- Email: [email protected]
Denervation-induced inflammation in the rat
Neuroscience Letters, 119 (1990) 37~t0
37
Elsevier Scientific Publishers Ireland Ltd. NSL 07246
Denervation-induced inflammation in the rat Jon D. Levine, Terence J. Coderre, D e b o r a h M. White, Walter E. Finkbeiner and Allan I. Basbaum Departments of Anatomy, Medicine, Pathology and Physiology and Division of Neurosciences, University of California, San Francisco, CA 94143
(U.S.A.) (Received 18 May 1990; Accepted 29 June 1990)
Key words: Inflammation; Neurogenic; Denervation; Neutrophil; Axonal transport; Cutaneous We report that section of the sciatic and sapbenous nerves, in the hindlimb of the rat, evokes an inflammatory response in the denervated tissue that can be distinguished from the previously described peptide-mediated neurogenic inflammation. This novel form of neurogenic inflammation has a very delayed onset (9.75 _+2.1 h, mean + S.E.M., n = 8), persists for more than 30 h, and is characterized by a marked neutrophific cellular infiltrate. These features cannot be mimicked by electrical stimulation of the peripheral nerve and are not prevented by either prior application of local anesthetics to the nerve lesion site or by neonatal treatment with capsaicin.
It is well established that the peripheral nervous system can contribute to the inflammatory response. For example, antidromic activation of unmyelinated primary afferents [2-4] produces physiological changes characteristic of inflammation, including vasodilation, plasma extravasation and sensitization of nociceptors. It is presumed that peptides released from peripheral terminals of primary afferent nerves mediate this neurogenic inflammation. In previous studies in the rat we observed that peripheral nerve section was associated with a transient swelling of the denervated paw. Initially we assumed that nerve section evoked antidromic activation of primary afferents inducing the release of proinflammatory neuropeptides from the primary afferent terminals. In the present study we demonstrate that this phenomenon, in fact, represents a novel form of neurogenic inflammation. The experiments were performed on 250-350 g male Sprague-Dawley rats (Bantin and Kingman, Fremont CA). Under pentobarbital anesthesia (50 mg/kg, i.p.) the sciatic and saphenous branches of the femoral nerve were either exposed and cut or just exposed (sham surgery), at the level of the mid-thigh. To minimize any contribution that might result from trauma to the paralyzed limb, the rats were housed on a deep bedding of soft-
Correspondence: J.D. Levine, Division of Rheumatology, U-426 (Box 0724), University of California, San Francisco, San Francisco, CA 94143-0724, U.S.A. 0304-3940/90/$ 03.50 © 1990 Elsevier Scientific Publishers Ireland Ltd.
wood shavings. The rats were monitored every hour after surgery, for a period of 36 hrs, for changes in paw thickness. Swelling was graded on a 0-3 scale, in which: 0 = a normal paw; 0.5 (trace) = redness and minimal swelling around the ankle joint; 1 =mild swelling extending from the ankle to the proximal foot; 2 = moderate swelling over the ankle and entire foot, and; 3 = severe swelling over the ankle and entire foot. For latency determinations, swelling was considered present if the score was 0.5 (trace) or above. In a group of rats undergoing sciatic and saphenous nerve section we examined the histological changes in the skin and subcutaneous tissue of the denervated paws. At various times after nerve injury, the rats received an anesthetic overdose and the skin overlying the area of maximal swelling on the dorsum of the paw was excised. The tissue was fixed for 24 hours in 0.1 M phosphatebuffered 4 % paraformaldehyde, dehydrated and embedded in glycomethacrylate (GMA). Three micron sections were cut on a JB4 microtome and stained with hematoxylin, eosin and azure. Two experimental protocols were used to evaluate the contribution of nerve section-induced neuronal discharge to the observed swelling. In one group of anesthetized rats we studied the effect of blocking neural activity, with local anesthetic, distal to the nerve lesion. Bupivacaine (0.5 %) was injected subepineurally, in the sciatic and saphenous nerves, 5 min prior to cutting these nerves. In another group of anesthetized rats, sciatic and saphenous nerves were electrically stimulated at a fre-
38
quency of 5 Hz for 15 min, to mimic the neuronal activity induced by cutting the nerve. The stimulus intensity and frequency (4 V, 0.5 ms, 5 Hz) were demonstrated to produce maximal peaks in C-fiber compound action potentials recorded at a site distal to the stimulating electrode. The animal was then allowed to recover from the general anesthesia and paw swelling was scored for the next 36 hrs. A group of rats that had been treated 24 h after birth with capsaicin (100 mg/kg, s.c., in a vehicle of 50% DMSO) was also used to evaluate the contribution of unmyelinated afferents to the swelling response induced by nerve lesion. Since nerve section disrupts axonal transport, we also studied whether changes in transport contribute to the swelling induced by nerve lesion. Specifically, we evaluated the effect of nerve sections made at different distances from the hindpaw. In the proximal-lesion group, the sciatic nerve was cut at the level of the sciatic notch and the femoral nerve was cut at the branch point of the anterior and posterior femoral nerves. In the distallesion group, the sciatic nerve was cut 40 m m distal to the site of the proximal-lesion group, at a level 1 cm distal to the trifurcation of the sciatic nerve; the femoral nerve was cut at the trifurcation of its saphenous branch. This is also approximately 40 m m distal to the femoral nerve cut in the proximal-lesion group. Fig. 1 compares the time course of hindpaw swelling after combined saphenous and sciatic nerve-section, electrical stimulation, or sham surgery. In rats that had both nerves transected, hindpaws began to swell approximately 9 h post-lesion; swelling peaked by 22 h and persisted to at least 36 h. Sham surgery did not produce a delayed onset of swelling. Histological examination of subcutaneous tissue from
2.0[] SECTION • SHAM O ELECT.STIM. 1.5-
O o0 1.0-
0.5-
0.00
lO
20
TIME {HRS} Fig. I. Time course of the swelling response in rats after nerve section (n = 8), sham surgery (n = 8) or C-fiber strength nerve stimulation of sciatic and saphenous nerves (n = 6). These data are from the same group of rats as in Table I.
A,
Fig. 2. Photomicrograph of skin from the dorsum of the hindpaw of the sham-operated (control) and nerve sectioned rat, taken 24 h after surgery. This photomicrograph shows skin thickness in the hindpaw of sham-operated (A) and nerve-sectioned (B) rats. The thickness of the skin from the nerve-sectioned rat is approximately twice the thickness of the skin from the sham-lesioned rat.
swollen paws harvested from rats 24 h after nerve section demonstrated an acute inflammatory response. This was characterized by an increase in the thickness of subcutaneous tissue (Fig. 2) and an infiltrate of neutrophils (Fig. 3). In contrast, at 24 h after sham operation, we found no histological signs of acute inflammation in skin or subcutaneous tissue. It is unlikely that antidromic activation of primary afferent neurons contributes to the delayed swelling since application of bupivacaine to the nerves, prior to nerve section, at a site distal to the nerve section, did not prevent the swelling (Table I). Furthermore, electrical stimulation of the sciatic and saphenous nerves failed to induce the delayed swelling (Fig. 1). We also observed that neonatal treatment with capsaicin did not prevent swelling (Table I). Finally, the latency to onset of swelling in the proximal nerve lesion group (34.75 + 2.9 h, n = 8, mean + S.E.M.) was significantly longer than the latency to onset for the distal nerve lesion group 9 . 7 5 + 2 . 1 h (n = 8) (q4 = 6.9, P<0.001). In the present study we have shown that section of the
39 TABLE I MEAN + S.E.M. MAGNITUDE OF SWELLING AT 24 HOURS, IN THE HINDPAWS OF RATS IN WHICH THE SCIATIC AND FEMORAL NERVES WERE CUT, JUST EXPOSED (SHAM), TREATED WITH BUPIVACAINE BEFORE BEING CUT, OR STIMULATED AT C-FIBRE STRENGTH Kruskal-Wallis analysis of variance reveals a significant effect of treatment group (/-/4=31.8, P<0.001). Significant differences from the sham group are indicated by an asterisk (* P < 0.01). There is no significant difference between the nerve lesion group and the groups that received pretreatment with bupivieaine + nerve lesion or pretreatment with neonatal capsaicin + nerve lesion (P > 0.05).
Fig. 3. Higher magnification photomicrographs illustrating compactness of collagen and cellular infiltration in sham-lesioned (A) and nerve-lesioned (B, C) rats. The section in the middle photomicrograph (B) was chosen to contain a similar number and size of blood vessels, when compared to the section from the sham-lesioned rat (A). The skin from the nerve-lesioned rat demonstrates separation of collagen (edema), cellular aggregation in the venules, and dermal parenchyma infiltration.
sciatic and saphenous nerves elicits swelling and neutrophilic infiltration in the denervated tissue. Application o f local anesthetic, to block antidromic stimulation, or neonatal treatment with capsaicin, to ablate unmyelinated afferents, h a d no effect on swelling or the presence o f cellular infiltration p r o d u c e d by these nerve lesions. Furthermore, a l t h o u g h prolonged antidromic stimulation (electrical) o f the sciatic and saphenous nerves, at inten-
Experimental condition
Magnitude of swelling (ca. 24 h)*
Sham nerve lesion Nerve lesion Bupivicaine + nerve lesion Capsaicin + nerve lesion Nerve stimulation
0.00 + 1.47 + 1.38 _ 2.25 + 0.00 _
0.00 (n = 8) 0.18" (n=8) 0.16" (n=8) 0.2* (n = 10) 0.00 (n = 6)
sities that activated unmyelinated axons, typically induced a transient, trace swelling within minutes, it did not p r o d u c e the significant swelling with the delayed onset that characterized the nerve lesion-induced changes. These data suggest that the inflammation induced by nerve section is different from the neurogenic inflammation p r o d u c e d by antidromic activation o f C-fibers. The latter, for example, can be prevented by local anesthetic. Since the site o f nerve section (proximal vs distal) had a m a r k e d effect on latency to swelling, it is unlikely that the response results f r o m t r a u m a to the paralyzed paw. This new p h e n o m e n o n thus represents a novel f o r m o f neurogenic inflammation. The delayed onset to swelling for the m o r e proximal lesion suggests that swelling is precipitated by loss o f a factor that is usually transported distally in the nerve. That is, it appears that in n o r m a l animals, factor(s) released f r o m peripheral nerve m a y be continually inhibiting this cellular inflammation. F r o m the twenty-five h o u r difference in latency to effect, we calculate a transport rate o f approximately 1.6 mm/hr, compatible with the k n o w n rate o f slow axonal transport [1, 5, 6]. A n alternative hypothesis would be that factor(s) taken up at the cut proximal end o f the distal nerve segment, are transported distally, and then released by the nerve terminals, resulting in swelling and cellular infiltration. These could be distinguished by the use o f axonal transp o r t inhibitors. Unfortunately, at doses o f colchicine necessary to p r o d u c e blockade o f axonal transport, we consistently observed paralysis, suggesting that nerve injury had occurred. F u r t h e r studies are planned to investigate the nature o f the specific mediators that m o d u late this novel f o r m o f cellular neurogenic inflammation.
40 W e t h a n k Dr. Philip Heller for m a n y helpful suggestions d u r i n g the writing o f the m a n u s c r i p t . This w o r k was s u p p o r t e d b y N I H G r a n t A M 3 2 6 3 4 . J . D , L . is a R i t a A l l e n F o u n d a t i o n fellow. 1 Baitinger, C., Levine, J., Lorenz, T., Simon, C., Skene, P. and Willard, M., Characteristics of axonally transported proteins. In D.G. Weis (Ed.), Axoplasmic Transport, Springer, Berlin, 1982, pp. 110120. 2 Fitzgerald, M., The spread of sensitization ofpolymodal noeiceptors in the rabbit from nearby injury and by antidromic stimulation, J. Physiol., 297 (1979) 207-216.
3 Gamse, R., Holzer, P. and Lembeck, F., Decrease of substance P in primary afferent neurones and impairment of neurogenic plasma extravasation by capsaicin, Br. J. Pharmaeol., 68 (1980) 207-213. 4 Jancso, N., Jancso-Gabor, A. and Szoicsanyi, J., Direct evidence for neurogenie inflammation and its prevention by denervation and by pretreatment with capsaiein, Br. J. Pharmacol., 31 (1967) 138-151. 5 Kristensson, K., Retrograde transport of macromolecules in axons, Annu. Rev. Pharmacol. Toxicol., 18 (1978) 97-110. 6 Levine, J. and Willard, M., The composition and organization of axonally transported proteins in the retinal ganglion cells of the guinea pig, Brain Res., 194 (1980) 13%!54.
37
Elsevier Scientific Publishers Ireland Ltd. NSL 07246
Denervation-induced inflammation in the rat Jon D. Levine, Terence J. Coderre, D e b o r a h M. White, Walter E. Finkbeiner and Allan I. Basbaum Departments of Anatomy, Medicine, Pathology and Physiology and Division of Neurosciences, University of California, San Francisco, CA 94143
(U.S.A.) (Received 18 May 1990; Accepted 29 June 1990)
Key words: Inflammation; Neurogenic; Denervation; Neutrophil; Axonal transport; Cutaneous We report that section of the sciatic and sapbenous nerves, in the hindlimb of the rat, evokes an inflammatory response in the denervated tissue that can be distinguished from the previously described peptide-mediated neurogenic inflammation. This novel form of neurogenic inflammation has a very delayed onset (9.75 _+2.1 h, mean + S.E.M., n = 8), persists for more than 30 h, and is characterized by a marked neutrophific cellular infiltrate. These features cannot be mimicked by electrical stimulation of the peripheral nerve and are not prevented by either prior application of local anesthetics to the nerve lesion site or by neonatal treatment with capsaicin.
It is well established that the peripheral nervous system can contribute to the inflammatory response. For example, antidromic activation of unmyelinated primary afferents [2-4] produces physiological changes characteristic of inflammation, including vasodilation, plasma extravasation and sensitization of nociceptors. It is presumed that peptides released from peripheral terminals of primary afferent nerves mediate this neurogenic inflammation. In previous studies in the rat we observed that peripheral nerve section was associated with a transient swelling of the denervated paw. Initially we assumed that nerve section evoked antidromic activation of primary afferents inducing the release of proinflammatory neuropeptides from the primary afferent terminals. In the present study we demonstrate that this phenomenon, in fact, represents a novel form of neurogenic inflammation. The experiments were performed on 250-350 g male Sprague-Dawley rats (Bantin and Kingman, Fremont CA). Under pentobarbital anesthesia (50 mg/kg, i.p.) the sciatic and saphenous branches of the femoral nerve were either exposed and cut or just exposed (sham surgery), at the level of the mid-thigh. To minimize any contribution that might result from trauma to the paralyzed limb, the rats were housed on a deep bedding of soft-
Correspondence: J.D. Levine, Division of Rheumatology, U-426 (Box 0724), University of California, San Francisco, San Francisco, CA 94143-0724, U.S.A. 0304-3940/90/$ 03.50 © 1990 Elsevier Scientific Publishers Ireland Ltd.
wood shavings. The rats were monitored every hour after surgery, for a period of 36 hrs, for changes in paw thickness. Swelling was graded on a 0-3 scale, in which: 0 = a normal paw; 0.5 (trace) = redness and minimal swelling around the ankle joint; 1 =mild swelling extending from the ankle to the proximal foot; 2 = moderate swelling over the ankle and entire foot, and; 3 = severe swelling over the ankle and entire foot. For latency determinations, swelling was considered present if the score was 0.5 (trace) or above. In a group of rats undergoing sciatic and saphenous nerve section we examined the histological changes in the skin and subcutaneous tissue of the denervated paws. At various times after nerve injury, the rats received an anesthetic overdose and the skin overlying the area of maximal swelling on the dorsum of the paw was excised. The tissue was fixed for 24 hours in 0.1 M phosphatebuffered 4 % paraformaldehyde, dehydrated and embedded in glycomethacrylate (GMA). Three micron sections were cut on a JB4 microtome and stained with hematoxylin, eosin and azure. Two experimental protocols were used to evaluate the contribution of nerve section-induced neuronal discharge to the observed swelling. In one group of anesthetized rats we studied the effect of blocking neural activity, with local anesthetic, distal to the nerve lesion. Bupivacaine (0.5 %) was injected subepineurally, in the sciatic and saphenous nerves, 5 min prior to cutting these nerves. In another group of anesthetized rats, sciatic and saphenous nerves were electrically stimulated at a fre-
38
quency of 5 Hz for 15 min, to mimic the neuronal activity induced by cutting the nerve. The stimulus intensity and frequency (4 V, 0.5 ms, 5 Hz) were demonstrated to produce maximal peaks in C-fiber compound action potentials recorded at a site distal to the stimulating electrode. The animal was then allowed to recover from the general anesthesia and paw swelling was scored for the next 36 hrs. A group of rats that had been treated 24 h after birth with capsaicin (100 mg/kg, s.c., in a vehicle of 50% DMSO) was also used to evaluate the contribution of unmyelinated afferents to the swelling response induced by nerve lesion. Since nerve section disrupts axonal transport, we also studied whether changes in transport contribute to the swelling induced by nerve lesion. Specifically, we evaluated the effect of nerve sections made at different distances from the hindpaw. In the proximal-lesion group, the sciatic nerve was cut at the level of the sciatic notch and the femoral nerve was cut at the branch point of the anterior and posterior femoral nerves. In the distallesion group, the sciatic nerve was cut 40 m m distal to the site of the proximal-lesion group, at a level 1 cm distal to the trifurcation of the sciatic nerve; the femoral nerve was cut at the trifurcation of its saphenous branch. This is also approximately 40 m m distal to the femoral nerve cut in the proximal-lesion group. Fig. 1 compares the time course of hindpaw swelling after combined saphenous and sciatic nerve-section, electrical stimulation, or sham surgery. In rats that had both nerves transected, hindpaws began to swell approximately 9 h post-lesion; swelling peaked by 22 h and persisted to at least 36 h. Sham surgery did not produce a delayed onset of swelling. Histological examination of subcutaneous tissue from
2.0[] SECTION • SHAM O ELECT.STIM. 1.5-
O o0 1.0-
0.5-
0.00
lO
20
TIME {HRS} Fig. I. Time course of the swelling response in rats after nerve section (n = 8), sham surgery (n = 8) or C-fiber strength nerve stimulation of sciatic and saphenous nerves (n = 6). These data are from the same group of rats as in Table I.
A,
Fig. 2. Photomicrograph of skin from the dorsum of the hindpaw of the sham-operated (control) and nerve sectioned rat, taken 24 h after surgery. This photomicrograph shows skin thickness in the hindpaw of sham-operated (A) and nerve-sectioned (B) rats. The thickness of the skin from the nerve-sectioned rat is approximately twice the thickness of the skin from the sham-lesioned rat.
swollen paws harvested from rats 24 h after nerve section demonstrated an acute inflammatory response. This was characterized by an increase in the thickness of subcutaneous tissue (Fig. 2) and an infiltrate of neutrophils (Fig. 3). In contrast, at 24 h after sham operation, we found no histological signs of acute inflammation in skin or subcutaneous tissue. It is unlikely that antidromic activation of primary afferent neurons contributes to the delayed swelling since application of bupivacaine to the nerves, prior to nerve section, at a site distal to the nerve section, did not prevent the swelling (Table I). Furthermore, electrical stimulation of the sciatic and saphenous nerves failed to induce the delayed swelling (Fig. 1). We also observed that neonatal treatment with capsaicin did not prevent swelling (Table I). Finally, the latency to onset of swelling in the proximal nerve lesion group (34.75 + 2.9 h, n = 8, mean + S.E.M.) was significantly longer than the latency to onset for the distal nerve lesion group 9 . 7 5 + 2 . 1 h (n = 8) (q4 = 6.9, P<0.001). In the present study we have shown that section of the
39 TABLE I MEAN + S.E.M. MAGNITUDE OF SWELLING AT 24 HOURS, IN THE HINDPAWS OF RATS IN WHICH THE SCIATIC AND FEMORAL NERVES WERE CUT, JUST EXPOSED (SHAM), TREATED WITH BUPIVACAINE BEFORE BEING CUT, OR STIMULATED AT C-FIBRE STRENGTH Kruskal-Wallis analysis of variance reveals a significant effect of treatment group (/-/4=31.8, P<0.001). Significant differences from the sham group are indicated by an asterisk (* P < 0.01). There is no significant difference between the nerve lesion group and the groups that received pretreatment with bupivieaine + nerve lesion or pretreatment with neonatal capsaicin + nerve lesion (P > 0.05).
Fig. 3. Higher magnification photomicrographs illustrating compactness of collagen and cellular infiltration in sham-lesioned (A) and nerve-lesioned (B, C) rats. The section in the middle photomicrograph (B) was chosen to contain a similar number and size of blood vessels, when compared to the section from the sham-lesioned rat (A). The skin from the nerve-lesioned rat demonstrates separation of collagen (edema), cellular aggregation in the venules, and dermal parenchyma infiltration.
sciatic and saphenous nerves elicits swelling and neutrophilic infiltration in the denervated tissue. Application o f local anesthetic, to block antidromic stimulation, or neonatal treatment with capsaicin, to ablate unmyelinated afferents, h a d no effect on swelling or the presence o f cellular infiltration p r o d u c e d by these nerve lesions. Furthermore, a l t h o u g h prolonged antidromic stimulation (electrical) o f the sciatic and saphenous nerves, at inten-
Experimental condition
Magnitude of swelling (ca. 24 h)*
Sham nerve lesion Nerve lesion Bupivicaine + nerve lesion Capsaicin + nerve lesion Nerve stimulation
0.00 + 1.47 + 1.38 _ 2.25 + 0.00 _
0.00 (n = 8) 0.18" (n=8) 0.16" (n=8) 0.2* (n = 10) 0.00 (n = 6)
sities that activated unmyelinated axons, typically induced a transient, trace swelling within minutes, it did not p r o d u c e the significant swelling with the delayed onset that characterized the nerve lesion-induced changes. These data suggest that the inflammation induced by nerve section is different from the neurogenic inflammation p r o d u c e d by antidromic activation o f C-fibers. The latter, for example, can be prevented by local anesthetic. Since the site o f nerve section (proximal vs distal) had a m a r k e d effect on latency to swelling, it is unlikely that the response results f r o m t r a u m a to the paralyzed paw. This new p h e n o m e n o n thus represents a novel f o r m o f neurogenic inflammation. The delayed onset to swelling for the m o r e proximal lesion suggests that swelling is precipitated by loss o f a factor that is usually transported distally in the nerve. That is, it appears that in n o r m a l animals, factor(s) released f r o m peripheral nerve m a y be continually inhibiting this cellular inflammation. F r o m the twenty-five h o u r difference in latency to effect, we calculate a transport rate o f approximately 1.6 mm/hr, compatible with the k n o w n rate o f slow axonal transport [1, 5, 6]. A n alternative hypothesis would be that factor(s) taken up at the cut proximal end o f the distal nerve segment, are transported distally, and then released by the nerve terminals, resulting in swelling and cellular infiltration. These could be distinguished by the use o f axonal transp o r t inhibitors. Unfortunately, at doses o f colchicine necessary to p r o d u c e blockade o f axonal transport, we consistently observed paralysis, suggesting that nerve injury had occurred. F u r t h e r studies are planned to investigate the nature o f the specific mediators that m o d u late this novel f o r m o f cellular neurogenic inflammation.
40 W e t h a n k Dr. Philip Heller for m a n y helpful suggestions d u r i n g the writing o f the m a n u s c r i p t . This w o r k was s u p p o r t e d b y N I H G r a n t A M 3 2 6 3 4 . J . D , L . is a R i t a A l l e n F o u n d a t i o n fellow. 1 Baitinger, C., Levine, J., Lorenz, T., Simon, C., Skene, P. and Willard, M., Characteristics of axonally transported proteins. In D.G. Weis (Ed.), Axoplasmic Transport, Springer, Berlin, 1982, pp. 110120. 2 Fitzgerald, M., The spread of sensitization ofpolymodal noeiceptors in the rabbit from nearby injury and by antidromic stimulation, J. Physiol., 297 (1979) 207-216.
3 Gamse, R., Holzer, P. and Lembeck, F., Decrease of substance P in primary afferent neurones and impairment of neurogenic plasma extravasation by capsaicin, Br. J. Pharmaeol., 68 (1980) 207-213. 4 Jancso, N., Jancso-Gabor, A. and Szoicsanyi, J., Direct evidence for neurogenie inflammation and its prevention by denervation and by pretreatment with capsaiein, Br. J. Pharmacol., 31 (1967) 138-151. 5 Kristensson, K., Retrograde transport of macromolecules in axons, Annu. Rev. Pharmacol. Toxicol., 18 (1978) 97-110. 6 Levine, J. and Willard, M., The composition and organization of axonally transported proteins in the retinal ganglion cells of the guinea pig, Brain Res., 194 (1980) 13%!54.