Normal and injury-induced sympathetic innervation of rat dorsal root ganglia increases with ageкод для вставкиСкачать
THE JOURNAL OF COMPARATIVE NEUROLOGY 394:38–47 (1998) Normal and Injury-Induced Sympathetic Innervation of Rat Dorsal Root Ganglia Increases With Age MATT S. RAMER AND MARK A. BISBY* Department of Physiology, Queen’s University, Kingston, Ontario, Canada K7L 3N6 ABSTRACT In rats, partial injury to a peripheral nerve often leads to sympathetically maintained pain (SMP). In humans, this condition is especially apparent in the elderly. Nerve injury also causes perivascular sympathetic axons to sprout into the dorsal root ganglion (DRG), forming a possible anatomical substrate for SMP. Here, we describe the effects of chronic sciatic nerve constriction injury (CCI) in young (3 months) and old (16 months) rats on neuropathic pain behavior and on sympathetic sprouting in DRG. Behavioral tests assessed changes in thermal allodynia and hyperalgesia and in mechanical allodynia. We found that 1) sympathetic innervation of the DRG increased naturally with age, forming pericellular baskets mainly around large DRG neurons, and that sympathetic fibers were often associated with myelinated sensory axons; 2) sympathetic fiber density following CCI was also greater in old than in young rats; and 3) in old rats, thermal allodynia was less pronounced than in young rats, whereas thermal hyperalgesia and mechanical allodynia were more pronounced. These results highlight the possibility that sympathetic sprouting in the DRG is responsible for the sympathetic generation or maintenance of pain, especially in the elderly. J. Comp. Neurol. 394:38–47, 1998. r 1998 Wiley-Liss, Inc. Indexing terms: neuropathic pain; adrenergic, hyperalgesia; allodynia Neuropathic pain is a condition that can result from disease (e.g., diabetes, herpes zoster) or trauma to nervous tissue (thalamic infarct, spinal cord, or peripheral nerve injury) and that differs from ‘normal pain,‘ in that it can be evoked at much lower stimulus thresholds and is and often of greater severity. A recent review indicates that a significant number of elderly humans experience chronic pain generally, and this remains underdiagnosed and undertreated (Gagliese and Melzack, 1997). Neuropathic pain (one form of chronic pain) is especially apparent in the elderly (Gibson et al., 1994), particularly in cases of postherpetic or trigeminal neuralgia (Devor, 1991). It is interesting that the more painful conditions are a result of incomplete nerve injuries, which leave an intact connection to the periphery (Jänig, 1985). Several animal models of neuropathic pain have been employed in the search for underlying mechanisms, three of which are similar, in that they all involve partial injury to a peripheral nerve (Bennett and Xie, 1988; Seltzer et al., 1990; Kim and Chung, 1992). Painful sensations due to normally nonnoxious stimulation (allodynia) and an abnormally intense perception of pain due to noxious stimulation (hyperalgesia) are symptoms of neuropathic pain both in humans and in animal models. Also common to both animals and humans is the fact that stimulation of the sympathetic r 1998 WILEY-LISS, INC. nervous system often exacerbates the pain (Bennett, 1991), whereas surgical or pharmacological sympathectomy often relieves it, albeit to differing extents (Kim and Chung, 1991; Neil et al., 1991; Seltzer and Shir, 1991; Kim et al., 1993; Desmeules et al., 1995). With respect to sympathetically linked pain in the elderly, as Gagliese and Melzack (1997) point out, although a ‘significant majority of the elderly experience pain which may interfere with normal functioning,’ the study of pain in the elderly has been neglected. We have found no references in the literature to suggest that the proportion of elderly sympathetically maintained pain (SMP) patients differs from the general population. However, the fact that a majority of the elderly suffer from chronic pain may be a reflection of an increase in the prevalence of all types of pain. In rats (Chung et al., 1993, 1996; McLachlan et al., 1993) and in mice (Davis et al., 1994, Ramer et al., 1997), a Grant sponsor: Medical Research Council of Canada; Grant number: MT-5198. *Correspondence to: Mark A. Bisby, Department of Physiology, Queen’s University, Kingston, Ontario, Canada K7L 3N6. E-mail: firstname.lastname@example.org Received 1 July 1997; Revised 30 October 1997; Accepted 8 December 1997 ADRENERGIC SPROUTING IN DRG IN AGING possible anatomical substrate for the interaction between the sympathetic and sensory nervous systems has been described: in the dorsal root ganglia (DRG) of nerveinjured animals, sympathetic fibers can be found among DRG neurons and sometimes form baskets around largediameter neuronal profiles. Although there are conflicting reports on the amount of contact between sympathetic axons and DRG neurons following nerve injury (Devor et al., 1995; Chung et al., 1997), electrophysiological evidence supports this site as a point of communication between these neuronal types (McLachlan et al., 1993; Devor et al., 1994; Michaelis et al., 1996). The sprouting phenomenon occurs after both complete and partial sciatic nerve injuries. After partial nerve injury, sprouting is well established 2 weeks postoperatively, develops with a time course similar to that of allodynia and hyperalgesia (Chung et al., 1996; Ramer and Bisby, 1997), and lasts as long as 10 weeks following a complete nerve injury (McLachlan et al., 1993). Because of the prevalence of chronic pain in elderly humans (Gagliese and Melzack, 1997) and the potential importance of this sympathetic sprouting in neuropathic pain states, our objectives were to determine 1) whether there were changes in the sympathetic innervation of the normal, uninjured DRG with aging; 2) whether partial nerve injuries made in young and aging rats resulted in differences in sympathetic sprouting; and 3) whether there were changes in the behavioral responses to partial nerve injuries made in young and old rats. We have used the chronic constriction injury (CCI) method of Bennett and Xie (1988), because, in our hands, this model produces easily reproducible behavioral results and sympathetic sprouting in the DRG (Ramer and Bisby, 1997). There is also evidence that this model produces pain behaviors involving the sympathetic nervous system (Neil et al., 1991; Wakisaka et al., 1991; Desmeules et al., 1995; Kim et al., 1997). We show here that the invasion of the DRG by sympathetic axons occurs naturally with age in rats, and sympathetic sprouting is increased in aged rats along with increased thermal hyperalgesia and mechanoallodynia. MATERIALS AND METHODS Animal preparation These experiments were carried out in accordance with the guidelines of the Canadian Council on Animal Care, and the experimental procedures were approved by the Queen’s University Animal Care Committee. Twenty-eight female Sprague-Dawley rats were used, which were separated into five groups to study the effects of age and partial nerve injury on behavior and on sympathetic sprouting in the DRG. The groups differed in their starting age or whether or not they were injured: Group 3/0, for example, consisted of 3-month-old uninjured rats, whereas group 16/1 consisted of 16-month-old rats that had a partial nerve injury for 1 month. It should be noted that the large size difference between young and old rats precluded a blinded study. To determine the effect of age alone on pain-related behavior and sympathetic sprouting in the DRG, six 3-month-old rats (group 3/0), five 10-month-old rats (group 10/0), and six 16-month-old rats (group 16/0) were left unoperated. To determine what effect age had on chronic constriction injury-induced behavior and sympathetic sprouting, six 3-month-old rats (group 3/1) and five 16- 39 month old rats (group 16/1) were given a CCI, in accordance with the work of Bennett and Xie (1988; see below), and these rats survived for 1 month. The rats were housed in pairs on a 12-hour dark/12-hour light cycle and had access to food and water ad libitum. For the partial nerve injuries, rats were anesthetized with sodium pentobarbital (48 mg/kg body weight, i.p.), an incision was made in the upper left thigh, and the nerve injuries occurred where the sciatic nerve passes over the fused gemelli and obturator tendons. The CCI consisted of tying four loose 4-0 nonabsorbable silk ligatures around the sciatic nerve approximately 1 mm apart. The ligatures were slowly tightened until the first visible twitch in the leg or foot was observed. The incision was closed in layers with sutures, and the skin was closed with clips. Rats survived for 1 month before killing. Tissue preparation and histofluorescence The rats were deeply anesthetized with sodium pentobarbital and transcardially perfused with ice-cold phosphatebuffered saline, pH 7.5. The injured (left) and uninjured (right) L4 and L5 DRG were removed and immediately frozen on the same block. Sixteen-micrometer-thick frozen sections were cut, thaw-mounted, and dipped five times (1 second/dip) in sucrose-potassium phosphate-glyoxylic acid (SPG) solution (10.2 g sucrose, 4.8 g potassium phosphate monobasic, 1.5 g glyoxylic acid monohydrate per 150 ml H2O, pH 7.5). The slides were dried under cool air blowers, baked at 95°C for 2.5 minutes, and coverslipped with mineral oil. Sympathetic fibers were visualized under incident ultraviolet light with a 395–446 nm exciter filter and a .470 nm barrier filter. For each animal, four to eight randomly selected whole DRG sections were inspected for sympathetic fibers: In uninjured animals, L4 and L5 DRG sections were sampled; and, in nerve-injured animals, only L5 DRG sections were sampled. Images of the entire DRG sections were captured at 803 by using a digital-imaging camera (World Precision Instruments, Inc., Sarasota, FL), and all fields in the cell layers (that is, areas occupied mainly by neuronal profiles and not fiber bundles) of the DRG sections containing fluorescent fibers were captured at 3203. The total length of sympathetic axons in the high-power images was measured by manually tracing the on-screen image with a computer mouse, using an imageanalysis software package (Mocha; Jandel Inc., San Rafael, CA). It should be noted that this method consistently underestimated the total sympathetic axon length, because invading sympathetic fibers often ran in bundles of several axons. Care was taken not to include sympathetic fibers obviously surrounding blood vessels. These were rare in the cell layer in adult animals and more commonly occurred between fiber bundles or in the DRG capsule. Sympathetic sprout density was calculated by dividing the total sympathetic axon length in the DRG section by the area occupied by DRG neurons in that section. The sprout density was averaged for each animal, and uninjured groups (3/0, 10/0, and 16/0) were compared for significant differences by using a one-way analysis of variance (ANOVA) followed by Bonferroni’s t-test for pairwise differences. Significant differences in sprout density in nerveinjured rats were assessed by using paired t-tests to investigate left-right differences and by using a one-way ANOVA to investigate differences from rats killed preoperatively (again, followed by Bonferroni’s t-test). 40 M.S. RAMER AND M.A. BISBY Behavioral tests Three behavioral tests were carried out on groups 3/0, 16/0, 3/1, and 16/1. The first was an assessment of thermal nociceptive threshold or thermal allodynia (allodynia is defined as pain due to a stimulus that is not normally painful): Rats were placed on a glass floor, and latency to withdrawal from a radiant heat source applied to the sciatic territory of the hindpaws was measured (Hargreaves et al., 1988). This is a well-characterized method of detecting changes in sensitivity of hindpaw skin in which decreased time to withdrawal reflects decreased thermal nociceptive thresholds. We calibrated the intensity of the radiant heat source in a preliminary study, so that the average withdrawal latency for young uninjured rats was 10–12 seconds. Every rat was tested five times alternately on each hindpaw, allowing at least 2 minutes between tests, and withdrawal latency measurements were then averaged. Left and right withdrawal latencies in uninjured groups (3/0 and 16/0) were averaged, and the averaged values were compared by using a t-test. Paired t-tests were used to compare 1) postoperative (groups 3/1 and 16/1) ipsilateral to contralateral withdrawal latencies, and 2) ipsilateral and contralateral postoperative values to the corresponding preoperative values. We also measured the duration of the response to a painful thermal stimulus or thermal hyperalgesia. Hyperalgesia is defined as increased pain due to a normally noxious stimulus. Increased withdrawal durations are presumed to reflect longer lasting pain sensation following the thermal stimulus, and these measurements are useful, in that they provide information on a different aspect of thermal pain (duration) than measurements of withdrawal latency (threshold). The same paradigm was used as that used in the threshold test, except that the duration of withdrawal from a more intense stimulus (double the intensity of the thermal allodynia calibration) was measured, and each rat was tested three times per hindpaw with at least 15 minutes between tests. It should be noted that the duration of the stimulus required to elicit hindpaw withdrawals at this increased intensity (approximately 4 seconds) was not significantly different between any groups. This is important, because a comparison of duration of response between groups requires that the same suprathreshold stimulus be delivered. A flick of the hindpaw (i.e., without a sustained elevation) was assigned a duration of 0.5 seconds. The same statistical analysis that was used for thermal allodynia was also used here. A third test assessed the rats’ mechanical nocifensive (defensive reaction to painful stimulation) response frequency. An increase in response frequency to repetitive mechanical stimulation can be interpreted as a decrease in the threshold of mechanically induced pain (mechanoallodynia): The more often a defined mechanical stimulus elicits a nocifensive withdrawal response, the more painful that stimulus is for the animal. To carry out the test, rats were placed in a cage with a wire mesh floor, and a 45-g von Frey hair was applied to the sciatic territory of alternate hindpaws (ten times per hindpaw at a frequency of 1 Hz). If the animal lifted its paw in response to the von Frey hair, then the testing was suspended until the paw was placed back on the cage floor. The number of sustained responses (defined as maintained elevation for .2 seconds, or licking the hindpaw, or vocalization) per ten trials was recorded. We chose to quantify sustained responses, because these served as better indicators of nociception than brisk paw flicks, which may have reflected either pain or a nonnocifensive response to mechanical stimulation (e.g., tickle). The same tests described above were used to compare groups for significant differences. All statistical measures are presented as means 6 S.E.M. Photomicrographs Fluorescent sections were photographed and processed as 35-mm color slides (Kodak EL 135-24 film; Eastman Kodak, Rochester, NY). These were scanned (Agfa Arcus II scanner; Agfa-Gevaert N.V., Mortsel, Belgium), and the color of the digital images was corrected to match that of the scanned slide by using a Macintosh computer (Apple Computers, Cupertino, CA) and Adobe Illustrator software (Adobe Systems, Mountain View, CA). The images were not manipulated in any other way. Composites were assembled and printed either in color (Fig. 1) or in black and white (Fig. 2; Kodak xls 8600 PS Printer; Eastman Kodak). RESULTS Histofluorescence After killing the rats, the sciatic nerves were exposed to check the condition of the CCI. All ligatures were still in place, surrounded by fibrous connective tissue. There were no obvious differences in the appearance of the injured nerves between group 3/1 and group 16/1. The left and right L4 and L5 DRG from all rats were removed and processed for glyoxylic acid-induced fluorescence of monoamines. This technique is particularly useful, because orange autofluorescent, lipofuscin-containing DRG neurons; pale green, fluorescent myelin, and bright bluegreen, varicose sympathetic axons can all be easily distinguished. Sympathetic sprouting in uninjured and nerve-injured rats In 3-month-old uninjured rats, there were rarely any fluorescent monoamine-positive (sympathetic) fibers within the cell layer of the DRG. Rather, they were located mainly between fiber layers and around the perimeter of the DRG, almost always associated with blood vessels (Fig. 1A, arrowhead). When, on occasion, sympathetic fibers were found in the cell body layer, they took on a ‘railroad-track’ appearance or formed a ring around an empty space (Fig. 1C,E, arrowheads), both of which are characteristic of perivascular fibers when the blood vessel is cut in longitudinal or transverse section, respectively. In group 10/0 (n 5 5) and group 16/0 animals (n 5 6), lipofuscin granules were more numerous (Fig. 1), and the number of varicose sympathetic fibers in the cell body layer of the DRG increased compared to group 3/0 (n 5 6) rats (compare Fig. 1A with Fig. 1B). The increase in fluorescent fiber density was significantly different from group 3/0 animals for both group 10/0 and group 16/0 rats, and the sprout density was greater in the group 16/0 animals compared with the group 10/0 animals (Fig. 3). The fluorescent fibers in both injured and aged uninjured animals often formed basket-like structures around largediameter DRG neurons (Figs. 1C,D, 2C, arrows), as described previously (Chung et al., 1993, 1996; McLachlan et al., 1993). The distribution of sympathetic axons in the ganglia was not even: rather, there were often clusters of ADRENERGIC SPROUTING IN DRG IN AGING 41 Fig. 1. Glyoxylic acid-induced fluorescence of sympathetic axons in dorsal root ganglia (DRG) of 3-month-old and 16-month-old uninjured rats (A,B) and 1 month following partial sciatic nerve injury in 3-month-old (C) and 16-month-old (D) rats. Large-diameter DRG neurons surrounded by sympathetic axons at high power are also shown from 3-month-old (E) and 16-month-old (F) nerve-injured animals. Note the increased density of lipofuscin granules in the DRG neurons of the older animals (B,D,F). Perivascular sympathetic axons are rare in the cell body layer of 3-month-old DRG (A) but can be found often in the fiber layer (A, arrow). Blood vessel-associated sympathetic axons are indicated by thick arrows in A, C, and E. In aging animals, fluorescent fibers can be seen invading from the DRG perimeter (B, arrow). DRG neurons surrounded by sympathetic axons are indicated by thin arrows in C and D. Sympathetic sprout density increases with age. Scale bars 5 100 µm in A–D, 20 µm in E,F. large-diameter DRG neurons with sympathetic baskets near blood vessels or (more frequently) near the perimeter of ganglion sections, from which the sympathetic axons appeared to have grown (see, e.g., Fig. 1B). Many fluorescent fibers could be seen that extended from the vasculature or the DRG capsule deep into the parenchyma of the ganglion. Often, fibers invaded the DRG in association with myelinated axons, as shown in Figure 2A,B (from a group 10/0 rat and a group 16/1 rat, respectively). At the light microscope level, these abnormal arborizations, when basketing DRG neurons, often appeared to be associated more intimately with the very pale fluorescent glia sur- 42 M.S. RAMER AND M.A. BISBY Fig. 3. Quantitation of sympathetic sprouting in dorsal root ganglia (DRG) of the six groups of animals (mean and S.E.M.). Vertical axis indicates the sprout density expressed as length of fluorescent fibers per unit area of DRG occupied by cell bodies (for details, see Materials and Methods). On the left (unoperated groups), o indicates a statistically significant difference from all younger groups (P , 0.05; one-way analysis of variance; ANOVA). On the right (nerve-injured groups), an asterisk indicates significant differences from the corresponding unoperated groups (t-test), a plus sign indicates significant left-right differences (P , 0.05; paired t-test), and a cross indicates significant difference between nerve-injured young and aging animals (t-test). CCI, chronic constriction injury. Postoperatively, the sympathetic fiber density in the ipsilateral and contralateral DRG was compared with age-matched control values for both sets of animals (young and old; Fig. 3). In both age groups, a 1-month CCI induced significant sympathetic sprouting into ipsilateral and, to a lesser extent, contralateral ganglia (P , 0.001 for all groups relative to age-matched control sprout densities; one-way ANOVA). Sprout density was significantly increased in both the ipsilateral and contralateral ganglia of group 16/1 rats compared with the respective ganglia in group 3/1 rats (same test). In both young and old nerveinjured ganglia, many basketed DRG neurons could be found, and, as with uninjured animals, these occurred more frequently in aged animals (compare Fig. 1C with Fig. 1D). Behavior Fig. 2. Sympathetic axons associated with myelinated sensory axons in dorsal root ganglia (DRG) visualized histochemically with glyoxylic acid-induced fluorescence. At low power (A; from a 10-monthold, uninjured rat) and at higher power (B; from a 16-month-old, nerve-injured rat), sympathetic axons run along the outer surface of myelinated axons (node of Ranvier indicated by arrow in A). When sympathetic baskets form around large-diameter DRG neurons, these are often intimately associated with weakly autofluorescent satellite glia (arrows in C; from a 16-month-old, nerve-injured rat). Scale bars 5 100 µm in A, 20 µm in B,C. rounding the DRG neurons than with the neurons themselves (Fig. 2C). This was true both pre- and postoperatively. All nerve-injured rats were of normal weight for age, and their coats were well groomed. There were very slight motor deficits in all of the nerve-injured rats when they were observed at the end of the behavioral testing periods, including a greater extension of the injured hind limb when the rats were lifted by the tail, and the toes of the ipsilateral hindpaw were slightly flexed compared with the contralateral side. None of the nerve-injured rats showed the hind limb paralysis that is typical of total axotomy, indicating that the CCI-induced injury was indeed incomplete. Three of the six aging, nerve-injured rats self mutilated the injured hind limb. In one, this was so severe that it was killed early and was not included in the analysis. The results of the behavioral tests are shown in Figure 4. Preoperatively, both young (group 3/0; n 5 12) and aging (group 16/0; n 5 12) rats had a withdrawal latency to radiant heat of about 11–12 seconds (Fig. 4A). One month ADRENERGIC SPROUTING IN DRG IN AGING 43 following CCI, young rats (group 3/1; n 5 6) showed a marked decrease in withdrawal latency ipsilateral to the injury compared with preoperative values (P , 0.05; paired t-test), but not contralateral to the injury. There was a significant left-right difference as well (P , 0.05; paired t-test). Aging (group 16/1; n 5 5) rats showed a slight (but not significant) reduction in withdrawal latency ipsilateral to the injury 1 month following CCI. It should be noted that, although the data presented are those for rats at 1 month postoperative, intermittent testing at shorter postoperative intervals showed that aging rats never exhibited hyperalgesia within that time. Once rats responded to the radiant heat by withdrawal, aging, uninjured rats consistently held their hindpaws away from the floor longer than younger uninjured rats (Fig. 4B; 4.1 6 0.7 seconds for group 16/0, left side; 1.3 6 0.3 seconds for group 3/0, left side; P , 0.05; t-test). Postoperatively, the ipsilateral withdrawal duration was increased significantly in young rats (4.5 6 0.9 seconds), but it increased much more dramatically in older rats (17 6 4 seconds). We saw no contralateral increases in response duration. The frequency of withdrawal from a 45-g von Frey hair depended on age (Fig. 4C), with older, uninjured rats withdrawing on average more than younger uninjured rats (P , 0.05; t-test). Postoperatively, withdrawal frequency was elevated ipsilaterally in both young and old rats (P , 0.05; paired t-test). In old rats, withdrawal frequency (both ipsilaterally and contralaterally) was increased compared with control values (paired t-tests) and also compared with younger rats with the same injury (P , 0.05; ANOVA). DISCUSSION Fig. 4. A–C: Behavioral assessment of uninjured and nerveinjured (chronic constriction injury; CCI) young (3 months old; circles) and aging (16 months old; squares) rats. All points represent the mean and S.E.M. (note that some error bars are hidden behind the symbols). Postoperative measurements were taken 1 month following CCI to the left sciatic nerve. An asterisk indicates significant difference from preoperative values (paired t-test), a cross indicates significant difference between left (solid symbols) and right (open symbols) sides (paired t-test), and an x indicates significant difference between young (group 3/0 or 3/1) and aging (group 16/0 or 16/1) animals (t-test). In this study we have shown that 1) sympathetic innervation of rat DRG increases normally with age; 2) innervation density in response to CCI is also significantly greater in aged rats compared with younger rats; 3) aged, uninjured animals exhibit subtle but statistically significant increases in nocifensive behavior in response to some thermal and mechanical stimulation; and 4) aged, nerveinjured animals show increased duration of response to thermal stimulation and an increased frequency of withdrawal from nociceptive mechanical stimulation in response to a CCI. However, aged animals do not show a significantly decreased thermal nociceptive threshold following injury. An important technical consideration in the present experiment was whether or not the extent of the lesion was similar in the young and old rats. In rats with polyethylene cuffs of varying sizes placed around the sciatic nerve, the resulting behavior and nerve damage depended on the cuff diameter (Mosconi and Kruger, 1996). Similarly, the extent of sympathetic sprouting in DRG varies with the extent of sciatic nerve injury: In the first 2 weeks following CCI, sprout density is increased compared with rats that have completely transected sciatic nerves (Ramer and Bisby, 1997). The CCI is a relatively subjective lesion and varies between laboratories. Although a detailed histological analysis of the injured nerves was not carried out, we feel confident that in the present experiment, the extent of the lesion in young and old animals was similar for a number of reasons. First, the criterion for determining the tightness of the ligatures was met for all animals (the ligatures 44 were tightened slowly until a leg twitch was observed), and all lesions were performed by the same individual (M.S.R.). In this way, the induction of the lesion could be controlled. Second, the extent of the motor deficits resulting from the injury was identical for both groups. Third, the gross appearance of the injured nerve at the time of killing was similar. Finally, the persistence of an afferent connection to the sciatic territory (apparent during both early intermittent tests as well as at the end of the experiment) showed that the injury did not result in complete axotomy for either group. These factors are strongly suggestive that the injuries are indeed comparable. Sympathetic sprouting and neuropathic pain Although sympathetic-sensory coupling has been documented in human neuromas following nerve injury (Chabal et al., 1992), there are now several reports that show a connection between sympathetic invasion of the DRG and behavior associated with neuropathic pain following nerve injury in rats (Chung et al., 1993, 1996; Davis et al., 1994; Kim et al., 1996). Whether this association can be linked to the genesis and/or the maintenance of painful behavior is currently a matter of controversy, but previous work from our laboratory shows good temporal correlation between CCI-induced allodynia and hyperalgesia and sympathetic sprouting that accompanies the injury (Ramer and Bisby, 1997). In addition, the more rapid sprouting and sympathetic basket formation that occurs following a spinal nerve ligation (Chung et al., 1993, 1996; Kim et al., 1996) is accompanied by equally rapid establishment of mechanical and cold allodynia in that model of neuropathic pain compared with more distal nerve injuries (Kim et al., 1996). The present study clearly shows that there is a relationship between the extent of sympathetic innervation of the DRG and some types of abnormal painful behavior: Specifically, there seems to be a tight relationship between thermal hyperalgesia and mechanoallodynia (which develop in both young and old animals) and sympathetic sprouting, whereas thermal allodynia and sympathetic sprouting do not always coexist (for example, in old, nerve-injured animals). These findings have obvious implications in terms of the dependence of neuropathic pain on the sympathetic nervous system. However, different models of neuropathic pain resulting from partial nerve injuries are relieved to varying extents on surgical or chemical sympathectomy (Neil et al., 1991; Seltzer and Shir, 1991; Kim et al., 1993; Kinnman and Levine, 1995; Chung et al., 1996). Results from separate experiments employing distal partial sciatic nerve injuries (Seltzer and Shir, 1991; Wakisaka et al., 1991), proximal injuries (Chung et al., 1996), or both (Kim and Chung, 1992; Kim et al., 1997) indicate that the more proximal the partial nerve injury is to the DRG, the more effective sympathectomy is at relieving the pain. Not surprisingly, then, the dependence of distal CCI-induced pain on the sympathetic nervous system has also been debated (Wakisaka et al., 1991; Perrot et al., 1993; Desmeules et al., 1995). This inconsistency of the effects of sympathectomy after CCI may be due to the absence of a sympathetic component to the pain, but it is more likely due to the involvement of mechanisms in addition to but separate from the sympathetic nervous system in the maintenance of the pain. Not only does the ability of sympathectomy to relieve pain vary with distance of the lesion from the DRG, but so does the degree to which M.S. RAMER AND M.A. BISBY sympathetic axons invade the DRG: Distal lesions have been shown to have less pronounced effects in terms of sympathetic sprouting in the DRG than proximal lesions (Kim et al., 1996). Age effects Previous studies on the effect of age on partial nerve injury-induced neuropathic pain have yielded contradictory results: L5 spinal nerve ligation causes more pronounced symptoms in young rats (Chung et al., 1995), whereas ligation of one of the spinal nerves of the rat tail has a greater behavioral effect in aged rats (Kim et al., 1995). One report on the effect of age on behavior following chronic constriction of the sciatic nerve showed increased thresholds to thermal and mechanical stimulation in 107-day-old rats compared with 54- and 71-day-old rats (Tanck et al., 1992). However, the oldest rats in that study were the same age as the youngest rats in the present study (90–100 days). We report here for the first time the exacerbating effect of advanced age on CCI-induced pain. We also report that thermal hyperalgesia (but not thermal allodynia) and mechanoallodynia develop with age (albeit slightly), even in the absence of injury. This may well be analogous to the finding that older humans show decreased pain sensitivity (that is, no allodynia) to contact thermal stimulation but increased severity of thermal pain (hyperalgesia; Harkins et al., 1986). The failure of the aged rats to exhibit a decreased threshold for thermal stimulation following nerve injury may be related to changes in the thermal properties of the skin or to alterations in sensory fiber function with age, because antidromic electrical stimulation of the sciatic nerve elicits a decreased substance P-mediated, cutaneous, vascular response compared with young rats (Khalil et al., 1994), and substance P levels in DRG are significantly reduced with age (unpublished observation cited in Khalil et al., 1994). The finding that sympathetic fibers invade the DRG with age, even in the absence of injury, raises the question of the stimulus for sprouting following injury or during normal aging. In both cases, sympathetic axons arborize almost exclusively around large-diameter DRG neurons, and, in both cases, sympathetic axons invade the DRG from the vasculature and from the ganglion perimeter (McLachlan et al., 1993; present results), suggesting a common mechanism. What is the sprouting stimulus? The sprouting of sympathetic fibers from perivascular nerves and the DRG capsule may be due to nerve growth factor (NGF). When a peripheral nerve is injured, the distal stump undergoes Wallerian degeneration, which is characterized by Schwann cell proliferation, macrophage invasion, and the production of NGF. NGF is also elevated in denervated target tissues (Mearow et al., 1993), and NGF mRNA is elevated in the DRG following injury (Sebert and Shooter, 1993, Wells et al., 1994). We have previously proposed that a mechanism for sympathetic sprouting in the DRG relies on retrograde transport to the DRG of NGF (or some other product of Wallerian degeneration) produced in the distal stump of a partially injured nerve via axons spared by the injury. Retrograde-transported NGF would then effect perivascular sympathetic axon sprouting (Ramer and Bisby, 1997; Ramer et al., 1997). Others have also suggested an involvement of NGF ADRENERGIC SPROUTING IN DRG IN AGING (Davis et al., 1994; Zhou et al., 1996). The NGF-specific receptor trkA, which is present on sympathetic axons (for review, see Barbacid, 1994), very likely plays a role in the sprouting response of sympathetic fibers. The retrograde transport of NGF in DRG neurons also requires trkA, and, although the receptor is expressed mainly in small primary afferents, it is possible that the sympathetic baskets encapsulate a subpopulation of large-diameter neurons, which also express trkA (Wright and Snider, 1995). Whatever stimulatory agent is involved, it seems that the requirement is some product of the degenerating nerve, because behavioral manifestations of a CCI are suppressed in C57Bl/Wld (Wld) mice, a strain that is deficient in Wallerian degeneration (Myers et al., 1996) and in which sympathetic invasion of the DRG is also delayed (Ramer et al., 1997). The mechanism underlying the increased sympathetic innervation of the DRG in aged rats is unknown. Sympathetic sprouting and pain behavior in the absence of injury may be related to degenerative changes that are known to occur in aging rat DRG, such as sensory neuron involution (for review, see Devor, 1991). Age-related changes in sympathetic innervation seem to be tissue specific: Innervation of the cerebral vasculature decreases, whereas that of the tail artery increases (Andrews and Cowen, 1994). If this is true of the peripheral vasculature generally, then a side effect may be increased sprouting into the DRG. The increased innervation in the aged DRG following nerve injury may again be ascribed to NGF. First, aged sympathetic fibers have a slightly enhanced outgrowth ability when they are provided with a peripheral target (Gavazzi, 1995), and they respond to applied NGF with more widespread growth (Andrews and Cowen, 1994). Second, degenerating nerves of old rats have elevated NGF levels compared with young rats (Date et al., 1994). This may explain why sympathetic innervation of the DRG after CCI is more pronounced in the aged rats, but a more trivial explanation is simply that there are more sympathetic fibers resident within the DRG when the nerve is injured, providing a richer source of sprouts. Contralateral effects The mechanisms for the contralateral effects of CCI are unclear. In the present experiment, we found no significant contralateral changes in thermal nociceptive thresholds, nocifensive response duration, but the withdrawal frequency from mechanical stimulation was elevated slightly but significantly on the contralateral side of aging nerve-injured rats. Kupers et al. (1992) also noted saphenous-mediated bilateral changes in nocifensive behavior following a sciatic nerve injury. In addition, chronic constriction of the rat infraorbital nerves (IoN) results in increased sensitivity in the territory of both the injured and the contralateral IoN (Vos et al., 1994). Human cases have been reported in which allodynia is present bilaterally after a unilateral mechanical insult (Baron and Maier, 1995). Like the work of McLachlan et al. (1993) and our own previous work (Ramer and Bisby, 1997), we found that ganglia opposite the injury had been invaded by sympathetic axons. Other bilateral neuroanatomical changes have been observed in response to unilateral experimental manipulations: Bilateral changes in neuropeptide expression are seen after a unilateral intraplantar injection of NGF (Amann et al., 1996) and after unilateral adjuvantinduced arthritis (Donaldson et al., 1995). In addition, the 45 expression of the low-affinity NGF receptor (p75) increases bilaterally in DRG satellite cells after a sciatic nerve injury (Zhou et al., 1996). There remains no satisfactory explanation for these bilateral changes, but elevated targetderived factors (for example, from the skin; Mearow et al., 1993) or trophic factors from the degenerating peripheral nerve may be acting on DRG neurons and sympathetic axons in the contralateral ganglion via the circulation. Sympathetic-sensory coupling The mechanism by which sympathetic fibers in the DRG might increase pain sensitivity and/or pain severity is also unknown, but the ability of DRG neurons to express adrenergic receptors has been investigated (Gold et al., 1997), and it has been shown that some small-diameter DRG neurons up-regulate a receptors following injury (Sato and Perl, 1991; Nishiyama et al., 1993). Adrenergic a-2 receptors have been shown to be coupled to N-type calcium channels in cultures of previously axotomized rat sensory neurons, and application of norepinephrine had the greatest effects on small cells (possibly nociceptive) and on DRG neurons cultured from autotomizing rats (Abdulla and Smith, 1997). Alternatively, as suggested by Levine and colleagues (1986), increased levels of released norepinephrine may be acting on the sympathetic fibers themselves, causing the release of prostaglandins PGE2 and PGI2 in the vicinity of DRG axons (or cell bodies), resulting in increased afferent sensitivity. At any rate, there is an accumulating body of evidence that shows increased DRG neuron activity in response to sympathetic stimulation after nerve injury (McLachlan et al., 1993; Devor et al., 1994; Michaelis et al., 1996). The specific submodalities of pain influenced by the sympathetic nervous system remain unknown. The baskets of sympathetic fibers eventually form preferentially around large-diameter (low-threshold mechanosensory or possibly proprioceptive) DRG neurons, although some small neurons are initially surrounded (Chung et al., 1996). Injury induces neurochemical changes in the phenotype of large-diameter DRG neurons: Preprotachykinin (the precursor to the pain-transmitting peptide substance P) mRNA can be found in large DRG neurons following injury, but not in uninjured animals (Noguchi et al., 1994). In addition, large DRG neurons undergo a change in connectivity in the spinal cord following injury, sprouting into the more superficial laminae of the dorsal horn (which normally receives much of the input from small diameter nociceptive afferents; Woolf et al., 1992; Shortland and Woolf, 1993; Doubell et al., 1997). Increased responses of large DRG neurons to normally innocuous mechanical stimuli through sympathetic coupling, combined with phenotype changes and sprouting into the dorsal horn, could underlie some abnormal behaviors that are characteristic of neuropathic pain. CONCLUSIONS In conclusion, we show that sympathetic fibers accumulate in the DRG of uninjured animals as a function of age and that the CCI-induced sympathetic innervation of DRG of older animals is greater than that of younger animals. This work provides further evidence of a link between sympathetic sprouts in the sensory ganglia and behavioral anomalies associated with partial nerve injury. The importance of these phenomena, particularly with respect to 46 M.S. RAMER AND M.A. BISBY aging, is that they may provide insight into neuropathic pain in the elderly, who are especially afflicted by this debilitating clinical condition. ACKNOWLEDGMENTS We thank Serap Erdebil and Kevan McRae for fine technical assistance. M.S.R. receives an MRC studentship. LITERATURE CITED Abdulla, F.A. and P.A. Smith (1997) Ectopic a2-adrenoceptors couple to N-type Ca21 channels in axotomized rat sensory neurons. J. Neurosci. 17:1633–1641. Amann, R., D.J.S. Sirinathsinghji, J. Donnerer, I. Liebmann, and R. Schuligoi (1996) Stimulation by nerve growth factor of neuropeptide synthesis in the adult rat in vivo—Bilateral response to unilateral intraplantar injections. Neurosci. Lett. 203:171–174. Andrews, T.J. and T. Cowen (1994) In vivo infusion of NGF induces the organotypic regrowth of perivascular nerves following their atrophy in aged rats. J. Neurosci. 14:3048–3058. Barbacid, M., (1994) The Trk family of neurotrophin receptors. J. Neurobiol. 25:1386–1403. Baron, R. and C. Maier (1995) Painful neuropathy: C-nociceptor activity may not be necessary to maintain central mechanisms accounting for dynamic mechanical allodynia. Clin. J. Pain 11:63–69. Bennett, G.J. (1991) The role of the sympathetic nervous system in painful peripheral neuropathy. Pain 45:221–223. Bennett, G.J. and Y.K. Xie (1988) A peripheral mononeuropathy in rat that produces disorders of pain sensation like those seen in man. Pain 33:87–107. Chabal, C., L. Jacobson, L.C. Russel, and K.J. Burchiel (1992) Pain response to perineuromal injection of normal saline, epinephrine, and lidocaine in humans. Pain 49:9–12. Chung, J.M., Y. Choi, Y.W. Yoon, and H.S. Na (1995) Effects of age on behavioral signs of neuropathic pain in an experimental rat model. Neurosci. Lett. 183:54–57. Chung, J.S., B.H. Lee, Y.W. Yoon, and J.M. Chung (1996) Sympathetic sprouting in the dorsal root ganglia of the injured peripheral nerve in a rat neuropathic pain model. J. Comp. Neurol. 376:241–252. Chung, K., H.J. Kim, H.S. Na, M.J. Park, and J.M. Chung (1993) Abnormalities of sympathetic innervation in the area of an injured peripheral nerve in a rat model of neuropathic pain. Neurosci. Lett. 162:85–88. Chung, K., Y.W. Yoon, and J.M. Chung (1997) Sprouting sympathetic fibers form synaptic varicosities in the dorsal root ganglion of the rat with neuropathic injury. Brain Res. 751:275–280. Date, I., S. Furukawa, and T. Ohmoto (1994) Increased nerve growth factor level in the distal stump of transected sciatic nerve in relation to aging and its application for neural grafting. Exp. Neurol. 130:168–171. Davis, B.M., K.M. Albers, K.B. Seroogy, and D.M. Katz (1994) Overexpression of nerve growth factor in transgenic mice induces novel sympathetic projections to primary sensory neurons. J. Comp. Neurol. 349:464– 474. Desmeules, J.A., V. Kayser, J. Weil-Fuggaza, A. Bertrand, and G. Guilbaud (1995) Influence of the sympathetic nervous system in the development of abnormal pain-related behaviours in a rat model of neuropathic pain. Neuroscience 67:941–951. Devor, M. (1991) Chronic pain in the aged: Possible relation between neurogenesis, involution and pathophysiology in adult sensory ganglia. J. Basic Clin. Physiol. Pharmacol. 2:1–15. Devor, M., W. Janig, and M. Michaelis (1994) Modulation of activity in dorsal root ganglion neurons by sympathetic activation in nerve-injured rats. J. Neurophysiol. 71:38–47. Devor, M., V. Shinder, and R. Govrin-Lippmann (1995) Sympathetic sprouting in axotomized rat DRG: Ultrastructure. Soc. Neurosci. Abstr. 21:894 Donaldson, L.F., D.S. McQueen, and J.R. Seckl (1995) Neuropeptide gene expression and capsaicin-sensitive primary afferents: Maintenance and spread of adjuvant arthritis in the rat. J. Physiol. 486:473–482. Doubell, T.P., R.J. Mannion, and C.J. Woolf (1997) Intact sciatic myelinated primary afferent terminals collaterally sprout in the adult rat dorsal horn following section of a neighboring peripheral nerve. J. Comp. Neurol. 380:95–104. Gagliese, L. and R. Melzack (1997) Chronic pain in elderly people. Pain. 70:3–14. Gavazzi, I. (1995) Collateral sprouting and responsiveness to nerve growth factor of ageing neurons. Neurosci. Lett. 189:47–50. Gibson, S.J., B. Katz, T.M. Corran, M.J. Farrell, and R.D. Helme (1994) Pain in older persons. Disab. Rehabil. 16:127–139. Gold, M.S., S. Dastmalchi, and J.D. Levine (1997) a2-Adrenergic subtypes in rat dorsal root and superior cervical ganglion neurons. Pain. 69:179– 190. Hargreaves, K., R. Dubner, F. Brown, C. Flores, and J. Joris (1988) A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia. Pain 32:77–88. Harkins, S.W., D.D. Price, and M. Martelli (1986) Effects of age on pain perception: Thermonociception. J. Gerontol. 41:58–63. Janı̈g, W. (1985) Systemic and specific autonomic reactions in pain: Efferent, afferent and endocrine components. Eur. J. Anaesthesiol. 2:319–346. Khalil, Z., V. Ralevic, M. Bassirat, G.J. Dusting, and R.D. Helme (1994) Effects of ageing on sensory nerve function in rat skin. Brain Res. 641:265–272. Kim, H.J., H.S. Na, H.J. Nam, K.A. Park, S.K. Hong, and B.S. Kang (1996) Sprouting of sympathetic nerve fibers into the dorsal root ganglion following peripheral nerve injury depends on the injury site. Neurosci. Lett. 212:191–194. Kim, K.J., Y.W. Yoon, and J.M. Chung (1997) Comparison of three rodent neuropathic pain models. Exp. Brain Res. 113:200–206. Kim, S.H. and J.M. Chung (1991) Sympathectomy alleviates mechanical allodynia in an experimental animal model for neuropathy in the rat. Neurosci. Lett. 134:131–134. Kim, S.H. and J.M. Chung (1992) An experimental model for peripheral neuropathy produced by segmental spinal nerve ligation in the rat. Pain 50:355–363. Kim, S.H., H.S. Na, K. Sheen, and J.M. Chung (1993) Effects of sympathectomy on a rat model of peripheral neuropathy. Pain 55:85–92. Kim, Y.I., H.S. Na, Y.W. Yoon, S.H. Nahm, K.H. Ko, and S.K. Hong (1995) Mechanical allodynia is more strongly manifested in older rats in an experimental model of peripheral neuropathy. Neurosci. Lett. 199:158– 160. Kinnman, E. and J.D. Levine (1995) Involvement of the sympathetic postganglionic neuron in capsaicin-induced secondary hyperalgesia in the rat. Neuroscience 65:283–291. Kupers, R.C., D. Nuytten, M. De Castro-Costa, and J.M. Gybels (1992) A time course analysis of the changes in spontaneous and evoked behaviour in a rat model of neuropathic pain. Pain 50:101–111. Levine, J.D., Y.O. Taiwo, S.D. Collins, and J.K. Tam (1986) Noradrenaline hyperalgesia is mediated through interaction with sympathetic postganglionic neurone terminals rather than activation of primary afferent nociceptors. Nature 323:158–160. McLachlan, E.M., W. Janig, M. Devor, and M. Michaelis (1993) Peripheralnerve injury triggers moradrenergic sprouting within dorsal-root ganglia. Nature 363:543–546. Mearow, K.M., Y. Kril, and J. Diamond (1993) Increased NGF mRNA expression in denervated rat skin. Neuroreport 4:351–354. Michaelis, M., M. Devor, and W. Janı̈g (1996) Sympathetic modulation of activity in rat dorsal root ganglion neurons changes over time following peripheral nerve injury. J. Neurophysiol. 76:753–763. Mosconi, T. and L. Kruger (1996) Fixed-diameter polyethylene cuffs applied to the rat sciatic nerve induce a painful neuropathy: Ultrastructural morphometric analysis of axonal alterations. Pain 64:37–57. Myers, R.R., H.M. Heckman, and M. Rodriguez (1996) Reduced hyperalgesia in nerve-injured wld mice: Relationship to nerve fiber phagocytosis, axonal degeneration, and regeneration in normal mice. Exp. Neurol. 141:94–101. Neil, A., N. Attal, and G. Guilbaud (1991) Effects of guanethidine on sensitization to natural stimuli and self-mutilating behaviour in rats with a peripheral neuropathy. Brain Res. 565:237–246. Nishiyama, K., B.W. Brighton, D.F. Bossut, and E.R. Perl (1993) Peripheral nerve injury enhances alpha-2-adrenergic receptor expression by some DRG neurons. Soc. Neurosci. Abstr. 19:499. Noguchi, K., R. Dubner, M. DeLeon, E. Senba, and M.A. Ruda (1994) Axotomy induces preprotachykinin gene expression in a subpopulation of dorsal root ganglion neurons. J. Neurosci. Res. 37:596–603. Perrot, S., N. Attal, D. Ardid, and G. Guilbaud (1993) Are mechanical and cold allodynia in mononeuropathic and arthritic rats relieved by systemic treatment with calcitonin or guanethidine? Pain 52:41–47. ADRENERGIC SPROUTING IN DRG IN AGING Ramer, M.S. and M.A. Bisby (1997) Rapid sprouting of sympathetic axons in dorsal root ganglia of rats with a chronic constriction injury. Pain 270:237–244. Ramer, M.S., G.D. French, and M.A. Bisby (1997) Wallerian degeneration is required for both neuropathic pain and sympathetic sprouting into the DRG. Pain 72:71–78. Sato, J. and E.R. Perl (1991) Adrenergic excitation of cutaneous pain receptors induced by peripheral nerve injury. Science 251:1608–1610. Sebert, M.E. and E.M. Shooter (1993) Expression of mRNA for neurotrophic factors and their receptors in the rat dorsal root ganglion and sciatic nerve following nerve injury. J. Neurosci. Res. 36:357–367. Seltzer, Z. and Y. Shir (1991) Sympathetically-maintained causalgiform disorders in a model for neuropathic pain: A review. J. Basic Clin. Physiol. Pharmacol. 2:17–61. Seltzer, Z., R. Dubner, and Y. Shir (1990) Novel behavioral model of neuropathic pain disorders produced in rats by partial sciatic nerve injury. Pain 43:205–218. Shortland, P. and C.J. Woolf (1993) Chronic peripheral nerve section results in a rearrangement of the central arborizations of axotomized A-beta primary afferent neurons in the rat spinal cord. J. Comp. Neurol. 330:65–82. 47 Tanck, E.N., J.S. Kroin, R.J. McCarthy, R.D. Penn, and A.D. Ivankovich (1992) Effects of age and size on development of allodynia in a chronic pain model produced by sciatic nerve ligation in rats. Pain 51:313–316. Vos, B.P., A.M. Strassman, and R.J. Maciewicz (1994) Behavioral evidence of trigeminal neuropathic pain following chronic constriction injury to the rat’s infraorbital nerve. J. Neurosci. 14:2708–2723. Wakisaka, S., K.C. Kajander, and G.J. Bennett (1991) Abnormal skin temperature and abnormal sympathetic vasomotor innervation in an experimental painful peripheral neuropathy. Pain 46:299–313. Wells, M.R., U. Vaidya, and J.P. Schwartz (1994) Bilateral phasic increases in dorsal root ganglia nerve growth factor synthesis after unilateral sciatic nerve crush. Exp. Brain Res. 101:53–58. Woolf, C.J., P. Shortland, and R.J. Coggeshall (1992) Peripheral nerve injury triggers central sprouting of myelinated afferents. Nature. 355:75–77. Wright, D.E. and W.D. Snider (1995) Neurotrophin receptor mRNA expression defines distinct populations of neurons in rat dorsal root ganglia. J. Comp. Neurol. 351:329–328. Zhou, X.F., R.A. Rush, and E.M. McLachlan (1996) Differential expression of the p75 nerve growth factor receptor in glia and neurons of the rat dorsal root ganglia after peripheral nerve transection. J. Neurosci. 16:2901–2911.