Dorsal root ganglion compression as an animal model of sciatica and low back pain

Pain 77 (1998) 15–23

Research Papers

An experimental model for chronic compression of dorsal root ganglion produced by intervertebral foramen stenosis in the rat San-Jue Hu*, Jun-Ling Xing Institute of Neuroscience, The Fourth Military Medical University, The Sixth Clinical Center of CASP, Xi’an 710032, PR China Received 7 August 1996; received in revised form 17 March 1998; accepted 27 March 1998

Abstract Under anesthesia and sterile surgery, a small stainless steel rod (4 mm in length and 0.5–0.8 mm in diameter) was inserted into the L5 intervertebral foramen in the rat, developing intervertebral foramen stenosis and hence producing a chronic steady compression of the dorsal root ganglion (DRG). The hind paw on the injured side exhibited a significant reduction in the latency of foot withdrawal to noxious heat and manifested a persistent heat hyperalgesia 5–35 days after surgery. Injection of 1% carrageenan into the intervertebral foramen, presumably causing inflammation of the DRG, also produced hyperalgesia to heat on the hind paw of the injured side 5–21 days after surgery. Extracellular electrophysiological recordings from myelinated dorsal root fibers were performed in vivo. Spontaneous activity was present in 21.5% of the fibers recorded from DRG neurons injured with chronic compression in contrast to 1.98% from uninjured DRG neurons. The pattern of spontaneous activity was periodic and bursting in 75.3% of the spontaneously active fibers. These neurons had a greatly enhanced sensitivity to mechanical stimulation of the injured DRG and a prolonged after discharge. In response to TEA, topically applied to the DRG, excitatory responses were evoked in the injured, but not the uninjured, DRG neurons. Application of this experimental model may further our understanding of the neural mechanisms by which chronic compression of DRG induces low back pain and sciatica.  1998 International Association for the Study of Pain. Published by Elsevier Science B.V. Keywords: Dorsal root ganglion; Chronic compression; Hyperalgesia; Spontaneous activity

1. Introduction The chronic compression of the dorsal root ganglion (DRG) or its near nerve roots after vertebral injuries, intervertebral disc herniation or intervertebral foramen stenosis is an important factor causing low back pain and sciatica. Animal models designed to investigate the algogenic mechanisms consequent to injuries of the nerve or dorsal root, have employed experimental procedures such as dorsal root irritation or rhizotomy, cauda equina compression, segmental spinal nerve ligation and nerve root irritation (Howe et al., 1977; Wiesenfeld and Lindblom, 1980; Olmarker et al., 1991; Kim and Chung, 1992; Kawakami et al., 1994). However, none of these were aimed at determining the pathophysiological processes resulting from a chronic compression of the DRG. * Corresponding author. Tel.: +86 29 3251305; fax: +86 29 3246270.

We present a method of producing a chronic compression of the L5 DRG in rats by inserting a fine stainless steel rod into the L5 intervertebral foramen and thereby producing a stenosis of the intervertebral foramen. The affected hind limb becomes hyperalgesic to noxious heat stimuli and the L5 DRG neurons exhibit persistent spontaneous activity and enhanced responses to certain mechanical and chemical stimuli. We therefore believe that this animal model of a chronic compression of the DRG will increase our understanding of the neural mechanisms of low back pain and sciatica.

2. Methods 2.1. Animals and surgery Adult (200–350 g) Sprague–Dawley rats of both sexes

0304-3959/98/$19.00  1998 International Association for the Study of Pain. Published by Elsevier Science B.V. PII S0304-3959 (98 )0 0067-0

16

S.-J. Hu, J.-L. Xing / Pain 77 (1998) 15–23

were used under a University approved protocol. The sterilized surgical procedures were carried out under sodium pentobarbital anesthesia (40 mg/kg, i.p.). The skin was incised on the left side of the lumbar vertebrae between L4 and L6 and the left paraspinal muscles separated from the mammillary process and the transverse process at the L4–L6 level. In the first group of rats (the chronic compression group), the L5 intervertebral foramen was clearly exposed and a fine, L-shaped needle (about 0.6 mm in diameter) inserted about 4 mm into the L5 intervertebral foramen at a 30° angle with respect to the dorsal middle line and 10° with respect to the vertebral horizontal line. When the needle tip reached the DRG, the hind leg muscles of the operated side exhibited a slight, transient twitch. Then, the needle was withdrawn from the L5 intervertebral foramen and a stainless steel rod, 4 mm in length and 0.5–0.8 mm diameter, was inserted into the L5 intervertebral foramen along the path of the needle. This was intended to produce a steady compression against the L5 DRG (Fig. 1). With the intent of exerting a consistent pressure on the DRG, the diameter of the stainless steel rod was increased in relation to the weight of the rat, such that a diameter of 0.5–0.6 mm was selected for the rats of 200–250 g and diameters of 0.7 and 0.8 mm for rats weighing 250–300 and 300–350 g, respectively. Then the muscular layer and skin were sutured and antibiotics administered. In the second group of rats (the inflammation group), an injection of 15 ml of 1% carragee-

nan into the region around DRG through the L5 intervertebral foramen produced a local inflammation. In the third group of rats (the acute compression group), a stainless steel rod was inserted into the L5 intervertebral foramen for 30 s and then immediately removed. Lastly, in the fourth group, a sham surgery group of rats, the surgical procedure was identical to that for the chronic compression group, except that the stainless steel rod was not inserted into the interveterbral foramen. 2.2. Measurement of foot withdrawal latencies to noxious heat stimuli The rat was placed in an apparatus used to measure the latency of foot withdrawal to noxious heat stimuli (Hargreaves et al., 1988; Yang et al., 1993). A radiant heat source beneath the glass floor was focused on a 5 mm diameter spot on the posterior part of the plantar surface of the hind paw. The latency of foot withdrawal to the heat stimulus was measured as described previously. The hind paw on each side was alternately tested at interstimulus intervals of 5 min, five times on each paw. A difference score was calculated by subtracting the average latency of the control side from the average latency of the operated side. A negative difference score indicated a lower threshold at the operated side, i.e. hyperalgesia to noxious heat (Bennett and Xie, 1988). 2.3. Surgical exposure of DRG and the method of microfilament recording

Fig. 1. Schematic illustration of the method for producing a chronic compression of the DRG in rat. (A) Diagram of the position of the intervertebral foramen at the lumbar segment in the rat (lateral view). (B) Position and direction of a stainless steel rod inserted into the intervertebral foramen. (B1) Dorsal view of the 30° angle between the rod and the dorsal median line. (B2) Lateral view of the 10° angle between the rod and the vertebral lateral horizontal line.

Electrophysiological recordings were obtained 5–20 days after stainless steel rod insertion. Under chloralose-urethane anesthesia, the laminectomy was performed at L1–L2 and L4–L5 and at each of the two sites, a small pool prepared. In the L4–L5 pool, the L5 DRG on each side was sufficiently exposed, the stainless steel rod taken out. The spinal nerve was transected about 5–10 mm distal to DRG so that the discharges recorded from dorsal root fibers originated primarily from the ganglion and not from receptors in peripheral tissue. A pair of electrodes, used in the determination of conduction velocity, was placed on the spinal nerve distal to the ganglion and 25–30 mm to the recording electrode. During the experiment, the L4–L5 pool was filled with 1 ml of warm Kreb’s solution (35–37°C) containing, in mM: NaCl 150, KCl 5, CaCl2 2, MgCl2 1, glucose 10 and Tris 10 at a pH of 7.4. In the L1–L2 pool, the dorsal root of the L5 DRG was covered with warm paraffin oil (35–37°C). Under a microscope, a microfilament, about 30–50 mm in diameter, was teased away from the dorsal root of the normal or injured DRG and its proximal end placed on a fine platinum electrode (30 mm in diameter) for electrophysiological recording. The discharges were amplified, displayed on a memory oscilloscope (VC-11, Japan) and recorded by a computer. Discharges were assigned to a single unit if they were of the

S.-J. Hu, J.-L. Xing / Pain 77 (1998) 15–23

same shape and amplitude (Hu and Zhu, 1989). As stimulus current reached threshold for a spontaneously active fiber, a unit spike that had the same height and shape as the spontaneous activity was recorded and sometimes a brief afterdischarge recorded. These characteristics aided in identifying the spontaneously active unit in microfilaments containing a number of fibers (Wall and Devor, 1983). Conduction velocity of spontaneous active fibers was determined by dividing conduction distance by response latency to electrical stimulus. The number of spontaneously active nerve units in each microfilament was measured by counting the number of different spike heights in the ongoing discharge. The total number of conducting units in each microfilament was determined by measuring the number of all-or-none action potentials recruited as the stimulus current delivered to the spinal nerve was gradually raised (1 Hz, 0.1–0.2 ms, 0–7 mA; Fig. 2). The number of spontaneous active units divided by the total number of conducting units sampled yielded the percent incidence of spontaneous activity (Devor and Govrin-Lippmann, 1983; Wall and Devor, 1983).

17

times between the 5th and 42nd day after surgery. During this period, most of the difference scores for the chronic compression group were significantly greater than those for the sham surgery group, indicating the presence of cutaneous hyperalgesia. Most of the difference scores of the inflammation group were significantly greater than those of sham surgery group from 5 to 21 days following surgery but not significantly different beyond 24 days. This finding indicates that the time course of hyperalgesia was shorter for inflammation than for chronic compression. There were no significant differences between the difference scores for the acute compression and the sham surgery groups (Fig. 3).

2.4. Mechanical stimulation of DRG A small mechanical stimulator designed by our laboratory applied to the DRG a metal rod that exerted a steady force of 5 g for 10 s over a contact area of 0.79 mm2. The interval between mechanical stimuli was more than 5 min. 2.5. Test of the effects of chemical agents on activity in DRG neurons The ganglionic capsule and the inflammatory adhesive tissues on the injured DRG surface were carefully opened up and partially removed so that the Kreb’s solution, containing a given chemical agent in the L4– L5 pool, would be more likely to reach the neuronal somata within the DRG. Before applying the agent, the pool was washed twice with Kreb’s solution. The immersion time for each chemical agent was about 3–5 min and the interval between the applications of each chemical agent was more than 10 min.

3. Results 3.1. Latency of foot withdrawal due to noxious heat Before surgery, the latencies of foot withdrawal to noxious heat were measured for seven normal animals. The mean difference score was 0.15 ± 0.75 s (mean ± SE). During the month following surgery, none of the animals, including those in the chronic compression group (n = 6), inflammation group (n = 8), acute compression group (n = 5) or sham surgery group (n = 5) exhibited autotomy or obvious change in gait or posture. The latencies of hind foot withdrawal to noxious heat were tested at different

Fig. 2. Method of counting conducting units and measuring the conduction velocity of electrically evoked impulses in dorsal root fibers. (A) Spontaneous in a unit recorded from an L5 dorsal root filament. Three traces are superimposed. (B) Recruitment of activity in four fibers as the current of the stimulating electrode was gradually increased. Units 1, 2, 3 and 4 (the no. marked on the right of each action potential) were recruited into the compound action potential as all-or-none events as their thresholds (1.8, 2.4, 4.0 and 5.9 mA, respectively) were reached. Conduction distance was 30 mm. On the second trace (from the bottom), the action potential (marked no. 2) had the same height and shape as the unit with ongoing activity seen in (A) and its latency was 2.7 ms. Thus, the conduction velocity of this unit was 11.11 m/s. (C) Superimposed waveforms from the four units. As the stimulus current was increased above threshold, the latencies of previous recruited units shifted leftward. Time scale, 1 ms, voltage scale, 0.1 mV.

18

S.-J. Hu, J.-L. Xing / Pain 77 (1998) 15–23

Fig. 3. Time course of the differences in latency to noxious heat for the two hind paws (i.e. the difference score) in four groups of rats. Each data point provides the mean along with one SEM. ‘P’ on the abscissa represents the results of preoperative baseline tests from seven rats. *Significant difference between the given value and the corresponding value for the sham surgery group (P , 0.05 by Student’s t-test). K, Indicates that there are no corresponding values for the sham surgery group.

3.2. Incidence of spontaneous activity in neurons with cell bodies in the chronic compressed DRG Upon exposure of the compressed L5 DRG, 5–20 days after the initial surgery, three conditions were observed: (1) the stainless steel rod was in close contact with the DRG; (2) in most cases, the tissue surrounding the compressed DRG appeared slightly swollen and a little redder than that for the DRG on the contralateral side; (3) the injured DRG was often covered by inflamed tissue that adhered to the epineurium. When 162 microfilaments were dissected free from the dorsal roots of the injured DRGs, 922 active fibers responsive to spinal nerve stimulation were recorded, averaging 5.7 active fibers per microfilament. Of the total 922, 198 or 21.5% were spontaneously active. All the fibers had myelinated axons, as their conduction velocities were within 4.4– 38.5 m/s. Of the total 303 A-type fibers recorded from the dorsal roots of eight normal L5 DRGs, only six or 1.98% exhibited spontaneous activity (irregular and of low frequency). 3.3. Patterns of spontaneous discharges from the chronic compressed DRG neurons According to the dynamic features of interspike interval series the rhythm patterns of spontaneous discharge from the chronic compressed DRG neurons could be classified into three types (Ren et al., 1997): (1) periodic activity characterized by the interspike intervals that appeared repeatedly at regular intervals, including period one, period

two and so on; 33.3% (27/81) of the total number of spontaneous units fell into this category. (2) Non-periodic activity characterized by an irregular pattern of the interspike intervals (20/81 or 24.7%). (3) Bursting activity characterized by bursts separated by silent periods (34/81 or 42%). This type could be subdivided into regular bursting and irregular bursting (Fig. 4). 3.4. Responses of DRG neurons to mechanical stimulation of the ganglion When a mechanical force of 5 g in weight was gently applied for 10 s to the normal DRG (control group), discharges occurred in 16 of the 25 silent nerve units. The mean number of impulses evoked during the first 15 s after stimulus onset was 133.33 ± 20.46 (SEM). The mean duration of after discharge was 6.13 ± 2.0 s. No obvious response was elicited in the other nine units. When the same stimulus was applied to the injured DRG (chronic compression group), 17 of the 20 spontaneous active fibers tested exhibited a significant increase in activity. The mean number of impulses occurring during the first 15 s after stimulus onset, minus the mean number occurring during the interval 15 s before onset was 483.5 ± 67.3. The mean duration of after discharge was 49.3 ± 18.13 s (Fig. 5). Although the number of units responding to acute mechanical stimulation applied to the DRG was similar for the control and the chronic compression groups (16/25 vs. 17/ 20, P . 0.05 by chi-square test), the magnitude and the duration of evoked response (mean number of impulses

S.-J. Hu, J.-L. Xing / Pain 77 (1998) 15–23

19

Fig. 4. Patterns of spontaneous activities in dorsal root fibers from chronically compressed DRGs. (A) Activity with a single period. (B) Activity with two periods. (C) Non-periodic activity. (D) Regular bursting. (E) Irregular bursting. Time scale, (A,B,E) 200 ms; (C) 150 ms; (D) 800 ms. Voltage scale, 0.1 mV.

and mean duration) were each significantly different (P , 0.001, Student’s t-test). 3.5. Effects of TEA and EGTA on neuronal activity A topical application of 10 mM of TEA, evoked responses in most neurons tested in the chronic compression group (17/19) but in none of those tested in the normal DRG group (0/22) (Table 1). That is, the chronic compression resulted in an abnormal sensitivity to TEA. Although application of EGTA (10 mM, diluted in Ca2+-free Kreb’s solution) in the bath evoked or increased discharges in most normal and most injured DRG neurons tested (16/21 and 10/12, respectively) without any apparent difference (Figs. 6 and 7).

4. Discussion To our knowledge, there has been no practicable animal model of chronic compression of the DRG probably due to

the special location of the ganglion and the difficulty in maintaining a steady compression of it. In the present study, a mechanical compression of the DRG was achieved by inserting a stainless steel rod into a selected intervertebral foramen, without having to remove the vertebrae and expose the vertebral canal. The intervertebral foramen, having a limited size, became narrowed with the insertion of an object into it. This narrowing would be exacerbated by any swelling of the tissue surrounding the DRG, thus establishing a chronic, steady compression of the DRG. A local compression of the DRG can induce an intraneural edema and reduce the blood flow to the sensory nerve cell somata (Rydevik et al., 1988), which may be result in a change in neuronal excitability. The present study shows that the persistent hyperalgesia of the hind paw to noxious heat stimulus was caused by the chronic compression of the DRG, but not by an acute, compressive injury or a sham surgery. Recently, we observed that this chronic compression can also produce hyperalgesia to punctuate mechanical stimulation of the ganglion-injured hind paw (Song et al., 1997). Therefore, there is reason to believe that this model can

20

S.-J. Hu, J.-L. Xing / Pain 77 (1998) 15–23

Fig. 5. Responses evoked during mechanical stimulation of the DRG. Each arrow mark the onset of a force of 5 g in weight applied to the DRG for 10 s. (A) Peristimulus histograms of discharges from three silent units in normal DRG. (B) Peristimulus histograms of discharges from three spontaneous activity units in a chronically compressed DRG. Time scale, 2.0 min. Vertical mark, (A) 30 impulses; (B) 50 impulses. Bin width, 5 s.

reflect the sensory pathology course caused by chronic compression of the DRG neurons. However, we are uncertain of whether an intervertebral foramen stenosis can cause either cutaneous hyperalgesia in man or the symptoms of sciatica and low back pain. It would also be difficult to determine the presence of any sciatica or low back pain in the rat. However, it seems probable, based on following facts, that the present experimental model closely relates to sciatica and low back pain. First, in human patients, compression and irritation of the spinal nerve, dorsal root or the DRG is believed to be a main contributor to sciatica (radicular pain) and back pain (Farfan, 1979; Kirwan, 1989; Devor, 1996). In an early autopsy study it was found that in all cases of herniated lumber disc, the DRG was compressed, distorted, and manifested in various degrees of degeneration (Lindblom and Rexed, 1948). In the present study, the L5 DRG and possibly its near root in rats is persistently compressed by a fine stainless steel rod inserted into the L5 intervertebral foramen. Second, clinical

data have shown that the traumatized region of a dorsal root or the DRG was much more sensitive to mechanical stimulation than its normal, non-traumatized counterpart (Smyth and Wright, 1958; Murphy, 1967; Kuslich and Ulstro, 1991). Based on our observation that a mechanical force of 5 g in weight evokes a greater and more prolonged discharge rate in dorsal root fibers from a chronically compressed as opposed to a normal DRG, we conclude that the injured DRG neurons have developed an enhanced sensitivity to mechanical stimulation. Third, certain clinical observations strongly suggest that nerve impulses responsible for sciatica are generated in traumatized dorsal root axons and ganglia (Kuslich and Ulstro, 1991; Devor, 1996). The present study demonstrates that the DRG neurons injured with chronic compression generates a high incidence of spontaneous activity that is manifested in a variety of temporal patterns. These results suggest that the injured DRG neurons become an important source of abnormal spontaneous firing.

Table 1 Proportion of excitatory response units of DRG neurons to TEA and EGTA Injured DRGa

Normal DRG

TEA (10 mM) EGTA (10 mM) a

SU

SAU

ERR

SU

SAU

ERR

0/20 15/20

0/2 1/1

0/22 16/21

5/5 3/3

12/14 7/9

17/19 10/12

Only showing the data from the chronic compression DRG group. SU, silent units; SAU, spontaneous activity units; ERR, excitatory response units rate.

S.-J. Hu, J.-L. Xing / Pain 77 (1998) 15–23

21

Fig. 6. Discharges evoked by application of TEA and EGTA to the normal DRG. The interval between the two arrows represents the duration of application of 10 mM of TEA and 10 mM of EGTA (diluted with Ca2+-free Kreb’s solution). (A) A silent unit. (B) A spontaneously active unit. Time scale, 2.0 min. Vertical mark, 100 impulses. Bin width, 5 s.

Carrageenan can produce an inflammation of the tissue surrounding the nerve fiber and its ending, which results in nociceptor sensitization and hyperalgesia (Kocher et al., 1987; Eliav et al., 1996). In present study, injection of carrageenan into the intervertebral foramen presumably caused an inflammation of the DRG resulting in an ipsilateral hind paw hyperalgesia that is similar to that produced by an inflammation of the sciatic nerve (Eliav et al., 1996).The time course of hyperalgesia from DRG inflammation is shorter than that resulting from a chronic compression possibly because the chronic (granuloma) inflammatory responses induced by carrageenan may have gradually resolved within 7–14 days after injection (Rosa, 1972). In contrast, the chronic compression of DRG induced by the stainless steel rod inserted into the foramen could persist for a longer time. Therefore, injection of carrageenan through the intervertebral foramen into DRG as reported herein may be a convenient way for producing a transient inflammation of DRG. Inflammatory cells, such as macrophages, are present in the DRG after peripheral axotomy (Lu and Richardson, 1993). Some inflammatory mediators, such as bradykinin, prostaglandin and serotonin, enhance the excitability of normal and injured sensory neurons (Baccaglini and Hogan, 1983; Handwerker and Weeh, 1992; Walters and Ambron, 1995; Song et al., 1996), suggesting that inflammation is a critical factor for generating neural sensitization. Our observation of an inflammation of the tissues surrounding the surface of injured DRG after chronic compression, suggests the hypothesis that the stainless steel rod may have not only produced chronic compression of DRG but also a chemical irritation of DRG and hence a secondary inflammation of

the DRG. Therefore, it is very possible that both a direct, persistent compression and a secondary inflammation of the DRG play a role in generating the hyperexcitability of DRG neurons, which, in turn, might contribute to the hyperalgesia. In the present study, only a very small percentage of neurons in the normal DRG exhibited spontaneous discharges in accordance with a previous report by Wall and Devor (1983). In contrast, the incidence of spontaneous activity in chronically compressed DRG neurons was considerably higher than that resulting from a complete transaction of sciatic nerve (Wall and Devor, 1983; Babbedge et al., 1996). Ectopic spontaneous activity of DRG origin may become an objective and quantitative index for studying mechanisms of neuronal hyperexcitability and hyperalgesia and for analyzing relevant intracellular factors (Hu et al., 1997) because ectopic spontaneous activity from injured nerves is the characteristic manifestation of neuronal hyperexcitability and may contribute to spontaneous pain and paraesthesiae (Dubner, 1991; Bennett, 1993; Sheen and Chung, 1993; Matzner and Devor, 1994; Yoon et al., 1996). An important characteristic of the spontaneous activity from chronically compressed DRG neurons is that the patterns of periodic and bursting firing were displayed by 75.3% of the spontaneous active units. This is distinctly different from that of some reports (Wall and Devor, 1983; Xie et al., 1995) although very similar to that described in another (Babbedge et al., 1996). Further research on the significance of the higher incidence of periodic and bursting rhythm patterns and the underlying mechanisms in the DRG are needed.

22

S.-J. Hu, J.-L. Xing / Pain 77 (1998) 15–23

Fig. 7. Discharges evoked by application of TEA and EGTA to the chronically compressed DRG. (A) A silent unit. (B,C) Spontaneously active units. Same format as in Fig. 6 except that the vertical mark is 25 impulses for (A) and (C) and 75 impulses for (B).

It is well known that bathing neurons in a Ca2+-free medium reduces the after-hyperpolarization such that the frequency and duration of discharges are increased (Madison and Nicoll, 1984; Shen et al., 1994). It is also reported that increasing the extracellular concentration of Ca2+ at a nerve injury site can increase discharges in a sciatic nerve constriction model (Xie et al., 1993). In the present study, EGTA, a Ca2+ chelator, was topically applied to the DRG to determine whether the effects on neuronal response differ for normal and chronically injured DRG neurons. The results indicate no difference between normal and injured neurons in their excitatory responses to the removal of Ca2+. Injured axons in peripheral nerve have a greater sensitivity to K+ channel blockers, such as TEA (Xie and Xiao, 1990). Kajander et al. (1992) believed that this results from an exposure of K+ channels in the axolemma due to

demyelination at the injured site. They hypothesized that the normal DRG neurons and the T-junction, where there is no myelin, may also exhibit a sensitivity to K+ channel blockers. The present study found that an excitatory response to TEA was produced in most of neurons in the injured, but not in normal, DRG neurons. This result shows that injured neurons develop a relatively specific sensitivity to TEA and suggests that exposure of the axolemma may not be a direct cause of TEA sensitivity.

Acknowledgements We thank Dr. Robert H. LaMotte for his helpful comments on the manuscript. This study was supported by Chines National Foundation of Natural Sciences grant 39670247.

S.-J. Hu, J.-L. Xing / Pain 77 (1998) 15–23

References Babbedge, R.C., Soper, A.J., Gentry, C.T., Hood, V.C., Campbell, E.A. and Urban, L., In vitro characterization of a peripheral afferent pathway of the rat after chronic sciatic nerve section, J. Neurophysiol., 76 (1996) 3169–3177. Baccaglini, P.L. and Hogan, P.G., Some sensory neurons in culture express characteristics of differentiated pain sensory cells, Proc. Natl. Acad. Sci. USA, 80 (1983) 594–598. Bennett, G.J., An animal model of neuropathic pain: a review, Muscle Nerve, 16 (1993) 1040–1048. Bennett, 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. Devor, M., Pain mechanisms and pain syndromes. In: J.H. Campbell (Ed.), Pain 1996 – An Updated Review, IASP Press, Seattle, WA, 1996, pp. 103–112. Devor, M. and Govrin-Lippmann, R., Axoplasmic transport block reduces atopic impulse generation in injured peripheral nerves, Pain, 16 (1983) 73–85. Dubner, R., Neuronal plasticity and pain following peripheral tissue inflammation and nerve injury. In: M.R. Bone et al. (Eds.), Proceeding of the VIth World Congress on Pain, Elsevier, Amsterdam, 1991, pp. 263–276. Eliav, E., Ruda, M.A. and Bennett, G.J., An experimental neuritis of the rat sciatic nerve produces ipsilateral hindpaw allodynia and hyperalgesia, Soc. Neurosci. Abstr., (1996) 856. Farfan, H.F., Mechanical factors in the genesis of low back pain, In: Bonica et al. (Eds.), Advances in Pain Research and Therapy, Vol. 3, Raven Press, New York, 1979, pp. 635–645. Handwerker, H.O. and Weeh, P.W., Nociceptors chemosensitivity and sensitization by chemical agents. In: W.D. Willis (Ed.), Hyperalgesia and Allodynia, Raven Press, New York, 1992, pp. 107–115. Hargreaves, K., Dubner, R., Brown, F., Flores, C. and Joris, J., A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia, Pain, 32 (1988) 77–88. Howe, J.F., Loeser, J.D. and Calvin, W.H., Mechanosensitivity of dorsal root ganglion and chronically injured axons: a physiological basis for the radicular pain of nerve root compression, Pain, 3 (1977) 25–45. Hu, S.J. and Zhu, J., Sympathetic facilitation of sustained discharges of polymodal nociceptors, Pain, 38 (1989) 85–90. Hu, S.J., Song, X.J., Greenqust, K.W., Zhang, J.M. and LaMotte, R.H., Protein kinase A (PKA) mediates ectopic spontaneous activity in chronically compressed dorsal root ganglion, Soc. Neurosci. Abstr., 23 (1997) 1259. Kajander, K.C., Wakisaka, S. and Bennett, G.J., Spontaneous discharge originates in the dorsal root ganglion at the onset of a painful peripheral neuropathy in the rat, Neurosci. Lett., 138 (1992) 225–228. Kawakami, M., Weinstein, J.N., Chatani, K., Spratt, K.F., Meller, S.T. and Gebhart, G.F., Experimental lumber radiculopathy, behavior and histologic changes in a model of radicular pain after spinal nerve root irritation with chronic gut ligatures in the rat, Spine, 19 (1994) 1795–1802. Kim, S.H. and Chung, J.M., An experimental model for peripheral neuropathy produced by segmental spinal nerve ligation in rat, Pain, 50 (1992) 355–363. Kirwan, E.O.’G., Back pain. In: P.D. Wall and R. Melzak (Eds.), Textbook of Pain, Churchill Livingstone, New York, 1989, pp. 335–340. Kocher, L., Anton, F., Reeh, P.W. and Handerwerk, H.O., The effect of carrageenan induced inflammation of the sensitivity of unmyelinated skin nociceptors in the rat, Pain, 29 (1987) 363–373. Kuslich, S.D. and Ulstro, C.L., The tissue origin of low back pain and sciatica, Orth. Clin. North Am., 22 (1991) 181–187.

23

Lindblom, K. and Rexed, B., Spinal nerve injury in dorsolateral protrusions of lumbar discs, J. Neurosurg., 5 (1948) 413–432. Lu, X. and Richardson, P.M., Responses of macrophages in rat dorsal root ganglia following peripheral nerve injury, J. Neurocytol., 22 (1993) 334–341. Madison, D.V. and Nicoll, R.A., Control of the repetitive discharge of rat CA1 pyramidal neurons in vitro, J. Physiol., 354 (1984) 319– 331. Matzner, O. and Devor, M., Hyperexcitability at sites of nerve injury depends on voltage sensitivity Na+ channels, J. Neurophysiol., 72 (1994) 349–359. Murphy, F., Sources and patterns of pain in disc disease, Clin. Neurosurg., 15 (1967) 343–351. Olmarker, K., Holm, S., Rosenqvist, A.L. and Rdevik, B., Experimental nerve root compression, a model of acute, graded compression of the porcine cauda equina and an analysis of neural and vascular anatomy, Spine, 16 (1991) 61–69. Ren, W., Hu, S.J., Zhan, B.J., Wang, F.Z., Gong, Y.F. and Xu, J.X., Period-adding bifurcation with chaos in the interspike intervals generated by an experimental neural pacemaker, Int. J. Bifur. Chaos, 7 (1997) 709–712. Rosa, M.D., Biological properties of carrageenan, J. Pharm., 24 (1972) 89–102. Rydevik, B.L., Myers, R.R. and Powell, H.C., Tissue fluid pressure in the dorsal root ganglion: an experimental study on the effects of compression, 34th Annual Meeting, Orthopedic Research Society, 1988, p. 135. Sheen, K. and Chung, J.M., Signs of neuropathic pain depend on signals from injured nerve fibers in a rat model, Brain Res., 610 (1993) 62–68. Shen, E., Wu, S.Y. and Dun, N.J., Spontaneous and transmitter-induced rhythmic activity in neonatal rat sympathetic preganglionic neurons in vitro, J. Neurophysiol., 71 (1994) 1197–1205. Smyth, M.J. and Wright, V., Sciatica and the intervertebral disc, J. Bone Jt. Surg., 40A (1958) 1404–1418. Song, X.-J., Hu, S.-J., Zhang, J.-M., Greenquist, K.W. and LaMotte, R.H., Cutaneous hyperalgesia during chronic compression of dorsal root ganglion, Soc. Neurosci. Abstr., 23 (1997) 1259. Song, X.-J., Zhang, J.-M. and LaMotte, R.H., Excitation of dorsal root ganglion cells by inflammatory mediators and glutamate in nerveinjured rat, Soc. Neurosci. Abstr., 22 (1996) 1804. Wall, P.D. and Devor, M., Sensory afferent impulses originate from dorsal root ganglia as well as from the periphery in normal and nerve injured rats, Pain, 17 (1983) 321–339. Walters, E.T. and Ambron, R.T., Long-term alterations induced by injury and by 5-HT in aplysia sensory neurons: convergent pathways and common signals?, Trends Neurosci., 18 (1995) 137–142. Wiesenfeld, Z. and Lindblom, U., Behavioral and electrophysiological effects of various types of peripheral nerve lesions in the rat: a comparison of possible models for chronic pain, Pain, 8 (1980) 285–298. Xie, Y.K. and Xiao, W.H., Electrophysiological evidence for hyperalgesia in the peripheral neuropathy, Sci. China, Ser. B, 33 (1990) 663– 672. Xie, Y.K., Xiao, W.H. and Li, H.Q., The relationship between new ion channels and ectopic discharges from a region of nerve injury, Sci. China, 36 (1993) 68–74. Xie, Y.K., Zhang, J.M., Petersen, M. and LaMotte, R.H., Functional changes in dorsal root ganglion cells after chronic nerve constriction in the rat, J. Neurophysiol, 73 (1995) 1811–1820. Yang, D.H., Zhao, Q., Li, L. and Hu, S.J., A method for measuring paw withdrawal latency to thermal nociception in rats, J. Fourth Milit. Med. Univ., 14 (1993) 101–103. Yoon, Y.W., Na, H.S. and Chung, J.M., Contributions of injured and intact afferents to neuropathic pain in an experimental rat model, Pain, 64 (1996) 27–36.