- Email: [email protected]
Protective effects of ex vivo graft radiation and tacrolimus on syngeneic transplanted rat small bowel motility
Protective effects of ex vivo graft radiation and tacrolimus on syngeneic transplanted rat small bowel motility Nicolas T. Schwarz, MD, Atsunori Nakao, MD, Michael A. Nalesnik, MD, Jörg C. Kalff, MD, Noriko Murase, MD, and Anthony J. Bauer, PhD, Pittsburgh, Pa
Background. Intestinal transplantation is unduly complicated by the nontolerogenic properties of the gut-associated lymphoid tissue. Because simultaneous graft irradiation and bone marrow infusion significantly prolong the survival of the small bowel transplanted animal, our objective was to determine the functional motility effects of the immune modulating, graft irradiation procedure in the presence and absence of tacrolimus immunosuppression. Methods. Four groups of syngeneic orthotopic small bowel transplanted animals were studied 48 hours after operations (untreated, tacrolimus, ex vivo graft irradiation, and tacrolimus + irradiation) and compared with controls. Histologic analysis was performed for mucosal apoptosis and neutrophilic infiltration into the muscularis externa. Gastrointestinal in vivo transit and in vitro circular muscle strip contractions were quantified in response to bethanechol (0.3-300 µmol/L). Results.Graft irradiation ex vivo alone or in the presence of tacrolimus significantly increases (> 10-fold) the number of apoptotic mucosal cells after transplantation. Functional measurements showed that transplantation resulted in a significant delay in gastrointestinal transit and a decrease in muscle strip contractility. Tacrolimus and graft irradiation significantly ameliorated the transplantinduced dysfunction. Conclusions.Given the endowed propensity of mucosal regeneration, the immunologic and functional benefits of ex vivo graft irradiation appear to outweigh the detrimental effects to the mucosa. (Surgery 2002;131:413-23.) From the Departments of Surgery and Medicine/Gastroenterology, University of Pittsburgh Medical Center, Pittsburgh, Pa
CLINICAL SMALL INTESTINAL TRANSPLANTATION is a procedure that could potentially improve the quality of life of an estimated 20,000 patients in the United States who are currently on total parenteral nutrition at a cost in excess of $280 per day.1 Progress in immunosuppression, organ preservation, and surgical techniques has proven that the transplanted intestine can provide life-sustaining nutrients to the recipient if it can successfully endure the transplant process and the continuous insults of rejection.2-4 However, in comparison with other organs, the intestine remains the Supported by National Institutes of Health grants R01-GM58241, R01-DK-AI54232, and P50-GM-53789, and a grant from the Deutsche Forschungsgemeinschaft Schw 745 1/1 to Nicolas T. Schwarz. Accepted for publication December 9, 2001. Reprint requests: Anthony J. Bauer, PhD, Department of Medicine/Gastroenterology, S-849 Scaife Hall, 3550 Terrace St, University of Pittsburgh Medical School, Pittsburgh, PA 15261. Copyright © 2002 by Mosby, Inc. All rights reserved. 0039-6060/2002/$35.00 + 0 doi:10.1067/msy.2002.122372
11/56/122372
“Achilles heel” of transplants. We have hypothesized that a possible reason for this is the unique immunologic complexities of the gut-associated lymphoid tissue (GALT). In contrast to other organs, the intestine is a rich source of mature, relatively nontolerogenic leukocytes that predispose the graft to rejection and the recipient to graftversus-host disease.5 Recently, we have shown that modification of intestinal passenger leukocytes improves the outcome of the transplanted intestinal allograft.6 This immunomodification was accomplished by the ex vivo irradiation of the intestinal allograft at doses designed to eradicate the mature nontolerogenic lymphoid elements from the intestine.6 In addition, the procedure consisted of a simultaneous intravenous infusion of donor bone marrow cells to replace the irradiated intestinal mature lymphocytes. Hypothetically, this procedure should result in a microchimeric animal with immunotolerant leukocytes, a condition that favors graft acceptance. Indeed, the combined procedure of allograft irradiation and bone marrow infusion SURGERY 413
414 Schwarz et al
increased the survival of the animals that underwent small bowel transplantation to more than 150 days compared with 51 days for untreated allotransplanted animals. The combined procedure also prevented graft-versus-host disease.6 Although irradiation can efficiently eliminate mature lymphocytes from allografts, it can also produce unpropitious immediate and late damaging effects to the endothelium, mucosa, nerves, and smooth muscle of a graft. Because immunomodulation of intestinal transplantation with ex vivo graft irradiation and simultaneous donor bone marrow infusion appears to have potential therapeutic benefits in regards to rejection,6 the objectives of this study were to determine if the acute beneficial effects of irradiation-induced immunomodulation on graft survival are attained at the expense of alterations in intestinal motility and mucosal structure. We addressed these objectives by using a syngeneic orthotopic small intestinal transplant model in which the individual and combined effects of tacrolimus and ex vivo graft irradiation were determined on routine histopathologic examination, in vivo gastrointestinal transit, in vitro circular muscle contractility, and leukocyte recruitment. Syngeneic transplants were used to avoid any complications that could arise from allogeneic immune reactions. METHODS Small intestinal transplantation. Male Lewis rats (RT1l) (Harlan Sprague Dawley, Indianapolis, Ind) weighing 200 to 300 g were used. The animals were housed in the University of Pittsburgh animal care facilities. While the rats were under methoxyflurane anesthesia, orthotopic small intestinal transplantation with caval drainage was performed in a syngeneic combination. Briefly, the small intestine of the donor was isolated from the ligament of Treitz to the ileocecal valve, with its vascular pedicle consisting of the superior mesenteric artery with a piece of aorta and the superior mesenteric vein. Then the graft vascular bed was perfused with chilled lactated Ringer’s solution and the intestinal lumen was irrigated with 20 mL of cold neomycinsulfate (Sigma, St Louis, Mo) saline solution. The graft was stored briefly in cold Ringer’s solution during the preparation of the recipient. An end-toside anastomosis between the graft aorta and the recipient’s infrarenal aorta, and between the graft vein and the recipient’s vena cava, was performed with 10-0 polybutester (Novafil) sutures. The time required to complete the vascular anastomosis was 20 to 30 minutes. The entire recipient intestine was removed, and enteric continuity was restored by
Surgery April 2002
proximal and distal end-to-end intestinal anastomoses. Animals were prophylactically given the antibiotic cefamandole nafate (20 mg/d). All animals were killed 48 hours after transplantation. Experimental groups. Five groups of animals were examined in this study, 4 of which received syngeneic small bowel transplants. Group 1 consisted of unoperated controls. Transplant recipients in group 2 did not receive any treatment and served as transplant controls. In group 3, tacrolimus (FK506; Fujisawa Pharmaceutical, Osaka, Japan) was intramuscularly administrated to recipients 4 and 28 hours after transplantation at a dosage of 1.0 mg/kg. In recipients of group 4, the harvested intestinal graft, stored in cold lactate Ringer’s solution in a 50 mL Falcon tube, was ex vivo irradiated at 10 Gy from a 137Cs source (Gammacell 1000 Elite; Nordion International, Ontario, Canada) and transplanted. Total time to achieve this irradiation dose was 2.5 minutes. Animals in experimental group 5 received a combined treatment of tacrolimus and ex vivo graft irradiation. Histopathologic analysis. Specimens from each intestinal graft including Peyer’s patches and mesenteric lymph nodes were taken and fixed in formalin. Samples were embedded in paraffin and sliced as 6-µm cross-sections of the entire gut wall, then stained with hematoxylin and eosin. Mucosal apoptotic cells were counted in hematoxylin and eosin cross-sections as a routine histopathologic analysis. Slides were reviewed blindly by one of the authors (M.A.N) without knowledge of the experimental groups. Whole-mounts of the intestinal muscularis were investigated for the presence of recruited neutrophils. Mid-jejunal segments were cut from the bowel and immersed in Krebs Ringer’s buffer (KRB) in a chilled and Sylgard- (Dow Corning, Midland, Mich) bottomed glass dish, as described previously.7 The length and width of each jejunal segment were measured with a caliper, and then the segment was gently pinned down along the mesenteric border. The bowel was opened along the mesentery and stretched to 150% of the length and 250% of the width. The opened segments of jejunum were fixed in 100% ethanol for 10 minutes. Each segment was washed twice in KRB, and mucosa and submucosa were stripped off under microscopic observation (Wild, Heerbrugg, Switzerland). The mucosa-free muscularis wholemounts were finally cut into 1 3–cm pieces and used for staining procedures. Polymorphonuclear neutrophils were visualized by a myeloperoxidase stain: freshly prepared whole-mounts were immersed in a mixture of 10 mg of Hanker-Yates
Surgery Volume 131, Number 4
reagent (Polysciences, Warrington, Pa), 10 mL of KRB, and 100 µL of 3% hydrogen peroxide for 20 minutes. The reaction was stopped with cold KRB. Whole-mounts were cover-slipped and inspected by light microscopy after staining (Nikon FXA; Fryer, Huntley, Ill). Leukocytes were counted in 5 randomly chosen areas in each specimen at a magnification of 200. Intestinal muscle function. In vivo gastrointestinal transit was measured in controls and transplanted animals at 48 hours postoperatively by evaluating the intestinal location of fluorescein isothiocyanate (FITC)-labeled dextran (70000 MW). Animals were lightly anesthetized with methoxyflurane and fed FITC-labeled dextran (200 µL of 25 mg/mL stock solution) orally. Two hours after administration, the contents of the entire small bowel, divided into 10 equal segments, cecum, proximal colon, and distal colon were vortexed with 2 mL of saline solution to obtain a supernatant containing the fluorescent marker that was present within the lumen of each bowel segment. The supernatant was centrifuged at 7000 rpm to pellet the intestinal chyme and duplicate aliquots of the cleared supernatant were read in a CytoFluor II (PerSeptive Biosystems, Framingham, Mass) multi-well fluorescence plate reader (excitation 485 ± 30 and emission 530 ± 20 nm) for quantification of the fluorescent signal in each bowel segment. A median histogram of the fluorescence was then plotted for analysis of transit. This provided an accurate, nonradioactive measurement of liquid gastrointestinal transit without further surgical intervention. 8 The data were also calculated to obtain the geometric center (GC) of distribution of fluorescein-labeled dextran to quantify transit with the use of the following formula: GC = Σ (% of total fluorescent signal per segment segment number)/100. The GC value reflects the transit of fluorescently labeled dextran down the gastrointestinal tract, with higher values indicating more distal distribution (segment 1 = stomach; segment 15 = distal colon). In vitro circular muscle mechanical activity was assessed with full thickness strips obtained from the middle jejunum. Muscle strips were prepared from an opened segment of the jejunum by making 2 cuts (1 10 mm) parallel to the circular muscle layer. Each strip was then fixed in a mechanical organ chamber with 1 end pinned to the Sylgard floor and the other attached to an isometric force transducer (WPI, Sarasota, Fla) by 4-0 silk sutures.
Schwarz et al 415
Each chamber was continuously perfused with a preheated, preoxygenated KRB solution that was constantly monitored and maintained at 37° ± 0.5°C. Once in the organ chamber, each strip was allowed to equilibrate for 1 hour. It was then stretched in small increments to determine the length at which maximal spontaneous contractions of the muscle occurred (L0). After determination of L0, dose-response curves were generated with sequentially increasing doses of bethanechol (0.1300 µmol/L) for 10 minutes with intervening KRB wash periods of 15 minutes. Contractions between different muscle strips were normalized by converting grams of contraction to grams/millimeters2 of tissue. This was done by determining the cross-sectional area by the following equation (muscle density assumed to be 1.03 mg/mm3): mm2 = (wet muscle weight [in mg]/muscle length [in mm] muscle density [mg/mm3]).9,10 Contractions were recorded, measured, and stored in a computer with an A/D hardware and software package (Biopac Systems, Santa Barbara, Calif). Statistical analysis. Results are expressed as mean ± SEM. Data were statistically analyzed with an unpaired Student t test. A probability level of P < .05 was considered significant. RESULTS General observations. All transplanted animals throughout the study period had a healthy appearance and were able to drink within 6 hours after the operation. Animal body weight changed to a similar extent in all 4 groups of animals after syngeneic transplantation (P > .05, N = 7). All the animals showed an approximate 5% decrease in weight because of the transplantation procedure regardless of treatment. This amount of body weight loss is typical after intestinal transplantation in the rat.6 At the time the animals were killed, the 2-day transplanted grafts looked slightly edematous with a rosy appearance macroscopically. Routine pathologic analysis showed essentially normal intestinal architecture. The mucosa of the untreated transplanted grafts 48 hours after transplantation had only 0.5 ± 0.58 apoptotic cells per 10 glands. Similarly, animals treated postoperatively with tacrolimus showed 0.7 ± 0.58 apoptotic cells per 10 glands. In contrast, animals treated with radiation alone or in combination with tacrolimus displayed a significant increase in apoptotic cells 48 hours after transplantation (P < .01, 6.3 ± 3.23, and 7.7 ± 2.14 cells per 10 glands, respectively). In vivo transit after transplantation. As we have previously shown,8 normal gastrointestinal transit of nonabsorbable FITC-labeled dextran accumu-
416 Schwarz et al
Surgery April 2002
Fig 1. Transplantation significantly delays gastrointestinal transit. Transit histograms for distribution of nonabsorbable FITC-labeled dextran along gastrointestinal tract 2 hours after oral administration. In control animals (dark bars), most of fluorescent marker accumulated in last 3 segments of small bowel (sb8, sb9, sb10). In contrast, transplantation caused significant delay in the transit of the fluorescence. Two days after transplantation (cross-hatched bars), orally fed marker was distributed primarily in proximal regions of the small bowel (sb2, sb3, and sb4). Data represent averaged percent distribution of fluorescence intensity from 4 animals. SB-Tx, Small bowl transplantation; st, stomach; sb, small bowel; cm, cecum; c, colon.
lates in the last 3 distal segments of the small intestine after 2 hours of oral administration demonstrating a GC of 9.1 ± 0.32 (Fig 1). Forty-eight hours after syngeneic transplantation of the small bowel, 2-hour transit of the FITC-labeled dextran was significantly delayed (GC = 4.3 ± 0.17). In these untreated transplanted animals, the fluorescent marker emptied out of the stomach and made it past the site of bowel anastomosis (located in the first jejunal segment). However, the marker traversed the bowel only to the proximal regions of the jejunum (Fig 1). In the untreated syngeneic transplanted animal, transit returned to a normal pattern of fluorescent marker distribution after 8 days (GC = 8.9 ± 0.27, N = 4). Tacrolimus has been reported to lessen ischemia and reperfusion injury to the liver,11 in addition to its primary usage as an immunosuppressive agent for allogeneic organ transplantation. Therefore, in this syngeneic model, we investigated the functional in vivo effects of tacrolimus on syngeneic transplant- induced intestinal ileus. As shown in Fig 2, A, tacrolimus significantly improved the transplantinduced delay in FITC-labeled dextran transit through the small intestine. Under treatment with this immunosuppressive agent, gastrointestinal transit improved to closely resemble that observed in control unoperated animals with the majority of the fluorescent signal accumulating again in the terminal regions of the small intestine (GC = 8.8 ± 0.64). The objective of the next series of experiments
was to determine the effect of low dose graft irradiation on transit in the absence of the rejection phenomenon. Ex vivo graft irradiation unexpectedly improved transit in the transplanted bowel (Fig 2, B) compared with the untreated group. In fact, like tacrolimus, irradiation significantly averted the transplantation-induced delay in transit with movement of the marker to the distal regions of the small intestine over a period of 2 hours (GC = 9.5 ± 0.98). Additionally, we determined the combined effects of tacrolimus and graft irradiation on transit, but no apparent combined supplemental protective effect was observed (Fig 2, C). The combined treatment group displayed a GC of 7.8 ± 0.93, which was a significant improvement over that of the untreated transplanted animals. However, because transit had returned to near control levels with individual treatments, an additive beneficial effect would have been hard to demonstrate. In vitro muscle contractility after transplantation. Alterations in intestinal muscle function in vitro because of syngeneic transplantation and the modifying effects of tacrolimus and graft irradiation were assessed on jejunal circular muscle contractions. Circular muscle strips from control Lewis rats generated spontaneous contractions at a frequency of 34.3 ± 0.5 per minute, and mean spontaneous contractile force was calculated to be 0.019 ± 0.0083 g per mm2 per second (N = 4). Muscles from syngeneic transplanted grafts harvested on postoperative day 2 generated spontaneous contractions at a similar
Surgery Volume 131, Number 4
Schwarz et al 417
Fig 2. Tacrolimus and ex vivo graft irradiation significantly prevented delay in gastrointestinal transit caused by transplantation (gray bars, N = 4 each). Transit histograms for distribution of nonabsorbable FITC-labeled dextran along gastrointestinal tract 2 hours after oral administration. In transplanted tacrolimus treated animals (A, dark bars), most fluorescent marker accumulated in last segments of small bowel, which was similar to control unoperated animals. Additionally, ex vivo graft irradiation alone (B, dark bars) or combined with tacrolimus (C, dark bars) significantly prevented delay in marker movement down transplanted intestine. SBTx, Small bowel transplantation; st, stomach; sb, small bowel; cm, cecum; c, colon.
frequency (32.7 ± 0.8/min), but the amplitude of spontaneous contractions was significantly decreased compared with controls (0.003 ± 0.0095 g/mm2/s, N = 6) (leading contractile traces from Fig 3, A and B). In addition to spontaneous activity, bethanechol-stimulated dose response curves also showed significant impairment of the transplanted muscularis (1.059 ± 0.0485 vs 0.456 ± 0.0558 g/mm2/s with a bethanechol concentration of 100 µmol/L g/mm2/s, N = 4) (Fig 4). Correlating with the beneficial effect of
tacrolimus on in vivo transit measurements, in vitro bethanechol-stimulated muscle contractility of tacrolimus-treated animals was significantly improved compared with the untreated transplanted muscularis, as shown in the raw mechanical trace in Fig 3, C. Similarly, bethanechol-stimulated dose response curves showed significant improvement in animals treated with tacrolimus compared with muscle activity from untreated grafts (Fig 4). Statistical analysis of these data showed that not only was there a significant improvement com-
418 Schwarz et al
Surgery April 2002
Fig 3. Representative in vitro jejunal circular smooth muscle contractile activity recorded in response to bethanechol (100 µmol/L) from control, transplanted, and tacrolimus treated transplanted animals. Control muscles exhibited robust spontaneous activity and large phasic and tonic responses to bethanechol at higher concentrations (A). Spontaneous activity and response to bethanechol was significantly suppressed in muscles harvested from untreated transplanted animals (B). Tacrolimus treatment of the transplanted animals showed a significant improvement in spontaneous and bethanechol-stimulated contractions (C). SB-Tx, Small bowel transplantation.
pared with untreated animals, but that bethanechol responses at higher concentrations were not significantly different from muscles of control animals. Next, the effect of graft irradiation was tested to determine its effect on in vitro muscle strip contractility. As predicted by the transit measurements and shown in Fig 4, graft irradiation resulted in a significant improvement in muscle strip contractility after syngeneic transplantation compared with the muscles harvested from the untreated transplanted grafts. Additionally, the combined effects of tacrolimus and graft irradiation on circular muscle contractility were determined. The combined treatment of the graft also resulted in improved muscle performance. However, as shown in Fig 4, both irradiated-treated groups did not reach the
level of the controls or the graft muscles treated by tacrolimus alone. Leukocyte recruitment. Fig 5 shows that syngeneic transplantation in the absence of rejection results in the massive recruitment of neutrophils into the intestinal muscularis 48 hours after transplantation. We have previously shown a correlation and direct mechanistic link between neutrophilic infiltration and muscle contractility in a model of intestinal postoperative ileus.7 Therefore, we sought to determine if tacrolimus and graft irradiation would result in a decrease in the leukocytic infiltrate that is initiated by the trauma of transplantation. Counts of neutrophils demonstrated that transplantation caused a 14-fold increase in the presence of neutrophils within the graft mus-
Surgery Volume 131, Number 4
Schwarz et al 419
Fig 4. Effect of tacrolimus and ex vivo graft irradiation on in vitro–generated standard bethanechol dose response curves recorded from jejunal circular smooth muscle strips (N = 4 each). Control jejunal circular muscle strips showed a dose-dependent increase in contractile area in response to bethanechol (filled square). This activity was significantly diminished in muscles harvested 48 hours after transplantation (filled circle). Tacrolimus (open circle) treatment of the transplanted animal completely reversed the suppression caused by transplantation. Significant improvement was also measured in transplanted animals after ex vivo graft irradiation, either alone (diamond), or in combination with tacrolimus administration (inverted triangle). SB-Tx, Small bowel transplanted.
cularis externa. Furthermore, as shown in Fig 6, the administration of tacrolimus and graft irradiation, alone or in combination, significantly decreased the recruitment and extravasation of polymorphonuclear neutrophils into the jejunal intestinal muscularis. DISCUSSION A successful small intestinal transplant would be the ultimate therapeutic option for patients with intestinal failure. Accordingly, the transplanted donor intestine needs to efficiently digest and absorb a wide variety of nutrients to sustain life. Although complex steps including enzymatic digestion and epithelial absorption are important to proper digestion, these steps are dependent on the generation of specific spatial and temporal motility patterns to mix and propel the ingested food at a rate that facilitates digestion and absorption. The major obstacle to successful intestinal transplantation continues to be graft rejection. 2,4 And although the mucosa has a remarkable ability to regenerate, the historical accumulation of rejection episodes are stored within the vasculature, intestinal muscularis, and enteric nervous system,12,13 structures which are crucial to motility. Undoubtedly, recent advances in immunosuppressive treatments have improved the outcome of intestinal transplantation. 2,4 However, heavy immunosuppression is still required for small bowel transplantation and, therefore, substantial concern remains over subsequent infectious sus-
ceptibility and lymphoproliferative disease.2,14 Our recent development of an immunomodulating strategy that makes use of ex vivo graft irradiation and the simultaneous injection of donor bone marrow is aimed to solve these problems and to improve the outcome of small bowel transplantation.6 The rationale for this immunomodification procedure is based on the theoretical concept of microchimeric symbiosis.15 The facilitation of microchimerism in clinical organ transplantation with adjunct donor bone marrow infusion is currently being evaluated.16 Although still controversial, some evidence supports that leukocytic chimerism is beneficial to the success of both liver and bone marrow transplantation.17 However, in the case of the small intestine, the nontolerogenic properties of the mature leukocytes in the GALT uniquely predispose the transplanted bowel to immunologic difficulties.4 Our immunomodification procedure is designed to remove the nontolerogenic GALT leukocytes with irradiation and gain the benefit of microchimerism through a simultaneous bone marrow infusion. In a recent publication, with the use of the immunomodification model, we have demonstrated that this procedure creates a chimeric animal with immunotolerant leukocytes. The use of this procedure resulted in a significantly improved outcome for the intestinal allografts and also the prevention of graft-versus-host disease.6 Thus, this approach seems immunologically sound. Although irradiation can efficiently eliminate
420 Schwarz et al
Surgery April 2002
Fig 5. Typical histochemically stained full thickness muscularis whole-mounts with use of the Hanker-Yates reagent for the presence of myeloperoxidase positive leukocytes extravasated into muscularis. Panel A shows typical muscularis whole-mount from normal animal, which was virtually devoid of myeloperoxidase positive leukocytes. Panel B shows presence of numerous extravasated leukocytes within muscularis whole-mount obtained from animal 48 hours after intestinal transplantation. Original magnification for both wholemounts was 100. SB-Tx, Small bowel transplanted.
mature lymphocytes from allografts, it also has the potential to produce unpropitious immediate and late damaging effects to the vasculature, mucosa, enteric nervous system, and smooth muscle. The complications of oncologic radiotherapy have been extensively studied during the treatment of malignant neoplasms with maximally tolerated irradiation doses, usually exceeding 20 Gy. These studies show early and late complications in the gastrointestinal tract, especially in the small intestine, one of the most radiosensitive organs. Early radiation effects in the intestine reflect epithelial stem cell injury and later progressive endarteritis develops. Irradiation is also potentially harmful to the neuromuscular apparatus of the intestine, causing a disturbance in normal motility patterns.18,19 However, ex vivo irradiation of an excised organ is considered to be less harmful compared with irradiation of intact tissues, and the single 10 Gy dose applied
in this study is below the dosages commonly used in oncology. In this study, we observed that ex vivo irradiation did significantly increase the number of apoptotic enterocytes over non-irradiated control grafts 48 hours after transplantation. However, because of the regenerative capacity of the mucosa, this did not seem to be a major limiting factor of this procedure. Progressive vascular complications have also been detected with oncologic doses of irradiation and are well known to be associated with late morbidity.20,21 As with other radiation-induced injuries, radiationinduced arteriopathy is basically dose dependent. However, the ex vivo dose of 10 Gy used in this study did not induce alterations in the vasculature in the syngeneic transplantation model.6 Potentially of greater importance would be the detrimental effect of irradiation on the neuromuscular apparatus of the transplanted intestine and
Surgery Volume 131, Number 4
Schwarz et al 421
Fig 6. Histogram quantifying the number of extravasated neutrophils within full thickness whole-mounts of jejunal muscularis externa of specimens from controls, small bowel-transplanted (SB-Tx), tacrolimus treated, graft ex vivo irradiated, and combined graft irradiation with tacrolimus treatment. As reflected in histological analysis, transplantation results in significant cellular inflammatory response within the muscularis. Immunomodulation of animals with tacrolimus or graft irradiation, either alone or in combination, significantly decreased number of myeloperoxidase positive cells, which extravasated into muscularis in response to transplantation. However, the muscularis of all transplanted animal groups had significantly more myeloperoxidase cells present per 200 field compared with control. Asterisk, Compared with control; single dagger, compared with untreated SB-Tx, P < .05, N = 4 each.
the induction of sustained motility abnormalities. Changes in intestinal motility during single or fractionated doses of abdominal irradiation have been reported in large and small animal models. Radiation-induced disturbances in motility that occur relatively early after exposure have been hypothesized to be responsible for gastrointestinal symptoms such as nausea, vomiting, and diarrhea.18,19,22 However, our in vivo gastrointestinal transit experiments show that graft irradiation and tacrolimus treatment significantly improved transit. In the treated animals, the orally administered nonabsorbable fluorescent marker moved down the gastrointestinal tract and reached to the last segments of the ileum, displaying a distribution pattern similar to nonmanipulated control animals. On the other hand, untreated animals showed a dramatic suppression of the in vivo movement of the nonabsorbable fluorescent marker down the gastrointestinal tract. These data show that the intestinal transplantation procedure itself causes a significant dysfunction in normal intestinal motility, which is at least partially ameliorated with graft irradiation and tacrolimus. Similarly, the in vitro contractile responses of isolated jejunal circular smooth muscle strips in response to bethanechol were significantly improved by irradiation and tacrolimus treatment, either alone or in combination. These results may further suggest that, in contrast to the mucosa,
muscle function is more severely injured by the transplantation procedures than by ex vivo graft irradiation with 10 Gy. The beneficial effects of irradiation in this study, in contrast to reports in the literature,18,19,22 may be explained by the fact that ex vivo irradiation of an excised organ is considered to be less harmful compared with in vivo irradiation of tissues and the low level of irradiation that we used. As shown here, and in our previous studies, muscle dysfunction and abnormal intestinal motility temporally correlates with the extravasation of neutrophils and monocytes into the intestinal muscle layer.7 Therefore, it was not surprising to observe that the functional improvement of the treated transplanted grafts was associated with a decreased number of neutrophils within the graft muscularis externa. Tacrolimus is a potent immunosuppressive reagent for the prevention of allograft rejection. It has also been shown to have protective effects in ischemia and reperfusion injuries of the heart, liver, and intestine.11,23,24 The mechanism of this protective effect appears to work through the blockade of nuclear factor-κB activation, which results in a blunting of the increased expression of intercellular adhesion molecule-1 mRNA and, therefore, a subsequent reduction in the number of recruited leukocytes compared with the number in untreated animals.11 Additionally in the intestine, tacrolimus was shown to reduce elevated
422 Schwarz et al
mucosal Leukotriene B4 levels, which are normally observed after intestinal reperfusion. A growing body of literature suggests that tacrolimus may be an important agent in modulating neutrophil infiltration in acute inflammatory conditions.24-27 The mechanisms that are responsible for the observed significant ameliorative effects of ex vivo graft irradiation are currently unknown. However, it has been shown that low-level laser irradiation will reduce the expression of the inflammatory mediator cyclooxygenase-2 and the subsequent production of prostaglandin E2 in gingival tissues.28 We can also speculate that irradiation will result in a decrease in the expression of the inducible synthases inducible macrophage-type nitric oxide synthase and cyclooxygenase-2 within the donor intestinal muscularis. Because nitric oxide and prostanoids play a significant role in the inflamed gut,29-33 a decrease in the release of these potent inflammatory and kinetically active factors from the irradiated resident muscularis macrophages and a decreased number of infiltrating leukocytes would result in improved intestinal function.
Surgery April 2002
4.
5.
6.
7.
8.
9.
10.
CONCLUSION In summary, graft irradiation ex vivo alone or in the presence of tacrolimus produced a significant increase in the number of apoptotic enterocytes after transplantation. However, functional motility measurements showed that treatment with tacrolimus and ex vivo graft irradiation, either alone or combined, significantly ameliorated the transplant-induced dysfunction in motility. We conclude from this and our previous studies that immunomodulation of both the graft with ex vivo irradiation and of the host with donor bone marrow appears to result in substantial benefit for graft and host survival, and significantly improved intestinal motility. And given the endowed propensity of the mucosa for regeneration, the early immunologic and functional benefits of ex vivo graft irradiation appear to outweigh the immediate detrimental effects to the mucosa. Although these short-term irradiation results are encouraging, the prolonged effects of graft irradiation have yet to be determined. REFERENCES 1. Howard L, Ament M, Fleming CR, Shike M, Steiger E. Current use and clinical outcome of home parenteral and enteral nutrition therapies in the United States. Gastroenterology 1995;109:355-65. 2. Grant D. Intestinal transplantation: 1997 report of the international registry. Intestinal Transplant Registry. Transplantation 1999;67:1061-4. 3. DiMartini A, Rovera GM, Graham TO, Furukawa H, Todo S, Funovits M, et al. Quality of life after small intestinal trans-
11.
12.
13.
14.
15.
16.
17.
18.
19.
plantation and among home parenteral nutrition patients. JPEN J Parenter Enteral Nutr 1998;22:357-62. Abu-Elmagd K, Reyes J, Todo S, Rao A, Lee R, Irish W, et al. Clinical intestinal transplantation: new perspectives and immunologic considerations. J Am Coll Surg 1998;186: 512-5. Murase N, Starzl TE, Tanabe M, Fujisaki S, Miyazawa H, Ye Q, et al. Variable chimerism, graft-versus-host disease, and tolerance after different kinds of cell and whole organ transplantation from Lewis to brown Norway rats. Transplantation 1995;60:158-71. Murase N, Ye Q, Nalesnik MA, Demetris AJ, Abu-Elmagd K, Reyes J, et al. Immunomodulation for intestinal transplantation by allograft irradiation, adjunct donor bone marrow infusion, or both. Transplantation 2000;70:1632-41. Kalff JC, Carlos TM, Schraut WH, Billiar TR, Simmons RL, Bauer AJ. Surgically induced leukocytic infiltrates within the rat intestinal muscularis mediate postoperative ileus. Gastroenterology 1999;117:378-87. Kalff JC, Buchholz BM, Eskandari MK, Hierholzer C, Schraut WH, Simmons RL, et al. Biphasic response to gut manipulation and temporal correlation of cellular infiltrates and muscle dysfunction in rat. Surgery 1999;126: 498-509. Eskandari MK, Kalff JC, Billiar TR, Lee KK, Bauer AJ. LPSinduced muscularis macrophage nitric oxide suppresses rat jejunal circular muscle activity. Am J Physiol 1999;277: G478-86. Kalff JC, Schraut WH, Billiar TR, Simmons RL, Bauer AJ. Role of inducible nitric oxide synthase in postoperative intestinal smooth muscle dysfunction in rodents. Gastroenterology 2000;118:316-27. Squadrito F, Altavilla D, Squadrito G, Saitta A, Deodato B, Arlotta M, et al. Tacrolimus limits polymorphonuclear leucocyte accumulation and protects against myocardial ischaemiareperfusion injury. J Mol Cell Cardiol 2000; 32:429-40. Heeckt PF, Lee KKW, Halfter WM, Schraut WH, Bauer AJ. Functional impairment of enteric smooth muscle and nerves caused by chronic intestinal allograft rejection regresses after FK506 rescue. Transplantation 1995;59: 159-64. Heeckt PF, Halfter W, Schraut WH, Bauer AJ. Subclinical chronic rejection leads to an early destruction of the intrinsic nervous system in rat small bowel transplants. Am J Physiol 1997;36:G413-21. Sigurdsson L, Reyes J, Kocoshis SA, Mazariegos G, AbuElmagd K, Green M. Bacteremia after intestinal transplantation in children correlates temporally with rejection or gastrointestinal lymphoproliferative disease. Transplantation 2000;70:302-5. Starzl TE, Demetris AJ, Murase N, Ildstad S, Ricordi C, Trucco M. Cell migration, chimerism, and graft acceptance. Lancet 1992;339:1579-82. Fontes P, Rao AS, Demetris AJ, Zeevi A, Trucco M, Carroll P, et al. Bone marrow augmentation of donor-cell chimerism in kidney, liver, heart, and pancreas islet transplantation. Lancet 1994;344:151-55. Thomson AW, Lu L, Murase N, Demetris AJ, Rao AS. Microchimerism, dendritic cell progenitors and transplantation. Stem Cells 1995;13:622-39. Fraser R, Frisby C, Blackshaw LA, Schirmer M, Howarth G, Yeoh E. Small intestinal dysmotility following abdominal irradiation in the rat small intestine. Neurogastroenterol Motil 1998;10:413-9. Erickson BA, Otterson MF, Moulder JE, Sarna SK. Altered motility causes the early gastrointestinal toxicity of irradiation. Int J Radiat Oncol Biol Phys 1994;28:905-12.
Surgery Volume 131, Number 4 20. Quarmby S, Kumar P, Kumar S. Radiation-induced normal tissue injury: role of adhesion molecules in leukocyteendothelial cell interactions. Int J Cancer 1999; 82:385-95. 21. Hopewell JW, Calvo W, Jaenke R, Reinhold HS, Robbins ME, Whitehouse EM. Microvasculature and radiation damage. Recent Results Cancer Res 1993;130:1-16. 22. Summers RW, Maves BV, Reeves RD, Arjes LJ, Oberley LW. Irradiation increases superoxide dismutase in rat intestinal smooth muscle. Free Radic Biol Med 1989;6:261-70. 23. Garcia-Criado FJ, Palma-Vargas JM, Valdunciel-Garcia JJ, Toledo AH, Misawa K, Gomez-Alonso A, et al. Tacrolimus (FK506) down-regulates free radical tissue levels, serum cytokines, and neutrophil infiltration after severe liver ischemia. Transplantation 1997;64:594-8. 24. Kubes P, Hunter J, Granger DN. Effects of cyclosporin A and FK506 on ischemia/reperfusion-induced neutrophil infiltration in the cat. Dig Dis Sci 1991;36:1469-72. 25. Hamasaki Y, Kobayashi I, Matsumoto S, Zaitu M, Muro E, Ichimaru T, et al. Inhibition of leukotriene production by FK506 in rat basophilic leukemia-1 cells. Pharmacology 1995;50:137-145. 26. Hamajima E, Sugiyama S, Hoshino H, Goto H, Tsukamoto Y, Ozawa T. Effects of FK506, an immunosuppressive agent, on genesis of water-immersion stress-induced gastric lesions in rats. Dig Dis Sci 1994;39:713-20.
Schwarz et al 423
27. Kubes P, Hunter J, Granger DN. Effects of cyclosporin A and FK506 on ischemia/reperfusion-induced neutrophil infiltration in the cat. Dig Dis Sci 1991;36:1469-72. 28. Sakurai Y, Yamaguchi M, Abiko Y. Inhibitory effect of lowlevel laser irradiation on LPS-stimulated prostaglandin E2 production and cyclooxygenase-2 in human gingival fibroblasts. Eur J Oral Sci 2000;108:29-34. 29. Kalff JC, Schraut WH, Billiar TR, Simmons RL, Bauer AJ. Role of inducible nitric oxide synthase in postoperative intestinal smooth muscle dysfunction in rodents. Gastroenterology 2000;118:316-27. 30. Eskandari MK, Kalff JC, Billiar TR, Lee KK, Bauer AJ. LPSinduced muscularis macrophage nitric oxide suppresses rat jejunal circular muscle activity. Am J Physiol 1999;277:G478-86. 31. Hori M, Kita M, Torihashi S, Miyamoto S, Won KJ, Sato K, et al. Upregulation of iNOS by COX-2 in muscularis resident macrophage of rat intestine stimulated with LPS. Am J Physiol Gastrointest Liver Physiol 2001;280:G930-8. 32. Barbara G, De Giorgio R, Deng Y, Vallance B, Blennerhassett P, Collins SM. Role of immunologic factors and cyclooxygenase 2 in persistent postinfective enteric muscle dysfunction in mice. Gastroenterology 2001;120:1729-36. 33. Schwarz NT, Kalff JC, Turler A, Engel BM, Watkins SC, Billiar TR, et al. Prostanoid production via COX-2 as a causative mechanism of rodent postoperative ileus. Gastroenterology 2001;121:1354-71.
Background. Intestinal transplantation is unduly complicated by the nontolerogenic properties of the gut-associated lymphoid tissue. Because simultaneous graft irradiation and bone marrow infusion significantly prolong the survival of the small bowel transplanted animal, our objective was to determine the functional motility effects of the immune modulating, graft irradiation procedure in the presence and absence of tacrolimus immunosuppression. Methods. Four groups of syngeneic orthotopic small bowel transplanted animals were studied 48 hours after operations (untreated, tacrolimus, ex vivo graft irradiation, and tacrolimus + irradiation) and compared with controls. Histologic analysis was performed for mucosal apoptosis and neutrophilic infiltration into the muscularis externa. Gastrointestinal in vivo transit and in vitro circular muscle strip contractions were quantified in response to bethanechol (0.3-300 µmol/L). Results.Graft irradiation ex vivo alone or in the presence of tacrolimus significantly increases (> 10-fold) the number of apoptotic mucosal cells after transplantation. Functional measurements showed that transplantation resulted in a significant delay in gastrointestinal transit and a decrease in muscle strip contractility. Tacrolimus and graft irradiation significantly ameliorated the transplantinduced dysfunction. Conclusions.Given the endowed propensity of mucosal regeneration, the immunologic and functional benefits of ex vivo graft irradiation appear to outweigh the detrimental effects to the mucosa. (Surgery 2002;131:413-23.) From the Departments of Surgery and Medicine/Gastroenterology, University of Pittsburgh Medical Center, Pittsburgh, Pa
CLINICAL SMALL INTESTINAL TRANSPLANTATION is a procedure that could potentially improve the quality of life of an estimated 20,000 patients in the United States who are currently on total parenteral nutrition at a cost in excess of $280 per day.1 Progress in immunosuppression, organ preservation, and surgical techniques has proven that the transplanted intestine can provide life-sustaining nutrients to the recipient if it can successfully endure the transplant process and the continuous insults of rejection.2-4 However, in comparison with other organs, the intestine remains the Supported by National Institutes of Health grants R01-GM58241, R01-DK-AI54232, and P50-GM-53789, and a grant from the Deutsche Forschungsgemeinschaft Schw 745 1/1 to Nicolas T. Schwarz. Accepted for publication December 9, 2001. Reprint requests: Anthony J. Bauer, PhD, Department of Medicine/Gastroenterology, S-849 Scaife Hall, 3550 Terrace St, University of Pittsburgh Medical School, Pittsburgh, PA 15261. Copyright © 2002 by Mosby, Inc. All rights reserved. 0039-6060/2002/$35.00 + 0 doi:10.1067/msy.2002.122372
11/56/122372
“Achilles heel” of transplants. We have hypothesized that a possible reason for this is the unique immunologic complexities of the gut-associated lymphoid tissue (GALT). In contrast to other organs, the intestine is a rich source of mature, relatively nontolerogenic leukocytes that predispose the graft to rejection and the recipient to graftversus-host disease.5 Recently, we have shown that modification of intestinal passenger leukocytes improves the outcome of the transplanted intestinal allograft.6 This immunomodification was accomplished by the ex vivo irradiation of the intestinal allograft at doses designed to eradicate the mature nontolerogenic lymphoid elements from the intestine.6 In addition, the procedure consisted of a simultaneous intravenous infusion of donor bone marrow cells to replace the irradiated intestinal mature lymphocytes. Hypothetically, this procedure should result in a microchimeric animal with immunotolerant leukocytes, a condition that favors graft acceptance. Indeed, the combined procedure of allograft irradiation and bone marrow infusion SURGERY 413
414 Schwarz et al
increased the survival of the animals that underwent small bowel transplantation to more than 150 days compared with 51 days for untreated allotransplanted animals. The combined procedure also prevented graft-versus-host disease.6 Although irradiation can efficiently eliminate mature lymphocytes from allografts, it can also produce unpropitious immediate and late damaging effects to the endothelium, mucosa, nerves, and smooth muscle of a graft. Because immunomodulation of intestinal transplantation with ex vivo graft irradiation and simultaneous donor bone marrow infusion appears to have potential therapeutic benefits in regards to rejection,6 the objectives of this study were to determine if the acute beneficial effects of irradiation-induced immunomodulation on graft survival are attained at the expense of alterations in intestinal motility and mucosal structure. We addressed these objectives by using a syngeneic orthotopic small intestinal transplant model in which the individual and combined effects of tacrolimus and ex vivo graft irradiation were determined on routine histopathologic examination, in vivo gastrointestinal transit, in vitro circular muscle contractility, and leukocyte recruitment. Syngeneic transplants were used to avoid any complications that could arise from allogeneic immune reactions. METHODS Small intestinal transplantation. Male Lewis rats (RT1l) (Harlan Sprague Dawley, Indianapolis, Ind) weighing 200 to 300 g were used. The animals were housed in the University of Pittsburgh animal care facilities. While the rats were under methoxyflurane anesthesia, orthotopic small intestinal transplantation with caval drainage was performed in a syngeneic combination. Briefly, the small intestine of the donor was isolated from the ligament of Treitz to the ileocecal valve, with its vascular pedicle consisting of the superior mesenteric artery with a piece of aorta and the superior mesenteric vein. Then the graft vascular bed was perfused with chilled lactated Ringer’s solution and the intestinal lumen was irrigated with 20 mL of cold neomycinsulfate (Sigma, St Louis, Mo) saline solution. The graft was stored briefly in cold Ringer’s solution during the preparation of the recipient. An end-toside anastomosis between the graft aorta and the recipient’s infrarenal aorta, and between the graft vein and the recipient’s vena cava, was performed with 10-0 polybutester (Novafil) sutures. The time required to complete the vascular anastomosis was 20 to 30 minutes. The entire recipient intestine was removed, and enteric continuity was restored by
Surgery April 2002
proximal and distal end-to-end intestinal anastomoses. Animals were prophylactically given the antibiotic cefamandole nafate (20 mg/d). All animals were killed 48 hours after transplantation. Experimental groups. Five groups of animals were examined in this study, 4 of which received syngeneic small bowel transplants. Group 1 consisted of unoperated controls. Transplant recipients in group 2 did not receive any treatment and served as transplant controls. In group 3, tacrolimus (FK506; Fujisawa Pharmaceutical, Osaka, Japan) was intramuscularly administrated to recipients 4 and 28 hours after transplantation at a dosage of 1.0 mg/kg. In recipients of group 4, the harvested intestinal graft, stored in cold lactate Ringer’s solution in a 50 mL Falcon tube, was ex vivo irradiated at 10 Gy from a 137Cs source (Gammacell 1000 Elite; Nordion International, Ontario, Canada) and transplanted. Total time to achieve this irradiation dose was 2.5 minutes. Animals in experimental group 5 received a combined treatment of tacrolimus and ex vivo graft irradiation. Histopathologic analysis. Specimens from each intestinal graft including Peyer’s patches and mesenteric lymph nodes were taken and fixed in formalin. Samples were embedded in paraffin and sliced as 6-µm cross-sections of the entire gut wall, then stained with hematoxylin and eosin. Mucosal apoptotic cells were counted in hematoxylin and eosin cross-sections as a routine histopathologic analysis. Slides were reviewed blindly by one of the authors (M.A.N) without knowledge of the experimental groups. Whole-mounts of the intestinal muscularis were investigated for the presence of recruited neutrophils. Mid-jejunal segments were cut from the bowel and immersed in Krebs Ringer’s buffer (KRB) in a chilled and Sylgard- (Dow Corning, Midland, Mich) bottomed glass dish, as described previously.7 The length and width of each jejunal segment were measured with a caliper, and then the segment was gently pinned down along the mesenteric border. The bowel was opened along the mesentery and stretched to 150% of the length and 250% of the width. The opened segments of jejunum were fixed in 100% ethanol for 10 minutes. Each segment was washed twice in KRB, and mucosa and submucosa were stripped off under microscopic observation (Wild, Heerbrugg, Switzerland). The mucosa-free muscularis wholemounts were finally cut into 1 3–cm pieces and used for staining procedures. Polymorphonuclear neutrophils were visualized by a myeloperoxidase stain: freshly prepared whole-mounts were immersed in a mixture of 10 mg of Hanker-Yates
Surgery Volume 131, Number 4
reagent (Polysciences, Warrington, Pa), 10 mL of KRB, and 100 µL of 3% hydrogen peroxide for 20 minutes. The reaction was stopped with cold KRB. Whole-mounts were cover-slipped and inspected by light microscopy after staining (Nikon FXA; Fryer, Huntley, Ill). Leukocytes were counted in 5 randomly chosen areas in each specimen at a magnification of 200. Intestinal muscle function. In vivo gastrointestinal transit was measured in controls and transplanted animals at 48 hours postoperatively by evaluating the intestinal location of fluorescein isothiocyanate (FITC)-labeled dextran (70000 MW). Animals were lightly anesthetized with methoxyflurane and fed FITC-labeled dextran (200 µL of 25 mg/mL stock solution) orally. Two hours after administration, the contents of the entire small bowel, divided into 10 equal segments, cecum, proximal colon, and distal colon were vortexed with 2 mL of saline solution to obtain a supernatant containing the fluorescent marker that was present within the lumen of each bowel segment. The supernatant was centrifuged at 7000 rpm to pellet the intestinal chyme and duplicate aliquots of the cleared supernatant were read in a CytoFluor II (PerSeptive Biosystems, Framingham, Mass) multi-well fluorescence plate reader (excitation 485 ± 30 and emission 530 ± 20 nm) for quantification of the fluorescent signal in each bowel segment. A median histogram of the fluorescence was then plotted for analysis of transit. This provided an accurate, nonradioactive measurement of liquid gastrointestinal transit without further surgical intervention. 8 The data were also calculated to obtain the geometric center (GC) of distribution of fluorescein-labeled dextran to quantify transit with the use of the following formula: GC = Σ (% of total fluorescent signal per segment segment number)/100. The GC value reflects the transit of fluorescently labeled dextran down the gastrointestinal tract, with higher values indicating more distal distribution (segment 1 = stomach; segment 15 = distal colon). In vitro circular muscle mechanical activity was assessed with full thickness strips obtained from the middle jejunum. Muscle strips were prepared from an opened segment of the jejunum by making 2 cuts (1 10 mm) parallel to the circular muscle layer. Each strip was then fixed in a mechanical organ chamber with 1 end pinned to the Sylgard floor and the other attached to an isometric force transducer (WPI, Sarasota, Fla) by 4-0 silk sutures.
Schwarz et al 415
Each chamber was continuously perfused with a preheated, preoxygenated KRB solution that was constantly monitored and maintained at 37° ± 0.5°C. Once in the organ chamber, each strip was allowed to equilibrate for 1 hour. It was then stretched in small increments to determine the length at which maximal spontaneous contractions of the muscle occurred (L0). After determination of L0, dose-response curves were generated with sequentially increasing doses of bethanechol (0.1300 µmol/L) for 10 minutes with intervening KRB wash periods of 15 minutes. Contractions between different muscle strips were normalized by converting grams of contraction to grams/millimeters2 of tissue. This was done by determining the cross-sectional area by the following equation (muscle density assumed to be 1.03 mg/mm3): mm2 = (wet muscle weight [in mg]/muscle length [in mm] muscle density [mg/mm3]).9,10 Contractions were recorded, measured, and stored in a computer with an A/D hardware and software package (Biopac Systems, Santa Barbara, Calif). Statistical analysis. Results are expressed as mean ± SEM. Data were statistically analyzed with an unpaired Student t test. A probability level of P < .05 was considered significant. RESULTS General observations. All transplanted animals throughout the study period had a healthy appearance and were able to drink within 6 hours after the operation. Animal body weight changed to a similar extent in all 4 groups of animals after syngeneic transplantation (P > .05, N = 7). All the animals showed an approximate 5% decrease in weight because of the transplantation procedure regardless of treatment. This amount of body weight loss is typical after intestinal transplantation in the rat.6 At the time the animals were killed, the 2-day transplanted grafts looked slightly edematous with a rosy appearance macroscopically. Routine pathologic analysis showed essentially normal intestinal architecture. The mucosa of the untreated transplanted grafts 48 hours after transplantation had only 0.5 ± 0.58 apoptotic cells per 10 glands. Similarly, animals treated postoperatively with tacrolimus showed 0.7 ± 0.58 apoptotic cells per 10 glands. In contrast, animals treated with radiation alone or in combination with tacrolimus displayed a significant increase in apoptotic cells 48 hours after transplantation (P < .01, 6.3 ± 3.23, and 7.7 ± 2.14 cells per 10 glands, respectively). In vivo transit after transplantation. As we have previously shown,8 normal gastrointestinal transit of nonabsorbable FITC-labeled dextran accumu-
416 Schwarz et al
Surgery April 2002
Fig 1. Transplantation significantly delays gastrointestinal transit. Transit histograms for distribution of nonabsorbable FITC-labeled dextran along gastrointestinal tract 2 hours after oral administration. In control animals (dark bars), most of fluorescent marker accumulated in last 3 segments of small bowel (sb8, sb9, sb10). In contrast, transplantation caused significant delay in the transit of the fluorescence. Two days after transplantation (cross-hatched bars), orally fed marker was distributed primarily in proximal regions of the small bowel (sb2, sb3, and sb4). Data represent averaged percent distribution of fluorescence intensity from 4 animals. SB-Tx, Small bowl transplantation; st, stomach; sb, small bowel; cm, cecum; c, colon.
lates in the last 3 distal segments of the small intestine after 2 hours of oral administration demonstrating a GC of 9.1 ± 0.32 (Fig 1). Forty-eight hours after syngeneic transplantation of the small bowel, 2-hour transit of the FITC-labeled dextran was significantly delayed (GC = 4.3 ± 0.17). In these untreated transplanted animals, the fluorescent marker emptied out of the stomach and made it past the site of bowel anastomosis (located in the first jejunal segment). However, the marker traversed the bowel only to the proximal regions of the jejunum (Fig 1). In the untreated syngeneic transplanted animal, transit returned to a normal pattern of fluorescent marker distribution after 8 days (GC = 8.9 ± 0.27, N = 4). Tacrolimus has been reported to lessen ischemia and reperfusion injury to the liver,11 in addition to its primary usage as an immunosuppressive agent for allogeneic organ transplantation. Therefore, in this syngeneic model, we investigated the functional in vivo effects of tacrolimus on syngeneic transplant- induced intestinal ileus. As shown in Fig 2, A, tacrolimus significantly improved the transplantinduced delay in FITC-labeled dextran transit through the small intestine. Under treatment with this immunosuppressive agent, gastrointestinal transit improved to closely resemble that observed in control unoperated animals with the majority of the fluorescent signal accumulating again in the terminal regions of the small intestine (GC = 8.8 ± 0.64). The objective of the next series of experiments
was to determine the effect of low dose graft irradiation on transit in the absence of the rejection phenomenon. Ex vivo graft irradiation unexpectedly improved transit in the transplanted bowel (Fig 2, B) compared with the untreated group. In fact, like tacrolimus, irradiation significantly averted the transplantation-induced delay in transit with movement of the marker to the distal regions of the small intestine over a period of 2 hours (GC = 9.5 ± 0.98). Additionally, we determined the combined effects of tacrolimus and graft irradiation on transit, but no apparent combined supplemental protective effect was observed (Fig 2, C). The combined treatment group displayed a GC of 7.8 ± 0.93, which was a significant improvement over that of the untreated transplanted animals. However, because transit had returned to near control levels with individual treatments, an additive beneficial effect would have been hard to demonstrate. In vitro muscle contractility after transplantation. Alterations in intestinal muscle function in vitro because of syngeneic transplantation and the modifying effects of tacrolimus and graft irradiation were assessed on jejunal circular muscle contractions. Circular muscle strips from control Lewis rats generated spontaneous contractions at a frequency of 34.3 ± 0.5 per minute, and mean spontaneous contractile force was calculated to be 0.019 ± 0.0083 g per mm2 per second (N = 4). Muscles from syngeneic transplanted grafts harvested on postoperative day 2 generated spontaneous contractions at a similar
Surgery Volume 131, Number 4
Schwarz et al 417
Fig 2. Tacrolimus and ex vivo graft irradiation significantly prevented delay in gastrointestinal transit caused by transplantation (gray bars, N = 4 each). Transit histograms for distribution of nonabsorbable FITC-labeled dextran along gastrointestinal tract 2 hours after oral administration. In transplanted tacrolimus treated animals (A, dark bars), most fluorescent marker accumulated in last segments of small bowel, which was similar to control unoperated animals. Additionally, ex vivo graft irradiation alone (B, dark bars) or combined with tacrolimus (C, dark bars) significantly prevented delay in marker movement down transplanted intestine. SBTx, Small bowel transplantation; st, stomach; sb, small bowel; cm, cecum; c, colon.
frequency (32.7 ± 0.8/min), but the amplitude of spontaneous contractions was significantly decreased compared with controls (0.003 ± 0.0095 g/mm2/s, N = 6) (leading contractile traces from Fig 3, A and B). In addition to spontaneous activity, bethanechol-stimulated dose response curves also showed significant impairment of the transplanted muscularis (1.059 ± 0.0485 vs 0.456 ± 0.0558 g/mm2/s with a bethanechol concentration of 100 µmol/L g/mm2/s, N = 4) (Fig 4). Correlating with the beneficial effect of
tacrolimus on in vivo transit measurements, in vitro bethanechol-stimulated muscle contractility of tacrolimus-treated animals was significantly improved compared with the untreated transplanted muscularis, as shown in the raw mechanical trace in Fig 3, C. Similarly, bethanechol-stimulated dose response curves showed significant improvement in animals treated with tacrolimus compared with muscle activity from untreated grafts (Fig 4). Statistical analysis of these data showed that not only was there a significant improvement com-
418 Schwarz et al
Surgery April 2002
Fig 3. Representative in vitro jejunal circular smooth muscle contractile activity recorded in response to bethanechol (100 µmol/L) from control, transplanted, and tacrolimus treated transplanted animals. Control muscles exhibited robust spontaneous activity and large phasic and tonic responses to bethanechol at higher concentrations (A). Spontaneous activity and response to bethanechol was significantly suppressed in muscles harvested from untreated transplanted animals (B). Tacrolimus treatment of the transplanted animals showed a significant improvement in spontaneous and bethanechol-stimulated contractions (C). SB-Tx, Small bowel transplantation.
pared with untreated animals, but that bethanechol responses at higher concentrations were not significantly different from muscles of control animals. Next, the effect of graft irradiation was tested to determine its effect on in vitro muscle strip contractility. As predicted by the transit measurements and shown in Fig 4, graft irradiation resulted in a significant improvement in muscle strip contractility after syngeneic transplantation compared with the muscles harvested from the untreated transplanted grafts. Additionally, the combined effects of tacrolimus and graft irradiation on circular muscle contractility were determined. The combined treatment of the graft also resulted in improved muscle performance. However, as shown in Fig 4, both irradiated-treated groups did not reach the
level of the controls or the graft muscles treated by tacrolimus alone. Leukocyte recruitment. Fig 5 shows that syngeneic transplantation in the absence of rejection results in the massive recruitment of neutrophils into the intestinal muscularis 48 hours after transplantation. We have previously shown a correlation and direct mechanistic link between neutrophilic infiltration and muscle contractility in a model of intestinal postoperative ileus.7 Therefore, we sought to determine if tacrolimus and graft irradiation would result in a decrease in the leukocytic infiltrate that is initiated by the trauma of transplantation. Counts of neutrophils demonstrated that transplantation caused a 14-fold increase in the presence of neutrophils within the graft mus-
Surgery Volume 131, Number 4
Schwarz et al 419
Fig 4. Effect of tacrolimus and ex vivo graft irradiation on in vitro–generated standard bethanechol dose response curves recorded from jejunal circular smooth muscle strips (N = 4 each). Control jejunal circular muscle strips showed a dose-dependent increase in contractile area in response to bethanechol (filled square). This activity was significantly diminished in muscles harvested 48 hours after transplantation (filled circle). Tacrolimus (open circle) treatment of the transplanted animal completely reversed the suppression caused by transplantation. Significant improvement was also measured in transplanted animals after ex vivo graft irradiation, either alone (diamond), or in combination with tacrolimus administration (inverted triangle). SB-Tx, Small bowel transplanted.
cularis externa. Furthermore, as shown in Fig 6, the administration of tacrolimus and graft irradiation, alone or in combination, significantly decreased the recruitment and extravasation of polymorphonuclear neutrophils into the jejunal intestinal muscularis. DISCUSSION A successful small intestinal transplant would be the ultimate therapeutic option for patients with intestinal failure. Accordingly, the transplanted donor intestine needs to efficiently digest and absorb a wide variety of nutrients to sustain life. Although complex steps including enzymatic digestion and epithelial absorption are important to proper digestion, these steps are dependent on the generation of specific spatial and temporal motility patterns to mix and propel the ingested food at a rate that facilitates digestion and absorption. The major obstacle to successful intestinal transplantation continues to be graft rejection. 2,4 And although the mucosa has a remarkable ability to regenerate, the historical accumulation of rejection episodes are stored within the vasculature, intestinal muscularis, and enteric nervous system,12,13 structures which are crucial to motility. Undoubtedly, recent advances in immunosuppressive treatments have improved the outcome of intestinal transplantation. 2,4 However, heavy immunosuppression is still required for small bowel transplantation and, therefore, substantial concern remains over subsequent infectious sus-
ceptibility and lymphoproliferative disease.2,14 Our recent development of an immunomodulating strategy that makes use of ex vivo graft irradiation and the simultaneous injection of donor bone marrow is aimed to solve these problems and to improve the outcome of small bowel transplantation.6 The rationale for this immunomodification procedure is based on the theoretical concept of microchimeric symbiosis.15 The facilitation of microchimerism in clinical organ transplantation with adjunct donor bone marrow infusion is currently being evaluated.16 Although still controversial, some evidence supports that leukocytic chimerism is beneficial to the success of both liver and bone marrow transplantation.17 However, in the case of the small intestine, the nontolerogenic properties of the mature leukocytes in the GALT uniquely predispose the transplanted bowel to immunologic difficulties.4 Our immunomodification procedure is designed to remove the nontolerogenic GALT leukocytes with irradiation and gain the benefit of microchimerism through a simultaneous bone marrow infusion. In a recent publication, with the use of the immunomodification model, we have demonstrated that this procedure creates a chimeric animal with immunotolerant leukocytes. The use of this procedure resulted in a significantly improved outcome for the intestinal allografts and also the prevention of graft-versus-host disease.6 Thus, this approach seems immunologically sound. Although irradiation can efficiently eliminate
420 Schwarz et al
Surgery April 2002
Fig 5. Typical histochemically stained full thickness muscularis whole-mounts with use of the Hanker-Yates reagent for the presence of myeloperoxidase positive leukocytes extravasated into muscularis. Panel A shows typical muscularis whole-mount from normal animal, which was virtually devoid of myeloperoxidase positive leukocytes. Panel B shows presence of numerous extravasated leukocytes within muscularis whole-mount obtained from animal 48 hours after intestinal transplantation. Original magnification for both wholemounts was 100. SB-Tx, Small bowel transplanted.
mature lymphocytes from allografts, it also has the potential to produce unpropitious immediate and late damaging effects to the vasculature, mucosa, enteric nervous system, and smooth muscle. The complications of oncologic radiotherapy have been extensively studied during the treatment of malignant neoplasms with maximally tolerated irradiation doses, usually exceeding 20 Gy. These studies show early and late complications in the gastrointestinal tract, especially in the small intestine, one of the most radiosensitive organs. Early radiation effects in the intestine reflect epithelial stem cell injury and later progressive endarteritis develops. Irradiation is also potentially harmful to the neuromuscular apparatus of the intestine, causing a disturbance in normal motility patterns.18,19 However, ex vivo irradiation of an excised organ is considered to be less harmful compared with irradiation of intact tissues, and the single 10 Gy dose applied
in this study is below the dosages commonly used in oncology. In this study, we observed that ex vivo irradiation did significantly increase the number of apoptotic enterocytes over non-irradiated control grafts 48 hours after transplantation. However, because of the regenerative capacity of the mucosa, this did not seem to be a major limiting factor of this procedure. Progressive vascular complications have also been detected with oncologic doses of irradiation and are well known to be associated with late morbidity.20,21 As with other radiation-induced injuries, radiationinduced arteriopathy is basically dose dependent. However, the ex vivo dose of 10 Gy used in this study did not induce alterations in the vasculature in the syngeneic transplantation model.6 Potentially of greater importance would be the detrimental effect of irradiation on the neuromuscular apparatus of the transplanted intestine and
Surgery Volume 131, Number 4
Schwarz et al 421
Fig 6. Histogram quantifying the number of extravasated neutrophils within full thickness whole-mounts of jejunal muscularis externa of specimens from controls, small bowel-transplanted (SB-Tx), tacrolimus treated, graft ex vivo irradiated, and combined graft irradiation with tacrolimus treatment. As reflected in histological analysis, transplantation results in significant cellular inflammatory response within the muscularis. Immunomodulation of animals with tacrolimus or graft irradiation, either alone or in combination, significantly decreased number of myeloperoxidase positive cells, which extravasated into muscularis in response to transplantation. However, the muscularis of all transplanted animal groups had significantly more myeloperoxidase cells present per 200 field compared with control. Asterisk, Compared with control; single dagger, compared with untreated SB-Tx, P < .05, N = 4 each.
the induction of sustained motility abnormalities. Changes in intestinal motility during single or fractionated doses of abdominal irradiation have been reported in large and small animal models. Radiation-induced disturbances in motility that occur relatively early after exposure have been hypothesized to be responsible for gastrointestinal symptoms such as nausea, vomiting, and diarrhea.18,19,22 However, our in vivo gastrointestinal transit experiments show that graft irradiation and tacrolimus treatment significantly improved transit. In the treated animals, the orally administered nonabsorbable fluorescent marker moved down the gastrointestinal tract and reached to the last segments of the ileum, displaying a distribution pattern similar to nonmanipulated control animals. On the other hand, untreated animals showed a dramatic suppression of the in vivo movement of the nonabsorbable fluorescent marker down the gastrointestinal tract. These data show that the intestinal transplantation procedure itself causes a significant dysfunction in normal intestinal motility, which is at least partially ameliorated with graft irradiation and tacrolimus. Similarly, the in vitro contractile responses of isolated jejunal circular smooth muscle strips in response to bethanechol were significantly improved by irradiation and tacrolimus treatment, either alone or in combination. These results may further suggest that, in contrast to the mucosa,
muscle function is more severely injured by the transplantation procedures than by ex vivo graft irradiation with 10 Gy. The beneficial effects of irradiation in this study, in contrast to reports in the literature,18,19,22 may be explained by the fact that ex vivo irradiation of an excised organ is considered to be less harmful compared with in vivo irradiation of tissues and the low level of irradiation that we used. As shown here, and in our previous studies, muscle dysfunction and abnormal intestinal motility temporally correlates with the extravasation of neutrophils and monocytes into the intestinal muscle layer.7 Therefore, it was not surprising to observe that the functional improvement of the treated transplanted grafts was associated with a decreased number of neutrophils within the graft muscularis externa. Tacrolimus is a potent immunosuppressive reagent for the prevention of allograft rejection. It has also been shown to have protective effects in ischemia and reperfusion injuries of the heart, liver, and intestine.11,23,24 The mechanism of this protective effect appears to work through the blockade of nuclear factor-κB activation, which results in a blunting of the increased expression of intercellular adhesion molecule-1 mRNA and, therefore, a subsequent reduction in the number of recruited leukocytes compared with the number in untreated animals.11 Additionally in the intestine, tacrolimus was shown to reduce elevated
422 Schwarz et al
mucosal Leukotriene B4 levels, which are normally observed after intestinal reperfusion. A growing body of literature suggests that tacrolimus may be an important agent in modulating neutrophil infiltration in acute inflammatory conditions.24-27 The mechanisms that are responsible for the observed significant ameliorative effects of ex vivo graft irradiation are currently unknown. However, it has been shown that low-level laser irradiation will reduce the expression of the inflammatory mediator cyclooxygenase-2 and the subsequent production of prostaglandin E2 in gingival tissues.28 We can also speculate that irradiation will result in a decrease in the expression of the inducible synthases inducible macrophage-type nitric oxide synthase and cyclooxygenase-2 within the donor intestinal muscularis. Because nitric oxide and prostanoids play a significant role in the inflamed gut,29-33 a decrease in the release of these potent inflammatory and kinetically active factors from the irradiated resident muscularis macrophages and a decreased number of infiltrating leukocytes would result in improved intestinal function.
Surgery April 2002
4.
5.
6.
7.
8.
9.
10.
CONCLUSION In summary, graft irradiation ex vivo alone or in the presence of tacrolimus produced a significant increase in the number of apoptotic enterocytes after transplantation. However, functional motility measurements showed that treatment with tacrolimus and ex vivo graft irradiation, either alone or combined, significantly ameliorated the transplant-induced dysfunction in motility. We conclude from this and our previous studies that immunomodulation of both the graft with ex vivo irradiation and of the host with donor bone marrow appears to result in substantial benefit for graft and host survival, and significantly improved intestinal motility. And given the endowed propensity of the mucosa for regeneration, the early immunologic and functional benefits of ex vivo graft irradiation appear to outweigh the immediate detrimental effects to the mucosa. Although these short-term irradiation results are encouraging, the prolonged effects of graft irradiation have yet to be determined. REFERENCES 1. Howard L, Ament M, Fleming CR, Shike M, Steiger E. Current use and clinical outcome of home parenteral and enteral nutrition therapies in the United States. Gastroenterology 1995;109:355-65. 2. Grant D. Intestinal transplantation: 1997 report of the international registry. Intestinal Transplant Registry. Transplantation 1999;67:1061-4. 3. DiMartini A, Rovera GM, Graham TO, Furukawa H, Todo S, Funovits M, et al. Quality of life after small intestinal trans-
11.
12.
13.
14.
15.
16.
17.
18.
19.
plantation and among home parenteral nutrition patients. JPEN J Parenter Enteral Nutr 1998;22:357-62. Abu-Elmagd K, Reyes J, Todo S, Rao A, Lee R, Irish W, et al. Clinical intestinal transplantation: new perspectives and immunologic considerations. J Am Coll Surg 1998;186: 512-5. Murase N, Starzl TE, Tanabe M, Fujisaki S, Miyazawa H, Ye Q, et al. Variable chimerism, graft-versus-host disease, and tolerance after different kinds of cell and whole organ transplantation from Lewis to brown Norway rats. Transplantation 1995;60:158-71. Murase N, Ye Q, Nalesnik MA, Demetris AJ, Abu-Elmagd K, Reyes J, et al. Immunomodulation for intestinal transplantation by allograft irradiation, adjunct donor bone marrow infusion, or both. Transplantation 2000;70:1632-41. Kalff JC, Carlos TM, Schraut WH, Billiar TR, Simmons RL, Bauer AJ. Surgically induced leukocytic infiltrates within the rat intestinal muscularis mediate postoperative ileus. Gastroenterology 1999;117:378-87. Kalff JC, Buchholz BM, Eskandari MK, Hierholzer C, Schraut WH, Simmons RL, et al. Biphasic response to gut manipulation and temporal correlation of cellular infiltrates and muscle dysfunction in rat. Surgery 1999;126: 498-509. Eskandari MK, Kalff JC, Billiar TR, Lee KK, Bauer AJ. LPSinduced muscularis macrophage nitric oxide suppresses rat jejunal circular muscle activity. Am J Physiol 1999;277: G478-86. Kalff JC, Schraut WH, Billiar TR, Simmons RL, Bauer AJ. Role of inducible nitric oxide synthase in postoperative intestinal smooth muscle dysfunction in rodents. Gastroenterology 2000;118:316-27. Squadrito F, Altavilla D, Squadrito G, Saitta A, Deodato B, Arlotta M, et al. Tacrolimus limits polymorphonuclear leucocyte accumulation and protects against myocardial ischaemiareperfusion injury. J Mol Cell Cardiol 2000; 32:429-40. Heeckt PF, Lee KKW, Halfter WM, Schraut WH, Bauer AJ. Functional impairment of enteric smooth muscle and nerves caused by chronic intestinal allograft rejection regresses after FK506 rescue. Transplantation 1995;59: 159-64. Heeckt PF, Halfter W, Schraut WH, Bauer AJ. Subclinical chronic rejection leads to an early destruction of the intrinsic nervous system in rat small bowel transplants. Am J Physiol 1997;36:G413-21. Sigurdsson L, Reyes J, Kocoshis SA, Mazariegos G, AbuElmagd K, Green M. Bacteremia after intestinal transplantation in children correlates temporally with rejection or gastrointestinal lymphoproliferative disease. Transplantation 2000;70:302-5. Starzl TE, Demetris AJ, Murase N, Ildstad S, Ricordi C, Trucco M. Cell migration, chimerism, and graft acceptance. Lancet 1992;339:1579-82. Fontes P, Rao AS, Demetris AJ, Zeevi A, Trucco M, Carroll P, et al. Bone marrow augmentation of donor-cell chimerism in kidney, liver, heart, and pancreas islet transplantation. Lancet 1994;344:151-55. Thomson AW, Lu L, Murase N, Demetris AJ, Rao AS. Microchimerism, dendritic cell progenitors and transplantation. Stem Cells 1995;13:622-39. Fraser R, Frisby C, Blackshaw LA, Schirmer M, Howarth G, Yeoh E. Small intestinal dysmotility following abdominal irradiation in the rat small intestine. Neurogastroenterol Motil 1998;10:413-9. Erickson BA, Otterson MF, Moulder JE, Sarna SK. Altered motility causes the early gastrointestinal toxicity of irradiation. Int J Radiat Oncol Biol Phys 1994;28:905-12.
Surgery Volume 131, Number 4 20. Quarmby S, Kumar P, Kumar S. Radiation-induced normal tissue injury: role of adhesion molecules in leukocyteendothelial cell interactions. Int J Cancer 1999; 82:385-95. 21. Hopewell JW, Calvo W, Jaenke R, Reinhold HS, Robbins ME, Whitehouse EM. Microvasculature and radiation damage. Recent Results Cancer Res 1993;130:1-16. 22. Summers RW, Maves BV, Reeves RD, Arjes LJ, Oberley LW. Irradiation increases superoxide dismutase in rat intestinal smooth muscle. Free Radic Biol Med 1989;6:261-70. 23. Garcia-Criado FJ, Palma-Vargas JM, Valdunciel-Garcia JJ, Toledo AH, Misawa K, Gomez-Alonso A, et al. Tacrolimus (FK506) down-regulates free radical tissue levels, serum cytokines, and neutrophil infiltration after severe liver ischemia. Transplantation 1997;64:594-8. 24. Kubes P, Hunter J, Granger DN. Effects of cyclosporin A and FK506 on ischemia/reperfusion-induced neutrophil infiltration in the cat. Dig Dis Sci 1991;36:1469-72. 25. Hamasaki Y, Kobayashi I, Matsumoto S, Zaitu M, Muro E, Ichimaru T, et al. Inhibition of leukotriene production by FK506 in rat basophilic leukemia-1 cells. Pharmacology 1995;50:137-145. 26. Hamajima E, Sugiyama S, Hoshino H, Goto H, Tsukamoto Y, Ozawa T. Effects of FK506, an immunosuppressive agent, on genesis of water-immersion stress-induced gastric lesions in rats. Dig Dis Sci 1994;39:713-20.
Schwarz et al 423
27. Kubes P, Hunter J, Granger DN. Effects of cyclosporin A and FK506 on ischemia/reperfusion-induced neutrophil infiltration in the cat. Dig Dis Sci 1991;36:1469-72. 28. Sakurai Y, Yamaguchi M, Abiko Y. Inhibitory effect of lowlevel laser irradiation on LPS-stimulated prostaglandin E2 production and cyclooxygenase-2 in human gingival fibroblasts. Eur J Oral Sci 2000;108:29-34. 29. Kalff JC, Schraut WH, Billiar TR, Simmons RL, Bauer AJ. Role of inducible nitric oxide synthase in postoperative intestinal smooth muscle dysfunction in rodents. Gastroenterology 2000;118:316-27. 30. Eskandari MK, Kalff JC, Billiar TR, Lee KK, Bauer AJ. LPSinduced muscularis macrophage nitric oxide suppresses rat jejunal circular muscle activity. Am J Physiol 1999;277:G478-86. 31. Hori M, Kita M, Torihashi S, Miyamoto S, Won KJ, Sato K, et al. Upregulation of iNOS by COX-2 in muscularis resident macrophage of rat intestine stimulated with LPS. Am J Physiol Gastrointest Liver Physiol 2001;280:G930-8. 32. Barbara G, De Giorgio R, Deng Y, Vallance B, Blennerhassett P, Collins SM. Role of immunologic factors and cyclooxygenase 2 in persistent postinfective enteric muscle dysfunction in mice. Gastroenterology 2001;120:1729-36. 33. Schwarz NT, Kalff JC, Turler A, Engel BM, Watkins SC, Billiar TR, et al. Prostanoid production via COX-2 as a causative mechanism of rodent postoperative ileus. Gastroenterology 2001;121:1354-71.