HIGH-FREQUENCY ELECTRIC nerve stimulation (HFS)

Transcription

1 1575 ARTICLES The Relationship Between Dorsal Horn Neurotransmitter Content and Allodynia in Neuropathic Rats Treated With High-Frequency Transcutaneous Electric Nerve Stimulation David L. Somers, PhD, F. Richard Clemente, PhD ABSTRACT. Somers DL, Clemente FR. The relationship between dorsal horn neurotransmitter content and allodynia in neuropathic rats treated with high-frequency transcutaneous electric nerve stimulation. Arch Phys Med Rehabil 2003;84: Objective: To examine the relation between axon terminal neurotransmitter content in the dorsal horn and allodynia in neuropathic rats treated with high-frequency transcutaneous electric nerve stimulation (TENS). Design: A completely randomized experimental design. Two groups of rats received a chronic constriction injury to the right sciatic nerve, and 2 groups did not. The rats were either treated or not treated with TENS. Setting: Research laboratory. Animals: Adult male Sprague-Dawley rats ( g). Interventions: TENS was delivered daily for 1 hour to the chronic constriction injury rats or to the uninjured rats through self-adhesive electrodes applied to the skin innervated by the right dorsal rami of lumbar spinal nerves 1 to 6. Main Outcome Measures: Thermal and mechanical pain thresholds were assessed bilaterally in the hind paws of all rats twice before the chronic constriction injury surgery (baseline) and then 12 days after the surgery. An analogous time frame of assessment was used for rats that did not have chronic constriction injury surgery. Thermal and mechanical allodynia were expressed as difference scores between the pain thresholds of the right and left hind paws. These values were normalized to differences that existed between the 2 paws at baseline. The amino acid content of dorsal horn axon terminals was assessed bilaterally with high-pressure liquid chromatography, and values were normalized to wet weight. Results: The mean level of thermal and mechanical allodynia did not differ between the TENS-treated and untreated rats with chronic constriction injury. However, there was a significant relation between the dorsal horn, axon terminal content of glutamate (adjusted R 2.45, P.01) and glycine (adjusted R 2.51, P.005) and the magnitude of mechanical allodynia present in TENS-treated chronic constriction injury rats, but not in any other group. As axon terminal glutamate From the Department of Physical Therapy, John G. Rangos Sr. School of Health Sciences, Duquesne University, Pittsburgh, PA. Presented as a poster at the Society for Neuroscience s 30th Annual Meeting, November 9, 2000, New Orleans, LA. Supported by the National Institutes of Health (grant no. R15 NS ) and the Hunkele Foundation. Empi Epix XL transcutaneous electric nerve stimulators were supplied by Empi Inc. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Reprint requests to David L. Somers, PhD, Dept of Physical Therapy, Duquesne University, 111 Health Sciences Bldg, Pittsburgh, PA somers@duq.edu /03/ $30.00/0 doi: /s (03) and glycine decreased in the right dorsal horn and increased in the left, mechanical allodynia was reduced or absent. When this trend was reversed, mechanical allodynia was more severe. Daily TENS also reduced the mean axon terminal content of aspartate, glutamate, and glycine bilaterally in the chronic constriction injury rats from the level observed in untreated neuropathic rats (P.05). Conclusion: The variability in responsiveness of mechanical allodynia to daily TENS treatment in neuropathic rats is related to the axon terminal content of glutamate and glycine in the dorsal horn. These findings may help explain a similar variability in humans when TENS is used to treat neuropathic pain. Key Words: Amino acid neurotransmitters; Analgesia; Complex regional pain syndrome type II; Pain; Rehabilitation; Transcutaneous electric nerve stimulation by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation HIGH-FREQUENCY ELECTRIC nerve stimulation (HFS) is a therapeutic modality used to treat painful conditions. 1 The treatment is delivered at a frequency of 10 to 100Hz and uses short pulse durations ( s) at an amplitude that is detectable but that does not produce muscle contraction. 2 HFS can be delivered through surface electrodes that are placed on the skin, either in the vicinity of the perceived pain or on skin innervated by the same spinal cord segments as in the painful area. 2 In addition to this transcutaneous application, HFS is also delivered directly to peripheral nerves, 3-6 the dorsal roots of peripheral nerves, 3,7 or the dorsal aspect of the spinal cord Although the efficacy of HFS in relieving some types of pain is questionable, 11 the treatment can relieve pain that is associated with nerve injury. In patients with chronic pain, it relieves neuropathic pain more than it relieves pain of nonneuronal origin. 5,12,13 However, even among those with neurogenic pain, HFS is not always helpful, and the response to the treatment is often quite varied. For example, application of HFS transcutaneously (TENS) to people with painful diabetic neuropathy did not reduce pain in 3 of 15 treated subjects. 14 Moreover, the subjects who experienced pain relief were variable in their responses to the treatment. Four subjects improved by a single pain grade, 8 improved by 2 pain grades, and 3 improved by 3 pain grades under author-defined criteria for pain assessment. 14 Similar results were reported when TENS was used to treat chronic pain produced by isolated nerve injury. Application of TENS to skin, either over the painful area or proximal to the painful area, produced considerable pain relief in 53% to 62% of patients treated but gave only moderate, equivocal, or no relief in 38% to 47% of treated subjects. 15,16 The variable response of neuropathic pain to HFS delivered transcutaneously is also reported when the modality is delivered to peripheral nerves or to the spinal cord. When HFS was delivered to a relevant peripheral nerve in people with complex

2 1576 MECHANISMS FOR VARIABLE RESPONSE TO TENS, Somers regional pain syndrome type II, it produced good to fair pain reduction in 63% of the subjects but poor or no pain reduction in 27%. 17 Likewise, when a trial HFS treatment to the dorsal aspect of the spinal cord (DSCS) was tested on 30 patients with peripheral neuropathy, 19 responded favorably enough to warrant implantation of the electrodes, whereas 11 subjects did not respond. Of the 19 who received DSCS, 6 reported excellent pain relief, 8 had good pain relief, and 6 reported poor pain relief. 18 These findings suggest that, regardless of the mode of application, the responsiveness of neuropathic pain to HFS treatment is characteristically variable. The reason for this variability is unknown, but recent investigations using animal models of neuropathic pain have found the same variability in rats with painful neuropathy. For example, rats treated daily with TENS, beginning immediately after a chronic constriction injury to the sciatic nerve, did not as a group develop thermal allodynia. 19 However, there was considerable variability in this TENStreated group; some rats developed thermal allodynia, whereas other rats did not. 19 There was also variability in the response to HFS in rats with chronic constriction injury that received DSCS. Stimulation of these rats transiently relieved mechanical allodynia, 20 but this relief was not consistent across all the rats. In rats that developed mechanical allodynia after chronic constriction injury, only 10% had reduced allodynia after DSCS. 21 Although many factors may be related to the variable effectiveness of HFS in the treatment of neuropathic pain, 22 in rats this variability appears to be related to neurotransmission in the spinal cord. Rats with mechanical allodynia after partial ligation to the sciatic nerve had either reduced or no change in allodynia when they received DSCS. 21 In rats in which DSCSalleviated allodynia, the treatment reduced the dorsal horn extracellular level of glutamate and aspartate and increased the extracellular level of -aminobutyric acid (GABA). 23 In rats in which DSCS failed to reduce allodynia, the extracellular content of these same neurotransmitters in the dorsal horn was unchanged. It is unknown whether this relation between allodynia and dorsal horn neurotransmitter content exists when HFS is delivered transcutaneously. The primary purpose of our study was to determine whether the variability in allodynia s response to TENS in neuropathic rats is related to neurotransmitter content in the dorsal horn. If this relation is present, we also wanted to determine what the strength of it is between different dorsal horn neurotransmitters and the level of allodynia in TENS-treated rats with chronic constriction injury. For comparison, the relation between allodynia and neurotransmitter content was examined in untreated chronic constriction injury and TENS-treated control rats with no chronic constriction injury. METHODS Production of Chronic Constriction Injury This study was approved by the Institutional Animal Care and Use Committee of Duquesne University. Adult male Sprague-Dawley rats ( g) were deeply anesthetized with sodium pentobarbital (50mg/kg) by intraperitoneal injection. Four chromic gut sutures were applied to the right sciatic nerve according to the procedure of Bennett and Xie. 24 At the conclusion of an experiment, the sciatic nerve was reexposed, and the integrity of the sutures was confirmed. Behavioral Testing The mechanical pain threshold was assessed with calibrated Semmes-Weinstein monofilaments. 25,26 The filaments (.41, 1.2, 3.63, 8.51, 15.13g a ) were used to give graded pressure to the plantar surface of the hind paws. Care was taken to keep the stimulations in the midregion of the foot between the toes and the heel, because this area is innervated exclusively by the sciatic nerve. 27 Rats were placed on a metal grid and covered with a Plexiglas cover. The lowest-caliber filament (.41g) was pushed onto the plantar surface of the right hind paw 5 times until the filament bent; 2 to 3 seconds separated each push. Five minutes later, the left paw of the same rat was similarly assessed. Not less than 5 minutes after the left paw was assessed, the entire procedure was repeated. In this way, the right and left paws were pushed by the.41-g filament 10 times, and the number of withdrawals was recorded. In ascending order, each filament was so tested until the rat withdrew from all 10 pushes of a single-sized filament or until the largestcaliber filament was tested. To estimate the force from which the rat withdrew 50% of the time, force (in grams) was regressed on withdrawal frequency for all 5 filaments by using simple linear regression. The force at which the rat withdrew a paw on 5 of 10 pushes (50% withdrawal threshold) was predicted by using the regression equation. If the 50% withdrawal threshold so calculated was greater than 15.13g, the highest-caliber filament used (15.13g) was recorded as the gram force. If the rat failed to respond to any of the filaments, 15.13g was recorded as the 50% withdrawal threshold. We have found the intratester reliability of this method to be quite good. 19 The same investigator always performed this test. Thermal pain threshold was assessed by withdrawal latency from a radiant heat source (University Anesthesia Research Group, University of California, San Diego), as originally described by Hargreaves et al. 28 Radiant heat was applied to both paws 5 times, with 5 minutes separating each irradiation. The latency to withdraw (in hundredths of a second) was recorded for each irradiation. Twenty seconds was the maximum irradiation time. The 5 latency measurements obtained for each paw were averaged to obtain a mean latency for the right and left paws. Again, we have found the intratester reliability of this method to be quite good, 19 and, again, the same investigator always performed this test. Electric Stimulation TENS was applied to rats through self-adhesive surface electrodes (Empi SoftTouch 9000 b ) with an Empi Epix XL transcutaneous electric nerve stimulator. b The frequency of stimulation was 100Hz, and the intensity was 80% of that needed to produce a visible muscle contraction ( 30 to 40 A). The surface electrode placement was on the denuded, and presumably uninvolved, skin overlying the right paraspinal musculature. 19 The stimulated skin is innervated by the dorsal rami of lumbar spinal cord segments 1 to This placement spans the dermatomes of spinal cord segments that also innervate the painful right paw. 27,29 TENS commenced on the day of surgery while the rats were still under the influence of sodium pentobarbital. On subsequent days, TENS was administered while the rats were lightly anesthetized with halothane (4%, maintained at 0.2% 0.5%). Determination of Dorsal Horn Amino Acid Content To assess dorsal horn axon terminal (synaptosomal) levels of amino acids, rats were anesthetized with sodium pentobarbital. The L1 to L6 segments of the spinal cord were exposed and frozen in situ with powdered dry ice. Rats were then killed by cardiac puncture, and the frozen lumbar segments of the spinal cord were removed in block and stored at 70 C. The L4 and L5 segments of the spinal cord were isolated, cut into 300- m-

3 MECHANISMS FOR VARIABLE RESPONSE TO TENS, Somers 1577 thick sections, thaw-mounted onto glass slides, and quickly refrozen on dry ice. The right and left dorsal horn was then removed with a frozen micropunch c (diameter,.76mm). Punched tissue was collected in frozen microcentrifuge tubes, suspended in ice-cold.32mol/l sucrose, and homogenized. Synaptosomes were isolated with differential centrifugation at 2 C, 30 and amino acids were labeled with the AccQ-Tag fluorescent labeling system. d Labeled aspartate, glutamate, glycine, and GABA were quantified independently from right and left dorsal horns by using high-pressure liquid chromatography with fluorescent detection and were expressed as picomoles per milligram wet weight. Experimental Design We used 4 groups of neuropathic rats to determine whether the variability in responsiveness of allodynia to TENS is related to axon terminal amino acid content in the dorsal horn. Two of the groups received the chronic constriction injury procedure to the right sciatic nerve. The first group of ligated rats (TENS-treated chronic constriction injury; n 15) received TENS for 90 minutes beginning immediately after the chronic constriction injury surgery and then daily (60min) for the next 11 days. The second group of ligated rats was treated the same as the first, including halothane administration, except that no TENS was administered (chronic constriction injury; n 16). The third group of rats received only TENS, with no ligation to the sciatic nerve, using the same procedure described for the TENS-treated chronic constriction injury group (TENS controls; n 10). The fourth group (n 11) served as naive controls, although they also received halothane. All of the rats were assessed for mechanical and thermal pain thresholds 3 days after their arrival. A second baseline measurement was taken 1 day after the first, and the 2 measurements were averaged to give a single baseline pain threshold for each assessment. All rats were subsequently assessed 12 days after surgery or on the analogous day if they did not have chronic constriction injury surgery. The postsurgery assessment was performed once, not less than 12 hours after the treatment that occurred on day 11. There was no assessment performed on day 11. After the last assessment, the rats were killed, and their spinal cords were removed for analysis. Data Analysis Thermal and mechanical pain thresholds for the 12-day assessment were used to calculate normalized difference scores with the following formulas: Mechanical 50% withdraw threshold right paw 50% withdraw threshold left paw) baseline 50% withdraw threshold right paw baseline 50% withdraw threshold leftpaw) normalized 50% withdraw threshold Thermal withdraw latency right paw withdraw latency left paw baseline withdraw latency right paw baseline withdraw latency left paw normalized withdraw latency The more negative the normalized difference score, the lower the pain threshold in the right paw when compared with the left paw. Mean mechanical normalized difference scores were then compared among the 4 groups by using a single-factor analysis of variance (ANOVA). A Scheffé test was used for pairwise, post hoc comparisons of the groups. The same analyses were used to examine the mean withdrawal latency among the 4 groups. The mean synaptosomal content of amino acids was compared among the 4 groups by using a 2-factor ANOVA, with group and side of the rat as the factors. Separate 2-factor ANOVAs were performed for each amino acid examined. Because there was no significant side or interaction effect for any neurotransmitter examined, amino acid contents in the right and left sides were averaged, and mean synaptosomal contents for the 4 groups (there was a significant group effect) were then compared by using a single-factor ANOVA and a Scheffé test for pairwise post hoc comparison. The level was P less than or equal to.05 for all statistics used to compare mean values. To determine whether there was a relation between synaptosomal amino acid content in the dorsal horn and normalized difference scores of the right paw, we performed multiple regressions. Four backward, stepwise, multiple regressions were completed on each group. The first regression examined the relation between the normalized 50% withdrawal threshold (dependent variable) and excitatory amino acids (independent variables). Specifically, the independent variables in this first regression were the synaptosomal aspartate content and the glutamate content of the right and left dorsal horns. The second regression performed on the same group of rats examined the relation between the normalized 50% withdrawal threshold (dependent variable) and inhibitory amino acids (independent variables). Specifically, the independent variables in this regression were the right and left dorsal horn, synaptosomal content of GABA, and right and left dorsal horn content of glycine. The third regression examined the relation between normalized withdrawal latency and excitatory amino acids, and the fourth regression examined the relation between normalized withdrawal latency and inhibitory amino acids. The 4 regressions were performed for all 4 groups. The rejection criteria for removing an independent variable from the regression equation was P greater than.10. The level for the ANOVA of a regression equation was P less than or equal to.05. When an equation met this criterion, it was subsequently evaluated for an appropriate fit with the data (standardized residuals were not 2.5); for compliance with the assumptions of linear regression analysis (presence of homoscedasticity, normal distribution of residuals, absence of correlation between residuals, a linear relation between independent and dependent variables); and for the absence of significant multicolinearity between independent variables (largest variance inflation factor not 10, average variance inflation factor not 1, variance proportions for independent variables distributed to different dimensions). A regression equation was considered valid only if it met all of these criteria. Finally, all regression equations were assessed for outliers and overly influential cases (evaluation of standardized residuals and Cooks distance 1 for all cases, respectively). There were no outliers or overly influential cases in any regression equation. All statistical analyses were performed with SPSS, version 11, e for Windows. RESULTS Contralateral Effects of Chronic Constriction Injury on Paw Sensitivities It is possible that the paw contralateral to the chronic constriction injury surgery experienced an alteration in thermal and

4 1578 MECHANISMS FOR VARIABLE RESPONSE TO TENS, Somers TENS delivered to the rats of the control group that did not have chronic constriction injury did not alter mechanical or thermal sensitivities; there was no difference between TENS control and control rats in mean normalized 50% withdrawal threshold or withdrawal latency. Fig 1. (A) Mechanical and (B) thermal pain thresholds in the left paws of rats with chronic constriction injury (CCI) that were or were not treated with TENS. Bars are the mean standard error of the mean (SEM) 50% withdrawal threshold (WT) for mechanical pain threshold and withdrawal latency (WL) for thermal pain threshold. The thermal pain threshold in rats with chronic constriction injury was reduced at day 12 compared with baseline measurement 1 but was not reduced when compared with baseline measurement 2 (single-factor, repeated-measures ANOVA, Scheffé post hoc comparisons; P<.05). *Significant difference from 1 of the 2 baseline measurements. mechanical pain thresholds. 31 Because this would complicate the interpretation of normalized values, we used repeatedmeasures ANOVAs to determine whether the left paw of the rats with chronic constriction injury changed from baseline values over time. The mean 50% withdrawal threshold in the left paw did not change significantly from baseline values in TENS-treated chronic constriction injury or in chronic constriction injury rats (fig 1A). Likewise, mean withdrawal latency in the left paw did not change significantly over time from baseline values in the TENS-treated chronic constriction injury rats, but it did change in the chronic constriction injury rats (fig 1B). The mean withdrawal latency of the left paw at 12 days in chronic constriction injury rats was diminished when compared with the mean withdrawal latency of the first baseline measurement, but not when compared with the mean withdrawal latency of the second baseline measurement. Thermal and Mechanical Allodynia Both the TENS-treated and the nontreated chronic constriction injury rats developed mechanical and thermal allodynia (fig 2). The mean normalized 50% withdrawal threshold and withdrawal latency of both groups was significantly more negative than that observed in the control rats. Moreover, daily TENS was ineffective at reducing either mechanical or thermal allodynia. There was no difference in the mean normalized 50% withdrawal threshold and withdrawal latency between TENS-treated and nontreated chronic constriction injury rats. Mean Axon Terminal Dorsal Horn Content of Amino Acids Chronic constriction injury bilaterally increased the mean content of glutamate and aspartate in axon terminals of the dorsal horn (fig 3A). The mean synaptosomal content of glutamate and aspartate was increased by 44% and 46%, respectively, in the dorsal horns of chronic constriction injury rats when compared with control rats. This chronic constriction injury induced increase was blocked by daily application of TENS (fig 3A). The mean synaptosomal content of glutamate and aspartate in the dorsal horns of TENS-treated chronic constriction injury rats did not increase over the content of these neurotransmitters in control rats. Moreover, the content of glutamate and aspartate in the TENS-treated chronic constriction injury rats was significantly lower than the mean content of these neurotransmitters in untreated, chronic constriction injury rats. Chronic constriction injury did not alter the mean content of the inhibitory neurotransmitters GABA and glycine in the axon terminals of the dorsal horn (fig 3B). Although there was a tendency for bilateral mean synaptosomal GABA content to be reduced and glycine content to be increased when chronic constriction injury rats were compared with control rats, the changes were not significant. However, the mean glycine content in TENS-treated chronic constriction injury rats was lower than that observed in chronic constriction injury rats (fig 3B). TENS delivered daily to the control rats without chronic constriction injury bilaterally altered the mean content of glutamate and GABA in axon terminals from the dorsal horn (fig 3). The mean synaptosomal content of glutamate was increased by 44% and that of GABA was reduced by 38% when TENStreated control rats were compared with control rats. Relation Between Axon Terminal Content of Neurotransmitters and Right Paw Allodynia There was no relation between dorsal horn axon terminal content of amino acids and normalized withdrawal latency of Fig 2. Mechanical and thermal allodynia in the right paw. Bars represent mean SEM normalized 50% withdrawal threshold from calibrated Semmes-Weinstein monofilaments in grams (left) and mean SEM normalized withdrawal latency from a radiant heat source in seconds (right). The more negative a mean value, the greater the reduction in thermal or mechanical pain threshold. *Significant difference from naive control rats (single-factor ANOVA, Scheffé post hoc comparisons; P<.05). Rats with chronic constriction injury that were treated or not treated with TENS did not differ in normalized pain thresholds. Abbreviation: CO, controls.

5 MECHANISMS FOR VARIABLE RESPONSE TO TENS, Somers 1579 content and normalized 50% withdrawal thresholds (table 1). As the content of glutamate in synaptosomes from the right dorsal horn diminished (.44) and that from the left dorsal horn increased (.78), the normalized 50% withdrawal threshold increased (reduced tactile allodynia). Similarly, as the content of glycine in synaptosomes from the right dorsal horn diminished (.43) and that from the left dorsal horn increased (.82), the normalized 50% withdrawal threshold increased (reduced tactile allodynia). For both glutamate and glycine, left dorsal horn synaptosomal content was more strongly associated with normalized 50% withdrawal threshold in the right paw than was right dorsal horn content. DISCUSSION Fig 3. Dorsal horn, synaptosomal content of amino acid neurotransmitters. A 2-factor ANOVA showed a significant effect for group, but not for side of the rat or interaction between group and the side of the rat. Therefore, bars represent the mean synaptosomal content SEM (pmol/mg wet weight) for (A) excitatory amino acids and (B) inhibitory amino acids when right and left dorsal horns were averaged together. *Significant difference from control rats; significant difference from rats with chronic constriction injury (single-factor ANOVA, Scheffé post hoc comparisons; P<.05). the right paw in any of the 4 groups. Likewise, there was no relation between axon terminal content of amino acids and normalized 50% withdrawal threshold in the chronic constriction injury, control, and TENS-treated control groups. However, in TENS-treated chronic constriction injury rats, there was a relation between synaptosomal glutamate and glycine Relation of Dorsal Horn Amino Acid Content to Allodynia in Chronic Constriction Injury Rats Chronic constriction injury bilaterally increased the mean content of aspartate and glutamate over control values in axon terminals of the dorsal horn. This is consistent with the increased whole tissue and extracellular mean content of aspartate and glutamate that is also present in the dorsal horn of neuropathic rats. 23,32 Although there was a trend for rats with chronic constriction injury to have lower mean axon terminal levels of GABA and higher levels of glycine, these changes were not significant. However, at least for GABA, the immunoreactive 33 and extracellular 21 mean content of the neurotransmitter was also depressed in the dorsal horns of neuropathic rats. Thus, mean aspartate and glutamate content are increased and mean GABA is likely depressed in the dorsal horns of rats with painful neuropathy. There is substantial evidence that implicates these chronic constriction injury induced changes of neurotransmitter levels in the development and maintenance of neuropathic pain. Excitatory amino acids are involved in pain neurotransmission 34,35 and are increased in neuropathic conditions, 23,32 and antagonism of glutamate receptors reduces the development or expression of neuropathic pain in rats and humans On the other hand, the administration of GABA receptor agonists abolishes neuropathy-induced thermal allodynia, 40 GABA is diminished in neuropathic conditions, 21,33 and antagonism of Table 1: Relation Between Normalized 50% Withdrawal Threshold and Dorsal Horn Synaptosomal Content of Neurotransmitters in TENS-Treated Rats With Chronic Constriction Injury Category of Neurotransmitter Regression Statistics and Independent Variables P Excitatory amino acids Model statistics R 2.53; adj R SEE 3.14 Neurotransmitters retained in stepwise regression* Glutamate in right dorsal horn ( ) Glutamate in left dorsal horn ( ) Inhibitory amino acids Model statistics R 2.58; adj R SEE 2.95 Neurotransmitters retained in stepwise regression Glycine in right dorsal horn ( ) Glycine in left dorsal horn ( ) Abbreviation: adj, adjusted; SEE, standard error of estimate. * As the content of glutamate in the synaptosomes from the right dorsal horn diminishes and from the left dorsal horn increases, the normalized 50% withdrawal threshold increases (reduced tactile allodynia). As the content of glycine in the synaptosomes from the right dorsal horn diminishes and from the left dorsal horn increases, the normalized 50% withdrawal threshold increases (reduced tactile allodynia).

6 1580 MECHANISMS FOR VARIABLE RESPONSE TO TENS, Somers GABA receptors augments the development or expression of neuropathic pain in rats. 41 Because increased aspartate and glutamate and diminished GABA content in the dorsal horn are associated with the expression of neuropathic pain, we suspected that, on this backdrop of altered neurotransmitter content, there might be some variability that could explain the different levels of allodynia in rats with chronic constriction injury. Specifically, we suspected that as the dorsal horn increase in aspartate and glutamate and the reduction in GABA content became more severe, the magnitude of allodynia would increase. Conversely, as the dorsal horn increase in aspartate and glutamate and the reduction in GABA content became less severe, the magnitude of allodynia would diminish. This did not happen, and there was no relation between the axon terminal content of aspartate, glutamate, GABA, or glycine and allodynia expression in the painful paw. Thus, although increased aspartate and glutamate and diminished GABA content in the dorsal horn play a role in allodynia, subsequent variability within these already altered levels does not predict the magnitude of allodynia. Mean Dorsal Horn Amino Acid Content and Allodynia in TENS-Treated Rats Daily TENS applied to the ipsilateral dorsal rami of spinal nerves L1 to L6 did not reduce tactile or thermal allodynia in the painful paw of rats with chronic constriction injury. This is contrary to the reduction in thermal allodynia that we 19 reported when TENS was delivered in the same fashion to chronic constriction injury rats. Although unexpected, this variability appears to be characteristic of HFS when used to treat neuropathic pain. For example, rats that develop mechanical allodynia after either chronic constriction injury or partial nerve ligation to the sciatic nerve experience a reduction in mechanical allodynia during and immediately after DSCS. 20 In subsequent experiments, it was discovered that in rats with chronic constriction injury 42 or in rats with partial nerve ligation, 21 the response to DSCS can be more variable; some rats experience no reduction in mechanical allodynia during or after the treatment. Regardless of a precedent for variable responsiveness to different forms of HFS, the inability of daily TENS to reduce tactile or thermal allodynia in our rats with chronic constriction injury was unexpected. This was particularly so because TENS altered the dorsal horn content of excitatory amino acids in a way that is reminiscent of similar treatments that reduce allodynia. For example, DSCS reduces extracellular aspartate and glutamate content in the ipsilateral dorsal horn of neuropathic rats. 23 This change occurs only in rats that also experience a concomitant reduction of mechanical allodynia in the painful paw. Like DSCS, daily TENS reduces the axon terminal aspartate and glutamate content in the dorsal horns of chronic constriction injury rats. However, no concomitant reduction in mean mechanical allodynia occurred in the painful paws of these TENS-treated rats. Why daily TENS did not reduce allodynia even though it reduced the dorsal horn content of aspartate and glutamate is not known. However, the effect of HFS on dorsal horn inhibitory neurotransmitters may provide an explanation. The reported reduction in mechanical allodynia in neuropathic rats receiving DSCS was accompanied by a progressive increase in extracellular GABA in the dorsal horn ipsilateral to the painful paw. 23 This alteration and the decrease in aspartate and glutamate already mentioned appear to be necessary for the treatment to be effective, because none of these alterations occurred when mechanical allodynia was unchanged by the stimulation. It is possible that daily TENS failed to alter mean mechanical allodynia because it did not increase the mean dorsal horn content of GABA. This notion is consistent with the ability of exogenously applied GABA to affect the responsiveness of neuropathic rats to DSCS. Allodynic rats that were initially unresponsive to DSCS stimulation experienced a reduction in allodynia when exogenous GABA was applied to the dorsal horn during the stimulation. 42 If these unresponsive rats are similar to those in the TENS-treated chronic constriction injury group, then the inability of daily TENS to increase the dorsal horn content of GABA may be responsible for the failure of the treatment to consistently reduce allodynia. The mean axon terminal glycine content in the dorsal horn of TENS-treated rats with chronic constriction injury was reduced when compared with untreated rats with chronic constriction injury. The role this change plays in the failure of daily TENS to reduce mean allodynia is difficult to discern, because the mean glycine content in both chronic constriction injury and TENS-treated chronic constriction injury rats was statistically unchanged from that in control rats. Although difficult to interpret, the reduction of mean synaptosomal glycine content in TENS-treated chronic constriction injury rats relative to that in untreated chronic constriction injury rats may well be meaningful, because the glycinergic system is implicated in the modulation of neuropathic pain. 43,44 To determine whether these TENS-induced alterations in the dorsal horns of rats with chronic constriction injury were unique to rats with neuropathy, the effect of TENS on rats that did not receive a chronic constriction injury was also examined. For aspartate, glutamate, and glycine, it appears that daily TENS produces effects on rats with chronic constriction injury that are specific to the neuropathy. Although TENS delivered daily to rats with chronic constriction injury bilaterally decreased to control levels the chronic constriction injury induced increase in dorsal horn content of these neurotransmitters, TENS delivered to healthy rats bilaterally increased over control levels the dorsal horn content of these same neurotransmitters (albeit not significantly for aspartate and glycine). Thus, daily TENS produced changes in the mean dorsal horn neurotransmitter content of rats with chronic constriction injury that are unique to rats with neuropathy. Relation of Dorsal Horn Amino Acid Content to Allodynia in TENS-Treated Rats With Chronic Constriction Injury In addition to making an assessment of mean axon terminal content of neurotransmitters, we examined the relation between dorsal horn amino acid content and allodynia, using multiple regression. There was a significant relation between axon terminal content and mechanical allodynia in TENS-treated rats with chronic constriction injury that was not present in other groups. Because there was no relation between mechanical allodynia and neurotransmitter content in the untreated rats with chronic constriction injury, it appears that daily TENS induced a relationship between these variables where none had existed. The TENS-induced relationship that developed between dorsal horn neurotransmitter content and magnitude of allodynia in neuropathic rats may explain, in part, the variable responsiveness of mechanical allodynia to daily TENS. Within the rats of the TENS-treated chronic constriction injury group, 8 rats developed what could be called severe allodynia (eg, normalized 50% withdrawal threshold 8g). However, there were also 2 rats with moderate allodynia (eg, normalized 50% withdrawal threshold 4 to 7g) and 5 rats with low or no allodynia (eg, normalized 50% withdrawal threshold 0 to 3g).

7 MECHANISMS FOR VARIABLE RESPONSE TO TENS, Somers 1581 The mechanical allodynia within this entire spectrum of rats was significantly related to the axon terminal content of glutamate and glycine. As the content of either glutamate or glycine diminished in the right dorsal horn and increased in the left dorsal horn, mechanical allodynia in the right neuropathic paw decreased. Conversely, as the dorsal horn content of either glutamate or glycine increased in the right dorsal horn and decreased in the left dorsal horn, mechanical allodynia increased. Therefore, the content of neurotransmitter in dorsal horn axon terminals predicted the magnitude of allodynia that a TENS-treated rat with chronic constriction injury would experience. In fact, between 45% and 51% of the variability in mechanical allodynia was associated with the content of these 2 neurotransmitters within the dorsal horns of the spinal cord (adjusted R 2 glutamate.45; adjusted R 2 glycine.51). At least in the dorsal horn ipsilateral to the nerve injury (right), this relationship was consistent with what would be expected for excitatory amino acids, given the established role for increased aspartate and glutamate in the development and maintenance of neuropathic pain. Accordingly, as glutamate levels decreased in the right dorsal horn of HFS-treated rats with chronic constriction injury (extracellular 23 or axon terminal), mechanical allodynia was reduced in the right paw. However, the relationship reported here between the axon terminal content of glycine in the right dorsal horn and mechanical allodynia was not consistent with the role of this neurotransmitter in neuropathic pain development. Chronic intrathecal administration of glycine to rats with chronic constriction injury beginning before the surgery reduced the development of mechanical allodynia, 43 suggesting that as glycine content increases in the dorsal horn, mechanical allodynia will diminish. Here, we report exactly the opposite relationship; as the dorsal horn content of glycine diminished ipsilateral to the nerve injury, mechanical allodynia became less severe. Although the reason for this apparent paradox is unknown, the responsiveness of rat dorsal horn nociceptive neurons to iontophoretic application of different levels of glycine is likewise paradoxical. Application of low amounts of glycine augments evoked activity in dorsal horn nociceptive neurons, whereas application of high amounts of glycine to the same neurons inhibits evoked activity. 45 This bivariate response appears to depend on which glycine receptor is being stimulated. 45 Application of low amounts of glycine is believed to stimulate high-affinity allosteric glycine-binding sites on the N-methyl-D-aspartate (NMDA) receptor, increasing the evoked activity of nociceptive neurons. Application of higher amounts of glycine stimulates low-affinity strychnine-sensitive glycine receptors, inhibiting evoked activity of nociceptive neurons. It is not known whether a similar mechanism might explain why both decreasing (TENS application) and increasing the content of glycine (intrathecal application of glycine) in the ipsilateral dorsal horn of rats with chronic constriction injury is associated with reduced allodynia. However, it is tempting to speculate that TENS reduces mechanical allodynia by reducing glycine stimulation of the NMDA receptor. Indeed, reduced stimulation of this receptor by administration of an antagonist specific for the NMDA glycine site diminishes mechanical allodynia in rats with chronic constriction injury. 46 On the other hand, intrathecal administration of glycine may reduce allodynia by increasing the stimulation of the strychnine-sensitive inhibitory glycine receptor. Regardless of the mechanism involved, our results indicated that the glycine content in the dorsal horn ipsilateral to a nerve injury is critical to an understanding of the variable responsiveness of mechanical allodynia to TENS treatment. In fact, of the 2 neurotransmitters related to the effectiveness of the modality, glycine content (adjusted R 2.51), more so than glutamate (adjusted R 2.45), was predictive of the level of allodynia that can be expected after daily TENS treatment. The axon terminal content of glutamate and glycine in the left dorsal horn contributed most heavily to the variability in mechanical allodynia of TENS-treated rats with chronic constriction injury (.78 to.82 for the left dorsal horn;.43 to.44 for the right dorsal horn). This was so even though daily TENS was delivered to the dorsal rami of right-sided spinal nerves. Thus, although the treatment was delivered ipsilateral to the nerve injury, the best predictor of TENS effectiveness was neurotransmitter content in the contralateral dorsal horn. Although alterations in the dorsal horn contralateral to a nerve injury are well documented, 47 only recently has a modulatory role for contralateral neurotransmission been implicated in neuropathic pain. Accordingly, in the dorsal horns of rats with chronic constriction injury, wide dynamic range neurons identified ipsilateral to the nerve injury are responsive to stimulation of skin innervated by spinal nerves located contralateral to the nerve injury. 48 Noxious stimulation in these same contralateral receptive fields reduces the responsiveness of ipsilateral dorsal horn neurons to stimulation of the painful paw. 49 This modulation involves spinal cord commissural neurons because the longitudinal section of the spinal cord greatly attenuates the modulating effect of contralateral noxious stimulation. 49 Although daily TENS is not a painful stimulation, these recent studies indicate that spinal cord commissural neurons can modulate pain-mediating neurons of the dorsal horn. Together with our findings, these results suggest that the contralateral or bilateral application of daily TENS should be studied. It is possible that such an application would improve the effectiveness of the modality in reducing pain. CONCLUSIONS There was variability in the level of mechanical allodynia observed in TENS-treated rats with chronic constriction injury. Such a variable response to TENS is typical when it is used to treat neuropathic pain in humans At least in rats, this variable responsiveness is related to axon terminal content of neurotransmitters in the dorsal horn. Of the 4 neurotransmitters that we examined, glutamate and glycine were best able to predict the responsiveness of mechanical allodynia to TENS. Moreover, the application of TENS to dorsal rami located ipsilateral to the nerve injury produced a relation between mechanical allodynia and the content of these neurotransmitters in both dorsal horns. Thus, the effectiveness of TENS as it was delivered in this study may be augmented by bilateral or contralateral application. In fact, bilateral, daily TENS treatment to the dorsal rami of a woman with severe diabetic neuropathy greatly reduced pain in both of her lower extremities. 50 Our present and future similar investigations should continue to improve the effectiveness of this noninvasive treatment modality. Acknowledgment: We thank the Empi Corporation for supplying the TENS units used in the study. References 1. Johnson MI, Ashton CH, Thompson JW. Long-term use of transcutaneous electrical nerve stimulation at New Castle Pain Relief Clinic. J R Soc Med 1992;85: Barr JO. Transcutaneous electrical nerve stimulation for pain management. In: Nelson RM, Currier DP, editors. Clinical electrotherapy. Norwalk (CT): Appleton & Lange; p Long DM. Electrical stimulation for the relief of pain from chronic nerve injury. J Neurosurg 1973;39:

8 1582 MECHANISMS FOR VARIABLE RESPONSE TO TENS, Somers 4. Picaza JA, Cannon BW, Hunter SE, Boyd AS, Guma J, Maurer D. Pain suppression by peripheral nerve stimulation. Part II. Observations with implanted devices. Surg Neurol 1975;4: Campbell JN, Long DM. Peripheral nerve stimulation in the treatment of intractable pain. J Neurosurg 1976;45: Weiner RL. The future of peripheral nerve neurostimulation. Neurol Res 2000;22: Shimoji K, Hokari T, Kano T, et al. Management of intractable pain with percutaneous epidural spinal cord stimulation: differences in pain-relieving effects among diseases and sites of pain. Anesth Analg 1993;77: Burchiel KJ, Anderson VC, Brown FD, et al. Prospective, multicenter study of spinal cord stimulation for relief of chronic back and extremity pain. Spine 1996;21: Kemler MA, Barendse GA, van Kleef M, et al. Spinal cord stimulation in patients with chronic reflex sympathetic dystrophy. N Engl J Med 2000;343: Linderoth B, Meyerson BA. Central nervous system stimulation for neuropathic pain. In: Hansson PT, Fields HL, Hill RG, editors. Progress in pain research and management. Seattle: IASP Pr; p Marchand S, Charest J, Li J, Chenard JR, Lavignolle B, Laurencelle L. Is TENS purely a placebo effect? A controlled study on chronic low back pain. Pain 1993;54: Johansson F, Almay BG, Von Knorring L, Terenius L. Predictors for the outcome of treatment with high frequency transcutaneous electrical nerve stimulation in patients with chronic pain. Pain 1980;9: White PF, Li S, Chiu W. Electroanalgesia: its role in acute and chronic pain management. Anesth Analg 2001;92: Kumar D, Marshall HJ. Diabetic peripheral neuropathy: amelioration of pain with transcutaneous electrostimulation. Diabetes Care 1997;20: Cauthen JC, Renner EJ. Transcutaneous and peripheral nerve stimulation for chronic pain states. Surg Neurol 1975;4: Meyler WJ, de Jongste MJ, Rolf CA. Clinical evaluation of pain treatment with electrostimulation: a study on TENS in patients with different pain syndromes. Clin J Pain 1994;10: Hassenbusch SJ, Stanton-Hicks M, Schoppa D, Walsh JG, Covington EC. Long-term results of peripheral nerve stimulation for reflex sympathetic dystrophy. J Neurosurg 1996;84: Kumar K, Toth C, Nath RK. Spinal cord stimulation for chronic pain in peripheral neuropathy. Surg Neurol 1996;46: Somers DL, Clemente FR. High-frequency transcutaneous electrical nerve stimulation alters thermal but not mechanical allodynia following chronic constriction injury of the rat sciatic nerve. Arch Phys Med Rehabil 1998;79: Meyerson BA, Ren B, Herregodts P, Linderoth B. Spinal cord stimulation in animal models of mononeuropathy: effects on the withdrawal response and flexor reflex. Pain 1995;61: Stiller CO, Cui JG, O Connor T, Brodin E, Meyerson BA, Linderoth B. Release of gamma aminobutyric acid in the dorsal horn and suppression of tactile allodynia by spinal cord stimulation in mononeuropathic rats. Neurosurgery 1996;39: Burchiel KJ, Anderson VC, Wilson BJ, Olson KA, Shatin D. Prognostic factors of spinal cord stimulation for chronic back and leg pain. Neurosurgery 1995;36: Cui JG, O Connor T, Ungerstedt U, Linderoth B, Meyerson BA. Spinal cord stimulation attenuates augmented dorsal horn release of excitatory amino acids in mononeuropathy via a GABAergic mechanism. Pain 1997;73: Bennett GJ, Xie YK. A peripheral mononeuropathy in rats that produces disorders of pain sensation like those seen in man. Pain 1988;33: Kim SH, Chung JM. An experimental model for peripheral neuropathy produced by segmental spinal nerve ligation in the rat. Pain 1992;50: Kingery WS, Castellote JM, Wang EE. A loose ligature-induced mononeuropathy produces hyperalgesias mediated by both the injured sciatic nerve and the adjacent saphenous nerve. Pain 1993;55: Swett JE, Woolf CJ. The somatotopic organization of primary afferent terminals in the superficial lamina of the dorsal horn of the rat spinal cord. J Comp Neurol 1985;231: Hargreaves K, Dubner R, Brown F, Flores C, Joris J. A new and sensitive method for measuring thermal nociceptive in cutaneous hyperalgesia. Pain 1988;32: Takahashi Y, Nakajiama Y, Sakamoto T. Dermatomal mapping of the rat hindlimb by electrical stimulation of the spinal nerves. Neurosci Lett 1994;168: Somers DL, Beckstead RM. Chronic methionine sulfoximine administration reduces synaptosomal aspartate and glutamate in rat striatum. Neurosci Lett 1990;115: Attal N, Kayser JV, Guilbaud G. Further evidence for pain related behaviors in a model of unilateral peripheral mononeuropathy. Pain 1990;41: Kawamata M, Omote K. Involvement of increased excitatory amino acids and intracellular Ca 2 concentration in the spinal dorsal horn in an animal model of neuropathic pain. Pain 1996; 68: Ibuki T, Hama AT, Wang XT, Pappas GD, Sagen J. Loss of GABA-immunoreactivity in the spinal dorsal horn of rats with peripheral nerve injury and promotion of recovery by adrenal medullary grafts. Neuroscience 1996;76: Dougherty PM, Willis WD. Enhanced responses of spinothalamic tract neurons to excitatory amino acids accompany capsaicininduced sensitization in the monkey. J Neurosci 1992;12: Radhakrishnan V, Henry JL. Excitatory amino acid receptor mediation of sensory inputs to functionally identified dorsal horn neurons in cat spinal cord. Neuroscience 1993;55: Dogrul A, Ossipov MH, Lai J, Malan TP, Porreca F. Peripheral antihyperalgesic activity of SIB-1757, a metabotropic glutamate receptor (mglur5) antagonist, in experimental neuropathic pain in rats. Neurosci Lett 2000;292: Leung A, Wallace MS, Ridgeway B, Yaksh TL. Concentrationeffect relationship of intravenous alfentanil and ketamine on peripheral neurosensory thresholds, allodynia and hyperalgesia of neuropathic pain. Pain 2001;91: Mao J, Price DD, Hayes RL, Lu J, Mayer DJ. Differential roles of NMDA and non-nmda receptor activation in induction and maintenance of thermal hyperalgesia in rats with painful peripheral mononeuropathy. Brain Res 1992;598: Mao J, Price DD, Hayes RL, Lu J, Mayer DJ, Frenk H. Intrathecal treatment with dextrorphan or ketamine potently reduces painrelated behaviors in a rat model of peripheral mononeuropathy. Brain Res 1993;605: Hwang JH, Yaksh TL. The effect of spinal GABA receptor agonists on tactile allodynia in a surgically-induced neuropathic pain model in the rat. Pain 1997;70: Yamamoto T, Yaksh TL. Effects of intrathecal strychnine and bicuculline on nerve compression-induced thermal hyperalgesia and selective antagonism by MK-801. Pain 1993;54: Cui JG, Linderoth B, Meyerson BA. Effects of spinal cord stimulation on touch-evoked allodynia involve GABAergic mechanisms. An experimental study in the mononeuropathic rat. Pain 1996;66: Huang W, Simpson RK. Long-term intrathecal administration of glycine prevents mechanical hyperalgesia in a rat model of neuropathic pain. Neurol Res 2000;22: Simpson RK, Gondo M, Robertson CS, Goodman JC. Reduction and thermal hyperalgesia by intrathecal administration of glycine and related compounds. Neurochem Res 1997;22: Budai D, Wilcox GL, Larson AA. Enhancement of NMDAevoked neuronal activity by glycine in the rat spinal cord in vivo. Neurosci Lett 1992;135: Quartaroli M, Fasdelli N, Bettelini L, Maraia G, Corsi M. GV196771A, an NMDA receptor/glycine site antagonist, attenuates mechanical allodynia in neuropathic rats and reduces tolerance induced by morphine in mice. Eur J Pharmacol 2001;430: Koltzenburg M, Wall PD, McMahon SB. Does the right side know what the left is doing? Trends Neurosci 1999;22: Sotgiu ML, Biella G. Spinal neuron sensitization facilitates contralateral inputs in rats with peripheral mononeuropathy. Neurosci Lett 1998;241: