Secondary Logo

Journal Logo

Soy-Containing Diet Suppresses Chronic Neuropathic Sensory Disorders in Rats

Shir, Yoram MD*,; Sheth, Rishi MD†,; Campbell, James N. MD†,; Raja, Srinivasa N. DMD‡,; Seltzer, Ze’ev MD§

doi: 10.1097/00000539-200104000-00042
Regional Anesthesia And Pain Medicine: Research Report

Partial sciatic nerve ligation (PSL) in rodents produces chronic neuropathic sensory disorders resembling neuropathic pain in humans. We previously reported that levels of allodynia and hyperalgesia after PSL injury were markedly attenuated by consumption of soy-containing diets. Here we aimed to show that dietary effect on pain behavior is not specific to a certain laboratory. For this purpose, experiments were conducted in a different laboratory (Baltimore rather than Jerusalem) and a different rat strain (Wistar rather than Sabra), with additional and different testing methods (radiant heat from a lamp rather than a CO2 laser). Rats were fed two soy-free diets and a soy-containing one for 28 days. The sensitivity of rats to nonnoxious and noxious stimuli was determined before PSL injury, and levels of neuropathic sensory disorders were determined after it. We found that consuming the soy-containing diet prevented development of tactile and heat allodynia, but not mechanical hyperalgesia. This dietary effect was not correlated with calorie intake and weight gain or dietary concentration of fat and carbohydrates. We conclude that, regardless of experimental site, diet markedly affects chronic neuropathic sensory disorders in rats and should be standardized in animal models of pain.

*Department of Anesthesiology and Pain Relief Unit, Hadassah University Hospital, Jerusalem, Israel; †Department of Neurosurgery, Johns Hopkins Medical Institutions, Baltimore, Maryland; ‡Department of Anesthesiology, Johns Hopkins Medical Institutions, Baltimore, Maryland; and §Department of Physiology, Faculties of Medicine and Dental Medicine, Hebrew University, Jerusalem, Israel

December 12, 2000.

Implications: Levels of chronic sensory disorders in a rat model of allodynia and hyperalgesia after partial sciatic nerve ligation depend on the consumption of a soy-containing diet. Further studies are needed to determine the role of diet in humans with chronic pain.

Address correspondence and reprint requests to Yoram Shir, MD, Department of Anesthesiology and Pain Relief Unit, Hadassah University Hospital, Jerusalem 91120, Israel. Address e-mail to

Partial sciatic nerve ligation (PSL) in rats produces chronic neuropathic sensory disorders (CNSD) (1). Levels of allodynia and hyperalgesia in this model vary considerably across laboratories. Attempting to identify the variable at play, we found that diet markedly modifies levels of CNSD in this model (2). Specifically, consumption of a diet containing soy protein suppressed CNSD after PSL injury. We also found that diet affects CNSD levels in another neuropathic model produced in rodents by total hindpaw denervation (3). However, it is possible that these observations were idiosyncratic to a particular experimental design and not a generalized phenomenon. Indeed, differences in modes of behavior other than pain were found across different laboratories (4). Therefore, here we examined whether the effect of diet on CNSD was demonstrable across two laboratories, differing in their vivaria, rat strain, composition of diet, and sensory testing methods used to determine levels of sensory disorders. We also examined whether diet has an effect on baseline touch and acute pain sensitivity in intact animals (5). Finally, food palatability and calorie intake affect acute (6,7) and persistent inflammatory pain (8). Therefore, we examined whether the calorie intake of a certain diet under test is responsible for suppression of the sensitivity of intact rats to noxious stimuli and of CNSD in the PSL model.

Back to Top | Article Outline


Experiments were conducted at Johns Hopkins University, Baltimore, MD, and conformed to institutional regulations for animal experimentation and those of the International Association for the Study of Pain (9).

We used eight groups of male Wistar rats (n = 6–11 rats per group) (Harlan, Indianapolis, IN) weighing 225–275 g at the beginning of the experiment. Rats were housed under standard colony conditions (3 rats per cage; ambient temperature kept at 22°C ± 0.5°C; water and food supplied ad libitum; day/night cycle was lights switched on at 7:00 am and off on 7:00 pm). Tests were performed between 10:00 am and 5:00 pm.

We tested three diets:

  • 1. RMH (RMH-1000; PMI Feeds, St. Louis, MO), a balanced diet comprising 14% protein (85% from soy), 67.5% carbohydrates, 6% fat, 4.5% fiber, 8% ash, vitamins, and trace elements. This diet contained 70 μg phytoestrogen per gram of diet and provided 3.3 kcal/g.
  • 2. CAS (Bio-Serv Co., Frenchtown, NJ), an artificial balanced diet (10) comprising 16% casein protein, 65% carbohydrates (50% starch, 15% sucrose), 5% canola oil (composed of 6% saturated, 36% polyunsaturated, and 58% monounsaturated fatty acids), 5% fiber, 3.5% ash, vitamins, and trace elements. This diet contained 8 μg phytoestrogen per gram of diet and provided 3.8 kcal/g.
  • 3. BC, an unbalanced diet made of white bread and fresh cucumbers from a local store. The caloric and nutritional values of cucumbers were considered to be negligible. White bread is 8.2% protein, 50% carbohydrates (43% starch, 7% sucrose), 3.6% fat (composed of 30% saturated, 15% polyunsaturated, and 55% monounsaturated fatty acids), 3.4% fiber, 2.1% ash, vitamins, and trace elements. This diet contained 16 μg phytoestrogen per gram and provided 2.7 kcal/g.

Rats were fed RMH diet from weaning until the beginning of the experiment, then switched to one of the experimental diets for 28 days. Three Control groups of rats (n = 6 per group) were fed RMH, CAS, and BC and designated CON-RMH, CON-BC, and CON-CAS. Five rat groups, fed RMH, CAS, or BC diets, underwent a PSL injury on Day 14 (PSL-RMH, PSL-CAS, and PSL-BC groups). The former two groups were replicated. Because their results were not significantly different from the first run, data of the two runs were pooled.

Daily dietary consumption of each cage was calculated by weighing the diet remaining in the food hopper and subtracting it from the amount given the day before (no allowance was made for spillage). Daily calorie intake per rat was calculated by multiplying the daily weight of consumed diet per cage by its caloric value and dividing it by the number of rats per cage. Rats’ weights were recorded at the beginning of experiment, before nerve injury, and at the end of the experiment.

With the rats under inhaled anesthesia, the right sciatic nerve was exposed near the trochanter (1). By using a minineedle, an 8-0 silk suture was inserted in the middle of the nerve, trapping in a tight ligation the dorsal – of the nerve thickness. The wound was closed with muscle sutures and skin staples.

The responses of PSL-injured rats to mechanical and thermal stimuli were determined bilaterally 1 day before PSL surgery (Day 13 of experimental feeding) and repeated on Days 3, 8, and 14 thereafter (1). Intact rats were tested on Days 13, 16, 22, and 28 of the experiment, matching those of PSL-injured rats. The experimenter was unaware of the type of diet consumed by rats under testing.

Withdrawal threshold to touch was measured with a set of eight calibrated von Frey hairs (weighing 0.3, 1.1, 2.8, 4.4, 6.4, 9.5, 11, and 20 g). Each rat was placed in a chamber with a mesh metal floor, covered by an opaque plastic dome. After 5 min of acclimation, the tested hair was indented five times, at a rate of two per second, in the midplantar skin of the paw until it bowed. If below threshold, the stimulus intensity was increased by using the next hair. At threshold, rats responded by paw elevation.

Response to pinprick was tested in the same chamber by a single pricking of the midplantar area with a sharpened wooden stick. Duration of paw lifting was recorded with a stopwatch, with a cutoff of 30 s.

Response to noxious heat was tested with the Hargreaves method (11). Rats were tested in a chamber with a glass floor kept at 32°C. Heat was emitted from a 150 W lamp, automatically shutting off when the animal withdrew. We recorded the time lapsing from stimulus onset until paw lifting (“withdrawal threshold”) and the duration of paw lifting until it was replaced on the floor (“response duration”), with a cutoff of 30 s. Each paw was tested three times, allowing an interval of 2 min between tests to avoid sensitization. Withdrawal threshold and response duration of individual rats were calculated as the average of three trials.

Weight gain of individual rats and mean calorie intake per rat were calculated separately for 14 days before nerve injury and for 14 days after surgery. Sensitivity of intact rats to noxious and innocuous stimuli was calculated as the average of both hindpaws for each testing day. A grand average of the four testing sessions was performed for each group, creating a single score. One-way analysis of variance (ANOVA) for main effect of diet on the withdrawal threshold to touch was done by using the Kruskal-Wallis test. Post hoc differences between two diets were calculated with the Mann-Whitney U-test. Differences in the withdrawal threshold to heat and the response duration to pinprick and heat were assessed by using ANOVA, and t-test for post hoc analyses.

Behavioral scores of PSL-injured rats were calculated for each hindpaw separately. For each group, a grand average was calculated for the three postoperative testing sessions, creating a single score per group. P < 0.05 was used as the level of significance, corrected when appropriate by using the Bonferroni adjustment.

Back to Top | Article Outline


The average variance in the weight gain per cage, along the experiment, was small (5.5% of the average rat weight). Weight gain of PSL-BC, PSL-CAS, and PSL-RMH rats significantly differed a day before nerve injury (P < 0.0001) and at the end of the experiment (P < 0.0001). Weight gain of PSL-BC rats was significantly smaller than that of the PSL-RMH or PSL-CAS groups, before (P < 0.0001) and after nerve injury (P < 0.0001) (Fig. 1). Weight gain and calorie consumption of PSL-RMH and PSL-CAS groups were not significantly different before (P = 0.1) or after PSL injury (P = 0.1). PSL-BC rats consumed more calories but gained less weight than the PSL-CAS or PSL-RMH groups (Fig. 1).

Figure 1

Figure 1

Tactile withdrawal threshold and response duration to pinprick or noxious heat of intact rats were not significantly different among the three diet groups (P > 0.3 for all three comparisons) (Figs. 2, 3, 5). Withdrawal threshold to heat was significantly affected by type of diet (P < 0.0001) (Fig. 4). Consumption of bread and cucumbers for 28 days produced significant hypoalgesia compared with RMH and CAS diets (CON-BC versus CON-CAS or CON-RMH, P < 0.0001).

Figure 2

Figure 2

Figure 3

Figure 3

Figure 5

Figure 5

Figure 4

Figure 4

Nerve injury produced sensory abnormalities in all three dietary groups, lasting at least 14 days after surgery. Compared with intact rats, all three PSL-injured groups developed a significant tactile allodynia, because their withdrawal threshold to touch was significantly lower after nerve injury (PSL-RMH versus CON-RMH, P = 0.0004; PSL-CAS versus CON-CAS, P = 0.0002; PSL-BC versus CON-BC, P = 0.001) (Fig. 2). Diet significantly affected levels of tactile allodynia after PSL injury (P = 0.017). Consumption of RMH, a diet containing soy, produced significantly lower levels of tactile allodynia than diets devoid of soy (PSL-RMH versus PSL-CAS, P = 0.01; PSL-RMH versus PSL-BC, P = 0.039). Levels of allodynia to touch in rats fed diets devoid of soy were not significantly different (PSL-CAS versus PSL-BC, P = 0.9).

Compared with intact rats, all three PSL-injured groups developed a significant mechanical hyperalgesia, because their response duration to pinprick was significantly prolonged after nerve injury (PSL-RMH versus CON-RMH, P = 0.001; PSL-CAS versus CON-CAS, P = 0.0002; PSL-BC versus CON-BC, P = 0.01) (Fig. 3). Levels of mechanical hyperalgesia did not depend on the type of diet (P = 0.2).

Compared with intact rats, a significant heat allodynia developed in the two groups fed a diet devoid of soy (CAS and BC), because their withdrawal threshold to heat was significantly lower after nerve injury (PSL-CAS versus CON-CAS, P = 0.0003; PSL-BC versus CON-BC, P < 0.0001). However, consuming RMH had not changed the time needed to elicit paw withdrawal (PSL-RMH versus CON-RMH, P = 0.08) (Fig. 4). Diet had a significant effect on withdrawal threshold to heat after PSL injury (P < 0.0001). This effect was more robust in rats consuming diets devoid of soy compared with rats fed a diet containing soy (PSL-BC versus PSL-RMH, P = 0.0002; PSL-CAS versus PSL-RMH, P = 0.0006). Levels of heat allodynia in the two groups fed diets devoid of soy did not differ significantly (PSL-CAS versus PSL-BC, P = 0.5).

A significantly prolonged response duration to heat developed in all nerve-injured groups, compared with intact rats (PSL-RMH versus CON-RMH, P = 0.0005; PSL-CAS versus CON-CAS, P = 0.0001; PSL-BC versus CON-BC, P = 0.0004) (Fig. 5). The response duration did not depend on the type of diet (P = 0.3).

Of PSL-CAS and PSL-BC rats, 43% and 36%, respectively, expressed tactile allodynia on the contralateral intact hindpaw, compared with only 17% of PSL-RMH rats. In addition, 44% and 20% of PSL-BC and PSL-CAS rats, respectively, presented heat allodynia on the intact hindpaw, whereas none of the rats of the PSL-RMH group did so. Thus, consumption of soy-containing diet prevented development of allodynia on the mirror image side as well.

Back to Top | Article Outline


Surprisingly, very little is known about the effect of diet on chronic pain in animals (3,12) or humans (13). The main results of this experiment were that the type of diet rats consumed played an important role in neuropathic pain levels after partial denervation. Specifically, we found that a soy-containing diet significantly blunted tactile and heat allodynia after PSL injury, but not mechanical hyperalgesia. The analgesic properties of soy were not limited to a single laboratory because we reaffirmed its effect in a different laboratory and experimental design.

The only difference between our previous results in Jerusalem (2) and this study in Baltimore is that soy diet did not prevent the development of increased response duration to heat in the latter. This difference may be caused by the different rat strain used in the two studies (Sabra and Wistar, respectively). Genetic effects in the PSL and related models have been demonstrated in rats and mice (14–16). In addition, the differing results could be caused by different methods for production of heat (CO2 laser versus the Hargreaves method). Because in the latter method, rats withdraw from the stimulus at threshold, the level of sensitization it produces is lower compared with the CO2 laser.

In agreement with other reports, heat pain sensitivity of intact rats was significantly affected by diet. Increased amounts of dietary sucrose reportedly attenuated heat pain (6,7,17) and paw inflammation (8), an effect putatively mediated by endogenous opioids (18). However, in this experiment, RMH and casein diets contained equal concentrations of carbohydrates (67.5% and 65%, respectively), yet PSL-injured rats fed these diets differed significantly in their tactile and heat allodynia. Moreover, BC diet contained 15%–17% fewer carbohydrates than CAS or RMH, but levels of allodynia and hyperalgesia in rats fed BC did not differ significantly from CAS. Thus, carbohydrates did not affect the reduction in tactile and heat threshold in the two soy-free groups.

Increasing fat content in the diet attenuates acute heat pain (5). It is unlikely, however, that in this experiment, dietary fat content affected levels of CNSD. The CAS and BC diets differed both in their fatty acids concentration and composition. Yet similar levels of CNSD developed in these two groups after nerve injury. Moreover, CAS and RMH contained similar fat levels but expressed significantly different levels of allodynia.

Consumption of highly palatable foods increases calorie intake (19) and attenuates responses of rodents to acute pain (6,7). In this experiment, rats fed BC consumed significantly more calories than rats fed the RMH or CAS diets. It is not clear whether the increased calorie consumption of rats fed the BC diet was caused by its increased palatability or to its being an unbalanced diet. Nevertheless, similar to previous studies, the withdrawal threshold to noxious heat of intact BC rats was significantly higher compared with the other two groups (Fig. 4) (6,7). Higher calorie intake had no effect, however, on the development of the PSL model, because BC and CAS rats, significantly differing in their calorie intake, developed similar neuropathic pain levels.

Younger rats tend to develop more pain in other models of neuropathy, arguably because of lower body weight [reviewed by Zeltser and Seltzer (20) and Bennett et al. (21)]. This was never examined, however, in the PSL model. BC-fed rats gained significantly less weight than the other two groups, but their levels of CNSD after PSL injury were not significantly different than CAS-fed rats. Thus, body weight in Wistar rats is not associated with levels of neuropathic disorders in this model.

It is likely that soy is the dietary ingredient responsible for suppression of CNSD in the PSL model. RMH and CAS diets differed mainly in the protein source, i.e., soy protein in RMH versus casein protein in CAS. Nevertheless, these diets were not identical, suggesting that other ingredients could suppress CNSD in this model. However, when rats were fed identical diets differing only in their protein, i.e., soy versus casein, soy-based diet was associated with suppressed levels of allodynia and hyperalgesia (2). Of the various ingredients in soy protein that could prevent the development of CNSD, phytoestrogens and certain amino acids are the most plausible candidates. Phytoestrogens are plant compounds possessing estrogenic properties (22). Because estrogens may augment antinociception (23), soy phytoestrogens could mediate the suppression of CNSD after PSL injury. Our recent results in Wistar rats partially support this suggestion (unpublished data). However, the correlation between dietary and plasma levels of amino acids in rats fed the RMH, CAS, and BC diets—and CNSD in these rats—was not significant for any amino acid (unpublished data). Thus, a possible role for certain amino acids in the suppression of CNSD in the PSL model has yet to be found.

We conclude that the effect of diet on the development of CNSD in the PSL model was evident across two different laboratories, rat strains, heat-producing stimulators, diet types, diet providers, and vivaria. Food vendors supply rat chows standardized for the relative content of basic nutritional ingredients, but not for the source and type of these ingredients. Therefore, standardized, artificial balanced diets are recommended for rat experiments that use the PSL model, and probably for other pain models. Neuropathic pain in humans is only partially understood and many times is refractory to treatment. Future studies will show whether dietary manipulation can alleviate pain in humans with these syndromes.

We are grateful to Marshall Devor for commenting on this manuscript. This project was supported in part by the US-Israel Binational Science Foundation 98203, NIH grant 26363, and the Blaustein Foundation for Pain Research.

Back to Top | Article Outline


1. Seltzer Z, Dubner R, Shir Y. A novel behavioral model of neuropathic pain disorders produced in rats by partial sciatic nerve injury. Pain 1990; 43: 205–18.
2. Shir Y, Ratner A, Raja SN, et al. Neuropathic pain following partial nerve injury in rats is suppressed by dietary soy. Neurosci Lett 1998; 240: 73–6.
3. Shir Y, Ratner A, Seltzer Z. Diet can modify autotomy behavior in rats following peripheral neurectomy. Neurosci Lett 1997; 236: 71–4.
4. Crabbe JC, Wahlsten D, Dudek B. Genetics of mouse behavior: interactions with laboratory environment. Science 1999; 284: 1670–2.
5. Yehuda S, Carraso RL. Modulation of learning, pain thresholds and thermoregulation in the rats by preparations of free purified alpha linolenic and linolenic acids: determination of their optimal ratio. Proc Natl Acad Sci U S A 1993; 90: 10345–9.
6. Blass E, Fitzgerald E, Kehoe P. Interactions between sucrose, pain and isolation distress. Pharmacol Biochem Behav 1987; 26: 483–9.
7. Zhang T, Reid K, Acuff CG, et al. Cardiovascular and analgesic effects of a highly palatable diet in spontaneously hypertensive and Wistar-Kyoto rats. Pharmacol Biochem Behav 1994; 48: 57–61.
8. Ren K, Blass EM, Zhou Q, Dubner R. Suckling and sucrose ingestion suppress persistent hyperalgesia and spinal Fos expression after forepaw inflammation in infant rats. Proc Natl Acad Sci U S A 1997; 94: 1471–5.
9. Zimmermann M. Ethical guidelines for investigations of experimental pain in conscious animals. Pain 1983; 16: 109–10.
10. Anastasia JV, Braun BL, Smith KT. General and histopathological results of a two-year study of rats fed semi-purified diets containing casein and soya protein. Food Chem Toxicol 1990; 28: 147–56.
11. Hargreaves K, Dubner R, Brown F, et al. Hargreaves: a new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia. Pain 1988; 32: 77–88.
12. Abbott FV, Young SN. The effect of tryptophan supplementation on autotomy induced by nerve lesions in rats. Pharmacol Biochem Behav 1991; 40: 301–4.
13. King RB. Pain and tryptophan. J Neurosurg 1980; 53: 44–52.
14. Seltzer Z, Shir Y. Sympathetically-maintained causalgiform disorders in a model of neuropathic pain: a review. J Basic Clin Physiol Pharmacol 1991; 2: 18–56.
15. Devor M, Raber P. Heritability of symptoms in an experimental model of neuropathic pain. Pain 1990; 42: 51–68.
16. Mogil JS, Wilson SG, Bon K, et al. Heritability of nociception. I. Responses of 11 inbred mouse strains on 12 measures of nociception. Pain 1999; 80: 67–82.
17. Frye CA, Bock BC, Kanarek RB. Hormonal milieu affects tailflick latency in female rats and may be attenuated by access to sucrose. Physiol Behav 1992; 52: 699–706.
18. Marks-Kaufman R, Hamm MW, Barbato GF. The effects of dietary sucrose on opiate receptor binding in genetically obese (ob/ob) and lean mice. J Am Coll Nutr 1989; 8: 9–14.
19. Jakubczak LF. Age differences in the effects of palatability of diet on regulation of calorie intake and body weight of rats. J Gerontol 1977; 32: 49–57.
20. Zeltser R, Seltzer Z. A practical guide for the use of animal models for neuropathic pain. In: Boivie J, Hansson P, Lindblom U, eds. Touch, temperature and pain. Seattle: IASP Publications, 1994: 337–79.
21. Bennett GJ, Chung JM, Seltzer Z. Animal models for painful neuropathies. Protocols Neurosci. In press.
22. Aldercreutz H, Mazur W. Phytoestrogens and western diseases. Ann Med 1997; 29: 95–120.
23. Wardlaw S, Thompson L, Frantz A. Effects of sex steroids on brain beta-endorphin. Brain Res 1982; 245: 327–31.
© 2001 International Anesthesia Research Society