Mechanical and thermal allodynia in chronic central pain following spinal cord injury

P 68 (1996) 97–107 @ 1996 International Association for the Study of Pain. 0304-3959/96/

a

97

$15,00

PAIN 3224

Mechanicaland thermalallodyniain chroniccentral pain followingspinalcord injury Marc D. Christensen, Alex W. Everhart, Jason T. Pickelman and Claire E. Hulsebosch* Marine Biomedical Institute and Depurtmenl qfAnatomy and Neuroscience, The University of Texus Medical Brunch, Galveston, TX 77555-1069 (USA) (Received 26 January 1996, revised version received 21 June 1996, accepted 18 July 1996)

Summary

Spinalcord injury (SCI) resuIts in variable motor recoveries and chronic central pain syndromes develop in the majority of SCI patients. To provide a basis for further studies, we report a new rodent model of chronic central pain following spinal cord trauma. Male Sprague–Dawley rats (N= 10) were hemisectioned at T13 and were tested both preoperatively and postoperatively and compared to sham-operated controls (N= 10) for locomotor function, and mechanical and thermal thresholds of both paw withdrawal and supraspinal responses. Results support the development and persistence of allodynia which persists for 160 days. Locomotor function was tested using the Basso, Beattie and Bresnahan (BBB) open field test and only the limb ipsilateral to the hemisection was affected, demonstrating acute flaccid paralysis with motor recovery which approached normal values by postoperative day (POD) 15. Prior to the hemisection,the rats showed little to no paw withdrawal response to von Frey stimulation of 4.41 mN or 9.41 mN in both forelimbsand hindlimbs.Postoperatively,responsesin both ipsilateral and contralateral forelimbs and hindlimbs increased over time and the increase was statistically significant compared to intra-animal presurgical and sham control values (P< 0.05). There were no significant side-to-side differences in limb responses preoperatively or beyond POD 15. The forelimbs and hindlimbs responded to von Frey hair strengths of 122 mN preoperatively and postoperatively with similar withdrawal frequencies that were not statistically significant. Preoperatively, the paw withdrawal latency to heat stimuli was 22.9 * 3.0 (mean * SE) and 20.1 * 3.1 sec for the hindlimbs and forelimbs, respectively. Postoperatively, the mean hindlimb and forelimb latency of paw withdrawals decreased to 11.9* 1.8 and 9.2* 2.5 see, respectively. This decrease in thermal thresholds is statistically significant when compared to intra-animal preoperative and sham control values (P< 0.05). These data indicate that somatosensory thresholds for non-noxious mechanical and radiant heat which elicit paw withdrawal (flexor reflex) are significantly lowered following SCI. To further support the development and persistence of chronic pain following hemisection, supraspinal responses such as paw lick, head turns, attacking the stimulus, and vocalizations were elicited in response to mechanical and thermal stimuli and were statistically significant compared to presurgical intra-animal or sham control values (P < 0.05). Hemisected animals vocalized to von Frey hair bending forces of 49.8 with a mean of 6.0* 1.2 times out of 10 stimuli compared to intra-animal presurgical and sham control values of zero. Supraspinal responses of hemisected animals to thermal stimuli occurred at lower temperatures that were statistically significant compared to sham control or preoperative values (P< 0.05). These chronic changes in thresholds to both mechanical and thermal stimuli represent the development and persistence of mechanical and thermal allodynia after SCI. Key words: Central pain; Thermal hyperalgesia; Mechanical allodynia; Spinal cord injury; Hemisection

Introduction Spinal cord injuries (SCI) result in a devastating loss of function below the level of the lesion. Typically, in a complete lesion there is long term bilateral loss of volitional motor control and insensibility to sensory stimulation ensues. In both complete and partial spinal lesions, there are variable motor recoveries and chronic central pain syndromes de* Cfwre.rporrdingauthor: Claire E. Huk.ebosch, Ph.D., 301 University Boulevard, The University of Texas Medical Branch, Galveston, TX 77555-1069, USA, Tel.: (409) 772-2939; Fax: (409) 762-9382; E-mail: hulsebosch@mbiarr,utmb.edu PII S0304-3959(96)03224-

1

velop in the majority of spinal cord injured patients (White 1966; Schliep 1978; Jack and Lloyd 1983; Boivie 1984; Davidoff et al. 1987; Nogues 1987; Beric et al. 1988; Davidoff and Roth 1991). Once spinal shock has subsided, reflexes return, such as the flexor withdrawal response, and various pain syndromes develop, usually within months following injury (Richards et al. 1980). The pain syndromes or dysesthesias (disturbing somatic sensations) can be divided into two broad categories based on the dependency of the pain on peripheral stimuli: (1) persistent or spontaneous pain, which occurs independent of peripheral stimuli, occurs spontaneously and intermittently, persists but waxes and wanes, and is described as numbness (while not painful, a

98

disturbing somatic sensation), burning, cutting, piercing or electric-like (from Davidoff and Roth 1991); (2) stimulus evoked pain, which occurs in response to either a normally non-noxious (allodynia) or noxious (hyperalgesia) stimuli (Merskey and Bogduk 1994). In a partial injury, such as hemisection, either spontaneous or evoked pain may be subserved by the remaining andlor modified spinal cord circuitry: cranially, caudally, and through the intact cord adjacent to the lesion site (Willis 1982). These dysesthesias compound the necrologic deficits and contribute to the already compromised quality of life following SCI. Unfortunately, the abnormal pain sensations resulting from SCI have remained refractory to clinical treatments despite a variety of therapeutic strategies, including neurosurgical, pharmacological and behavioral interventions (Balazy 1992). The failure of therapeutic strategies to treat dysesthesias of SCI is due to the lack of attention given to mechanisms which elicit chronic pain following SCI. Reproducible mammalian models of chronic central pain have not been well developed to provide a basis for the study of pain mechanisms. Our goals were to develop a rodent model of chronic central pain after SCI which would have the following characteristics: (1) ease in performance from laboratory to laboratory (does not require impact machine or lasers which are used in other models of SCI); (2) reproducibility in terms of the quality of nociception tested (i.e. by both mechanical and thermal tests); (3) reproducibility in terms of the number of lesioned animals displaying the altered nociceptive thresholds (Xu et al. 1992 report only 44% of their ischemic SCI model develop symptoms consistent with chronic pain); (4) assayability for significant motor recovery as determined by the open field test first developed by Tarlov and Klinger (1954) and recently modified into the Basso, Beattie and Bresnahan (BBB) Locomotor Rating Scale (Basso et al. 1995); (5) produced by direct cord injury and not ischemia (Xu et al. 1992); (6) careful characterization of a model of central pain for pharmacological tests toward amelioration of the pain syndromes associated with SCI while preserving other somatosensory function. The purpose of this paper is to present a hemisection model for the development of chronic central pain after SCI and (1) to demonstrate the onset and time course of mechanical allodynia, (2) to demonstrate the onset and time course of thermal allodynia, (3) to determine the onset and time course of supraspinal measures of chronic pain, (4) to quantitate the open field test scores used in the BBB Locomotor Rating Scale, and (5) to compare lesion sites to outcome measures. Materials and methods Surgical procedures All procedures involving rats were reviewed by the local Animal Care and Use Committee and were consistent with the guidelines of the hrter-

rmtiorml Association for the Study of Pain and the NIH Guide for the Care and Use of Laboratory Animals. Male Sprague–Dawley rats (N= 20), 200-250 g, were obtained from Harlan Sprague–Dawley, Inc. and housed with a light/dark cycle of 12 h: 12 h where the dark cycle began at 0700 h. Behavioral tests occurred in the morning. Since the rats are nocturnal animals, the tests occurred during their ‘awake’ period in the circadian rhythm. For the hemisection surgery, rats were deeply anesthetized by intraperitoneal injection of sodium pentobarbital (0.5 mglkg) as determined by the absence of the tail and paw withdrawal reflex to pinch and the absence of the comeal blink reflex. The spinal cord was hemisected at T13 on the left side by the following procedure: following palpation of the dorsal surface to locate the cranial borders of the sacrum and the dorsal spinous processes of the lower thoracic and lumbar vertebrae, the T1 1–12 Iaminae were determined by counting spinous processes from the sacrum. The surgical field was shaved and prepared with betadine, a longitudinal incision was made exposing several segments, a Iaminectomy was performed at two vertebral segments, T1 I-T12, the lumbar enlargement was identified with accompanying dorsal vessel, and the spinal cord was hemisected just cranial to the L1 dorsal root entry zone with a no. 15 scalpel blade without damage to the major dorsal vessel or vascular branches. A tuberculin syringe witb a 28 gauge needle was placed dorsi-ventrally at the midline of the cord and the syringe was backfilled to remove blood while the needle was pulled laterally to ensure the completeness of the hemisection. The musculature and the fascia were then sutured and the skin was apposed by autoclips. The rats were eating and drinking within 3 h of surgery. To ensure the general health of the rats, the animals were weighed daily. Weight loss was minimal, occurred acutely over the first postoperative 2 days and was not greater than 5% of the total body weight. By the third postoperative day, the animals demonstrated normal weight gains. Another sign of distress often seen in moderate to severe SC] in rodents is cessation of grooming which is easily detectable by a cutaneous build-up of sediment around the ocular orbit. No rats demonstrated a lack of grooming using this criteria as well as empirical observation of the onset of grooming behavior several hours after surgery. Although not shown, it should be noted that right-sided hemisections were not statistically significant from left-sided hemisections in terms of the data collected and reported below. Of the 20 rats in the study, 10 received hemisections and 10 served as sham controls. At the termination of the behavioral testing, the hemisected and sham control groups were anesthetized as above, perfused transcardially with 4% paraformaldehyde, the spinal cords removed from TIO to L3, blocked and embedded in OCT. The cords were sectioned transversely with a cryostat and used for immunocytochemical studies reported elsewhere. The maximal lesion site was determined and the extent of the lesion into the contralateral cord was recorded by sbading the lesioned area onto an idealized spinal cord.

Behavioral procedures Locomotor function was observed and recorded using the modified open field test first developed by Tarlov and Klinger (1954) and recently modified into the Basso, Beattie and Bresnahan (BBB) Locomotor Rating Scale (Basso et al. 1995) to ensure that a motor recovery occurs and does not impair the somatosensory behavioral tests. The score is based on locomotor ability following experimental SCI in rodent models. Briefly, the BBB scale is a 21 point scale which ranges from O, which is no observable hindlimb movement, to 21 which is consistent and coordinated gait with parallel paw placement of the hindlimb and consistent trunk stability, Scores from O to 7 rank the early phase of recovery with tbe return of isolated joint movements of tbe three joints (hip, knee, ankle), scores from 8 to 13 describe the intermediate recovery phase with the return of paw placement, stepping and forelimb-hindlimb coordination, and scores 14 to 21 rank the late phase of recovery with the return of toe clearance during the step phase, predominant paw position, trunk stability and tail position. Thus, the BBB score for spinally injured rats allows assessment of hindlimb recovery. Left and right bindlimbs were assessed

99 separately, allowing detection of asymmetrical recovery. Spinally injured animals that displayed acute contmlateral hindlimb motor deficits were eliminated from the study. The score was tabulated and considered to be im indicator of motor recovery after spinal cord hemisection. Behavioral tests representing mechanical and thermal allodynia were performed preoperatively and postoperatively for both forelimbs and hindlimbs. The preoperative testing began 7 days prior to surgery and was used to establish both individual and group baseline behaviors. Prior to the onset of behavioral testing, all animals were environmentally acclimated to the plexiglass cubicle testing apparatus for 4 h daily for 3 days. During testing, the data from each limb was collected independently. The tests were performed postoperatively on alternate days for up to 7 weeks. Mechanical allodynia of the glabrous skin of the paw was quantified by measuring the number of brisk paw withdrawals in response to normally innocuous mechanical stimuli (Choi et al. 1994). The subthreshold mechanical stimuli were von Frey filamentswith bending forcesof 4.41 mN and 9.41 mN. In addition, a supra-threshold mechanical stimulus, a von Frey filament with a bending force of 122 mN, was used. To perform this test, rats are placed inside a clear plexiglass box (8 X 8 X 18 cm) on an elevated, tine metal screen and acclimated for 60 min prior to testing. The von Frey filament was applied from underneath the metal mesh floor, through the mesh, to the plantar surface of the glabrous skin of the paw for each limb. A single trial consisted of 10 applications of von Frey filament, applied once every 3.4 sec. A response is defined as a withdrawal of the stimulated paw. Any accompanying attention of the rat to the stimulus by head turning, disengaging behavior from ongoing activity, attempting to bite the von Frey hair, etc. was recorded. The mean occurrence of paw withdrawal in each of the trials was taken for each limb in each rat as a repeated measure and was expressed as number of responses where O indicated no paw withdrawal and 10 indicated the maximum number of paw withdrawals. Thermal allodynia was tested by measuring the latency of paw withdrawal in response to a noxious radiant heat stimuli developed by Hargreaves et al. (1988) and used in neuropathy models (Bennet and Xie 1988). To perform this test, rats are placed inside a clear plexiglass box (8 x 8 x 18 cm) on an elevated, glass plate and acclimated for 60 min prior to testing. A radiant heat stimulus is applied by aiming a beam of light through a 3 mm tube to the plantar surface of the glabrous skin of the paw in each limb. The light beam was turned off automatically by a photocell when the rat lifted its limb, allowing the measurement of time between the start of the radiant stimulus and the paw withdrawal. This time interval is defined as the paw withdrawal latency. Each trial consisted of three tests taken every 45 min. During a test, an individual rat was tested at 5 min intervals between each forelimb and hindlimb. The mean latency of foot withdrawal in each of the three trials was taken for each limb of each rat as the number of seconds to paw withdrawal. The maximum number of seconds before the stimulus was terminated was set at 35 sec to avoid tissue damage due to prolonged exposure to the radiant heat stimulus. The temperature of the radiant heat stimulus was determined by measuring the temperature of the testing surface using a copper thermocouple and was found to increase in time. At the start of the test, the surface temperature was 23.5°C and at 35 sec the surface temperature was 44.5”C. Therefore, since the radiant heat source heated the glass surface over time, a paw withdrawal latency change was associated with a different temperature of stimulus (see Himta et al. 1990 for other issues to consider). Since this method was not precise enough to distinguish the temperature of the delivered stimulus and we were interested in supraspinal responses, we developed the following measures. Suprmpirad responses to thermal allodynia were evaluated using a newly developed thermal heat application device constructed using materials and instruction obtained from Physitemp, Inc. (Clifton, NJ). Four 2 x 4 inch heat pads were placed on a flat surface, juxtaposed to each other. These pads were coupled to a power source and a thermocouple device. The temperature of each pad was individually controlled by a toggle switch. The testing temperature was preset on the digital power source. The rats were then housed in a plexiglass cubicle in which the floor was covered with the heat pads. A graded and increasing heat

stimulus was applied to individual limbs by selecting the appropriate toggle. The rate at which the temperature increased could be empirically adjusted so that the rate of change was rapid and would induce supraspinal responses. Supraspinal responses consisted of a brisk paw withdrawal with a lick or head turn and guarding of the paw, Each trial was repeated three times at 45 min intervals. The temperature at which supraspinal behavior occurred was recorded. Vocalizations are another indicator that noxious mechanical stimuli have been detected supraspinally. The number of vocalization were recorded and the stimulus sites at which vocalizations could be evoked were mapped on a drawing representing the body of a rat. Mechanical stimuli were applied with a von Frey filament with bending force of 49,8 mN to various locations on the trunk. The number of vocalizations out of 10 trials per site were counted. Sprague–Dawley rats normally do not vocalize or otherwise respond at this stimulus strength. Another measure of chronic pain, although very controversial, is the onset of self-mutilation, in which an animal chews off parts of a limb, such as the digits. Autotomy is a term used by some to indicate selfmutilation behavior, frequently associated with the denervation of a limb (Wall et al. 1979). However, this term in the invertebrate and lower vertebrate literature refers to the reflex or spontaneous shedding of an appendage (Dorland’s Medical Dictionary, 26th Edition); as for example the tail of a lizard or the limb of a crustacean. Our model does not result in shedding of the limb of the rodent. Thus, to reflect the nature of the behavioral occurrence we use the term self-mutilation, Self-mutilation was recorded and if found to progress more than 2 days the animal was eliminated from the experimental group.

Statistical analysis All statistical tests were evaluated at the alpha level of significance of 0.05, by two-tailed analyses. The data from these procedures were tested for statistical significance using ANOVA followed by multiple comparisons and where appropriate either the paired Student’s r-test (between sides within each animal or in comparisons of behavioral test results before and after surgery with the same animal) or the Student’s r-test (for between group comparisons). All data management mrd statistical analyses were performed using SAS (1992) statistical procedures. All values are reported either as means * SE or means * SD as indicated.

Results The extent of the hemisection lesion was assessed from histological sections. In general, the lesion was confined unilaterally and included the dorsal column system, Lissauer’s tract, both lateral and ventral column systems, and the gray matter ipsilaterally. In some cases, the lesion extended partially into the dorsal column system of the contralateral side as indicated in Fig. 1. Motor behavior Following the spinal hemisection, rats were allowed to explore in an open field test environment. The rats were scored using an expanded and modified Tarlov scale referred to as the BBB scale (Basso et al. 1995). Sham control rats consistently scored 21 on all days. Only the scores from the hindlimb on the hemisected side are reported since there were no observable differences in locomotor function for any other limb (Fig. 2). As can be seen from this figure, the early phase of recovery began on POD 6 where as by POD 12, all

100

Hm” —

9

w Fig. 1. A schematicdiagramof the extentof the surgical lesion to the T13 spinal cord as determinedin histological sections of the animals used in this study. The minimum lesion extended to the area is indicated by horizontal lines and the maximum lesion is indicated by vertical lines.

animals had entered the last phase of recovery with few to no toe drags. There were no significant changes in behavior from POD 17 in the hemisected group. On the 27th day, the mean * SD of the hemisected group was 18 * 0.8 compared to a value of 21 t O for the sham control groups which was statistically significant (P e 0.05). Mechanical allodynia

empirically that rats that were not acclimated were restless and attempted to escape the plexiglass cubicles which would confound any experiments, including nociceptive tests. Prior to the hemisection, the hemisected rats and sham control rats showed no response to von Frey stimulation of 4.41 mN in the hindlimbs or forelimbs. Presurgical vahtes were not statistically different from sham controls tested for up to 160 days (Figs. 3 and 4). A ‘no response’ to the mechanical stimulation is defined as no paw withdrawal response and in addition the animal did not attend to the stimulus by turning its head, disengaging in ongoing behavioral activity, or any other behavioral index that would indicate that the stimulus was threshold or above. PostSurgically, the paw withdrawals are often accompanied by aversive behaviors such as abrupt head turns, vocalizations and some animals would attempt to bite the von Frey hair. The hindlimb ipsilateral to the hemisection showed a delay of several days until the return of the paw withdrawal response which was due to the inability of lower limb function as can be seen by the results of the BBB scale. Although there was no statistical difference between left and right limbs in the hemisected group beyond POD 10, there continued to be statistically significant differences in hemisected hindlimb and forelimb responses compared to sham controls (Pc O.05). The decreased threshold to mechanical stimulation as measured by paw withdrawal occurred in both ipsilateral and contralateral hindlimbs and forelimbs.

After environmental acclimation, the rats were essentially stationary in the plexiglass test cubicles. We have found

10 9

22 1 20 i18 i 16 i

I

=

8

.: z :6

7

F -o~ –&–?-

-10 -5

Hind Ipsilateral Hind Contralateral Control Hind Left Control Hind Right

0

T-

5 10 15 20 25 30 35 40 45 50

160

Days After Hemisection -10

-5

0

I

I

I

I

I

5

10

15

20

25

30

Days After Hemisection Fig. 2. A graph of the mean k SD for the BBB motor score for the limb on the hemisected side (N= 10). Note that the recovery of limb function reached maximum values by 20 days. No other limb demonstrated changes in the behavioral score from presurgical data and these data are not shown. Since little if any behavioral changes were seen after 27 days postoperatively, these data are not shown.

Fig. 3. Time course of the mean* SE of paw withdrawals in both left and right hindlimbs in controls (Control Hind Left, Control Hind Right) and in the ipsilateral and contralateral hindlimb of hemisected rats (Hind Ipsilatemf, Hind Contmfateral) in response to a mechanical stimulus of 4.41 mN applied to the glabrous skin of the paw. Prior to henrisection, the stimulus is below the threshold of detection; however, 10 days atler the hemisection, both hindlimbs demonstrate an increase in the number of paw withdrawals which persists for the duration of the behavioral tests. This increase is statistically significant (P < 0.05).

101

To extend the data on the mechanical stimulus parameters, we also collected behavioral responses based on tests with a von Frey hair of 941 mN. In general, preoperatively and in sham controls, this stimulus strength was just above the threshold of detection for the forelimbs and hindlimbs (Figs. 5 and 6). Preoperatively, seven out of ten animals demonstrated forelimb paw withdrawal two or less times and the remaining animals did not respond at all to the 9.41 mN stimulus on the 3 test days prior to surgical intervention and were not statistically different from sham controls tested for 160 days (Fig. 6). Following the hemisection, the paw withdrawal frequencies increased in both forelimbs and hindlimbs and this increase is statistically significant when compared to sham control values (P< 0.05). There were no statistically significant side to side differences in withdrawal frequencies in the forelimbs or the hindlimbs in the hemisected group and only occasional statistically significant differences in the forelimb side to side comparisons in the sham control group. Unlike the von Frey hair with bending forces of 4.41 mN and 9.41 mN, application of a 122.09 mN bending force consistently elicited paw withdrawals prior to hemisection. This occurred with a mean* SE frequency of 8.6 k 0.3 and 4.6 ~ O.Z?paw withdrawals in the forelimbs and hindlimbs, respectively. The forelimb responses decreased on POD 1; however, by POD 6 the postoperative paw withdrawal frequencies were not significantly different from preoperative 10 9~

8-

.: G

~-

9

4 -4 -4 0

-10 -5

Fore Ipsilateral Fore Contralateral Control Fore Left Control Fore Right

I

-

160

5 10 15 20 25 30 35 40 45 50 Days After Hemisection

Fig, 5. Time course of the mean &SE of hindlimb paw withdrawals graphed m in Fig. 3, in response to a mechanical stimulus of 9,41 mN applied tothe glabrous skin of the paw. Priorto hemisection,the stimulus is below the threshold of detection; however, by 15 days after the hemisection, both hindlimbs demonstrate an increase in the number of paw withdrawals which persists for the duration of the behavioral tests, This increase is statistically significant when compared to sham control ~alues (P< 0.05)

,, /&

.L --u~ ~ ~

9

TT

0

10

/+ -u+ ++ ~

10

~

8

.:

7

Fore Ipsilateral Fore Contralateral Control Fore Left Control Fore Right

T

1 0+

~ti -10 -5

I

0

5 10 15 20 25 30 35 40 45 50

160

Days After Hemisection Fig, 4. Time course of the mean * SE of paw withdrawals in both left and right forelimbs in controls (Control Fore Left, Control Fore Right) and in the ipsilaterrd and contmlateml forelimb of hemisected rats (Fore Ipsilateml, Fore Contmlateml) in response to a mechanical stimulus of 4.41 mN applied to the glabrousskin of the paw. Prior to hemisection, the stimulus is below the threshold of detection; however, 15 days after the hemisection, both forelimbs demonstrate an increase in the number of paw withdrawals which persists for the duration of the behavioral tests. This increase is statistically significant (P < 0,05),

I

I

I

I

[

I

1

I

I

-lfJ -5“ O 5 10 15 20 25 30 35 40 45 50

//+

160

Days After Hemisection

Fig. 6. Time course of the mean * SE of forelimb paw withdrawals graphed as in Fig, 4, in response to a mechanical stimulus of 9,41 mN applied to the glabrous skin of the paw. Prior to hemisection, the stimulus is at the threshold of detection; however, 15 days after the hemisection, both forelimbs demonstrate an increase in the number of paw withdrawals which persists for the duration of the behavioral tests. This increase is statistically significant when compared to sham control values (P< 0,05),

102

L

0

(

-6 ~

-1

1

2

3

6

9

1315

19202224272931

24363841

43454849

Days After Hemisection

Fig. 7. Mean k SE of paw withdrawal latency measured in seconds (See) in both hindlimbs, ipsilateral turd contrafateral to the hemisection, in response to a radiant thermal stimulus applied to the glabrous skin of the paw, Postsurgical responses demonstrate a decrease in paw withdrawal latency for both hindlimbs which persists for the duration of the behavioral tests. This decrease is statistically significant when compared to sham control values (dashed line) of 22.9 * 2.8 (P < 0.05).

values. This persisted for 160 days until the end of testing. There were no statistically significant side to side differences in forelimb withdrawal frequencies. Similar trends occurred in the hindlimbs except there was a side to side difference in withdrawal frequencies until the ipsilateral hindlimb demonstrated motor recovery. Thermal allodynia Paw withdrawal latencies to a radiant heat source were measured and recorded as described above. Preoperatively, the mean paw withdrawal latency to the heat stimulus was 22.9 * 3.0 sec for the hindlimbs and 20.1 * 3.1 sec for the forelimbs and these responses were not statistically different when compared to sham control values of 22.9 * 2.8 and 20.3 * 2.9 for hindlimbs and forelimbs, respectively (Figs. 7 and 8). Following the hemisection, paw withdrawal latencies were significantly lower when compared to the preoperative mean latency times for both hindlimbs and forelimbs. Postoperatively, the mean hindlimb latency of paw withdrawal decreased to 11.9* 1.8 sec. This decrease is statistically significant when compared to preoperative and sham control values (P < 0.05). This statistically significant difference persisted for up to 7 weeks. There were no significant differences when ipsilateral and contralateral hindlimb paw withdrawals were compared at each time point, both preoperatively and postoperatively. Similar trends were evident in the paw withdrawal latencies measured in the forelimbs. In 15% of the trials for forelimb paw withdrawal to radiant heat, the rats would lick and shake their paws following paw withdrawal. The paw licking and shaking behavior was seldom exhibited in the hindlimb trials with the radiant heat source. Further thermal testing was done in order to assess supraspinal responses to noxious heat stimuli. The temperature at which the test animals withdrew their paws accompanied by supraspinal behavior such as paw lick was recorded using the Physitemp thermal pad stimulators. In the forelimbs,

control values of 50.9 ~ 0.4°C and 51.5 &0.3°C in the right and left limbs induced the responses, respectively (Fig. 9). These values decreased to 46.7 &0.6°C and 47.8 + 0.4”C, for the contra- and ipsilateral limbs following hemisection between POD 50 and 75. Animals tested up to POD 160 continued to have decreased paw withdrawal thresholds in the forelimbs bilaterally: 45.7 f 0.9°C and 46.5 ~ 1.4 see, contraand ipsilateral, respectively. These values are statistically significant (P< 0.05). Hindlimb control values of 53.6 ~ 0.3°C and 53.1 t 0.3°C were necessary to cause supraspinal responses. Following spinal hemisection, these values decreased to 50.6 ~ 0.4°C and 50.1 k 0.6°C for POD 17-50 in the contra- and ipsilateral hindlimbs, respectively. These values were statistically different compared to preoperative and sham control vaiues (P< 0.05). Animals tested up to POD 160 continued to have decreased thermal thresholds in the hindlimbs bilaterally with mean values of 50.2 ~ 1.0 and 51.0 f 0.1 for ipsilateral and contralateral, respectively. These responses were statistically different from sham control values (P< 0.05). Vocalization scores The vocalization scores were recorded in all animals preoperatively and after partial spinalization. For all animals, no vocalizations or other indications of a nociceptive response occurred during preoperative tests with a mechanical stimuIus strength of 49.8 mN. Similarly, sham controls did not vocalize to the stimulus strength during the entire testing period. Following the operation, for several days, all hemisected animals demonstrated an acute period of vocalization that occurred in response to mechanical stimulation. The most predominant vocalizations occurred within the first 2448 h and were characteristically different in timber (louder) and pitch (greater range of pitch within single vocalization) than subsequent vocalizations. Acutely, the hemisected animals vocalized during handling and would frequently attempt to bite the handler; whereas, the sham control animals 30$

+

Ipsilateral

+

C0ntralater61

25-

c ; 20 , : $! ~ 15 -

———————

———————

——————————————.

~ ~m 10 & ~

i 5-

‘~

-6 -4 -1 0

1

3

6

9 13 15 192022

2427

29 31 34 363841

43454849

Days After Hemisection

Fig. 8. Mean* SE of paw withdrawal latency in both forelimbs, ipsilateral and contralateraI to the hemisection, in response to a thermal stimulus applied to the glabrous skin of the paw. Postsurgical responses demonstrate a decrease in paw withdrawal latency for both forelimbs which persists for the duration of the behavioral tests. This decrease is statistically significant when compared to sham control values (dashed line) of 20,3 Y 2,9 (P < 0.05),

103 55

“.

I

B 54 = = 53

Controls Hemisected(50-65 dayS Hemisected(160 days)

52 $ 51 @ g 50 E g 49 48 47

termined empirically that more animals, approximately 30%, display similar behavior with respect to the penis. This behavior occurred for only a few days and was resolved. It was not clear if the penile tissue damage was due to excessive grooming by licking or due to self-mutilation, since both could occur as a result of the dysesthetic nature of the penis. The pain pathways to the penis are partially intact since the principal sensory innervation is C-fiber mediated and is bilateral from the dorsal nerve of the penis (Hulsebosch and Coggeshall 1982), and reach supraspinal pathway as indicated by vocalizations to touch with an applicator stick to administer topical antibiotics.

IliL

Discussion

46 45 c

g

c 8 p

L? Fig. 9. Graph compares the mean * SE of temperatures in “C at which supraspinal responses were elicited following a thermal heat stimulus applied to the paw of the forelimbs (Fore) or hindlimbs (Hind). Note that in sham control animals, the temperatures required to elicit this response are higher compared to hemisected animals tested from 50 to 65 days and tested again at 160 days postsurgery. These responses were maintained long-term in chronically injured animals and are statistically significant (P< 0.05) compared to control values.

did not display this behavior. On PODS7–21, the number of vocalizations was consistent in timber and pitch and was increased compared to preoperative intra-animal controls. There were individual variations as to the specific dermatomes which, when stimulated, evoked a vocalization response but the following generalizations can be made. Vocalizations occurred in response to stimulation of the skin rostrally, caudally, and bilaterally on the dorsal and lateral aspect of the animals for 5–7 segments rostral and caudal to the segmental site of the surgery. On POD 25, a mean vocalization score of 6.0 * 1.2 was obtained (Fig. 10). These values were statistically different from preoperative and sham control values of vocalization scores of O(F’< 0.05).

The results of the tests for mechanical allodynia (innocuous stimuli become noxious) as determined by increased number of paw withdrawals and increased number of vocalizations consistently demonstrated that both parameters produced nociceptive behaviors that are in agreement with the development of altered thresholds to mechanical stimulation. It is important to note that both forelimbs and hindlimbs demonstrated mechanical allodynia to these stimuli, despite the localization of thehemisectiontotheT13 spinal segment. The mechanical stimuli used in this test were von Frey hairs with weak bending forces of 4.41 mN or 9.41 mN to the glabrous skin of the rat paw. In human perceptual terms, these forces are described as very faint touch and faint touch,

9

.

8 7 6 I

I

Self-mutilation Only one rat originally selected for use in the present study demonstrated an observable self-mutilation of the hindlimb ipsilateral to the hemisection, and was subsequently eliminated from the current study. In other hemisection studies in our laboratories, we have found that approximately 10% of the rats displayed self-mutilation of the hindlimb contralateral to the hemisection. Once the self-mutilation behavior was initiated, it did not subside, requiring that the animal be euthanized for humane reasons. However, we de-

-10 -5

0

5

10

15 20

25

30

35 40

45

50

Days After Hemisection Fig. 10,Graphof themean* SEof thenumberof vocalizationsoutof 10 applications of a mechanical stimulus of 49,8 mN bending force. Note that prior to hemisection, this stimulus was below the threshold of detectability. Following surgery, the number of vocalizations increased and this increase persisted for the duration of the experiment.

104

respectively, and are perceived as innocuous. In agreement with the human perceptual data, the 4.41 mN von Frey hair was interpreted to be innocuous prior to the hemisection in both forelimbs and hindlimbs due to the absence of paw withdrawal behavior. In contrast to the human perceptual data, the 9.41 mN von Frey hair occasionally elicited paw withdrawal in the forelimbs of rats prior to hemisection. Since the majority of paw withdrawals are accompanied by aversive behaviors (abrupt head turning, biting the von Frey hair), these observations support the interpretation that, at least in the forelimbs of rats, this stimulus is noxious. This indicates the difference in thresholds at different body points when exposed to similar stimuli. However, both forelimbs and hindlimbs responded by increased frequency of paw withdrawal accompanied by aversive behaviors after the hemisection and this response was well established in all limbs by 14days. The application of 122.09mN bending forces to the glabrous skin of forelimbs and hindlimbs; however, was consistently noxious prior to hemisection. The forelimb responses were similar to baseline values after POD 6 but the hindlimb responses were increased compared to baseline values on the contralateral limb only. This discrepancy is difficult to explain but may represent the differences in somatosensory processing of high threshold mechanoreceptor stimuli when compared to low threshold mechanical stimuli (e.g. 4.41 mN). It is possible that nociceptive mechanical stimuli are mediated by tract cells which have increased electrical activity following contralateral hemisection. Further studies are necessary to elucidate the mechanism responsible for these changes. The thermal stimuli also elicited paw withdrawal responses, but in contrast to the mechanical stimuli, the paw withdrawal response was consistently elicited prior to hemisection. Since the rats displayed avoidance behavior that was consistent with the application of nociceptive stimuli, the radiant heat is interpreted to be noxious. In human perceptual terms, the radiant heat stimuli would best be described as an uncomfortably warm sensation. After the hemisection, the paw withdrawal response demonstrated a decreased latency in the hindlimbs. As in the development of mechanical allodynia, the thermal allodynia was prominent in both forelimbs and hindlimbs. It might be argued that the finding of enhanced paw withdrawal response, which is a flexor reflex, is not useful as an indicator of nociceptive behavior even though accompanied by supraspinally mediated aversive behavior. This line of reasoning is due to the fact that in a hemisection model, the surgical intervention severs some pathways involved in supraspinal perception and it follows that the usefulness of this model as a pain model might be questioned. It must be remembered that this reflex occurs independent of ascending or descending pathways since complete transections or isolating the spinal cord surgically results in heightened flexor reflex activity. This would lead to the development of behavioral changes which merely reflect heightened reflexes of the classical upper motor neuron lesion, partially due to the loss

of descending inhibition leading to enhanced segmental spinal excitability. While this can account for the acute and chronic effects of increased withdrawal of the hindlimb ipsiIateral to the lesion, it does not explain the attenuated mechanical and radiant heat thresholds for limb withdrawal to peripheral stimuli on the contralateral hindlimb and bilaterally on the forelimbs. The added behavior of licking and attending to each hindlimb, head turns and vocalizations following the stimulus application addresses the involvement of supraspinal transmission of the same primary afferent volley which elicited the withdrawal reflex. Finally, it must be remembered that normal reflex stepping during locomotion is best elicited by innocuous, proprioceptive stimuli. By contrast, the flexor reflex is elicited by nociceptive stimuli and results in a disruption of the alternation between flexion and extension that characterizes locomotion. Indeed, the nociceptive induced flexion is followed by extension only when the noxious stimulation is stopped (Willis 1982). Perhaps the most convincing tests with regard to the development of a chronic central pain state, are those that involve supraspinal pathways, such as vocalization, licking, head turning and self-mutilation. Each of these responses will be discussed in turn. Animals that vocalize in response to an evoked stimulus must receive activation of supraspinal centers. The initial increase in postoperative vocalizations was considered to be a result of primary allodynia from peripheral inflammatory processes which are unavoidable consequences with surgical intervention. Persistent and significant increases in vocalizations during the late postoperative period, after wound healing had occurred, were interpreted to represent the development of allodynia secondary to SCI and therefore indicative of the development of a central pain state. It is noteworthy that the cutaneous receptive fields of the allodynic area are located on the dorsal surfaces, rostral and caudal to the hemisection site, bilaterally. The ventral surface of the rats is for some reason resistant to the development of allodynia. The fact that the animals vocalized bilaterally in areas subserved by the spinal cord caudal to the lesions implies again that nociceptive information accesses supraspinal processing from either side of the injured cord. Another set of convincing data supporting the development of pain states following spinal hemisection were the supraspinal responses to thermal stimuli. The foot pads of each limb for each test animal were heated rapidly enough to elicit a paw withdrawal coupled with a supraspinal response. These responses included a paw lick, head turns, guarding, vocalization and moving away from the thermal stimulus. The temperature at which supraspinal responses occurred was significantly higher for all controls as well as for prehemisection test animals compared to those with spinal injury. Following spinal cord injury, supraspinal responses consistent with the development of pain occurred at lower thresholds for thermal stimuli in support of the development of thermal allodynia. In the case of hemisection of the cord, there remain several anatomic substrates for transmission of peripherally

105

evoked nociceptive information caudal to the lesion site both ipsilaterally and contralaterally to supraspinal centers. These pathways include the anterolateral spinal thalamic tract system, which decussates and mediates intact ipsilateral nociception caudal to the lesioned segment. The contralateral caudally affected area following anterolateral spinothalamic tractotomy can still be subserved by the uninterrupted spinocervicothalamic tract (Downie et al. 1988) and by the bilaterally projecting spinoreticuhr tract (Schoenen and Grant 1990) since both mediate nociception. In addition, there are short propriospinal pathways which are able to relay information to supraspinal centers from one side to the other since animals given thoracic over-hemisections on opposite sides of the cord, demonstrate supraspinal responses to noxious stimuli to the hindlimb (Basbaum 1973). An example of a neuroanatomic substrate for the unexpected forelimb behavior are the numerous individual nociceptive spinothalamic cells clustered in the first few cervical spinal segments which respond to peripheral noxious stimuli from the whole body (Smith et al. 1991; Smith and Hedge 1992). Finally, laminae I neurons, which are activated by nociceptive, thermoreceptive activity, are known to project ipsilaterally and bilaterally to regions in the brainstem that control behavioral states (Craig 1991, 1995). Thus, there are several intact pathways that provide neuroanatomic substrates for the transmission of hindlimb, as well as forelimb, somatosensation to supraspinal centers following spinal hemisection. Self-mutilation (or autophagia) as a measure of dysesthesias, which are disturbing somatic sensations, is a subject of much controversy in pain models (Rodin and Kruger 1984; Coderre et al. 1986). It should be noted that non-verbal humans and those with mental deficits demonstrate selfmutilation (Levitt 1985, 1987, 1988) that is similar to that observed in subhuman mammalian models (Coderre et al. 1986). Reports of central spinal cord lesions that result in hindlimb self-mutilation exist in the literature. Selfmutilation in the hindlimb contralateral to a thoracic hemisection or anterolateral cordotomy in monkeys occurs in 69~o of the animals (Levitt and Levitt 1981) and is interpreted to indicate the existence of painful dysesthesias (Levitt 1985). In our rodent model of spinal hemisection, approximately 10% of the rodents display contralateral hindlimb selfmutilation and thus are screened from further studies. Several investigators have reported self-mutilation after spinal hemisection with the results of 22-42Y0 displaying contralateral self-mutilation (Basbaum 1973; Saade et al. 1990; Xu et al. 1992). In a few cases, the ipsilateral limb was also involved (Levitt and Levitt 1981;Xu et al. 1992).Although the precise mechanism which triggers the self-mutilation behavior is unknown, it has been suggested that a specific group of antagonists of the NKI receptors may mediate this behavior (Wilcox 1988) and may give insight into the development of dysesthesias. Only one other well characterized mammalian model of chronic central pain developed after a spinal lesion exists to our knowledge. In that model, a laser is directed at the spinal

cord dorsal horn at a time when a solution containing a photopigment is in the circulation. The laser beam causes an intravascular photochemical reaction that produces occlusion of blood vessels and subsequent spinal cord ischemia (Hao et al. 1991; Xu et al. 1992). The allodynia that results from the lesion is then used to test the effects of drugs administered systemically to the animal. Our approach differs from that of Xu et. al. in the following ways: (1) a hemisection is a surgical lesion which spares the contralateral cord and minimizes variability; (2) responses to a variety of somatic stimuli are examined, since future pharmacological interventions will almost certainly selectively alter one quality of somatic sensation (e.g. heat and not touch); (3) forelimb and hindlimb behaviors were assessed; and (4) changes in locomotion are assessed using the BBB Locomotor Rating Scale to rule out alterations in motor control that might abolish the nociceptive behavior and as a record of behavioral outcome measures to allow comparisons in future pharmacological studies and comparisons with other SCI models. In spinal cord hemisection injury in humans, the BrownSequard syndrome (BSS), chronic severe pain can result on both sides of the body below the lesion (Guillian and Garcia 1931; Riddoch 1938). The pure BSS is characterized by a diagnosis of ipsilateral hemiplegia and contralateral hypalgesia (Koehler and Endtz 1986). On the lesioned side, pain presents acutely, and lasts days to weeks before resolving, whereas, on the contralateral side, burning pain and dysesthesias develop over months and are evoked by certain somatosensory stimuli, such as joint movements and pressure (Wall and Melzack 1989). In some cases, hemisection results in ipsilateral hyperesthesia (Koehler and Endtz 1986). In addition, some chronic spinal cord injured patients experience a band or ‘girdle’ of hyperpathia and/or allodynia at the level of the sensory loss (Tasker and Dostrovsky 1989). We see a similar development of allodynia in a girdle region in the rodent model but the distribution involves several segments rostral and caudal to the level of injury. The development of a girdle region of mechanical allodynia has been previously reported in a spinal ischemic injury in rodents (Hao et al. 1991;Xu et al. 1992). The acute phase of the pure BSS can be explained by a variety of mechanisms, including release of spinal cord nociceptive processing from descending inhibition (Sweet 1991) and an increase in the extracellular concentrations of excitatory amino acids (EEAs) induced by the acute trauma (Faden and Simon 1988; Liu and McAdoo 1993). Mechanisms that cause the development and maintenance of the chronic pain state in the model presented in the present manuscript are the focus of future studies. Proposed mechanisms elicited as a result of pathway lesions include the unmasking of latent pathways (Basbaum and Wall 1976; Devor and Wall 1981), decreased inhibition from lesions of descending or segmental pathways (Lombard et al. 1979; Devor and Wall 1981), the development of central sensitization or denervation hypersensitivity (Wright and Roberts 1978; Nakata et al. 1979; Willis 1993), or by plastic changes in

106

molecula and/or neuroanatomic pathways (McNeill et al. 1990, 1991). It is of importance to behaviorally characterize any nociceptive model in terms of the somatosensory stimuli that elicit the nociceptive response. This not only allows comparisons between other models of peripheral pain and central pain, but more importantly classifies the response toward pharmacological studies with intrathecally or systemically given agonists or antagonists with the goal of selectively alleviating the pain response while preserving other somatosensory functions and locomotor function. The contribution of the present model is an analysis of a variety of somatosensory stimuli in a model of chronic pain after SCI that will provide the basis for investigations of interventions for the prevention or the amelioration of the nociceptive behavior independent of other somatosensory and of locomotor behaviors.

Acknowledgements We would like to acknowledge the technical assistance of Ms. Mitria Ziainia, the secretarial assistance of Ms. Debbie Pavlu and the critical reading of this manuscript by Drs. William D. Willis, Howard Eisenberg, Haring Nauta and Regino Perez-Polo. This project was supported by The Kent Waldrep National Paralysis Foundation, the RGK Foundation and NIH grant NS 11255.

References Balazy, T.E., Clinical managementof chronic pain in spinal cord injury, Clin. J. Pain, 8 (1992) 102-110. Basbaum, A.I., Conduction of the effects of noxious stimulation by shorttiber multisynaptic systems of the spinal cord in the rat, Exp. Neurol., 40 (1973) 699-716. Basbaum, A.L and Wall, P.D., Chronic changes in the response of cells in adult cat dorsal horn following partial deafferentation: the appearance of responding cells in a previously non-responding region, Brain Res., 116 (1976) 181-204. Basso, D.M., Beattie, MS. and Breshnahan, J.C., A sensitive and reliable locomotor rating scale for open field testing in rats, J. Neurotraumtr, 12 (1995) 1-21. Bennet, G.J. and Xie, Y.K., A peripheral mononeuropathy in rat that produces disorders of pain sensation like those seen in man, Pain, 33 (1988) 87-107. Beric, A,, Dimitrjevic, M.R. and Lindbloom, U., Central dysesthesias syndrome in spinal cord injury patients, Pain, 34 (1988) 39– 48, Boivie, J., Disturbances in cutaneous sensibility in patients with central pain caused by the spinal cord lesions of syringomyelia, Pain Suppl., 2 (1984) s82. Choi, C., Yoon, Y., Na, H., Kim, S. and Chung, J.M., Behavioral signs of ongoing pain and cold dodynia in a rat model of neuropathic pain, Pain, 59 (1994) 369-376. Coderre, T.J., Grimes, R.W. and Melzack, R., Deafferentation and chronic pain in animals: an evaluation of evidence suggesting mrtonomy is related to pain, Pain, 26 (1986) 61–84. Craig, A.D., Spinal distribution of ascending lamina I axons antero-

gradely labeled with Phuseolus vrd$aris Ieucoagglutinin (PHA-L) in the cat, J, Comp. NeuroL, 313 (1991) 377–393, Craig, A.D., Distribution of bminstem projections from spinal Iamina I neurons in the cat and the monkey, J. Comp. NeuroL, 361 (1995) 225-248. Davidoff, G. and Roth, E.J., Clinical characteristics of central (dysesthetic) pain in spinal cord injury patients, In: K.L. Casey (Ed.), Pain and Central Nervous System Disease: The Central Pain Syndromes, Raven Press, New York, 1991, pp. 77-83. Davidoff, G., Roth, E., Guarracini, M., Sliwa, J. and Yarkony, G,, Functional limiting dysesthetic pain syndrome among spinal cord injury patients: a cross sectional study, Pain, 29 (1987) 39-48. Devor, M. and Wall, PD., Plasticity in the spinal cord sensory map following peripheral nerve injury in rats, J. Neurosci., 1 (1981) 679684, Downie, J.W., Ferrington, D.G., Sorkin, L.S. and Willis, Jr., W.D., The primate spinocervicothalamic pathway: responses of cells of the lateraI cervical nucleus and spinocervical tract to innocuous and noxious stimuli, J, Neurophysiol., 59 ( 1988) 861-885. Faden, A.I. and Simon, R.P., A potential role for excitotoxins in the pathophysiology of spinal cord injury, Ann. NeuroL, 23 (1988) 623626. Guillian, G. and Garcia, R,, Le syndrome de brown-sequard d’origine traumatique, Ann, Med., 29 (1931) 361–385. Hargreaves, K,, Dubner, R., Brown, F., Flores, C. and Jons, J., A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia, Pain, 32 (1988) 77-88. Hao, J.X., Xu, X.J., Aldskogius, H., Seiger, A. and Wiesenfield-Hallin, Z., Allodynia-like effects in rat after ischemic spinal cord injury photochemically induced by Imer imadiation, Pain, 45 (1991) 175-” 185. Hirata, H., Pataky, A., Kajmrder, K., LaMotte, R,H. and Collins J.G., A model of peripheral mononeuropathy in the rat, Pain, 42 (1990) 253– 254. Hulsebosch, C,E. and Coggeshall, R.E., An analysis of the axon populations in the nerves to the pelvic viscera in the rat, J. Comp. NeuroL, 211 (1982) 1-10. Jack, T.M. andLloyd, J.W., Long term efficacy of surgical cordotomy in intractable nonmrdignant pain, Ann. R, COIL Surg. Engl., 65 (1983) 97-102. Koehler, P,J. and Endtz, L.J., The Brown-Sequard syndrome, Arch. Neurol., 43 (1986) 921–924. Levitt, M., Dysesthesias and self mutilation in humans and subhumans: a review of clinical and experimental studies, Brain Res. Rev., 10 (1985) 247-290, Levitt, M., The central cause of deafferentation dysesthesias, Cephalalgia SUPP1. 7,6 (1987) 49–52. Levitt, M,, Experimental deafferentation syndromes, Appl. Neurophysiol., 51 (1988) 128–135. Levitt, M. and Levitt, J.H., The deafferentation syndrome in monkeys: dysesthesias of spinal origin, Pain, 10 (1981) 129-148. Liu, D. and McAdoo, D.J., An experimental model combining microdialysis with electrophysiology, histology, and neurochemistry for exploring mechanisms of secondary damage in spinrd cord injury: effects of potassium, J. Neurotrauma, 10 (1993) 349-362. Lombard, M.C., N&shold, B.S. and Albe-Fessard, D., Deafferentation hypersensitivity in the rat after dorsal rhizotomy: possible animal model for chronic pain, Pain, 6 (1979) 163–174, McNeill, D,L,, Carlton, S.M., Coggeshall, R,E. and Hulsebosch, C.E., Denervation-induced intraspinal synaptogenesis of calcitonin generelated peptide containing primary afferent terminals, J. Comp. Neural,, 296 (1990) 263–268, McNeill, D.L,, Carlton, S.M. and Hulsebosch, C.E., ]ntmspinal sprouting of calcitonin gene-related peptide containing primary afferents after deafferentiation in the rat, Exp. Neurol., 114 (1991) 321–329. Metskey, H. and Bogduk, N., Classification of Chronic Pain, IASP Press, Seattle, WA, 1994,222 pp.

107 Nakata,Y., Kusaka,Y. and Segawa, T., Supersensitivity to substance P after dorsal root section, Life Sci., 24 (1979) 1651–1654. and syringobulbia. In: P.J. Vinken, G.W. Nogues, M.A., Syrhsgomyelia Bruyn and H.L. Klawans (Eds.), Handbook of Clinical Neurology, Vol. 50, Elsevier, Amsterdam, 1987, pp.305-314.

Richards,J.S.,Meredhh,R.L.,Nepomuceno, C.,Fine,P.R.andBennett, G.,Psycho-social aspectsofchronicpaininspinalcord injury, Pain, 8 (1980) 355-366. Riddoch, G., The clinical features of central pain, Lancet, 234 (1938) 1093–1098, 1150-1156, 1205-1209. Rodin, BE. and Kruger, L., Deafferentation in animals as a model for the study of pain: an alternative hypothesis, Brain Res. Rev., 7 (1984) 213-228. Saade, NE., Atweh, S.F., Jabbur, S.J. and Wall, P.D., Effects of lesions in the anterolateral cohsmns and dorsolateral funiculii on selfmutilation behavior in rats, Pain, 42 (1990) 313–322. SAS/STAT, User’s Guide, Vols. I and 2, Version 6, SAS Institute, Cary, NC, 1992. Schliep G., Syringomyelia curdsyringobrdbia. In: P.J. Vinken and G.W. Bruyn (Eds.), Handbook of Clinical Neurology, Vol. 32, North Holland, Amsterdam, 1978, pp. 255-327. Schoenen, J. and Grant G., Spinal cord: connections. In: G. Paxinos (Ed.), The Human Nervous System, Academic Press, California, 1990, pp. 77–92. Smith, M.V, and Hedge, Jr., C.J., Response properties of upper cervical spinothalamic neurons in cats: a possible explanation for the unusual sensory symptoms associated with upper cervical lesionsin humans, Spine, 17 (1992) S375–S382. Smith, M.V., Apkmian, A.V. and Hedge, Jr., C.J., Somatosensory response properties of contralaterally projecting spinothalamic and nonspinothalamic neurons in the second cervical segment of the cat, J. Neurophysiol., 66 (1991) 83-102.

Sweet, W.H., Drmfferentation syndromes in humans: a general discussion. [ss:B.S. Nashold Jr. and J. Ovelmen-Levitt (Eds.), Deafferentation Pain Syndromes: Pathophysiology and Treatment, Raven Press, New York, 1991, pp. 259–283. Tarlov, 1.M. and Klinger, H., Spinal cord compression studies, II: time limits for recovery after acute compression in dogs, Arch. Neurol. PsychoI., 71 (1954) 271-290. Tasker, RR, and Dostrovsky, J.O., Deafferentation and central pain. In: P.D. Wall and R. Melzack (Eds.), Textbook of Pain, 2nd edn., Churchill Livingstone, New York, 1989, pp. 154–180. Wall, P.D. and Melzack, R., Textbook of Pain, 2nd edn., Churchill Livingston, New York, 1989. Wall, P.D., Devor, M., hrbal, R., Scadding, J.W., Schonfeld, D., Seltzer, Z. and Tomkiewicz, M.M., Autotomy following peripheral nerve lesions: ex~rimental wraesthesia dolorosa, Pain, 7 (1979) 103-113. White, J., Cordotomy: assessment of its effectiveness and suggestions for its improvement, Cain Neurosurg., 13 (1966) 1–19. Wilcox, G,L., Pharmacological studies of grooming und scratching behavior elicited by spinal substance P and excitatory amino acids, Ann. N, Y. Acad. Sci., 525 (1988) 228–236. Willis, W.D., Progress in Sensory Physiology 3, Springer-Verlag, New York, 1982, 159 pp. Willis, Jr., W.D,,Mechanical allodynia: a role for sensitized nociceptive tract cells with convergent input from mechanoreceptors and nociceptors? APS J., 2 (1993) 23–33, Wright, D.M. and Roberts, M.H., Supersensitivity to substance P anaIogue following dorsal root section, Life Sci., 22 (1978) 19–24. Xu, X.-J., Hao, J.-X., Aldskogius, H., Seiger, A. 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 (1992) 279–290.