205
Pain, 43 (1990) 205-218 Elsevier
PAIN 01685
Basic Section A novel behavioral model of neuropathic pain disorders produced in rats by partial sciatic nerve injury Ze’ev
Seltzer
a, Ronald
Dubner
b and Yoram
Shir
’
’ Physiology Branch, Faculty of Dental Medtcine, Hebrew University of Jerusalem, Jerusalem (Israel), b Neurobiology and Anesthesiology Branch, National Institute of Denral Research, NIH, Bethesda, MD 20892 (U.S.A.), and ’ Department of Anesthesiology and Pain Clinic, Hadassah University Hospital, Jerusalem (Israel) (Received
14 February
1990, revision
received 6 June 1990, accepted
12 June 1990)
Partial nerve injury is the main cause of causalgiform pain disorders in humans. We present here a novel animal model of this condition. In rats we unilaterally ligated about half of the sciatic nerve high in the thigh. Within a few hours after the operation, and for several months thereafter, the rats developed guarding behavior of the ipsilateral hind paw and licked it often, suggesting the possibility of spontaneous pain. The plantar surface of the foot was evenly hyperesthetic to non-noxious and noxious stimuli. None of the rats autotomized. There was a sharp decrease in the withdrawal thresholds bilaterally in response to repetitive Von Frey hair stimulation at the plantar side. After a series of such stimuli in the operated side, light touch elicited aversive responses, suggesting allodynia to touch. The withdrawal thresholds to CO, laser heat pulses were markedly lowered bilaterally. Suprathreshold noxious heat pulses elicited exaggerated responses unilaterally, suggesting thermal hyperalgesia. Pin-prick evoked such exaggerated responses bilaterally (mechanical hyperalgesia). In a companion report, we show that these abnormalities critically depend on the sympathetic outflow. Based on the immediate onset and long-lasting perpetuation of similar symptoms, such as touch-evoked allodynia and hyperalgesia, and the resemblance of the contralateral phenomena to ‘mirror image’ pains in some humans with causalgia, we suggest that this preparation may serve as a model for syndromes of the causalgiform variety that are triggered by partial nerve injury and maintained by sympathetic activity. Summary
Key words:
Causalgia;
Partial
nerve injury;
Neuropathic
pain; Animal
Introduction Partial injury to some somatosensory nerves sometimes causes causalgia in humans [3,4,6,7,12, 14,18,19,23,25,36,37]. This neurological syndrome is typified by a rapid onset (usually within hours) of spontaneous constant burning pain that is exacerbated by several factors such as light touch (allodynia), temperature change, movement of the
Correspondence to: Dr. Ze’ev Seltzer, Physiology Branch, Faculty of Dental Medicine, Hebrew University of Jerusalem, POB 1172, 91010 Jerusalem, Israel. 0304-3959/90/$03.50
0 1990 Elsevier Science Publishers
model;
(Rat)
involved limb, visual and auditory stimuli and emotional disturbances [3,13,15,18]. Common to several of these factors is increased sympathetic efferent activity [3], and many of these patients are relieved by sympathectomy [13]. The pain in causalgia sometimes spreads beyond the cutaneous distribution of the damaged nerve [l&23,24]. In addition, it may appear in mirror image sites on the contralateral intact limb [12,14,19,23]. Patients have increased response to noxious stimuli (hyperalgesia) [23,24] and the response to such stimuli may be exaggerated (hyperpathic), with the pain being perceived as unbearably intense [18]. The combination of hyperalgesia
B.V. (Biomedical
Division)
with allodynia, i.e., pain response to low threshold stimuli, characterizes the hyperesthetic state of causalgia. If untreated, symptoms may last indefinitely [IS]. For a detailed description of this syndrome see refs. 7, 14, 23, 25, 28 and 37. Because of the unavailability of a suitable behavioral animal model of causalgia, its underlying pathophysiological mechanisms are largely unknown, even though this syndrome has been recognized for more than a century 1191. This is reflected by the numerous, but only partly substantiated, hypotheses which involve peripheral sensory [5,6,29,41], sympathetic [5,6,13,15,26,41] and spinal [14,26,36] structures in the etiology of this syndrome. Here we present a novel behavioral model of causalgia in rats. Disorders in nocifensive behavior following noxious and non-noxious stimuli begin hours after partial sciatic injury and last for at least 7 months. This model shows many of the symptoms that characterize causalgia in humans, including the complex combination of rapid onset, allodynia to touch, hyperalgesia, mirror image phenomena and dependence on the sympathetic outflow, which are pathognomonic of the syndrome. In the present report we describe the surgical procedure and postoperative (PO) behavioral observations that strongly suggest the presence of spontaneous pain. Furthermore, we detail the responses of the rats in a battery of sensory tests. In a previous report we have shown that it is unlikely that these disorders are the result of sensitization of primary afferent receptors 1311. Rather. the rapid initiation of these disorders and their contralaterai appearance suggest central reorganization. In an accompanying report [33], we present the evidence for the dependence of these sensory abnormalities on the sympathetic outflow. Using the data of these complementary papers we show that this animal preparation may serve as a model for sympathetically maintained pain (SMP). A summary of this work was presented elsewhere ]3%
Suf?p% Adult male Sabra rats (Wistar-derived, II = IOh), weighing 300.- 350 g at the time of operation. wrrc used in this study. The animals were kept in standard colony conditions (3 4 per metal cage with a floor covered with pine-wood shavings and sawdust; water and food were supplied ad lihiturn: temperature range: 22-24” C’; natural day/night cycle).
Surgery Under ether anesthesia and aseptic conditions the right sciatic nerve was exposed at high-thigh level. In sham operated rats the nerve was left intact and the wound was closed with 2 muscle sutures (4-O cotton) and 3-4 skin staples. In experimental animals the sciatic nerve underwent partial injury. Under a magnification of 25 x . the dorsum of the nerve was carefully freed from surrounding connective tissues at a site near the trochanter just distal to the point at which the posterior biceps semitendinosus (‘PBST’) nerve branches off the common sciatic nerve. Using honed (no. 5) jewellers’ forceps the nerve was fixed in its place by pinching the epineurium on its dorsal aspect, taking care not to press the nerve against underlying structures. An X-O silicontreated silk suture was inserted into the nerve with a 3/g curved, reversed-cutting mini-needle, and tightly ligated so that the dorsal l/3-1/2 of the nerve thickness was trapped in the ligature. The wound was then closed as in sham operated rats. In all rats the left leg and sciatic nerve were untouched. In a number of rats (n = 7) we examined whether the location of the injury along the sciatic nerve was of importance for this model. The sciatic nerve was injured as described above, but at a distal location, 3 mm proximal to the trifurcation of the sciatic nerve at the popliteal fossa. The paw was systematically mapped at a resolution of 223 mm with multiple CO, laser heat pulses (focused to a spot 0.1 mm in diameter). and with multiple skin pinches using a toothed forceps.
Methods This study was conducted according to the guidelines of the Ethical Committee of the International Association for the Study of Pain [43].
Behuviorul rests Postoperatively (PO) the animals were studied in the tests listed below and in Table I, and their
207 TABLE I
Test
Parameters
Non-noxious
Mechanical Repetitive touch Repetitive touch
Withdrawal threshold (in g) Signs of allodynia (e.g., escape, paw lick)
Noxious
Mechanical Pin-prick Thermal Laser heat pulses Laser heat pulses 5 a C cold plate 5 0 C cold plate
Withdraws (weak; normal; exa&gerated) Withdrawal threshold (in mCa1) Response duration to max. intensity (set) No. of hind paw flicks/5 min Signs of allodynia (e.g., escape, paw lick)
General
Gait Posture Autotomy
3-4 times, at a frequency of about 2 stimuli/set. At threshold, normal rats responded by a quick paw flick. Withdrawal threshold to heat was determined by beaming short pulses of infra-red radiation from a CO, laser (25-120 msec, 5 W, 40-150 mCa1, 1.5 mm in diameter). The CO2 laser was guided by a visible He/Ne aiming beam which illuminated the target with a red spot. Thresholds were measured by gradually increasing the duration of the pulses until the rat responded with a paw flick. The heat pulses were aimed alternately at the two hind paws. The threshold for a paw was calculated as the average of the threshold determined at 3 locations per paw: mid-toe no. 3, Ed-plantar area and heel. Preliminary experiments in 5 animals showed that when the sciatic nerve is partially injured as the sensibility is uniform described above, throughout the plantar surface of the foot (+/10%). We were not. able to locate hypoesthetic zones (Fig. 1). Therefore, in subsequent rats the 3 representative testing sites were used per paw. Response to a suprathreshoid heat pulse. Three CO, laser heat pulses of noxious but non-tissue damaging intensity (5 W, 120 msec, 150 mCa1) were aimed at the Ed-plantar area of each foot, alternating between the two hind paws. The time that lapsed between lifting the paw until replacing it on the floor was recorded. Two minutes were allowed between pulses to avoid sensitization. The response duration for each paw was calculated as the average of the 3 trials. Pin-prick. A single prick was given at the midplantar area using a sharply pointed plastic rod. The response of the rats was classified as: weak, if they responded with a sluggish withdrawal expressed by a faint paw lift; normal, if the response was a quick and robust paw flick, sometimes accompanied by a short paw lick; exaggerated, if the response was prolonged and ‘explosive,’ i.e., consisted of a ~mbination of a paw flick, vocalization, escape and prolonged oral treatment of the stimulated paw (licking, gentle biting and nail pulling). CoId piate. Rats were placed in a clear covered chamber 20 X 20 cm, 15 cm high, with a metal plate with a surface temperature of 5 (+/0.2)” C. The number of flicks of each hind paw repeated
BEHAVIORAL TESTS AND EXAMINED PARAMETERS
Limping Preferred weight-bearing side Self-mutilation of denervated areas
responses were compared to preoperative baseline values. Most rats were tested several times, up to 43 days PO. Some rats were retested 7 months PO (n = 9). Observation of gait and posture. The rat was placed in a chamber with a glass floor. This enabled the experimenter to observe the animal’s posture from underneath. The contact of the operated hind paw with the floor when standing or lying down was monitored. The animal was then placed in an open field and gait disorders, e.g., limping and paw placement, were recorded. Withdrawal threshold to touch was measured with a set of Von Frey hairs ranging from 0.25 to 20.0 g. The rat was placed in a chamber with a mesh metal floor, covered by an opaque plastic dome 20 x 30 cm, 10 cm high. The dome enabled the rat to walk freely but not to rear up on its hind limbs. Hence, the experimenter was able to reach the plantar surface of the paws from beneath, unobserved by the rat. Each hair was indented in the mid-plantar skin until it just bent. This was
20x
Results
and the time each paw was in the air was measured during an observation period of 5 min [l].
Site uf nerve injug Fig. 1 shows the cutaneous distribution of withdrawal thresholds to CO, noxious heat pulses. Fig. 1A and B describe a typical case in which the sciatic nerve was injured at a high-thigh level, as detailed above (Fig. 1A shows the distribution prior to the operation and Fig. 1 B 3 days PO). The thresholds to CO, laser heat pulses were similar throughout the plantar surface ( + / - 10%). When the location of the partial injury was shifted several millimeters distally on the common sciatic nerve
Statistics Group averages appear with the standard error of the mean. The significance level of the difference between means of various groups at certain postoperative times was examined with a 2tailed t test. The difference in the distribution of values in various groups (e.g., withdrawal thresholds to touch) at certain postoperative times was examined with the x2 test. Values of P < 0.05 were regarded as significant.
-C
A
80 80
Fig. 1. Mapping the sensitivity of the rat hind paw plantar surface in typical cases following partial sciatic nerve injury. A: spatial distribution of the withdrawal thresholds to heat pulses from a CO, laser stimulation (numbers are in mCal) 1 day prior to nerve injury. B: same hind paw mapped 3 days following ipsilateral partial sciatic injury which was performed just distal to the branchpoint of the PBST nerve. C: spatial ditribution in another rat, mapped 3 days following the same ipsilateral partial sciatic iquq which was performed in a distal location to that in B, just proximal to the sciatic trifurcation.
209
Fig. 2. Photographs, taken through a glass floor, of a rat 8 days following partial sciatic injury. A: posture while standing on its 4 limbs. Note that the rat bears less pressure on the right paw pads. B: the same rat during rest. Note that the operated hind limb rests cushioned on the scrotum. and the paw is kept in mid-air.
(as shown in a typical case in Fig. lC), hypersensitive areas with lowered thresholds, as shown in Fig. lB, were found. However, in between these areas islands of hypoesthesia and normoesthesia were often seen. Their location was unpredictable and they introduced high variability in the data. Signs of spontaneous pain Gait and posture, Within several minutes after the termination of the surgical anesthesia the rats were awake. An hour later they were fully alert and responsive. From this time the rats placed the paw of the operated side on the floor in a normal fashion, except for the two lateral toes which were flexed. Judging from the blanching of the weightbearing foot and toe pads pressed against the glass floor, the hind limb of the operated side bore less weight than the intact one. This disorder lasted
several weeks PO (Fig. 2A), suggesting that the contact with the floor appeared to be an aversive stimulus. Furthermore, beginning about 1 h PO and for a period of several weeks thereafter, the rats at rest clearly preferred the contralateral intact hind limb, as shown in Fig. 2B. The partially deafferented hind limb was usually placed on the cushioned scrotum, with the paw hanging in midair. This behavior was not observed in the sham operated rats. While resting, the partially deafferented rats, but not the sham operated ones paid considerable attention to the operated calf and paw, in a manner unlike usual grooming. They repeatedly licked the paw for many seconds, gently biting the skin and firmly pulling on the nails. However, in spite of this intense oral treatment, none of the rats selfmutilated the partially deafferented hind paw, as
100
3 b z
80
I
n q
Left
Right
60
3 2
40
20
o-2
2.5-5
WITHDRAWAL
5.5-l 0
THRESHOLD
10.520
(gm)
Post-OD. Timg
Post-O& Time
80
80
o-2
2.5-5
5.5-10
10.5-20
WITHDRAWAL THRESHOLD Fig. 3. A: withdrawal (n 2 23). Dark
thresholds (m g) to repetitive columns
designate
da-R
2.5-5
d54-R
5.5-10
10.5-20
WITHDRAWAL THRESHOLD (gm)
Von Frey hair htunulatlon
the left side and the striped
q
o-2
(gm)
dl-R
d27-R
60
60
rat\
n q
columns
in the plantar
surface of the hind paw of unoperated
the right aide. B.C‘: distribution
of the withdrawal
thresholds at 1. 8. 27 and 54 days PO of the right hind paw (III B) and left hind paw (in C) (n = 23 rats).
they usually do following total denervation [40]. When prodded to walk, neither limping was noticed, nor was hind paw eversion. Responses to non-noxious stimuli Withdruwal thresholds to touch. Fig. 3A shows the distribution of withdrawal thresholds of 23
rats to repetitive stimulation using Von Frey hairs. Prior to the partial nerve injury the majority of the thresholds clustered at the right side of the histogram. Fig. 3B and C show the shift to the left in the distribution of withdrawal thresholds 1, 8. 27 and 54 days after partial deafferentation. At day 27 PO, 14% of the thresholds of the right hind paw
211
!3
I
I
20 1 16.
16
12.
12
8-
a
4-
4
0I-
I
6
POSTOPERATIVE TIME (days)
lhr
1
2
3
5
8
11 14 16 213 33
POSTOPERATIVE TIME (days)
Fig. 4. Postoperative follow-up of the average withdrawal thresholds (in g) to repetitive Von Frey hair stimulation. ‘B’ designates the preoperative, baseline values. Asterisks in days 33 and 54 designate rats that were not checked previously. A: rats after a right sciatic partial injury (n = 23). B: rats after a right sham operation (n = 23). Solid boxes: right hind paw: open boxes: left hind paw.
were larger than 5 g, compared to 90% preoperatively (P < 0.0001). Similarly, 20% of the thresholds of the left hind paw were larger than 5
g at day 27 PO, compared to 95% preoperatively (P < 0.0001). Fig. 4 plots the PO changes in the average withdrawal thresholds. Sham operated rats
150 135
Right Lefl
z E
120
105
105
90
90
75
75
60
60 Bihrl
2
3
POSTOPERATIVE
5
8
11
TIME (days)
14
33
Blhr
1
2
3
5
6
11
14
POSTOPERATIVE TIME (days)
Fig. 5. Postoperative follow-up of the withdrawal thresholds to heat pulses from a CO, laser (in mCa1). A: rats after a right sciatic partial injury (n = 23). B: rats after a right sham operation (n = 23). Solid boxes: right hind paw: open boxes: left hind paw. ‘B’ designates the preoperative baseline values.
212
A 35
30
25
25
20
15
10
5 OB lhr
1
2
3
5
POSTOPERATIVE
8
11 14 21 33 43 TIME (days)
POSTOPERATIVE TIME (days)
Fig. 6. Postoperative follow-up of the response duration (in xc) to suprathreshold noxious law heat pulse. A: rata after a right partial sciatic nerve iqury (n = 23). B: rats after a right sham operation (n = 23). Solid boxes: right hind paw: open boxes: left hind paw. ‘B’ designates the preoperative baseline values.
showed a transient slight decrease in the thresholds which returned to baseline values in 2-3 days (Fig. 4B). However, from Fig. 4A it is evident that by 1 h after partial injury to the right sciatic nerve, average thresholds dropped sharply from 12.5 +/ - 1.6 to 1.8 +/- 0.4, in the right hind paw (P < 0.001, n = 23) and from 13.2 +/1.3 to 6.2 +/1.6 (P < 0.005, n = 23) in the contralateral intact hind paw. This bilateral threshold decrease remained unchanged for at least the follow-up period of 54 days shown in Fig. 4. In fact, when retested as long as 210 days PO thresholds were still found to be very low, bilaterally (2.2 +/- 0.7 for the right hind paw; P < 0.0001; and 3.8 +/1.6 for the left hind paw: P < 0.001: n = 9). Responses to noxious stimuli Withdrawal threshold to heat. Fig. 5B shows that in sham operated rats the withdrawal thresholds to noxious heat pulses did not change PO. However, the partially deafferented rats (Fig. 5A) presented a bilateral decrease in the threshold. As soon as 1 h after nerve injury, the operated hind paw reached a value of 65% of the baseline measure. By day 33 PO the thresholds reached a lower
level, bilaterally, after a gradual decline. In a second group of rats, the thresholds had still not returned to normal values 210 days PO (94.5 +/ - 5.4 mCa1 for the right hind paw, P -c 0.01: and 95.6 +/- 6.6 mCa1 for the left hind paw, P -c 0.01). Thus, partial deafferentation causes a rapid and long-lasting marked decrease in the withdrawal thresholds to heat stimuli. Furthermore, as for responses to touch stimuli, an immediate and long-lasting decrease in withdrawal threshold to heat stimuli was observed in an intact hind limb contralateral to the deafferentation. Response to suprathreshold heat. The response of unoperated or sham operated rats to a single pulse of 150 mCa1 was either a quick flick of the hind limb or a flick followed by a brief licking of the paw. The response of partially deafferented rats, in addition to a more robust flick, included a combination of some or all of the following behaviors: vocalization, attempts to escape, including running around in the testing chamber, intensive biting of the paw without wounding it, freezing for many seconds with the paw held in the mouth and eyes shut tight, pulling on the nails and intensive licking of the paw and the entire hind limb over
213
an area much larger than that stimulated
directly.
The oral treatment of the hind limb was unlike normal grooming. Since this behavioral repertoire evoked by a single stimulus comprised a variable combination of responses, we quantified it by measuring its duration. Fig. 6 shows data for the first 43 days PO. The response duration of sham operated rats (Fig. 6B) did not change PO. In contrast, partially deafferented rats (Fig. 6A) presented a significantly (P < 0.001) increased duration of response when stimulated at the right hind paw (the operated side). There was no change contralaterally. The peak ipsilateral increase occurred during the first week PO. Thereafter it declined to a plateau which was still more than double the baseline value, or that of the contralateral intact hind paw (P < 0.02 at day 43 PO). In a second group of rats (n = 9) studied 7 months PO, the response duration at the right hind paw was still double that of the left side (4.1 +/- 0.9 set and 2.0 +/1.4 set, respectively, P < 0.02). It is interesting to note that in some trials, when the intact contralateral hind paw was stimulated, immediately following a quick flick of this % 60-
s 2 0 5
50-
b
SHAM
Left Hindpaw
A
SHAM
- Right Hindpaw
-o-
PD rats - Left Hindpaw
t
PD rats - Right Hindpaw
40.
E
n
paw, the rats turned to the partially deafferented paw. They intensely licked and gently bit it and pulled on its nails, similar to the response following stimulation of the operated side. Pin-prick. Pin-prick also elicited exaggerated responses. The proportion of such exaggerated responses rapidly increased with PO time, reaching a peak at day 1.5 PO, when 54% of the responses were of this type (Fig. 7). It then declined to a plateau lasting at least 33 days, in which about 30% were exaggerated. There were also exaggerated responses to this stimulus in the contralateral intact paw, but they were less frequent than in the operated side (Fig. 7). Cold plate. When the rats (n = 28) were placed at day 14 PO on a 5°C cold plate for 5 min, the number of times they lifted the paw was 2.6 +/0.6 at the right hind paw and 1.6 +/- 0.3 at the left hind paw. This is in contrast to unoperated rats (n = 12) who withdrew significantly more times: 9.2 +/1.3 at the right hind paw (P < 0.001) and 8.9 +/1.3 at the left hind paw (P < 0.001). Variability of symptom expression. Only 1 of 37 partially deafferented rats (3%) did not show any of the sensory disorders described above. Decreased withdrawal thresholds to repetitive Von Frey hair stimulation appeared in the right hind paw in 89% (33/37) and in the left hind paw in 86% (32/37) of the partially deafferented rats. Decreased withdrawal thresholds to CO, laser pulses appeared in 97% (36/37) of the partially deafferented rats, both in the right and left hind paws. Increased response duration to a suprathreshold heat pulse was detected in the right hind paw in 78% (29/37) and in the left hind paw in 11% (4/37) of the partially deafferented rats.
Discussion B
lhr
1.5
4
POST-OPERATIVE
10 TIME
18
33
(days)
Fig. 7. Postoperative follow-up of the responses to a pin-prick at the mid-plantar area. The proportion (in I) of exaggerated responses is shown on the ordinate. Open symbols: left hind paw; solid symbols: right hind paw; triangles: sham operated rats (n = 23); boxes: partially deafferented (PD) rats (n = 23); ‘B’: baseline values.
Previous work has shown that the sciatic nerve provides about 85% of the hind paw’s skin sensory innervation [32]. This includes 100% of plantar skin where most of the sensory tests in the present study were made. Elimination of 30-50% of this information channel could therefore have resulted in dulling sensibility at the affected area. In fact,
214 TABLl:
11
COMPARISON THE PRESENT
BETWEEN THE SYMPTOMS ANIMAL MODEL
Parameter I. 2. 3. 4. 5. 6.
Cause Onset Duration Contralateral phenomena Relief by sympathectomy Allodynia evoked hy Fine touch
[IX] ANI)
-I-HP, St;NSORY
The present
l/3-1 /2 sciatic nerve tightly ligated Immediate (hours) at least 7 months
nerve injury immediate indefinite if untreated image’ pains cases . ~
model
+ markedly reduced withdrawal thresholds + markedly reduced withdrawal thresholds t- exaggerated responses to noxious heat ’ -t exaggerated responses to pin-prick h 4 tactile-defensive behavior
i
hyperalgesia to heat and exaggerated ’ responses to pin-prick. around in the cage. vocalization, prolonged licking and pulling on the nails. single mid-plantar noxious stimulus by oral treatment over an area much larger of the hind limb was unlike normal grooming (see footnote b).
to non-noxious that sensation
01,
+
ti -t
we observed hyperresponsivity noxious stimuli, suggesting amplified considerably.
DISORDERS
partial usually usually ‘mirror m most
unc(~rnrn~~n [21,34,3X3 -t +
9. Mechanical hyperalgesia 1 I. Spontaneous pam 12. Pain radiation
IN HUMANS
Causalgia
i
Cold 7. Touch-evoked hyperesthesia X. Heat-evoked hyperalgesia
” Hyperesthesia to fine-touch, h Attempts to escape, running ’ Rats typically respond to a directly. The oral treatment
OF CAUSAl.
Pain, 43 (1990) 205-218 Elsevier
PAIN 01685
Basic Section A novel behavioral model of neuropathic pain disorders produced in rats by partial sciatic nerve injury Ze’ev
Seltzer
a, Ronald
Dubner
b and Yoram
Shir
’
’ Physiology Branch, Faculty of Dental Medtcine, Hebrew University of Jerusalem, Jerusalem (Israel), b Neurobiology and Anesthesiology Branch, National Institute of Denral Research, NIH, Bethesda, MD 20892 (U.S.A.), and ’ Department of Anesthesiology and Pain Clinic, Hadassah University Hospital, Jerusalem (Israel) (Received
14 February
1990, revision
received 6 June 1990, accepted
12 June 1990)
Partial nerve injury is the main cause of causalgiform pain disorders in humans. We present here a novel animal model of this condition. In rats we unilaterally ligated about half of the sciatic nerve high in the thigh. Within a few hours after the operation, and for several months thereafter, the rats developed guarding behavior of the ipsilateral hind paw and licked it often, suggesting the possibility of spontaneous pain. The plantar surface of the foot was evenly hyperesthetic to non-noxious and noxious stimuli. None of the rats autotomized. There was a sharp decrease in the withdrawal thresholds bilaterally in response to repetitive Von Frey hair stimulation at the plantar side. After a series of such stimuli in the operated side, light touch elicited aversive responses, suggesting allodynia to touch. The withdrawal thresholds to CO, laser heat pulses were markedly lowered bilaterally. Suprathreshold noxious heat pulses elicited exaggerated responses unilaterally, suggesting thermal hyperalgesia. Pin-prick evoked such exaggerated responses bilaterally (mechanical hyperalgesia). In a companion report, we show that these abnormalities critically depend on the sympathetic outflow. Based on the immediate onset and long-lasting perpetuation of similar symptoms, such as touch-evoked allodynia and hyperalgesia, and the resemblance of the contralateral phenomena to ‘mirror image’ pains in some humans with causalgia, we suggest that this preparation may serve as a model for syndromes of the causalgiform variety that are triggered by partial nerve injury and maintained by sympathetic activity. Summary
Key words:
Causalgia;
Partial
nerve injury;
Neuropathic
pain; Animal
Introduction Partial injury to some somatosensory nerves sometimes causes causalgia in humans [3,4,6,7,12, 14,18,19,23,25,36,37]. This neurological syndrome is typified by a rapid onset (usually within hours) of spontaneous constant burning pain that is exacerbated by several factors such as light touch (allodynia), temperature change, movement of the
Correspondence to: Dr. Ze’ev Seltzer, Physiology Branch, Faculty of Dental Medicine, Hebrew University of Jerusalem, POB 1172, 91010 Jerusalem, Israel. 0304-3959/90/$03.50
0 1990 Elsevier Science Publishers
model;
(Rat)
involved limb, visual and auditory stimuli and emotional disturbances [3,13,15,18]. Common to several of these factors is increased sympathetic efferent activity [3], and many of these patients are relieved by sympathectomy [13]. The pain in causalgia sometimes spreads beyond the cutaneous distribution of the damaged nerve [l&23,24]. In addition, it may appear in mirror image sites on the contralateral intact limb [12,14,19,23]. Patients have increased response to noxious stimuli (hyperalgesia) [23,24] and the response to such stimuli may be exaggerated (hyperpathic), with the pain being perceived as unbearably intense [18]. The combination of hyperalgesia
B.V. (Biomedical
Division)
with allodynia, i.e., pain response to low threshold stimuli, characterizes the hyperesthetic state of causalgia. If untreated, symptoms may last indefinitely [IS]. For a detailed description of this syndrome see refs. 7, 14, 23, 25, 28 and 37. Because of the unavailability of a suitable behavioral animal model of causalgia, its underlying pathophysiological mechanisms are largely unknown, even though this syndrome has been recognized for more than a century 1191. This is reflected by the numerous, but only partly substantiated, hypotheses which involve peripheral sensory [5,6,29,41], sympathetic [5,6,13,15,26,41] and spinal [14,26,36] structures in the etiology of this syndrome. Here we present a novel behavioral model of causalgia in rats. Disorders in nocifensive behavior following noxious and non-noxious stimuli begin hours after partial sciatic injury and last for at least 7 months. This model shows many of the symptoms that characterize causalgia in humans, including the complex combination of rapid onset, allodynia to touch, hyperalgesia, mirror image phenomena and dependence on the sympathetic outflow, which are pathognomonic of the syndrome. In the present report we describe the surgical procedure and postoperative (PO) behavioral observations that strongly suggest the presence of spontaneous pain. Furthermore, we detail the responses of the rats in a battery of sensory tests. In a previous report we have shown that it is unlikely that these disorders are the result of sensitization of primary afferent receptors 1311. Rather. the rapid initiation of these disorders and their contralaterai appearance suggest central reorganization. In an accompanying report [33], we present the evidence for the dependence of these sensory abnormalities on the sympathetic outflow. Using the data of these complementary papers we show that this animal preparation may serve as a model for sympathetically maintained pain (SMP). A summary of this work was presented elsewhere ]3%
Suf?p% Adult male Sabra rats (Wistar-derived, II = IOh), weighing 300.- 350 g at the time of operation. wrrc used in this study. The animals were kept in standard colony conditions (3 4 per metal cage with a floor covered with pine-wood shavings and sawdust; water and food were supplied ad lihiturn: temperature range: 22-24” C’; natural day/night cycle).
Surgery Under ether anesthesia and aseptic conditions the right sciatic nerve was exposed at high-thigh level. In sham operated rats the nerve was left intact and the wound was closed with 2 muscle sutures (4-O cotton) and 3-4 skin staples. In experimental animals the sciatic nerve underwent partial injury. Under a magnification of 25 x . the dorsum of the nerve was carefully freed from surrounding connective tissues at a site near the trochanter just distal to the point at which the posterior biceps semitendinosus (‘PBST’) nerve branches off the common sciatic nerve. Using honed (no. 5) jewellers’ forceps the nerve was fixed in its place by pinching the epineurium on its dorsal aspect, taking care not to press the nerve against underlying structures. An X-O silicontreated silk suture was inserted into the nerve with a 3/g curved, reversed-cutting mini-needle, and tightly ligated so that the dorsal l/3-1/2 of the nerve thickness was trapped in the ligature. The wound was then closed as in sham operated rats. In all rats the left leg and sciatic nerve were untouched. In a number of rats (n = 7) we examined whether the location of the injury along the sciatic nerve was of importance for this model. The sciatic nerve was injured as described above, but at a distal location, 3 mm proximal to the trifurcation of the sciatic nerve at the popliteal fossa. The paw was systematically mapped at a resolution of 223 mm with multiple CO, laser heat pulses (focused to a spot 0.1 mm in diameter). and with multiple skin pinches using a toothed forceps.
Methods This study was conducted according to the guidelines of the Ethical Committee of the International Association for the Study of Pain [43].
Behuviorul rests Postoperatively (PO) the animals were studied in the tests listed below and in Table I, and their
207 TABLE I
Test
Parameters
Non-noxious
Mechanical Repetitive touch Repetitive touch
Withdrawal threshold (in g) Signs of allodynia (e.g., escape, paw lick)
Noxious
Mechanical Pin-prick Thermal Laser heat pulses Laser heat pulses 5 a C cold plate 5 0 C cold plate
Withdraws (weak; normal; exa&gerated) Withdrawal threshold (in mCa1) Response duration to max. intensity (set) No. of hind paw flicks/5 min Signs of allodynia (e.g., escape, paw lick)
General
Gait Posture Autotomy
3-4 times, at a frequency of about 2 stimuli/set. At threshold, normal rats responded by a quick paw flick. Withdrawal threshold to heat was determined by beaming short pulses of infra-red radiation from a CO, laser (25-120 msec, 5 W, 40-150 mCa1, 1.5 mm in diameter). The CO2 laser was guided by a visible He/Ne aiming beam which illuminated the target with a red spot. Thresholds were measured by gradually increasing the duration of the pulses until the rat responded with a paw flick. The heat pulses were aimed alternately at the two hind paws. The threshold for a paw was calculated as the average of the threshold determined at 3 locations per paw: mid-toe no. 3, Ed-plantar area and heel. Preliminary experiments in 5 animals showed that when the sciatic nerve is partially injured as the sensibility is uniform described above, throughout the plantar surface of the foot (+/10%). We were not. able to locate hypoesthetic zones (Fig. 1). Therefore, in subsequent rats the 3 representative testing sites were used per paw. Response to a suprathreshoid heat pulse. Three CO, laser heat pulses of noxious but non-tissue damaging intensity (5 W, 120 msec, 150 mCa1) were aimed at the Ed-plantar area of each foot, alternating between the two hind paws. The time that lapsed between lifting the paw until replacing it on the floor was recorded. Two minutes were allowed between pulses to avoid sensitization. The response duration for each paw was calculated as the average of the 3 trials. Pin-prick. A single prick was given at the midplantar area using a sharply pointed plastic rod. The response of the rats was classified as: weak, if they responded with a sluggish withdrawal expressed by a faint paw lift; normal, if the response was a quick and robust paw flick, sometimes accompanied by a short paw lick; exaggerated, if the response was prolonged and ‘explosive,’ i.e., consisted of a ~mbination of a paw flick, vocalization, escape and prolonged oral treatment of the stimulated paw (licking, gentle biting and nail pulling). CoId piate. Rats were placed in a clear covered chamber 20 X 20 cm, 15 cm high, with a metal plate with a surface temperature of 5 (+/0.2)” C. The number of flicks of each hind paw repeated
BEHAVIORAL TESTS AND EXAMINED PARAMETERS
Limping Preferred weight-bearing side Self-mutilation of denervated areas
responses were compared to preoperative baseline values. Most rats were tested several times, up to 43 days PO. Some rats were retested 7 months PO (n = 9). Observation of gait and posture. The rat was placed in a chamber with a glass floor. This enabled the experimenter to observe the animal’s posture from underneath. The contact of the operated hind paw with the floor when standing or lying down was monitored. The animal was then placed in an open field and gait disorders, e.g., limping and paw placement, were recorded. Withdrawal threshold to touch was measured with a set of Von Frey hairs ranging from 0.25 to 20.0 g. The rat was placed in a chamber with a mesh metal floor, covered by an opaque plastic dome 20 x 30 cm, 10 cm high. The dome enabled the rat to walk freely but not to rear up on its hind limbs. Hence, the experimenter was able to reach the plantar surface of the paws from beneath, unobserved by the rat. Each hair was indented in the mid-plantar skin until it just bent. This was
20x
Results
and the time each paw was in the air was measured during an observation period of 5 min [l].
Site uf nerve injug Fig. 1 shows the cutaneous distribution of withdrawal thresholds to CO, noxious heat pulses. Fig. 1A and B describe a typical case in which the sciatic nerve was injured at a high-thigh level, as detailed above (Fig. 1A shows the distribution prior to the operation and Fig. 1 B 3 days PO). The thresholds to CO, laser heat pulses were similar throughout the plantar surface ( + / - 10%). When the location of the partial injury was shifted several millimeters distally on the common sciatic nerve
Statistics Group averages appear with the standard error of the mean. The significance level of the difference between means of various groups at certain postoperative times was examined with a 2tailed t test. The difference in the distribution of values in various groups (e.g., withdrawal thresholds to touch) at certain postoperative times was examined with the x2 test. Values of P < 0.05 were regarded as significant.
-C
A
80 80
Fig. 1. Mapping the sensitivity of the rat hind paw plantar surface in typical cases following partial sciatic nerve injury. A: spatial distribution of the withdrawal thresholds to heat pulses from a CO, laser stimulation (numbers are in mCal) 1 day prior to nerve injury. B: same hind paw mapped 3 days following ipsilateral partial sciatic injury which was performed just distal to the branchpoint of the PBST nerve. C: spatial ditribution in another rat, mapped 3 days following the same ipsilateral partial sciatic iquq which was performed in a distal location to that in B, just proximal to the sciatic trifurcation.
209
Fig. 2. Photographs, taken through a glass floor, of a rat 8 days following partial sciatic injury. A: posture while standing on its 4 limbs. Note that the rat bears less pressure on the right paw pads. B: the same rat during rest. Note that the operated hind limb rests cushioned on the scrotum. and the paw is kept in mid-air.
(as shown in a typical case in Fig. lC), hypersensitive areas with lowered thresholds, as shown in Fig. lB, were found. However, in between these areas islands of hypoesthesia and normoesthesia were often seen. Their location was unpredictable and they introduced high variability in the data. Signs of spontaneous pain Gait and posture, Within several minutes after the termination of the surgical anesthesia the rats were awake. An hour later they were fully alert and responsive. From this time the rats placed the paw of the operated side on the floor in a normal fashion, except for the two lateral toes which were flexed. Judging from the blanching of the weightbearing foot and toe pads pressed against the glass floor, the hind limb of the operated side bore less weight than the intact one. This disorder lasted
several weeks PO (Fig. 2A), suggesting that the contact with the floor appeared to be an aversive stimulus. Furthermore, beginning about 1 h PO and for a period of several weeks thereafter, the rats at rest clearly preferred the contralateral intact hind limb, as shown in Fig. 2B. The partially deafferented hind limb was usually placed on the cushioned scrotum, with the paw hanging in midair. This behavior was not observed in the sham operated rats. While resting, the partially deafferented rats, but not the sham operated ones paid considerable attention to the operated calf and paw, in a manner unlike usual grooming. They repeatedly licked the paw for many seconds, gently biting the skin and firmly pulling on the nails. However, in spite of this intense oral treatment, none of the rats selfmutilated the partially deafferented hind paw, as
100
3 b z
80
I
n q
Left
Right
60
3 2
40
20
o-2
2.5-5
WITHDRAWAL
5.5-l 0
THRESHOLD
10.520
(gm)
Post-OD. Timg
Post-O& Time
80
80
o-2
2.5-5
5.5-10
10.5-20
WITHDRAWAL THRESHOLD Fig. 3. A: withdrawal (n 2 23). Dark
thresholds (m g) to repetitive columns
designate
da-R
2.5-5
d54-R
5.5-10
10.5-20
WITHDRAWAL THRESHOLD (gm)
Von Frey hair htunulatlon
the left side and the striped
q
o-2
(gm)
dl-R
d27-R
60
60
rat\
n q
columns
in the plantar
surface of the hind paw of unoperated
the right aide. B.C‘: distribution
of the withdrawal
thresholds at 1. 8. 27 and 54 days PO of the right hind paw (III B) and left hind paw (in C) (n = 23 rats).
they usually do following total denervation [40]. When prodded to walk, neither limping was noticed, nor was hind paw eversion. Responses to non-noxious stimuli Withdruwal thresholds to touch. Fig. 3A shows the distribution of withdrawal thresholds of 23
rats to repetitive stimulation using Von Frey hairs. Prior to the partial nerve injury the majority of the thresholds clustered at the right side of the histogram. Fig. 3B and C show the shift to the left in the distribution of withdrawal thresholds 1, 8. 27 and 54 days after partial deafferentation. At day 27 PO, 14% of the thresholds of the right hind paw
211
!3
I
I
20 1 16.
16
12.
12
8-
a
4-
4
0I-
I
6
POSTOPERATIVE TIME (days)
lhr
1
2
3
5
8
11 14 16 213 33
POSTOPERATIVE TIME (days)
Fig. 4. Postoperative follow-up of the average withdrawal thresholds (in g) to repetitive Von Frey hair stimulation. ‘B’ designates the preoperative, baseline values. Asterisks in days 33 and 54 designate rats that were not checked previously. A: rats after a right sciatic partial injury (n = 23). B: rats after a right sham operation (n = 23). Solid boxes: right hind paw: open boxes: left hind paw.
were larger than 5 g, compared to 90% preoperatively (P < 0.0001). Similarly, 20% of the thresholds of the left hind paw were larger than 5
g at day 27 PO, compared to 95% preoperatively (P < 0.0001). Fig. 4 plots the PO changes in the average withdrawal thresholds. Sham operated rats
150 135
Right Lefl
z E
120
105
105
90
90
75
75
60
60 Bihrl
2
3
POSTOPERATIVE
5
8
11
TIME (days)
14
33
Blhr
1
2
3
5
6
11
14
POSTOPERATIVE TIME (days)
Fig. 5. Postoperative follow-up of the withdrawal thresholds to heat pulses from a CO, laser (in mCa1). A: rats after a right sciatic partial injury (n = 23). B: rats after a right sham operation (n = 23). Solid boxes: right hind paw: open boxes: left hind paw. ‘B’ designates the preoperative baseline values.
212
A 35
30
25
25
20
15
10
5 OB lhr
1
2
3
5
POSTOPERATIVE
8
11 14 21 33 43 TIME (days)
POSTOPERATIVE TIME (days)
Fig. 6. Postoperative follow-up of the response duration (in xc) to suprathreshold noxious law heat pulse. A: rata after a right partial sciatic nerve iqury (n = 23). B: rats after a right sham operation (n = 23). Solid boxes: right hind paw: open boxes: left hind paw. ‘B’ designates the preoperative baseline values.
showed a transient slight decrease in the thresholds which returned to baseline values in 2-3 days (Fig. 4B). However, from Fig. 4A it is evident that by 1 h after partial injury to the right sciatic nerve, average thresholds dropped sharply from 12.5 +/ - 1.6 to 1.8 +/- 0.4, in the right hind paw (P < 0.001, n = 23) and from 13.2 +/1.3 to 6.2 +/1.6 (P < 0.005, n = 23) in the contralateral intact hind paw. This bilateral threshold decrease remained unchanged for at least the follow-up period of 54 days shown in Fig. 4. In fact, when retested as long as 210 days PO thresholds were still found to be very low, bilaterally (2.2 +/- 0.7 for the right hind paw; P < 0.0001; and 3.8 +/1.6 for the left hind paw: P < 0.001: n = 9). Responses to noxious stimuli Withdrawal threshold to heat. Fig. 5B shows that in sham operated rats the withdrawal thresholds to noxious heat pulses did not change PO. However, the partially deafferented rats (Fig. 5A) presented a bilateral decrease in the threshold. As soon as 1 h after nerve injury, the operated hind paw reached a value of 65% of the baseline measure. By day 33 PO the thresholds reached a lower
level, bilaterally, after a gradual decline. In a second group of rats, the thresholds had still not returned to normal values 210 days PO (94.5 +/ - 5.4 mCa1 for the right hind paw, P -c 0.01: and 95.6 +/- 6.6 mCa1 for the left hind paw, P -c 0.01). Thus, partial deafferentation causes a rapid and long-lasting marked decrease in the withdrawal thresholds to heat stimuli. Furthermore, as for responses to touch stimuli, an immediate and long-lasting decrease in withdrawal threshold to heat stimuli was observed in an intact hind limb contralateral to the deafferentation. Response to suprathreshold heat. The response of unoperated or sham operated rats to a single pulse of 150 mCa1 was either a quick flick of the hind limb or a flick followed by a brief licking of the paw. The response of partially deafferented rats, in addition to a more robust flick, included a combination of some or all of the following behaviors: vocalization, attempts to escape, including running around in the testing chamber, intensive biting of the paw without wounding it, freezing for many seconds with the paw held in the mouth and eyes shut tight, pulling on the nails and intensive licking of the paw and the entire hind limb over
213
an area much larger than that stimulated
directly.
The oral treatment of the hind limb was unlike normal grooming. Since this behavioral repertoire evoked by a single stimulus comprised a variable combination of responses, we quantified it by measuring its duration. Fig. 6 shows data for the first 43 days PO. The response duration of sham operated rats (Fig. 6B) did not change PO. In contrast, partially deafferented rats (Fig. 6A) presented a significantly (P < 0.001) increased duration of response when stimulated at the right hind paw (the operated side). There was no change contralaterally. The peak ipsilateral increase occurred during the first week PO. Thereafter it declined to a plateau which was still more than double the baseline value, or that of the contralateral intact hind paw (P < 0.02 at day 43 PO). In a second group of rats (n = 9) studied 7 months PO, the response duration at the right hind paw was still double that of the left side (4.1 +/- 0.9 set and 2.0 +/1.4 set, respectively, P < 0.02). It is interesting to note that in some trials, when the intact contralateral hind paw was stimulated, immediately following a quick flick of this % 60-
s 2 0 5
50-
b
SHAM
Left Hindpaw
A
SHAM
- Right Hindpaw
-o-
PD rats - Left Hindpaw
t
PD rats - Right Hindpaw
40.
E
n
paw, the rats turned to the partially deafferented paw. They intensely licked and gently bit it and pulled on its nails, similar to the response following stimulation of the operated side. Pin-prick. Pin-prick also elicited exaggerated responses. The proportion of such exaggerated responses rapidly increased with PO time, reaching a peak at day 1.5 PO, when 54% of the responses were of this type (Fig. 7). It then declined to a plateau lasting at least 33 days, in which about 30% were exaggerated. There were also exaggerated responses to this stimulus in the contralateral intact paw, but they were less frequent than in the operated side (Fig. 7). Cold plate. When the rats (n = 28) were placed at day 14 PO on a 5°C cold plate for 5 min, the number of times they lifted the paw was 2.6 +/0.6 at the right hind paw and 1.6 +/- 0.3 at the left hind paw. This is in contrast to unoperated rats (n = 12) who withdrew significantly more times: 9.2 +/1.3 at the right hind paw (P < 0.001) and 8.9 +/1.3 at the left hind paw (P < 0.001). Variability of symptom expression. Only 1 of 37 partially deafferented rats (3%) did not show any of the sensory disorders described above. Decreased withdrawal thresholds to repetitive Von Frey hair stimulation appeared in the right hind paw in 89% (33/37) and in the left hind paw in 86% (32/37) of the partially deafferented rats. Decreased withdrawal thresholds to CO, laser pulses appeared in 97% (36/37) of the partially deafferented rats, both in the right and left hind paws. Increased response duration to a suprathreshold heat pulse was detected in the right hind paw in 78% (29/37) and in the left hind paw in 11% (4/37) of the partially deafferented rats.
Discussion B
lhr
1.5
4
POST-OPERATIVE
10 TIME
18
33
(days)
Fig. 7. Postoperative follow-up of the responses to a pin-prick at the mid-plantar area. The proportion (in I) of exaggerated responses is shown on the ordinate. Open symbols: left hind paw; solid symbols: right hind paw; triangles: sham operated rats (n = 23); boxes: partially deafferented (PD) rats (n = 23); ‘B’: baseline values.
Previous work has shown that the sciatic nerve provides about 85% of the hind paw’s skin sensory innervation [32]. This includes 100% of plantar skin where most of the sensory tests in the present study were made. Elimination of 30-50% of this information channel could therefore have resulted in dulling sensibility at the affected area. In fact,
214 TABLl:
11
COMPARISON THE PRESENT
BETWEEN THE SYMPTOMS ANIMAL MODEL
Parameter I. 2. 3. 4. 5. 6.
Cause Onset Duration Contralateral phenomena Relief by sympathectomy Allodynia evoked hy Fine touch
[IX] ANI)
-I-HP, St;NSORY
The present
l/3-1 /2 sciatic nerve tightly ligated Immediate (hours) at least 7 months
nerve injury immediate indefinite if untreated image’ pains cases . ~
model
+ markedly reduced withdrawal thresholds + markedly reduced withdrawal thresholds t- exaggerated responses to noxious heat ’ -t exaggerated responses to pin-prick h 4 tactile-defensive behavior
i
hyperalgesia to heat and exaggerated ’ responses to pin-prick. around in the cage. vocalization, prolonged licking and pulling on the nails. single mid-plantar noxious stimulus by oral treatment over an area much larger of the hind limb was unlike normal grooming (see footnote b).
to non-noxious that sensation
01,
+
ti -t
we observed hyperresponsivity noxious stimuli, suggesting amplified considerably.
DISORDERS
partial usually usually ‘mirror m most
unc(~rnrn~~n [21,34,3X3 -t +
9. Mechanical hyperalgesia 1 I. Spontaneous pam 12. Pain radiation
IN HUMANS
Causalgia
i
Cold 7. Touch-evoked hyperesthesia X. Heat-evoked hyperalgesia
” Hyperesthesia to fine-touch, h Attempts to escape, running ’ Rats typically respond to a directly. The oral treatment
OF CAUSAl.
