant currentstimulator and the proximal end of the cut dorsal root branchwas carefully led into a second suction electrode for monopo-lar recording of dorsal root compound action potentials(CAPs) by means of an extracellular amplifier (Iso-Dam-D,WPI, Sarasota, Florida, USA). CAPs evoked by single pulsestimulation (0.1 ms, 0.01–3 mA) were amplified 1000�, bandpass filtered at 0.1 Hz–3 kHz, digitized at 10 kHz and storedin a computer using the pCLAMP 9 acquisition software(Molecular Devices, Union City, CA, USA).

2.8. Statistical analysis

All values are given as means ± SEM. Two-way ANOVAfollowed by Tukey test, t-test, paired t-test, Mann–WhitneyRank Sum test, Wilcoxon Signed Rank test (when appropriatewith Bonferroni adjustment), and Fisher’s Exact test were usedfor statistical comparisons. Since no significant difference wasfound between data obtained from sham-operated and naıveanimals, the results from the two groups were pooled and des-ignated as control.

2.9. Drugs

Drugs and their sources were as follows: (�)-bicucullinemethiodide (bicuculline, 10 lM), L-glutamic acid (1 mM),andstrychnine (4 lM) were from Sigma (Deisenhofen, Germany),D-2-amino-5-phosphonovaleric acid (D-AP5, 50 lM), 6-cyano-7-nitro-quinoxaline-2,3-dione (CNQX, 10 lM), andtetrodotoxin (TTX, 1 lM) were from Alexis (Grunstadt, Ger-many), and fura-2 AM (10 lM) and pluronic (Pluronic F-127,0.02%) were from Molecular Probes (Leiden, The Nether-lands). Stock solutions were prepared by diluting bicucullineand D-AP5 in distilled water, strychnine, CNQX, fura-2 AM,and pluronic in dimethyl sulphoxide (Sigma), glutamate in1 M HCl and TTX in acidic buffer (pH 4.8) and stored in ali-quots at �20 �C. Drugs were added to the superfusion solutionat defined concentrations as indicated.

D. Schoffnegger et al. / Pain 135 (2008) 300–310 303

3. Results

3.1. Behaviour

Animals were tested on two consecutive days beforeand at 2-day intervals following SNI or sham surgery.Animals with a SNI of one sciatic nerve developed ipsi-lateral hypersensitivity to innocuous mechanical vonFrey hair stimulation 24 h after the surgery, and thismechanical allodynia persisted undiminished through-out the testing period (9–11 days) (Fig. 1). The forcerequired to elicit paw withdrawal on the side of the oper-ation dropped from a mean baseline value of 14 ± 1 g(n = 29 animals) tested on two consecutive days beforethe operation, to 2.4 ± 0.3 g on day 9 after the surgery(P < 0.01, SNI ipsilateral before surgery compared toSNI ipsilateral 9 days post surgery; as well as 9 days postsurgery: SNI ipsilateral compared to SNI contralateraland to sham-treated animals; 2-way ANOVA followedby Tukey test). Hindpaw withdrawal thresholds contra-lateral to the surgery were assessed over the whole per-iod and did not change significantly from presurgicalbaseline or when compared to sham-operated animals(P > 0.05, at every post operative time point comparedto presurgical time points and to sham-operated hind-paws; 2-way ANOVA followed by Tukey test). With-drawal thresholds in sham-operated animals remainedunchanged from presurgical values (ipsilateral preopera-tive: 15 ± 1 g; ipsilateral on day 9 after surgery:

Days-4 -2 0 2 4 6 8 10

50 %

thre

shol

d (g

)

0

2

4

6

8

10

12

14

16

18

20

**** ** ** **

Neuropathic (n = 29 animals) Ipsilateral

Sham (n = 17 animals)

Contralateral Ipsilateral Contralateral

Fig. 1. Time course of mechanical allodynia after spared nerve injury(SNI). Withdrawal thresholds of the ipsilateral hindpaws to mechan-ical stimulation with von Frey hair in SNI rats are compared to thoseof sham-operated animals. The SNI and sham operation took place onday 0. A significant reduction in the 50% withdrawal thresholdindicates mechanical allodynia (**P < 0.01 in the Tukey test following2-way ANOVA).

18 ± 1 g; n = 17 animals, P > 0.05, 2-way ANOVA, con-tralateral preoperative: 15 ± 1 g; contralateral on day 9post surgery: 18 ± 1 g; n = 17 animals, P > 0.05, 2-wayANOVA followed by Tukey test).

3.2. Dorsal root stimulation at Ab-fibre intensity evokesCa2+-transients in superficial dorsal horn neurons

We first tested if electrical stimulation of neurons inthe predominantly non-nociceptive Ab-fibres in the dor-sal root would lead to excitation of the predominantlynociceptive superficial dorsal horn in SNI animals. Toensure that the stimulation excited Ab-fibres, but notthe higher threshold, nociceptive Ad- and C-fibres, wesimultaneously recorded CAPs from a dorsal root sidebranch (Fig. 2A). Ab-, Ad-, and C-fibres can be distin-guished by their conduction velocities and their electri-cal stimulation thresholds as illustrated in Fig. 2B andTable 1. Similar to previous reports, no significant differ-ences in stimulation thresholds and conduction veloci-ties were observed between dorsal roots from controland neuropathic animals (Table 1, P > 0.05; t-test,Mann–Whitney Rank Sum test) [28].

Stimulation at an intensity activating only Ab-fibres(17–21 lA, 0.1 ms, as verified by simultaneous recordingof CAPs) resulted in Ca2+-transients in spinal superficialdorsal horn neurons in 12% of the slices from controlanimals (n = 17), but in 67% of the slices from neuro-pathic animals (n = 12) (P < 0.01, Fisher Exact Test).Consistently, significantly more superficial dorsal hornneurons in slices of neuropathic than of control animalsreacted to Ab-fibre stimulation (P < 0.05, t-test,Fig. 2C). In contrast, the numbers of cells in spinalsuperficial dorsal horn reacting in response to C- andAd-fibre stimulation (C-fibre: 0.55–0.7 mA, Ad-fibre:50–70 lA) were not significantly different between con-trol and neuropathic animals (P > 0.05, t-test,Fig. 2C). The absolute number of neurons reacting tostimulation at C-fibre intensity varied between individ-ual slices. We thus normalized the data to the numberof cells reacting in response to C-fibre stimulation(Fig. 2D). The normalized percentage of superficial dor-sal horn neurons excited by Ab-fibre intensity was alsosignificantly different between control and neuropathicanimals (control: 3 ± 3%, neuropathic: 21 ± 5%,P < 0.05, Mann–Whitney Rank Sum test) while no dif-ference was found for stimulation at Ad-fibre intensity(control: 46 ± 7%, neuropathic: 37 ± 10%, P > 0.05,Mann–Whitney Rank Sum test).

3.3. Electrical stimulation and glutamate microinjection in

deep dorsal horn leads to excitation of superficial dorsal

horn neurons in neuropathic animals

To test the hypothesis that, in neuropathic pain, exci-tation spreads from the predominantly non-nociceptive

Fig. 2. Excitation of superficial dorsal horn neurons in vitro by primary afferent stimulation in control and in neuropathic animals. (A) Schematicdiagram of the setup. The distal end of the dorsal root was led into a suction electrode for stimulation with a constant current stimulator. Compoundaction potentials (CAPs) were recorded from a proximal side branch of the dorsal root by means of an extracellular amplifier. Ca2+-imaging wasperformed in the medial part of laminae I/II (box). (B) Representative Ab- (Ba), Ad- (Bb),and C-fibre (Bc) evoked compound action potentialsrecorded extracellularly from the dorsal root of a SNI rat. (C) Number of spinal laminae I/II reacting neurons per slice in response to primaryafferent stimulation at C-, Ad-, and Ab-fibre intensity. Every data point represents 1 slice. The number of cells reacting in response to stimulation atAb-fibre intensity was significantly higher in SNI than in control animals (*P < 0.05 in the t-test). (D) The percentage of laminae I/II neurons excitedby stimulation at Ab-fibre intensity normalized to the number of neurons excited by stimulation at C-fibre intensity also revealed a significantdifference between neuropathic and control animals (*P < 0.05 in the Mann–Whitney Rank Sum test).

304 D. Schoffnegger et al. / Pain 135 (2008) 300–310

deep dorsal horn to the predominantly nociceptivesuperficial dorsal horn, a focal electrical stimulus or glu-tamate microinjection was applied within the somato-

Table 1Activation threshold and conduction velocity of Ab-, Ad-, and C-fibres in n

Threshold (lA)

Ab Ad C

Control (n = 17) 13 ± 1 25 ± 1 263Neuropathic (n = 12) 14 ± 1 27 ± 1 240

There was no significant difference in the activation threshold nor the conductanimals and neuropathic animals (P > 0.05, t-test; Mann–Whitney Rank Su

topic area of the injured tibial or the intact sural nervein the deep dorsal horn (indicated by the tip of the stim-ulation electrode, Fig. 3A). The activity of superficial

europathic and control animals

Conduction velocitiy m s�1

Ab Ad C

± 13 17 ± 1 5.1 ± 0.2 0.6 ± 0.04± 18 15 ± 1 4.4 ± 0.5 0.6 ± 0.04

ion velocity of the different primary afferent fibre types between controlm test). Values are means ± SEM.

Control Neuropathic

Eltr

aec

icl

timul

tios

an

Electrical stimulation

0

10

20

30

40

50

Control (n = 81 slices) Neuropathic (n = 41 slices)

**

nII

%re

act i n

gce

lls i

LI/

Control (n =13 slices) Neuropathic (n = 16 slices)

Glutamate microinjection

**

nI

%re

act i n

gce

lls i

LI/I

Gl u

tam

ate

stim

ultio

na

50 µm

A

B

B1 B2

B4

C

DΔ

F/F

340

380B3

0

10

20

30

40

50

200 µm

Tibial Sural

0

10

20

30

40

50

60

%r

ngs

inI/I

Iea

cti

cell

L0

10

20

30

40

50

60

Electrical stimulationControl (n = 12 slices) Neuropathic (n = 11 slices)

****

Glutamate microinjection Control (n = 12 slices) Neuropathic (n = 11 slices)

Tibial nerve area

Sural nerve area

a

a

b

b

0.095

0.120

%r

ting

ls in

LI/II

e ac

cel

Fig. 3. Electrical stimulation and glutamate microinjection in deep dorsal horn excite superficial dorsal horn neurons in neuropathic animals. (A)Outline of the dorsal quadrant of a transverse spinal cord slice. The dashed line indicates the approximate border between laminae II and III. Thesuperficial dorsal horn regions in the somatotopic area of the transected tibial and the intact sural nerve selected for imaging are shown by the boxes.The sites of electrical stimulation or glutamate microinjection are indicated by the tip of the stimulation pipettes. (B) Slices were incubated with fura-2AM. Then, the ratio of the images captured at 340 and 380 nm illuminations was used to detect changes in intracellular Ca2+-concentration.Examples of ratio images of control animals (B1, B3) and neuropathic animals (B2, B4) 500 ms after stimulation are shown in pseudocolour.Pseudocolour images show values of the difference between the F340/F380 ratio images before stimulation and at the time of usual peak reaction tostimulation (0.5 s after stimulation). Red indicates areas that were excited by the stimulation, whereas blue indicates unexcited areas. Whilesuperficial dorsal horn neurons of control animals did not show Ca2+-transients in the frame after electrical stimulation (B1) or glutamatemicroinjection (B3) in deep dorsal horn, in neuropathic animals, numerous superficial dorsal horn neurons were excited following electrical(stimulation B2) and glutamate microinjection (B4). (C and D) Summary of the results showing that electrical stimulation and glutamatemicroinjection in the deep dorsal horn in the area of the tibial (C) or the sural nerve (D) excited superficial dorsal horn neurons significantly moreoften in neuropathic than in control animals (**P < 0.01 in the Mann–Whitney Rank Sum test).

D. Schoffnegger et al. / Pain 135 (2008) 300–310 305

dorsal horn neurons was monitored by Ca2+-imaging(regions indicated by the boxes, Fig. 3A).

Electrical stimulation was applied either in the termi-nation area of the injured or the intact nerve in the deepdorsal horn. This induced a significant Ca2+-transient inonly 8 ± 1% (n = 81 slices) and 4 ± 3% (n = 12 slices),respectively of the superficial dorsal horn neurons incontrol animals. In neuropathic animals, however,39 ± 5% (n = 41 slices) and 40 ± 8% (n = 11 slices),respectively of the superficial dorsal horn neurons withinthe imaging window were excited by stimulation in deepdorsal horn (Fig. 3B and C; P < 0.01, Mann–WhitneyRank Sum Test). Bath application of the AMPA/kai-

nate receptor antagonist CNQX (10 lM) virtually abol-ished the reaction of superficial dorsal horn neurons inresponse to an electrical stimulus in the deep dorsal hornof both, SNI and control animals (P < 0.01, t-test withBonferroni adjustment, Fig. 4A), demonstrating thatexcitation was due to synaptic transmission and notdirect stimulation of superficial dorsal horn neurons.Blockade of NMDA receptors with D-AP5 (50 lM)had no effect on the spread of excitation from the deepto the superficial dorsal horn in neuropathic or controlanimals (P > 0.05, t-test with Bonferroni adjustment,Fig. 4B), which is consistent with a mono- or oligosy-naptic excitation.

Before CNQX

After CNQX

Control (n = 16 slices) Neuropathic (n = 10 slices)

****

A AMPAR blockade

Before D-AP5

After D-AP5

Control (n = 22 slices)Neuropathic (n = 13 slices)

** **

B NMDAR blockade

Before bicu/strych

After bicu/strych

0

20

40

60

80

100

Neuropathic (n = 21 slices) Control (n = 43 slices)

**

** **

**

C GABA R and GlyR blockadeA

%r

ngs

inI/I

Iea

cti

cell

L

0

20

40

60

80

100

%r

ngs

inI/I

Iea

cti

cell

L

0

20

40

60

80

100

%re

a ct in

gce

l lsi n

LI/I

I

BeforeGABA

AfterGABA

0

20

40

60

80

100%

reac

ti ng

ells

inL I

/ IIc

Control (n = 15 slices)Neuropathic (n = 15 slices)

GABA application

****

D

Fig. 4. Participation of GABAA, glycine, AMPA/kainate, and NMDA receptors in the spread of excitation from spinal laminae III/IV to laminae I/II. All experiments were performed with electrical stimulation in laminae III/IV. (A) AMPA/kainate receptor blockade with CNQX (10 lM)abolished the spread of excitation from spinal laminae III/IV to I/II. (B) D-AP5 (50 lM) had no effect on the spread of excitation. (C) Application ofbicuculline (bicu, 10 lM) and strychnine (strych, 4 lM) induced an increase in the number of laminae I/II neurons reacting to stimulation in laminaeIII/IV in both control and neuropathic animals. (D) Application of GABA (1 mM) reduced the number of laminae I/II neurons reacting to electricalstimulation in laminae III/IV. (**P < 0.01 in the Wilcoxon Signed Rank test for before/after comparisons, or in the Mann–Whitney Rank Sum testfor all other comparisons). For every experimental design (A–C), all six possible comparisons were performed and corrected with the Bonferroniadjustment.

306 D. Schoffnegger et al. / Pain 135 (2008) 300–310

In support of this conclusion, glutamate microinjec-tion in the deep dorsal horn resulted in a Ca2+-transientin significantly more superficial dorsal horn neuronsfrom neuropathic than control animals (Fig. 3C andD, somatotopic area of the tibial nerve: control:9 ± 7% neurons, n = 13 slices, neuropathic: 38 ± 7%neurons, n = 16 slices; somatotopic area of the suralnerve: control: 4 ± 2% neurons, n = 12 slices, neuro-pathic: 46 ± 8% neurons, n = 11 slices; P < 0.01,Mann–Whitney Rank Sum Test). To test if excitationof superficial dorsal horn neurons requires transmissionvia action potentials from the deep to the superficialdorsal horn rather than diffusion of glutamate to super-ficial dorsal horn neurons, we bath-applied the actionpotential blocker tetrodotoxin (TTX, 1 lM) to theslices. After TTX, stimulus-related excitation in the

superficial dorsal horn of the spinal cord was virtuallyabolished (control: 2 ± 1% of the neurons, n = 13 slices;neuropathic: 2 ± 1% of the neurons, n = 16 slices). Afterbath application of TTX, neurons around the stimula-tion electrode still showed a significant Ca2+-transientsuggesting that neurons were activated also in the pres-ence of TTX.

3.4. Participation of GABA and glycine receptors in thespread of excitation from the deep dorsal horn to the

superficial dorsal horn in neuropathic animals

Electrical stimulation in the deep dorsal horn of thespinal cord was used to assess the participation of themain excitatory and inhibitory transmitter systems inthe spread of excitation to the superficial dorsal horn.

D. Schoffnegger et al. / Pain 135 (2008) 300–310 307

After pharmacological disinhibition by blockade ofGABAA and glycine receptors (bicuculline, 10 lM;strychnine, 4 lM), the number of superficial dorsal hornneurons in control animals that responded to electricalstimulation was significantly higher (before bicuculline/strychnine: 10 ± 2%, after bicuculline/strychnine:47 ± 6%, n = 43 slices; P < 0.01, paired t-test with Bon-ferroni adjustment, Fig. 4C) and similar to the numberof superficial dorsal horn neurons in SNI animals react-ing to the same stimulus in the absence of the blockers(42 ± 6%; P > 0.05, t-test with Bonferroni adjustment).Blockade of GABAA and glycine receptors in slices ofneuropathic animals induced a further significantincrease in the number of superficial dorsal horn neu-rons that reacted to the electrical stimulation(P < 0.01, t-test with Bonferroni adjustment, Fig. 4C).

After application of GABA (1 mM), in control ani-mals, superficial dorsal horn neurons no longerresponded to stimulation in the deep dorsal horn whichwas not significantly different to the situation before theapplication of GABA (before GABA: 9 ± 5% superficialdorsal horn neurons reacting to the stimulus in the deepdorsal horn, after GABA: 0 ± 0%, n = 15 slices;P > 0.05, paired t-test with Bonferroni adjustment,Fig. 4D). In neuropathic animals, activation of GABA

A1

A2

A3

A4

A5

A2A1 A3 A4 A5 A6

A6

1 min

F/

340

F 380

Δ F

/F34

038

0

A Control

Single neurons:

Aa Ac

Ab

0.095

0.120

Fig. 5. Blockade of GABAA and glycine receptors induces synchronous neactivity in slices from the lumbar spinal cord of control (A) or neuropathicstrychnine (4 lM) are shown. (A) Aa shows a 380 nm image of the region seldid not evoke simultaneous Ca2+-transients. (Ab) Original traces show the tCa2+-concentration of the neurons that are marked by arrows in Aa. Pseud(arrows, A1–A6) are displayed in Ac. Red indicates a high intracellular Ca2+

synchronized between neurons. (B) In the slice from a neuropathic animspontaneous Ca2+-transients that were synchronized between many of thPseudocolour F340/F380 ratio images of the time points marked in Bb (arrowtime of simultaneous Ca2+-transients and intracellular Ca2+-concentration w

receptors diminished spread of excitation significantly(before GABA: 57 ± 8%, after GABA: 5 ± 2%, n = 15slices; P < 0.01, paired t-test with Bonferroni adjustment).

3.5. Disinhibition induces synchronous network activity in

the superficial dorsal horn of neuropathic but not controlanimals

After the application of bicuculline and strychnine,the dorsal horn neuronal network of neuropathic ani-mals showed spontaneous synchronous activity in 67%of the tested slices (n = 9), while such activity wasobserved in only 5% of the slices (n = 19) in control ani-mals (P < 0.01, Fisher-Exact test). An example of syn-chronous network activity is displayed in Fig. 5. Ca2+-traces of some representative neurons from a slice takenfrom a control animal illustrate that, while spontaneousactivity is present, it is not synchronized among neurons(Fig. 5A). In contrast, neurons from a slice of a neuro-pathic animal show spontaneous synchronous activity(Fig. 5B).

In neuropathic animals, we also tested whether syn-chronous network activity evoked by application ofbicuculline and strychnine may cross-modality borders.We placed the center of the imaging window at the bor-

B1

B1

B2

B2

B3

B3

B4

B4B5

B5

B6

B6

B

Single neurons:

Neuropathic

1 min

F/

340

F 380

Ba

Bb

Bc

0.095

0.120

Δ F

/F34

038

0

twork activity in laminae I/II of SNI rats. Examples of the network(B) animals arising after preincubation with bicuculline (10 lM) and

ected for recording. In slices from control animals, spinal disinhibitionime course (10–15 min after bicuculline/strychnine application) of theocolour F340/F380 ratio images taken at the time points marked in Ab-concentration. While some spontaneous activity was seen, it was notal, blockade of GABAA, and glycine receptors induced repetitive

e recorded neurons, as illustrated in the original traces (Bb). (Bc)s, B1–B6) show that intracellular Ca2+-concentration was high at theas low between two transients in spinal dorsal horn neurons.

308 D. Schoffnegger et al. / Pain 135 (2008) 300–310

der of lamina II to lamina III to be able to record both,excitation of superficial and deep dorsal horn neurons.In 80% (4 out of 5 slices), synchronous network activityafter superfusion with GABAA and glycine receptorantagonists, did not only arise in superficial dorsal hornbut also in the deep dorsal horn.

4. Discussion

In this study we showed a violation of modality bor-ders in spinal cord dorsal horn of neuropathic but notcontrol animals. The spread of excitation from laminaIII into superficial dorsal horn laminae is mediated byglutamatergic synapses and may underlie some formsof Ab-fibre-mediated allodynia. In neuropathic animals,the superficial dorsal horn network is more prone tosynchronous network activity as compared to controlanimals possibly contributing to the phenomena of ‘lan-cinating’ pain.

4.1. Following peripheral nerve injury, stimulation of the

dorsal root at Ab-fibre intensity results in excitation ofneurons in the superficial dorsal horn

Activation of Ab-fibres rarely excites any neurons inthe superficial dorsal horn in normal animals [48]. Here,we showed that following SNI, stimulation of Ab-fibresinduced excitation of superficial spinal dorsal horn neu-rons significantly more often than in control animals.This is consistent with results from studies using otherneuropathic pain models that found a higher incidenceof Ab-fibre evoked input to individual neurons recordedin the superficial dorsal horn after nerve injury[1,27,28,34]. However, a previous study using the SNImodel found no increase in the incidence of lamina IIneurons with Ab-fibre evoked EPSCs [28]. The use ofimaging techniques rather than single cell recordingsenabled us to more broadly screen for Ab-fibre inputto the superficial dorsal horn. The hypothesis that Ab-fibre collaterals sprout into superficial laminae andestablish synaptic contact there [27,44,50] has been chal-lenged recently by electrophysiological and by immuno-histochemical data [20,34,43]. Next to sprouting, spinalinterneurons in deeper layers of the dorsal horn couldconvey non-nociceptive information from Ab-fibres toneurons in the superficial dorsal horn. This possibilitywas tested here directly.

4.2. Electrical and chemical stimulation in deep dorsal

horn lead to excitation of neurons in superficial dorsal

horn

Electrical stimulation or microinjection of glutamateinto the deep dorsal horn excited only few superficialdorsal horn neurons in control animals. In contrast,many more neurons in the superficial dorsal horn were

excited in SNI animals. Glutamate does not excite fibresof passage suggesting that interneurons near the injectionsite in lamina III but not sprouted Ab-fibres were respon-sible for the excitation of superficial dorsal horn neuronsin neuropathic animals. These results demonstrate a vio-lation of the modality border between the predominantlynon-nociceptive deep dorsal horn and the mainly noci-ceptive superficial dorsal horn following peripheral nerveinjury. Interestingly, in SNI animals, there is no differ-ence in the spread of excitation from deep to superficialdorsal horn between the areas where the injured branch(tibial) or the intact branch (sural) of the sciatic nerveis ending. This suggests, that not only the neuronal net-work in the area of the injured but also of the sparednerve undergo changes that allow the spread of excita-tion from deep to superficial dorsal horn. This is of con-siderable functional importance as only a violation of themodality borders in the territory of the intact nerve canaccount for touch-evoked pain, whereas a violation ofmodality borders in the area of the cut nerves shouldnot contribute to stimulus-evoked pain. In control ani-mals, a similar spread of excitation is induced by an acutepharmacological disinhibition. This suggests that synap-tic connections between deep and superficial dorsal hornnormally exist, but that they are actively depressed inhealthy animals. This hypothesis is supported by the factthat many lamina III neurons send axonal projections tolamina II in the spinal dorsal horn [42,46].

We, therefore, investigated the role of inhibitory andexcitatory synaptic transmission for spread of excitationfrom deep to superficial dorsal horn in neuropathicanimals.

4.3. GABAergic/glycinergic inhibition is not completely

abolished in neuropathic animals

The spread of excitation from deep dorsal horn tosuperficial dorsal horn in SNI animals was mimickedby disinhibition (blockage of GABAA and glycine recep-tors) in control animals. Since GABAergic inhibition isreduced in the dorsal horn of neuropathic animals[12,21,32], these data collectively suggest that reducedGABAergic and glycinergic inhibition may be responsi-ble for the spread of excitation over modality borders inSNI animals. After complete pharmacological blockadeof GABAA and glycine receptors, there still was agreater spread of excitation over modality borders inneuropathic animals as compared to control animals.Thus, apart from disinhibition, there must be additionalmechanisms facilitating the spread of excitation over thedeep/superficial dorsal horn border in neuropathic ani-mals. This could include enhanced membrane excitabil-ity or synaptic strength, or reduced inhibition by opioidsor monoamines [40].

Activation of the spinal GABAergic system by theapplication of GABA, blocked spread of excitation in

D. Schoffnegger et al. / Pain 135 (2008) 300–310 309

neuropathic animals. This is consistent with data showingthat intrathecal application of GABA receptor agonistsreduces mechanical allodynia in neuropathic animals[30]. The inhibitory effect of GABA on the Ca2+-signalin superficial dorsal horn neurons could be due to anincreased inhibition of the synaptic pathway connectingdeep and superficial dorsal horn neurons. GABA couldalso induce a hyperpolarisation of superficial dorsal hornneurons, resulting in reduced or abolished Ca2+ entrythrough voltage gated calcium channels.

4.4. Spread of excitation over modality borders is

dependent on AMPA/kainate receptors but not on NMDA

receptors

We applied NMDA receptor antagonists to testwhether mono/oligosynaptic or polysynaptic pathwayswere involved in the spread of excitation from the deepto the superficial dorsal horn layers. Blockade ofNMDA receptors had no effect on the spread of excita-tion over modality borders in SNI animals. This is con-sistent with previous behavioural data showing thatblockade of NMDA receptors does not relieve pain-related responses following SNI and thermal injury[14,33], and that in CCI animals blockade of NMDAreceptors diminishes heat hyperalgesia but has no effecton mechanical allodynia [47,53]. However, other studieshave found that NMDA receptor antagonists mayreduce pain behaviour in some animal models of neuro-pathic pain [7,25,52]. Blockade of NMDA receptorsmay also reduce allodynia evoked by intrathecal appli-cation of the GABAA receptor antagonist bicuculline[22]. The fact that in the present study the blockade ofNMDA receptors did not have an effect suggests thatmainly mono- or oligosynaptic pathways are involvedin the spread of excitation from the deep to the superfi-cial dorsal horn, as polysynaptic excitatory pathwaysare more prone to NMDA receptor blockage [8].

We also studied the involvement of AMPA/kainatereceptors in the spread of excitation from deep to super-ficial dorsal horn. Blockade of AMPA/kainate receptorswith CNQX abolished the spread of excitation overmodality borders, indicating that a glutamatergic synap-tic pathway is involved. Consistently, application ofspinal AMPA receptor antagonists abolishes the devel-opment of hyperalgesia and allodynia associated withperipheral nerve injury [4,23,37].

4.5. Synchronous network activity in spinal dorsal horn of

neuropathic animals after disinhibition

Synchronous network activity is present at embryonicand early neonatal developmental stages [17,18,36] pro-moting neuronal maturation, activity-dependent synap-tic wiring and network formation [5,15,19,24]. Whilerhythmic activity is present in individual nociceptive

and non-nociceptive spinal dorsal horn neurons[13,41], large scale synchronous activity was not seenin spinal cord slices of mature animals under controlconditions. Under conditions of reduced inhibitionand increased excitability, synchronous network activityarises, however, in the spinal dorsal horn of moremature animals [39]. Furthermore, after section of thedorsolateral funiculus, which contains many descendinginhibitory pathways, large amplitude spontaneous dor-sal root potentials arise in the spinal dorsal horn [29],possibly indicating rhythmic activity of dorsal horn neu-rons. In the present study, synchronous activity could bemore easily evoked in slices from neuropathic animals.This demonstrates that in neuropathic animals the dor-sal horn network is more prone to synchronous networkactivity than in control animals. This could underliesome form of shooting pain sensation.

Acknowledgements

Supported by grant P19367 from the Austrian Sci-ence Fund (FWF) to J.S.

References

[1] Baba H, Doubell TP, Woolf CJ. Peripheral inflammation facil-itates A b fiber-mediated synaptic input to the substantiagelatinosa of the adult rat spinal cord. J Neurosci 1999;19:859–67.

[2] Baba H, Ji R-R, Kohno T, Moore KA, Ataka T, Wakai A, et al.Removal of GABAergic inhibition facilitates polysynaptic A fiber-mediated excitatory transmission to the superficial spinal dorsalhorn. Mol Cell Neurosci 2003;24:818–30.

[3] Bester H, Beggs S, Woolf CJ. Changes in tactile stimuli-inducedbehavior and c-Fos expression in the superficial dorsal horn and inparabrachial nuclei after sciatic nerve crush. J Comp Neurol2000;428:45–61.

[4] Blackburn-Munro G, Bomholt SF, Erichsen HK. Behaviouraleffects of the novel AMPA/GluR5 selective receptor antagonistNS1209 after systemic administration in animal models ofexperimental pain. Neuropharmacology 2004;47:351–62.

[5] Bregestovski P, Spitzer N. Calcium in the function of the nervoussystem: new implications. Cell Calcium 2005;37:371–4.

[6] Chaplan SR, Bach FW, Pogrel JW, Chung JM, Yaksh TL.Quantitative assessment of tactile allodynia in the rat paw. JNeurosci Methods 1994;53:55–63.

[7] Chaplan SR, Malmberg AB, Yaksh TL. Efficacy of spinal NMDAreceptor antagonism in formalin hyperalgesia and nerve injuryevoked allodynia in the rat. J Pharmacol Exp Ther1997;280:829–38.

[8] Davies J, Watkins JC. Role of excitatory amino acid receptors inmono- and polysynaptic excitation in the cat spinal cord. ExpBrain Res 1983;49:280–90.

[9] Decosterd I, Woolf CJ. Spared nerve injury: an animal model ofpersistent peripheral neuropathic pain. Pain 2000;87:149–58.

[10] Dixon WJ. The up and down method for small samples. J Am StatAssoc 1965;60:967–78.

[11] Dodt H-U, Eder M, Frick A, Zieglgansberger W. Preciselylocalized LTD in the neocortex revealed by infrared-guided laserstimulation. Science 1999;286:110–3.

[12] Eaton MJ, Plunkett JA, Karmally S, Martinez MA, Montanez K.Changes in GAD- and GABA-immunoreactivity in the spinal

310 D. Schoffnegger et al. / Pain 135 (2008) 300–310

dorsal horn after peripheral nerve injury and promotion ofrecovery by lumbar transplant of immortalized serotonergicprecursors. J Chem Neuroanat 1998;16:57–72.

[13] Eblen-Zajjur AA, Sandkuhler J. Synchronicity of nociceptive andnon-nociceptive adjacent neurons in the spinal dorsal horn of therat: stimulus-induced plasticity. Neuroscience 1997;76:39–54.

[14] Erichsen HK, Blackburn-Munro G. Pharmacological character-isation of the spared nerve injury model of neuropathic pain. Pain2002;98:151–61.

[15] Feller MB. Spontaneous correlated activity in developing neuralcircuits. Neuron 1999;22:653–6.

[16] Field MJ, Bramwell S, Hughes J, Singh L. Detection of static anddynamic components of mechanical allodynia in rat models ofneuropathic pain: are they signalled by distinct primary sensoryneurones? Pain 1999;83:303–11.

[17] Garaschuk O, Hanse E, Konnerth A. Developmental profile andsynaptic origin of early network oscillations in the CA1 region ofrat neonatal hippocampus. J Physiol 1998;507:219–36.

[18] Garaschuk O, Linn J, Eilers J, Konnerth A. Large-scale oscilla-tory calcium waves in the immature cortex. Nat Neurosci2000;3:452–9.

[19] Hanse E, Durand GM, Garaschuk O, Konnerth A. Activity-dependent wiring of the developing hippocampal neuronal circuit.Semin Cell Dev Biol 1997;8:35–42.

[20] Hughes DI, Scott DT, Todd AJ, Riddell JS. Lack of evidence forsprouting of Ab afferents into the superficial laminas of the spinalcord dorsal horn after nerve section. J Neurosci 2003;23:9491–9.

[21] Ibuki T, Hama AT, Wang X-T, Pappas GD, Sagen J. Loss ofGABA-immunoreactivity in the spinal dorsal horn of rats withperipheral nerve injury and promotion of recovery by adrenalmedullary grafts. Neuroscience 1997;76:845–58.

[22] Ishikawa T, Marsala M, Sakabe T, Yaksh TL. Characterization ofspinal amino acid release and touch-evoked allodynia producedby spinal glycine or GABAA receptor antagonist. Neuroscience2000;95:781–6.

[23] Jones TL, Sorkin LS. Calcium-permeable a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid/kainate receptors mediate devel-opment, but not maintenance, of secondary allodynia evoked byfirst-degree burn in the rat. J Pharmacol Exp Ther2004;310:223–9.

[24] Katz LC, Shatz CJ. Synaptic activity and the construction ofcortical circuits. Science 1996;274:1133–8.

[25] Kim YI, Na HS, Yoon YW, Han HC, Ko KH, Hong SK. NMDAreceptors are important for both mechanical and thermalallodynia from peripheral nerve injury in rats. Neuroreport1997;8:2149–53.

[26] Koerber HR, Mirnics K, Brown PB, Mendell LM. Centralsprouting and functional plasticity of regenerated primary affer-ents. J Neurosci 1994;14:3655–71.

[27] Kohama I, Ishikawa K, Kocsis JD. Synaptic reorganization in thesubstantia gelatinosa after peripheral nerve neuroma formation:aberrant innervation of lamina II neurons by A b afferents. JNeurosci 2000;20:1538–49.

[28] Kohno T, Moore KA, Baba H, Woolf CJ. Peripheral nerve injuryalters excitatory synaptic transmission in lamina II of the ratdorsal horn. J Physiol 2003;548:131–8.

[29] Lidierth M, Wall PD. Synchronous inherent oscillations ofpotentials within the rat lumbar spinal cord. Neurosci Lett1996;220:25–8.

[30] Malan TP, Mata HP, Porreca F. Spinal GABAA and GABAB

receptor pharmacology in a rat model of neuropathic pain.Anesthesiology 2002;96:1161–7.

[31] Molander C, Xu Q, Grant G. The cytoarchitectonic organizationof the spinal cord in the rat. I. The lower thoracic and lumbosacralcord. J Comp Neurol 1984;230:133–41.

[32] Moore KA, Kohno T, Karchewski LA, Scholz J, Baba H, WoolfCJ. Partial peripheral nerve injury promotes a selective loss of

GABAergic inhibition in the superficial dorsal horn of the spinalcord. J Neurosci 2002;22:6724–31.

[33] Nozaki-Taguchi N, Yaksh TL. Pharmacology of spinal glutama-tergic receptors in post-thermal injury-evoked tactile allodyniaand thermal hyperalgesia. Anesthesiology 2002;96:617–26.

[34] Okamoto M, Baba H, Goldstein PA, Higashi H, Shimoji K,Yoshimura M. Functional reorganization of sensory pathways inthe rat spinal dorsal horn following peripheral nerve injury. JPhysiol 2001;532:241–50.

[35] Ossipov MH, Bian D, Malan Jr TP, Lai J, Porreca F. Lack ofinvolvement of capsaicin-sensitive primary afferents in nerve-liga-tion injury induced tactile allodynia in rats. Pain 1999;79:127–33.

[36] Peinado A. Immature neocortical neurons exist as extensivesyncitial networks linked by dendrodendritic electrical connec-tions. J Neurophysiol 2001;85:620–9.

[37] Pogatzki EM, Niemeier JS, Sorkin LS, Brennan TJ. Spinalglutamate receptor antagonists differentiate primary and second-ary mechanical hyperalgesia caused by incision. Pain2003;105:97–107.

[38] Ruscheweyh R, Sandkuhler J. Epileptiform activity in rat spinaldorsal horn in vitro has common features with neuropathic pain.Pain 2003;105:327–38.

[39] Ruscheweyh R, Sandkuhler J. Long-range oscillatory Ca2+ wavesin rat spinal dorsal horn. Eur J Neurosci 2005;22:1967–76.

[40] Sandkuhler J. Understanding LTP in pain pathways. Mol Pain2007;3:9.

[41] Sandkuhler J, Eblen-Zajjur AA. Identification and characteriza-tion of rhythmic nociceptive and non-nociceptive spinal dorsalhorn neurons in the rat. Neuroscience 1994;61:991–1006.

[42] Scheibel ME, Scheibel AB. Terminal axonal patterns in cat spinalcord. II. The dorsal horn. Brain Res 1968;9:32–58.

[43] Shehab SAS, Spike RC, Todd AJ. Evidence against cholera toxinB subunit as a reliable tracer for sprouting of primary afferentsfollowing peripheral nerve injury. Brain Res 2003;964:218–27.

[44] Shortland P, Woolf CJ. Chronic peripheral nerve section results ina rearrangement of the central axonal arborizations of axotomizedA beta primary afferent neurons in the rat spinal cord. J CompNeurol 1993;330:65–82.

[45] Sivilotti L, Woolf CJ. The contribution of GABAA and glycinereceptors to central sensitization: disinhibition and touch-evokedallodynia in the spinal cord. J Neurophysiol 1994;72:169–79.

[46] Szentagothai J. Neuronal and synaptic arrangement in thesubstantia gelatinosa rolandi. J Comp Neurol 1964;122:219–39.

[47] Tal M, Bennett GJ. Neuropathic pain sensations are differentiallysensitive to dextrorphan. Neuroreport 1994;5:1438–40.

[48] Willis Jr WD, Coggeshall RE. Sensory mechanisms of the spinalcord. Primary afferent neurons and the spinal dorsal horn. NewYork: Kluwer Academic/Plenum Publishers; 2004.

[49] Woolf CJ, Fitzgerald M. Somatotopic organization of cutaneousafferent terminals and dorsal horn neuronal receptive fields in thesuperficial and deep laminae of the rat lumbar spinal cord. JComp Neurol 1986;251:517–31.

[50] Woolf CJ, Shortland P, Coggeshall RE. Peripheral nerve injurytriggers central sprouting of myelinated afferents. Nature1992;355:75–8.

[51] Woolf CJ, Shortland P, Reynolds M, Ridings J, Doubell T,Coggeshall RE. Reorganization of central terminals of myelinatedprimary afferents in the rat dorsal horn following peripheralaxotomy. J Comp Neurol 1995;360:121–34.

[52] Yashpal K, Fisher K, Chabot J-G, Coderre TJ. Differential effectsof NMDA and group I mGluR antagonists on both nociceptionand spinal cord protein kinase C translocation in the formalin testand a model of neuropathic pain in rats. Pain 2001;94:17–29.

[53] Yoshimura M, Yonehara N. Alteration in sensitivity of ionotropicglutamate receptors and tachykinin receptors in spinal cordcontribute to development and maintenance of nerve injury-evoked neuropathic pain. Neurosci Res 2006;56:21–8.