- Email: [email protected]
Extrinsic surgical denervation inhibits Clostridium difficile toxin A-induced enteritis in rats
Neuroscience Letters 292 (2000) 95±98
www.elsevier.com/locate/neulet
Extrinsic surgical denervation inhibits Clostridium dif®cile toxin A-induced enteritis in rats Christopher R. Mantyh a, Douglas C. McVey b,c, Steven R. Vigna b,c,* a Department of Surgery, Durham VA Medical Center, Durham, NC 27710, USA Department of Cell Biology and Medicine, Box 3709, Duke University Medical Center, Durham, NC 27710, USA c Department of Medicine, Box 3709, Duke University Medical Center, Durham, NC 27710, USA
b
Received 20 June 2000; received in revised form 8 August 2000; accepted 9 August 2000
Abstract Clostridium dif®cile enteritis is caused by toxin A (TA) which stimulates substance P release and subsequent receptor activation. This receptor stimulation results in secretion, in¯ammation, and structural damage. However, it is unclear as to which subset of neurons is required to initiate substance P release following toxin stimulation. Five centimeter ileal segments were surgically denervated. After 10 days, three ileal loops were constructed in each rat: the denervated loop was injected intraluminally with 5 mg of TA and two intact loops were injected with TA or vehicle, respectively. Ileal secretion, myeloperoxidase activity, and histology were then assessed. Denervated ileal loops injected with TA had a 75% reduction in ileal secretion (P , 0:001), 92% reduction in myeloperoxidase activity (P , 0:01) and 96% reduction in histologic damage (P , 0:001) compared to innervated loops. There were no signi®cant differences between the denervated loops injected with TA and those injected with vehicle. Extrinsic surgical denervation results in protection of ileal loops from TA enteritis. Furthermore, these results exclude the participation of intrinsic enteric nerves in TA-induced ileal damage. Finally, this suggests that extrinsic primary sensory neurons mediate the effects of intraluminal TA in the ileum. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: In¯ammation; Toxin A; Sensory neurons; Capsaicin; Substance P
Clostridium dif®cile is an anaerobic bacterium that releases several toxins leading to acute colitis and enteritis. Toxin A is an exotoxin that has been isolated and puri®ed from C. dif®cile and causes intense intestinal in¯ammation when injected intraluminally. Toxin A-induced enteritis in experimental animals is highly reproducible and rapid allowing for careful dissection of the in¯ammatory pathway. Currently it is proposed that toxin A binds to brush border receptors on villous enterocytes and super®cial mucosal epithelial cells in the colon [11]. Following toxin A binding to surface epithelial cells, neurotransmitters such as substance P and calcitonin gene-related peptide (CGRP) are released to further perpetuate the in¯ammatory response [10,12]. Inhibition of toxin A-induced intestinal secretion and in¯ammation in the rat occurs if substance P receptor antagonists or CGRP receptor antagonists are administered prior to toxin A injection. Additionally, substance P receptor (neurokinin (NK)-1 receptor) knockout mice are protected * Corresponding author. Tel.: 11-919-684-5311; fax: 11-919286-5530. E-mail address: [email protected] (S.R. Vigna).
from the physiologic and histologic effects of toxin A [2]. Thus it appears that primary sensory neurons are required for initiation of in¯ammation in the toxin A-stimulated enteritis model. The role of sensory neurons in modulating the in¯ammatory cascade is further supported by the abolition of toxin A's effects after capsaicin injection [10,12]. Capsaicin is a potent excitotoxin that is the active ingredient in hot chili peppers and acts as an exogenous ligand of the vanilloid receptor. The capsaicin-sensitive vanilloid receptor has recently been cloned [3] and is found on nociceptive afferent neurons. Acute, small doses of capsaicin cause vanilloid receptor activation and subsequent neuropeptide release. Excessive or chronic administration of capsaicin causes eventual death of dorsal root ganglion neurons and degeneration of the afferent neuron [6]. However, not all extrinsic sensory nerves are killed by excessive capsaicin administration. It appears that newborn rat neurons are particularly sensitive to the neurodegenerative effects of capsaicin, while adult rats may only have 50% of the sensory nerves ablated by a massive capsaicin injection [7]. In addition, capsaicin is not sensory neuron-speci®c in the adult rats,
0304-3940/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 0 0) 01 45 1- 8
96
C.R. Mantyh et al. / Neuroscience Letters 292 (2000) 95±98
as demonstrated by central neuronal death following large dose of capsaicin [13]. There is an absence of immunostaining for the vanilloid receptors in these central neurons. Also peri-vagal application of capsaicin is often used to study gastrointestinal functions. However, recently it has been shown that the vanilloid receptor is absent from vagal afferents again questioning the selectivity of capsaicin [3]. Finally, capsaicin appears to release substance P from human monocytes and macrophages indicating that the vanilloid receptor may not be uniquely present on sensory neurons [5]. Therefore, capsaicin-induced death of sensory neurons remains a useful but not speci®c tool to study the physiologic roles of sensory neurons. A more selective technique to examine neuronal inputs to the intestine is to extrinsically denervate surgically a loop of small bowel in the manner ®rst described by Furness [4]. By preserving the blood supply but eliminating the extrinsic nerves, only the intrinsic nerves in the plexi remain. Following super-selective denervation, we examined the in¯ammatory response after toxin A administration in the denervated segment in comparison to rats with intact extrinsic enteric nerves. Six Sprague±Dawley rats (150±175 g; Charles River Laboratories, Raleigh, NC) were fasted overnight with free access to water and were then anesthetized with iso¯urane. A small midline incision was made and a loop of ileum 4±5 cm in length with a discrete vascular bundle was selected. Dissection under an operating microscope using ®ne forceps and scissors was then performed to remove all nerve ®bers bundle while preserving the accompanying artery and vein as previously described [4]. The vessels were then swabbed with 80% phenol in distilled water causing the remaining nerve ®bers to appear white and these were subsequently dissected away. The denervated segment of ileum was then marked with a loose 3-0 silk ligature around the vascular pedicle and the bowel returned to the abdomen. The incision was closed in two layers and the animals were supplied with ad libitum food and water. Following ten days to allow for complete axonal degeneration, the animals were then treated with toxin A. Toxin A was prepared from C. dif®cile VPI strain 10463 as previously described [14]. Brie¯y, toxin A was puri®ed by sequential ammonium sulfate precipitation, anion-exchange chromatography, and isoelectric precipitation. The toxin preparation was homogeneous by crossed immunoelectrophoresis and native polyacrylamide gel electrophoresis [9]. The denervated ileal segments were isolated and ileal loops were constructed as previously described [10]. Brie¯y, rats fasted overnight with free access to water were anesthetized by im injection of 67 mg/kg ketamine:33 mg/kg xylazine. A midline abdominal incision was made, the terminal ileum was identi®ed, and ileal loops 5 cm in length were constructed by ligation with 4-0 silk sutures, taking care not to disturb the vascular supply. Toxin A (5 mg) in 400 ml of phosphate-buffered saline (PBS) (pH 7.4) was injected into the ileal loop using a 27 ga syringe needle
as previously described [10,12]. Control rats were prepared similarly and their ileal loops were injected with PBS. The severity of mucosal histological damage was graded on formalin-®xed, paraf®n-embedded, haematoxylin & eosin (H&E)-stained sections using a scheme described previously [10]. The following features were assessed: epithelial cell damage, mucosal congestion and edema, and neutrophil margination and in®ltration of the lamina propria (Fig. 1). A score of 0±3 for each category, denoting increasingly severe abnormality, was assigned by a blinded reviewer. Myeloperoxidase (MPO) activity was measured as described previously [1]. Brie¯y, pieces of control and treated ileal loops were homogenized in 0.5% hexadecyltrimethylammonium bromide in 50 mM KH2P04 (pH 6), freeze/thawed three times, centrifuged at 48C for 2 min, and then the absorbancy of each supernatant was read at 460 nm at 0, 30 and 60 s after the addition of 2.9 ml of o-dianisidine dihydrochloride to 0.1 ml supernatant. The maximal change in absorbancy per min was used to calculate the units of MPO activity based on the molar absorbancy index of oxidized o-dianisidine of 1.13 £ 10 4 M 21 cm 21. The results are expressed as MPO U of activity/g of tissue wet weight.
Fig. 1. The effects of intraluminal toxin A administration in isolated ileal segments. H&E-stained sections of paraf®nembedded samples were prepared after ileal loops were exposed for 3 h to (a) 0.4 ml of vehicle (PBS), (b) 0.4 ml of PBS containing 5 mg of puri®ed toxin A, or (c) 0.4 ml of PBS containing 5 mg of toxin A in a section of ileum that had been extrinsically denervated 10 days prior to toxin A administration. Toxin A disrupts villous architecture, causes neutrophil in®ltration, and results in mucosal edema as manifested by wide spacing of the crypts. Denervated ileal segments displayed normal villi, no neutrophil margination, and minimal edema. Magni®cation £130.
C.R. Mantyh et al. / Neuroscience Letters 292 (2000) 95±98
The presence or absence of immunoreactive tyrosine hydroxylase was used to assess the completeness of surgical denervation. Following ®xation in 4% (w/v) paraformaldehyde solution in 100 mM phosphate buffer (pH 7.4) for 1 h at room temperature the specimens were stored overnight in 30% sucrose. Frozen sections (10 mm) were then cut on a cryostat and dried on SuperFrost Plus slides. The sections were then immersed in PBS for 5 min and washed in a buffer consisting of: sodium phosphate 10 mmol/l, NaCl 0.5 mol/l, carrageenan 0.7% (w/v), Triton X-100 0.2% (v/v), sodium azide 0.01% (w/v), pH 7.4. The sections were then washed again in PBS three times for 2 min each. The sections were overlaid with 20 ml of monoclonal mouse anti-rat tyrosine hydroxylase (Boehringer Mannheim) and incubated for 1.5 h in a humid chamber at room temperature. The slides were washed three times for 2 min each in PBS. The slides were then stained with 20 ml of anti-mouse-IgG-FITC and incubated for 1 h in a humid chamber. The slides were ®nally washed three times for 2 min in PBS and coverslipped. Results are expressed as mean ^ SEM. Differences among groups were examined by one-way analysis of variance (ANOVA) with the Dunnett's or Tukey±Kramer post tests, using GraphPad Prism version 3.00 for Windows (GraphPad Software, San Diego, CA). P-values of ,0.05 were considered signi®cant. Intraluminal toxin A consistently demonstrates intense destruction of villous architecture, mucosal edema, and neutrophil in®ltration (Fig. 1B). Surgical denervation nearly completely ablated the histologic damage caused by toxin A (Fig. 1C). No difference was found between vehicleinjected controls and rats surgically denervated then injected with toxin A in epithelial damage, congestion and edema, or neutrophil in®ltration (Fig. 2) (P , 0:001). In addition, two rats were denervated and then injected with vehicle alone; again no histologic damage was evident. When all the pathologic scores were combined, there was a collective 96% reduction in histologic damage following extrinsic denervation. Denervation of the ileal segment reduced toxin A-induced ileal secretion and neutrophil activation as measured by myeloperoxidase (MPO) activity. Surgical denervation reduces toxin A-stimulated secretion by 75% (Fig. 3; P , 0:01) and MPO activity by 92% (Fig. 4; P , 0:001). In order to con®rm complete extrinsic denervation, tyrosine hydroxylase (TH) staining was performed. TH is an essential enzyme in the synthetic pathway and consistent marker of noradrenergic axons [4]. Little to no TH staining was found in the myenteric or submucosal plexus following denervation. Toxin A is an exotoxin released from C. dif®cile that rapidly produces intense intestinal in¯ammation causing ¯uid secretion, tissue necrosis, and neutrophil in®ltration. The in¯ammatory response from highly puri®ed toxin A is extremely reproducible and reaches near maximal stimulation at 3 h. For these reasons toxin A-enteritis represents a
97
unique model to mechanistically study the in¯ammatory cascade in the intestine. It appears that the ®rst step in toxin A-induced enteritis occurs after it binds to brush border receptors on enterocytes. Toxin A binding sites have been found in hamster, rabbit, rat, and human intestine, although the exact molecular structure of the receptor has yet to be agreed on [11]. Following binding to the enterocyte, toxin A stimulates substance P and CGRP release, which then propagate the in¯ammatory response. Speci®c antagonists to the substance P and CGRP receptors ameliorate the in¯ammatory effects of toxin A [8,12]. In addition, rapid endocytosis of the substance P receptor occurs following toxin A stimulation [10], and mice constructed to be de®cient in the substance P receptor are protected from toxin A-induced enteritis [2]. Thus it has been postulated that toxin A stimulates the nervous system to release a host of pro-in¯ammatory neuropeptides which then stimulate their respective receptors to propagate the in¯ammatory cascade. Direct evidence demonstrating which nerves modulate the in¯ammatory response of toxin A is lacking. Previously adult rats that received large doses of capsaicin were found to be protected from toxin A's effects [12]. However, it has been also demonstrated that adult sensory denervation with capsaicin is not complete, with up to 50% of sensory nerves
Fig. 2. Quanti®cation of the protective effects of extrinsic denervation on toxin A-induced congestion and edema (a), epithelial damage (b), and neutrophil in®ltration (c). The values shown are mean ^ SEM; n 5. *P , 0:001 vs. toxin A.
98
C.R. Mantyh et al. / Neuroscience Letters 292 (2000) 95±98
peptides resulting in uncontrolled in¯ammation. This provocative ®nding and technique of highly selective intestinal denervation may provide not only an important tool to elucidate the pathway of neurogenic in¯ammation but may also prove to be therapeutic. We thank Laurie M. Neville of TechLab, Inc., Blacksburg, VA for preparation of toxin A. This work was supported by National Institutes of Health grant DK-50265. Fig. 3. Effects of extrinsic denervation on basal and toxin Ainduced ileal secretion. Toxin A-stimulated ileal secretion was signi®cantly inhibited by extrinsic denervation. The values shown are mean ^ SEM; n 5. *P , 0:01 vs. toxin A.
remaining intact [7]. Additionally, adult capsaicinization results in central neuronal death in neurons that do not possess the vanilloid receptor [13]. Finally, it has been reported that immune cells release substance P after capsaicin treatment, again suggesting that capsaicin is not speci®c for sensory neurons [5]. In this study we used the technique of Furness et al. [4] in which the extrinsic nerves to a segment of intestine are carefully removed while preserving the vasculature. This technique allows for complete removal of the extrinsic enteric nervous system while preserving the intrinsic nerves found in the enteric plexi. Complete axonal death was then tested 10 days post-operatively by staining for tyrosine hydroxylase. Little to no TH staining in the denervated segments was found indicating successful extrinsic denervation. After exposure to toxin A these denervated segments demonstrated no signi®cant difference in histologic damage than observed in vehicle-injected segments. Additionally, toxin A-stimulated ileal secretion and neutrophil in®ltration were also prevented in the denervated segments. These data indicate that extrinsic sensory nerves are directly responsible for propagating toxin A in¯ammation. In conclusion, toxin A, an exotoxin secreted by C. dif®cile, appears to activate brush border receptors which through a yet undetermined mechanism then stimulate extrinsic sensory nerves to release pro-in¯ammatory neuro-
Fig. 4. Effects of extrinsic denervation on basal and toxin Ainduced ileal myeloperoxidase (MPO) activity. Toxin A-stimulated MPO activity was signi®cantly inhibited by extrinsic denervation. The values shown are mean ^ SEM; n 5. *P , 0:001 vs. toxin A.
[1] Bradley, P.P., Priebat, D.A., Christensen, R.D. and Rothstein, G., Measurement of cutaneous in¯ammation: estimation of neutrophil content with an enzyme marker, J. Invest. Dermatol., 78 (1982) 206±209. [2] Castagliuolo, I., Riegler, M., Pasha, A., Nikulasson, S., Lu, B., Gerard, C. and Gerard, N.P., Neurokinin-1 (NK-1) receptor is required in Clostridium dif®cile-induced enteritis, J. Clin. Invest., 101 (1998) 1547±1550. [3] Caterina, M.J., Schumacher, M.A., Tominaga, M., Rosen, T.A., Levine, J.D. and Julius, D., The capsaicin receptor: a heat-activated ion channel in the pain pathway, Nature, 389 (1997) 816±824. [4] Furness, J.B., Johnson, P.J., Pompolo, S. and Bornstein, J.C., Evidence that enteric motility re¯exes can be initiated through entirely intrinsic mechanisms in the guinea pig small intestine, Neurogastroenterol. Mot., 7 (1995) 89±96. [5] Ho, W.Z., Lai, J.P., Zhu, X.H., Uvaydova, M. and Douglas, S.D., Human monocytes and macrophages express substance P and neurokinin-1 receptor, J. Immunol., 159 (1997) 5654±5660. [6] Holzer, P., Neural injury, repair, and adaptation in the GI tract. II. The elusive action of capsaicin on the vagus nerve. Am. J. Physiol., 275 (1998) G8±G13. [7] Holzer, P., Capsaicin: cellular targets, mechanisms of action, and selectivity for thin sensory neurons, Pharmacol. Rev., 43 (1991) 143±201. [8] Keates, A.C., Castagliuolo, I., Qiu, B., Nikulasson, S., Sengupta, A. and Pothoulakis, C., CGRP upregulation in dorsal root ganglia and ileal mucosa during Clostridium dif®cile toxin A-induced enteritis, Am. J. Physiol., 37 (1998) G196±G202. [9] Lyerly, D.M., Roberts, M.D. and Wilkins, T.D., Puri®cation and properties of toxins A and B of Clostridium dif®cile, FEMS Microbiol. Lett., 33 (1986) 31±35. [10] Mantyh, C.R., Pappas, T.N., Lapp, J.A., Washington, M.K., Neville, L.M., Ghilardi, J.R., Rogers, S.D., Mantyh, P.W. Vigna, S.R. Substance P activation of enteric neurons in response to intraluminal Clostridium dif®cile toxin A in the rat ileum, Gastroenterology, 111 (1996) 1272±1280. [11] Pothoulakis, C., Pathogenesis of Clostridium dif®cile-associated diarrhea, Eur. J. Gastroenterol. Hepatol., 8 (1996) 1041±1047. [12] Pothoulakis, C., Castagliuolo, I., LaMont, J.T., Jaffer, A., O'Keane, J.C., Snider, R.M. and Leeman, S.E., CP-96,345, a substance P antagonist, inhibits rat intestinal responses to Clostridium dif®cile toxin A but not cholera toxin, Proc. Natl. Acad. Sci. USA, 91 (1994) 947±951. [13] Ritter, S. and Dinh, T.T., Age related changes in capsaicininduced degeneration in rat brain, J. Comp. Neurol., 318 (1992) 103±116. [14] Sullivan, N.M., Pellett, S. and Wilkins, T.D., Puri®cation and characterization of toxins A and B of Clostridium dif®cile, Infect. Immun., 15 (1982) 1032±1040.
www.elsevier.com/locate/neulet
Extrinsic surgical denervation inhibits Clostridium dif®cile toxin A-induced enteritis in rats Christopher R. Mantyh a, Douglas C. McVey b,c, Steven R. Vigna b,c,* a Department of Surgery, Durham VA Medical Center, Durham, NC 27710, USA Department of Cell Biology and Medicine, Box 3709, Duke University Medical Center, Durham, NC 27710, USA c Department of Medicine, Box 3709, Duke University Medical Center, Durham, NC 27710, USA
b
Received 20 June 2000; received in revised form 8 August 2000; accepted 9 August 2000
Abstract Clostridium dif®cile enteritis is caused by toxin A (TA) which stimulates substance P release and subsequent receptor activation. This receptor stimulation results in secretion, in¯ammation, and structural damage. However, it is unclear as to which subset of neurons is required to initiate substance P release following toxin stimulation. Five centimeter ileal segments were surgically denervated. After 10 days, three ileal loops were constructed in each rat: the denervated loop was injected intraluminally with 5 mg of TA and two intact loops were injected with TA or vehicle, respectively. Ileal secretion, myeloperoxidase activity, and histology were then assessed. Denervated ileal loops injected with TA had a 75% reduction in ileal secretion (P , 0:001), 92% reduction in myeloperoxidase activity (P , 0:01) and 96% reduction in histologic damage (P , 0:001) compared to innervated loops. There were no signi®cant differences between the denervated loops injected with TA and those injected with vehicle. Extrinsic surgical denervation results in protection of ileal loops from TA enteritis. Furthermore, these results exclude the participation of intrinsic enteric nerves in TA-induced ileal damage. Finally, this suggests that extrinsic primary sensory neurons mediate the effects of intraluminal TA in the ileum. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: In¯ammation; Toxin A; Sensory neurons; Capsaicin; Substance P
Clostridium dif®cile is an anaerobic bacterium that releases several toxins leading to acute colitis and enteritis. Toxin A is an exotoxin that has been isolated and puri®ed from C. dif®cile and causes intense intestinal in¯ammation when injected intraluminally. Toxin A-induced enteritis in experimental animals is highly reproducible and rapid allowing for careful dissection of the in¯ammatory pathway. Currently it is proposed that toxin A binds to brush border receptors on villous enterocytes and super®cial mucosal epithelial cells in the colon [11]. Following toxin A binding to surface epithelial cells, neurotransmitters such as substance P and calcitonin gene-related peptide (CGRP) are released to further perpetuate the in¯ammatory response [10,12]. Inhibition of toxin A-induced intestinal secretion and in¯ammation in the rat occurs if substance P receptor antagonists or CGRP receptor antagonists are administered prior to toxin A injection. Additionally, substance P receptor (neurokinin (NK)-1 receptor) knockout mice are protected * Corresponding author. Tel.: 11-919-684-5311; fax: 11-919286-5530. E-mail address: [email protected] (S.R. Vigna).
from the physiologic and histologic effects of toxin A [2]. Thus it appears that primary sensory neurons are required for initiation of in¯ammation in the toxin A-stimulated enteritis model. The role of sensory neurons in modulating the in¯ammatory cascade is further supported by the abolition of toxin A's effects after capsaicin injection [10,12]. Capsaicin is a potent excitotoxin that is the active ingredient in hot chili peppers and acts as an exogenous ligand of the vanilloid receptor. The capsaicin-sensitive vanilloid receptor has recently been cloned [3] and is found on nociceptive afferent neurons. Acute, small doses of capsaicin cause vanilloid receptor activation and subsequent neuropeptide release. Excessive or chronic administration of capsaicin causes eventual death of dorsal root ganglion neurons and degeneration of the afferent neuron [6]. However, not all extrinsic sensory nerves are killed by excessive capsaicin administration. It appears that newborn rat neurons are particularly sensitive to the neurodegenerative effects of capsaicin, while adult rats may only have 50% of the sensory nerves ablated by a massive capsaicin injection [7]. In addition, capsaicin is not sensory neuron-speci®c in the adult rats,
0304-3940/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 0 0) 01 45 1- 8
96
C.R. Mantyh et al. / Neuroscience Letters 292 (2000) 95±98
as demonstrated by central neuronal death following large dose of capsaicin [13]. There is an absence of immunostaining for the vanilloid receptors in these central neurons. Also peri-vagal application of capsaicin is often used to study gastrointestinal functions. However, recently it has been shown that the vanilloid receptor is absent from vagal afferents again questioning the selectivity of capsaicin [3]. Finally, capsaicin appears to release substance P from human monocytes and macrophages indicating that the vanilloid receptor may not be uniquely present on sensory neurons [5]. Therefore, capsaicin-induced death of sensory neurons remains a useful but not speci®c tool to study the physiologic roles of sensory neurons. A more selective technique to examine neuronal inputs to the intestine is to extrinsically denervate surgically a loop of small bowel in the manner ®rst described by Furness [4]. By preserving the blood supply but eliminating the extrinsic nerves, only the intrinsic nerves in the plexi remain. Following super-selective denervation, we examined the in¯ammatory response after toxin A administration in the denervated segment in comparison to rats with intact extrinsic enteric nerves. Six Sprague±Dawley rats (150±175 g; Charles River Laboratories, Raleigh, NC) were fasted overnight with free access to water and were then anesthetized with iso¯urane. A small midline incision was made and a loop of ileum 4±5 cm in length with a discrete vascular bundle was selected. Dissection under an operating microscope using ®ne forceps and scissors was then performed to remove all nerve ®bers bundle while preserving the accompanying artery and vein as previously described [4]. The vessels were then swabbed with 80% phenol in distilled water causing the remaining nerve ®bers to appear white and these were subsequently dissected away. The denervated segment of ileum was then marked with a loose 3-0 silk ligature around the vascular pedicle and the bowel returned to the abdomen. The incision was closed in two layers and the animals were supplied with ad libitum food and water. Following ten days to allow for complete axonal degeneration, the animals were then treated with toxin A. Toxin A was prepared from C. dif®cile VPI strain 10463 as previously described [14]. Brie¯y, toxin A was puri®ed by sequential ammonium sulfate precipitation, anion-exchange chromatography, and isoelectric precipitation. The toxin preparation was homogeneous by crossed immunoelectrophoresis and native polyacrylamide gel electrophoresis [9]. The denervated ileal segments were isolated and ileal loops were constructed as previously described [10]. Brie¯y, rats fasted overnight with free access to water were anesthetized by im injection of 67 mg/kg ketamine:33 mg/kg xylazine. A midline abdominal incision was made, the terminal ileum was identi®ed, and ileal loops 5 cm in length were constructed by ligation with 4-0 silk sutures, taking care not to disturb the vascular supply. Toxin A (5 mg) in 400 ml of phosphate-buffered saline (PBS) (pH 7.4) was injected into the ileal loop using a 27 ga syringe needle
as previously described [10,12]. Control rats were prepared similarly and their ileal loops were injected with PBS. The severity of mucosal histological damage was graded on formalin-®xed, paraf®n-embedded, haematoxylin & eosin (H&E)-stained sections using a scheme described previously [10]. The following features were assessed: epithelial cell damage, mucosal congestion and edema, and neutrophil margination and in®ltration of the lamina propria (Fig. 1). A score of 0±3 for each category, denoting increasingly severe abnormality, was assigned by a blinded reviewer. Myeloperoxidase (MPO) activity was measured as described previously [1]. Brie¯y, pieces of control and treated ileal loops were homogenized in 0.5% hexadecyltrimethylammonium bromide in 50 mM KH2P04 (pH 6), freeze/thawed three times, centrifuged at 48C for 2 min, and then the absorbancy of each supernatant was read at 460 nm at 0, 30 and 60 s after the addition of 2.9 ml of o-dianisidine dihydrochloride to 0.1 ml supernatant. The maximal change in absorbancy per min was used to calculate the units of MPO activity based on the molar absorbancy index of oxidized o-dianisidine of 1.13 £ 10 4 M 21 cm 21. The results are expressed as MPO U of activity/g of tissue wet weight.
Fig. 1. The effects of intraluminal toxin A administration in isolated ileal segments. H&E-stained sections of paraf®nembedded samples were prepared after ileal loops were exposed for 3 h to (a) 0.4 ml of vehicle (PBS), (b) 0.4 ml of PBS containing 5 mg of puri®ed toxin A, or (c) 0.4 ml of PBS containing 5 mg of toxin A in a section of ileum that had been extrinsically denervated 10 days prior to toxin A administration. Toxin A disrupts villous architecture, causes neutrophil in®ltration, and results in mucosal edema as manifested by wide spacing of the crypts. Denervated ileal segments displayed normal villi, no neutrophil margination, and minimal edema. Magni®cation £130.
C.R. Mantyh et al. / Neuroscience Letters 292 (2000) 95±98
The presence or absence of immunoreactive tyrosine hydroxylase was used to assess the completeness of surgical denervation. Following ®xation in 4% (w/v) paraformaldehyde solution in 100 mM phosphate buffer (pH 7.4) for 1 h at room temperature the specimens were stored overnight in 30% sucrose. Frozen sections (10 mm) were then cut on a cryostat and dried on SuperFrost Plus slides. The sections were then immersed in PBS for 5 min and washed in a buffer consisting of: sodium phosphate 10 mmol/l, NaCl 0.5 mol/l, carrageenan 0.7% (w/v), Triton X-100 0.2% (v/v), sodium azide 0.01% (w/v), pH 7.4. The sections were then washed again in PBS three times for 2 min each. The sections were overlaid with 20 ml of monoclonal mouse anti-rat tyrosine hydroxylase (Boehringer Mannheim) and incubated for 1.5 h in a humid chamber at room temperature. The slides were washed three times for 2 min each in PBS. The slides were then stained with 20 ml of anti-mouse-IgG-FITC and incubated for 1 h in a humid chamber. The slides were ®nally washed three times for 2 min in PBS and coverslipped. Results are expressed as mean ^ SEM. Differences among groups were examined by one-way analysis of variance (ANOVA) with the Dunnett's or Tukey±Kramer post tests, using GraphPad Prism version 3.00 for Windows (GraphPad Software, San Diego, CA). P-values of ,0.05 were considered signi®cant. Intraluminal toxin A consistently demonstrates intense destruction of villous architecture, mucosal edema, and neutrophil in®ltration (Fig. 1B). Surgical denervation nearly completely ablated the histologic damage caused by toxin A (Fig. 1C). No difference was found between vehicleinjected controls and rats surgically denervated then injected with toxin A in epithelial damage, congestion and edema, or neutrophil in®ltration (Fig. 2) (P , 0:001). In addition, two rats were denervated and then injected with vehicle alone; again no histologic damage was evident. When all the pathologic scores were combined, there was a collective 96% reduction in histologic damage following extrinsic denervation. Denervation of the ileal segment reduced toxin A-induced ileal secretion and neutrophil activation as measured by myeloperoxidase (MPO) activity. Surgical denervation reduces toxin A-stimulated secretion by 75% (Fig. 3; P , 0:01) and MPO activity by 92% (Fig. 4; P , 0:001). In order to con®rm complete extrinsic denervation, tyrosine hydroxylase (TH) staining was performed. TH is an essential enzyme in the synthetic pathway and consistent marker of noradrenergic axons [4]. Little to no TH staining was found in the myenteric or submucosal plexus following denervation. Toxin A is an exotoxin released from C. dif®cile that rapidly produces intense intestinal in¯ammation causing ¯uid secretion, tissue necrosis, and neutrophil in®ltration. The in¯ammatory response from highly puri®ed toxin A is extremely reproducible and reaches near maximal stimulation at 3 h. For these reasons toxin A-enteritis represents a
97
unique model to mechanistically study the in¯ammatory cascade in the intestine. It appears that the ®rst step in toxin A-induced enteritis occurs after it binds to brush border receptors on enterocytes. Toxin A binding sites have been found in hamster, rabbit, rat, and human intestine, although the exact molecular structure of the receptor has yet to be agreed on [11]. Following binding to the enterocyte, toxin A stimulates substance P and CGRP release, which then propagate the in¯ammatory response. Speci®c antagonists to the substance P and CGRP receptors ameliorate the in¯ammatory effects of toxin A [8,12]. In addition, rapid endocytosis of the substance P receptor occurs following toxin A stimulation [10], and mice constructed to be de®cient in the substance P receptor are protected from toxin A-induced enteritis [2]. Thus it has been postulated that toxin A stimulates the nervous system to release a host of pro-in¯ammatory neuropeptides which then stimulate their respective receptors to propagate the in¯ammatory cascade. Direct evidence demonstrating which nerves modulate the in¯ammatory response of toxin A is lacking. Previously adult rats that received large doses of capsaicin were found to be protected from toxin A's effects [12]. However, it has been also demonstrated that adult sensory denervation with capsaicin is not complete, with up to 50% of sensory nerves
Fig. 2. Quanti®cation of the protective effects of extrinsic denervation on toxin A-induced congestion and edema (a), epithelial damage (b), and neutrophil in®ltration (c). The values shown are mean ^ SEM; n 5. *P , 0:001 vs. toxin A.
98
C.R. Mantyh et al. / Neuroscience Letters 292 (2000) 95±98
peptides resulting in uncontrolled in¯ammation. This provocative ®nding and technique of highly selective intestinal denervation may provide not only an important tool to elucidate the pathway of neurogenic in¯ammation but may also prove to be therapeutic. We thank Laurie M. Neville of TechLab, Inc., Blacksburg, VA for preparation of toxin A. This work was supported by National Institutes of Health grant DK-50265. Fig. 3. Effects of extrinsic denervation on basal and toxin Ainduced ileal secretion. Toxin A-stimulated ileal secretion was signi®cantly inhibited by extrinsic denervation. The values shown are mean ^ SEM; n 5. *P , 0:01 vs. toxin A.
remaining intact [7]. Additionally, adult capsaicinization results in central neuronal death in neurons that do not possess the vanilloid receptor [13]. Finally, it has been reported that immune cells release substance P after capsaicin treatment, again suggesting that capsaicin is not speci®c for sensory neurons [5]. In this study we used the technique of Furness et al. [4] in which the extrinsic nerves to a segment of intestine are carefully removed while preserving the vasculature. This technique allows for complete removal of the extrinsic enteric nervous system while preserving the intrinsic nerves found in the enteric plexi. Complete axonal death was then tested 10 days post-operatively by staining for tyrosine hydroxylase. Little to no TH staining in the denervated segments was found indicating successful extrinsic denervation. After exposure to toxin A these denervated segments demonstrated no signi®cant difference in histologic damage than observed in vehicle-injected segments. Additionally, toxin A-stimulated ileal secretion and neutrophil in®ltration were also prevented in the denervated segments. These data indicate that extrinsic sensory nerves are directly responsible for propagating toxin A in¯ammation. In conclusion, toxin A, an exotoxin secreted by C. dif®cile, appears to activate brush border receptors which through a yet undetermined mechanism then stimulate extrinsic sensory nerves to release pro-in¯ammatory neuro-
Fig. 4. Effects of extrinsic denervation on basal and toxin Ainduced ileal myeloperoxidase (MPO) activity. Toxin A-stimulated MPO activity was signi®cantly inhibited by extrinsic denervation. The values shown are mean ^ SEM; n 5. *P , 0:001 vs. toxin A.
[1] Bradley, P.P., Priebat, D.A., Christensen, R.D. and Rothstein, G., Measurement of cutaneous in¯ammation: estimation of neutrophil content with an enzyme marker, J. Invest. Dermatol., 78 (1982) 206±209. [2] Castagliuolo, I., Riegler, M., Pasha, A., Nikulasson, S., Lu, B., Gerard, C. and Gerard, N.P., Neurokinin-1 (NK-1) receptor is required in Clostridium dif®cile-induced enteritis, J. Clin. Invest., 101 (1998) 1547±1550. [3] Caterina, M.J., Schumacher, M.A., Tominaga, M., Rosen, T.A., Levine, J.D. and Julius, D., The capsaicin receptor: a heat-activated ion channel in the pain pathway, Nature, 389 (1997) 816±824. [4] Furness, J.B., Johnson, P.J., Pompolo, S. and Bornstein, J.C., Evidence that enteric motility re¯exes can be initiated through entirely intrinsic mechanisms in the guinea pig small intestine, Neurogastroenterol. Mot., 7 (1995) 89±96. [5] Ho, W.Z., Lai, J.P., Zhu, X.H., Uvaydova, M. and Douglas, S.D., Human monocytes and macrophages express substance P and neurokinin-1 receptor, J. Immunol., 159 (1997) 5654±5660. [6] Holzer, P., Neural injury, repair, and adaptation in the GI tract. II. The elusive action of capsaicin on the vagus nerve. Am. J. Physiol., 275 (1998) G8±G13. [7] Holzer, P., Capsaicin: cellular targets, mechanisms of action, and selectivity for thin sensory neurons, Pharmacol. Rev., 43 (1991) 143±201. [8] Keates, A.C., Castagliuolo, I., Qiu, B., Nikulasson, S., Sengupta, A. and Pothoulakis, C., CGRP upregulation in dorsal root ganglia and ileal mucosa during Clostridium dif®cile toxin A-induced enteritis, Am. J. Physiol., 37 (1998) G196±G202. [9] Lyerly, D.M., Roberts, M.D. and Wilkins, T.D., Puri®cation and properties of toxins A and B of Clostridium dif®cile, FEMS Microbiol. Lett., 33 (1986) 31±35. [10] Mantyh, C.R., Pappas, T.N., Lapp, J.A., Washington, M.K., Neville, L.M., Ghilardi, J.R., Rogers, S.D., Mantyh, P.W. Vigna, S.R. Substance P activation of enteric neurons in response to intraluminal Clostridium dif®cile toxin A in the rat ileum, Gastroenterology, 111 (1996) 1272±1280. [11] Pothoulakis, C., Pathogenesis of Clostridium dif®cile-associated diarrhea, Eur. J. Gastroenterol. Hepatol., 8 (1996) 1041±1047. [12] Pothoulakis, C., Castagliuolo, I., LaMont, J.T., Jaffer, A., O'Keane, J.C., Snider, R.M. and Leeman, S.E., CP-96,345, a substance P antagonist, inhibits rat intestinal responses to Clostridium dif®cile toxin A but not cholera toxin, Proc. Natl. Acad. Sci. USA, 91 (1994) 947±951. [13] Ritter, S. and Dinh, T.T., Age related changes in capsaicininduced degeneration in rat brain, J. Comp. Neurol., 318 (1992) 103±116. [14] Sullivan, N.M., Pellett, S. and Wilkins, T.D., Puri®cation and characterization of toxins A and B of Clostridium dif®cile, Infect. Immun., 15 (1982) 1032±1040.