Gut epithelial cell death by apoptosis is increased in the gut epithelium after severe burn associated with mucosal atrophy. We hypothesized that tumor necrosis factor (TNF)-α-TNF receptor (TNFR) interaction activates apoptosis in small bowel mucosal cells after severe burn. C57BL6 mice received a 30% total body surface area scald burn and were treated with neutralizing anti-TNF-α. The proximal small bowel was assessed for mucosal atrophy. Proliferation and apoptosis of mucosal cells were assessed by proliferative cell nuclear antigen-immunostaining and terminal deoxyuridine nick-end labeling assay, respectively. Mucosal height and mucosal cell number decreased after burn. Anti-TNF-α-treated mice showed significantly less mucosal atrophy. Proliferation of intestinal cells was not changed with burn or anti-TNF-α treatment. An over threefold increase in apoptotic cell number was seen after burn, which was diminished by anti-TNF-α treatment. Changes in gut mucosal homeostasis after severe burn are affected, in part, by the activation of apoptosis by TNF-α-TNFR interaction.
- small bowel mucosa
- mucosal atrophy
- tumor necrosis factor-α
gut mucosal homeostasis, and thus the morphological and functional integrity of the gut, is maintained by a balance between epithelial cell proliferation and cell death. Proliferation of gut epithelium occurs by mitosis in the intestinal crypts, whereas cell death occurs throughout the crypt and villus (19). Apoptosis is programmed death and removal of senescent or otherwise dysfunctional cells without inflammation. It can be considered an antagonistic and regulatory process to cellular proliferation by mitosis. Both apoptosis and mitosis are continually ongoing in live gut epithelium to maintain mucosal cellular balance. This delicate balance of mucosal cell mass can be influenced by exogenous and endogenous factors, such as nutritional depletion, chronic disease, or severe trauma (i.e., severe burn). We (19) previously showed increased gut epithelial cell death by apoptosis in the gut epithelium after severe burn, which was associated with mucosal atrophy. Potential mechanisms for these effects include the induction of epithelial cell death indirectly by relative hypoperfusion or directly through interaction of inflammatory mediators and their receptors located on gut epithelial cells. We (14) recently showed that burn-induced hypoperfusion of the gut is insufficient to induce apoptosis of gut epithelial cells. In other organs, programmed cell death can be induced by several membrane-bound ligand-receptor interactions, such as through Fas ligand (FasL)-Fas interaction or tumor necrosis factor-α-TNF receptor (TNF-α-TNFR) interaction (3). We therefore hypothesized that TNF-α-TNFR interaction activates apoptosis in small bowel mucosal cells after severe burn, and thus is a crucial element in changes in gut mucosa after severe burn.
MATERIALS AND METHODS
Adult male C57BL6 mice (Harlan Sprague-Dawley, Houston, TX) weighing 23 ± 2 g were housed individually in a temperature-controlled cubicle with a 12:12-h light-dark cycle. Mice were fed and received water ad libitum. The study was approved by the Animal Care and Use Committee of The University of Texas Medical Branch, Galveston, TX.
After 1 wk, the mice were randomly assigned to sham burn control, 30% total body surface area (TBSA) scald burn, and 30% TBSA scald burn with treatment by neutralizing hamster-anti-mouse-TNF-α antibody. Mice were anesthetized with methoxyflurane as inhalational agent (1–4%) and buprenorphine hydrochloride (0.1 mg/kg) given subcutaneously. The dorsum of the trunk was shaved, and a 30% TBSA burn was administered by placing the animals in a mold exposing an area of 4.2 × 2.9 cm of the back. The mold was placed in a column of 95 to 99°C steam for 6 s, which delivered a full thickness cutaneous burn demonstrated by histologic sections. Sham control animals were anesthetized, shaved, and placed in the mold without exposure to steam. Burned animals were resuscitated with 1 ml of 0.9% NaCl solution sq and 1 ml NaCl solution ip. Anti-TNF-α-treated animals received 200 μg neutralizing hamster anti-mouse-TNF-α antibody (BD Pharmingen, San Diego, CA) in 1 ml saline ip in addition to 1 ml resuscitation fluid sq. Mice were then returned to their cages. This time point was chosen in reference to our previously published findings (19) in which the maximum apoptotic reaction in small bowel was seen at 12 h. After sham burn or burn, all groups were given water ad libitum and fasted to avoid the confounding variable of different food intake in burned and unburned animals. Animals were killed at 12 h after injury by decapitation. The entire small bowel was excised, measured for length and divided in half. The proximal segment was flushed with ice cold saline, opened longitudinally, blotted dry on paper and immediately weighed. Thereafter, a 2-cm segment of the proximal end of the small bowel was taken and immediately fixed in 10% buffered formalin. This piece was used for histology, immunohistochemistry, and terminal deoxyuridine nick-end labeling (TUNEL) assay. Of the remaining 16 to 20 cm lengths, a 3-cm section was used for the determination of dry weight after desiccation at 50°C for 48 h. The rest of the bowel sample was snap frozen in liquid nitrogen and stored at −70°C. Additionally, measures of body weight and liver weights were obtained.
Formalin-fixed tissues were processed and embedded in paraffin. Three 3-μm sections were obtained of each tissue block at 40-μm intervals, deparaffinized, rehydrated in graded alcohol (100, 95, and 70%), and washed with deionized water. Hematoxylin and eosin staining was performed, and mucosal height, crypt depth, and villus height was determined by randomly selecting 10 complete villi from each section and measuring the distance from the base of the crypt to the villus tip, the base of the crypt to the crypt-villus junction, or from the crypt-villus junction to the villus tip, respectively. Values from the measured villi were averaged to reach individual mucosal height, crypt depth, or villus height measurements. For identification of apoptotic cells the TUNEL method (ApoTag; Oncor, San Francisco, CA) was used. Prepared sections were treated with proteinase K (20 μl/ml in PBS) to digest proteins; endogenous peroxidase activity was quenched with 2% H2O2 in PBS. Seventy-five microliters of equilibration buffer was placed on each section, and diluted TdT enzyme solution was applied and incubated at 37°C for 1 h. After incubation, the slides were placed in stop/wash buffer. Then 55 μl of antidigoxigenin peroxidase was added, and the slides were incubated for 30 min at room temperature. Sections were again washed, and diaminobenzidine-hydrogen peroxide was used for color development. Sections were then counterstained with 2% hematoxylin and mounted for examination. In each section, 10 full-length villi were randomly selected to count TUNEL-positive cells. Apoptotic cells were identified as cells with brown-stained nuclei, or as apoptotic bodies, which are fragments of apoptotic cells engulfed by neighboring epithelial cells. Intraepithelial lymphocytes were excluded by morphology. All epithelial cells within the villi were counted, and apoptosis was expressed as percentage of apoptotic cells of the total cells for each section. Values for all of the three sections were averaged to reach a percentage of apoptosis for the proximal gut of each animal.
Proliferation was quantified in a similar way with immunostaining for proliferative cell nuclear antigen (PCNA). Deparaffinized and rehydrated sections were incubated with a horseradish peroxidase conjugated PCNA-antibody (SC-56, Santa Cruz Biotechnology, Santa Cruz, CA) at a 1:50 dilution overnight at 4°C, followed by washing in PBS, and diaminobenzidine-hydrogen peroxide for color development. After counterstaining and mounting, PCNA-positive cells were counted on three sections for each animal as described earlier. All examinations were carried out by blinded observers (M. Spies and V. L. Chappell).
Serum collected at death was assayed for TNF-α by ELISA following the manufacturer's instructions (Biosource International, Camarillo, CA).
For RT-PCR, total cellular RNA was isolated from small bowel samples by acid guanidinium thiocyanate-phenol-chloroform extraction using TRIzol reagent (GIBCO-BRL, Rockville, MD), Samples were homogenized in TRIzol reagent on ice, and total RNA was extracted following the manufacturers′ instructions. Extracted RNA was quantitated by ultraviolet (UV) spectrophotometry and stored at −80°C for future analysis. The cDNA reaction as well as the PCR were performed with an optimized buffer and enzyme system (Titan One Tube RT-PCR System; Roche, Indianapolis, IN) according to the manufacturer's instruction. This system is designed to use avian myeloblastosis virus (AMV) RT for first-strand synthesis and the Expand high-fidelity blend of thermostabile DNA polymerases, which consists of Taq DNA polymerase and a proof reading polymerase, for the PCR part. The reaction was carried out in 50-μl volume containing 50–100 ng of the total RNA, 10 pM of forward and reverse primers specific for TNF-α (GenBank accession no. M11731, forward: 5′-AGC AAA CCA CCA AGT GGA GG-3′ and reverse: 5′-CAA GGT ACA ACC CAT CGG CT-3′), 1× PCR buffer with Mg2+, 0.2 mM 2-deoxynucleotide 5′-triphosphate, 5 mM dithiothreitol solution, 5–10 units of RNAse inhibitor, and 0.05 U/μl reaction of the enzyme mix (high-fidelity enzyme mix, RT, and AMV in storage buffer). An initial RT step was performed at 50°C for 30 min and 94°C for 2 min for one cycle, followed by 35 cycles (denaturation 94°C for 30 s, annealing at 59°C for 45 s, and extension at 68°C for 30 s), and finally, one cycle at 68°C for 7 min. In addition, a pair of primers was designed to amplify a portion of the mouse β-actin transcript that spans an exon/exon boundary (GenBank accession no. W82269, forward: 5′-CCT TCA ACA CCC AGC CAT GT-3′ and reverse: 5′-TGT GGA CCA CCA GAG GCA TAC-3′). β-Actin was used as a “housekeeping gene” to provide an internal marker for mRNA integrity within the experiment. PCR products were separated on (1% wt/vol) agarose gels, visualized by ethidium bromide staining under UV light. Image capture and density analysis of bands were done with the SynGene gel documentation system (SynGene-Synoptics, Cambridge, UK).
For evaluation of caspase-8 presence and activity, tissue samples were homogenized in lysis buffer. For Western blot, 25–30 μg of total protein from the tissue extract were separated on a 10% SDS-polyacrylamide gels under reducing conditions and transferred to nitrocellulose membranes (Hybond-C; Amersham Pharmacia Biotech, Piscataway, NJ) in a semidry blotting chamber. After blockage of nonspecific binding sites with 5% nonfat milk in PBS containing 0.1% Tween-20, membranes were incubated in 1:1,000 dilution of anticaspase-8 rabbit polyclonal antibody (Santa Cruz Biotechnology) for 2 h at room temperature. After extensive washing, the nitrocellulose membrane was incubated with anti-rabbit IgG conjugated with horseradish peroxidase (final concentration 1:2,000) for 90 min at room temperature. Bound antibodies were detected with enhanced chemiluminescence Western blotting detection reagents (Amersham Pharmacia Biotech) according to manufacturer's instructions. Image capturing and density analysis of bands were again performed using the SynGene gel documentation system (SynGene-Synoptics).
Caspase-8 activity was determined by a colorimetric assay (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions. In brief, tissue extract of 100 to 200 μg of total protein were incubated with 5 μl of a caspase-8 specific peptide conjugated with IETD-p-nitroaniline in reaction buffer at 37°C for 2 h; then light absorption was read at a wavelength of 405 nm together with controls. Results were compared between groups.
Statistical analysis of data was performed by one-way ANOVA with Tukey's test using a statistical software package (SigmaStat 2.03, SPSS, San Rafael, CA). Significance was accepted at P< 0.05.
Over the 12-h study period, no significant changes in body weights were seen. Total liver weights were not different between groups. Wet weights of the proximal small bowel decreased in burned and anti-TNF-α antibody-treated animals compared with controls (Table1). Dry weights of the proximal small bowel showed a similar decrease.
Histologic measurements showed significantly decreased mucosal height after severe burn (447 ± 30 vs. 602 ± 20 μm;P < 0.05). Animals treated with anti-TNF-α showed significantly less mucosal atrophy (533 ± 12 vs. 447 ± 30 μm, P < 0.05) than untreated animals (Table2 and Fig.1). Treatment with neutralizing antibody alone in unburned animals did not show morphological changes compared with unburned controls. When villus heights and crypt depth were analyzed separately, only villus height was decreased by burn and again was partially restored by anti-TNF-α, whereas crypt depth was not affected (Fig. 2). A similar pattern was seen in the total mucosal cell number. In burned animals, total mucosal cell number was significantly decreased. This effect is partially, but not completely, restored by anti-TNF-α treatment (Fig.3). Again, the main changes were seen in the villus and not in the crypt.
Proliferating intestinal cells, as identified by PCNA staining, were found in the crypt. However, the number of PCNA positive stained cells per crypt was not different among groups (Fig.4). TUNEL-positive cells and apoptotic bodies were much more likely to be found in sections of burned animals. An over threefold increase in apoptotic cell number was seen after burn. This response was diminished by anti-TNF-α treatment. However, this did not result in a complete return to unburned control values (Fig. 5).
Systemic TNF-α levels showed no significant differences among groups when tested 12 h after burn (data not shown). Gene expression levels for TNF-α in small bowel, determined by RT-PCR, showed no differences among groups (Fig.6). Caspase-8 expression and activity by Western blot and activity assay also showed no differences among groups (Fig. 7).
Our results show that the response of gut epithelium to severe burn is a decrease in small bowel weight, which occurs within 12 h of injury. These changes take place without concomitant changes in total body weight. This weight loss in the proximal small bowel is an indication that injury induces gut mucosal atrophy and that this atrophy occurs very early after injury (within 12 h). Histologic changes associated with this finding are loss of mucosal height and cell number. The changes after injury occur predominantly in the villus region. In this study severe burn did not induce any changes in intestinal cell proliferation. However, intestinal cell apoptosis occurred at a threefold higher rate in burned animals. Additionally, we found that atrophic changes in mucosal height and cell mass could be partially restored by treatment with a neutralizing anti-TNF-α antibody, implicating TNF as an effector of gut mucosal changes seen after severe burn. The effect is clearly more pronounced in the villus region and does not seem to occur in the crypt. Increased apoptosis in burned animals was diminished by anti-TNF-α treatment without changes in proliferation.
As previously shown by our group, the loss of gut epithelial cells after severe burn is due to an increase in apoptotic cell death (19). This was coupled with an increase in proliferation, indicating increased cell turnover after injury. Maximum response in gut epithelium in this model was seen at 12 h after burn, after which the response quickly diminishes (19). Although similar effects may also be seen at earlier or later time points, the most prominent changes were seen at 12 h, making this a valuable time point to investigate the role of TNF-α in gut epithelial apoptosis. Potential mechanisms for these events are changes in gut perfusion with burn or the systemic effect of inflammatory mediators released immediately following injury. Studies have shown that in rat models, ischemia alone is able to induce apoptosis in jejunal and ileal mucosa, which is further exacerbated by reperfusion effects (6, 10, 17). However, this appears not to be the cause for the changes observed in severe burn. When addressing this issue we showed in a recent study that significant gut hypoperfusion occurs with severe burn injury within the first 4 h (14). When inducing a comparable hypoperfusion in the gut of unburned animals, the apoptotic index did not differ from normal controls, indicating that hypoperfusion induced by burns is not sufficient by itself to cause the observed increase in apoptosis. This confirms that, in fact, the atrophic changes are induced by increased apoptosis of gut epithelial cells and shifts the point of interest to other potential initiation mechanisms of apoptosis.
Several soluble and membrane-bound factors are known to induce apoptosis. Activation of members of the TNF-receptor superfamily as well as associated proteins (TNFR, TRAIL, TRADD, FADD, etc.) has been implicated as a potential mechanism in many cell systems as an inductor of apoptosis (1). Garside et al. (4) showed that a single dose of TNF-α alone is able to induce typical small bowel pathology, which is manifested as crypt hyperplasia and villus atrophy with 15 min of application. In a study by Piguet et al. (13) neutralization of TNF-α in a murine acute graft-versus-host-disease (GVHD) model reduced target organ damage. Recently, Stuber et al. (16) showed in a murine GVHD-model that FasL-Fas interaction is not involved in the induction of apoptosis of the small intestinal mucosa. The neutralization of TNF-α reduced the amount of apoptosis and the extent of mucosal atrophy. Another study by Guy-Grand et al. (5) showed increased villous cell apoptosis in TNF-α treated normal animals, which appears to be dependent on the presence of intraepithelial lymphocytes. This differs from increased apoptosis in lymphatic tissues such as spleen and thymus found in reaction to burn injury, which was linked to increased FasL mRNA expression and increased caspase-3 activity (2, 3). Apparently the apoptotic response is triggered differently in distinct tissues.
Triggering of the apoptotic response by TNF-α in gut epithelium may be initiated by either increased systemic levels acting locally in the epithelium or by increased local production of TNF-α in the gut mucosa itself. Both mechanisms are valid options and could explain the observed phenomenon of increased gut epithelial apoptosis.
Systemic levels of TNF-α have been found to be elevated after burns by several authors (7, 8, 21, 22) with burn injury alone as well as with an additional second hit. Systemic effects of TNF-α neutralization were investigated by O'Riordain et al. (12) who described a time and dose dependency of survival in a burn and septic challenge model. Efficacy of treatment depended on intrinsic TNF-α production, showing greatest effects of neutralization with maximum TNF-α levels. Applying these results to our study led us to an early application of anti-TNF-α with knowledge of sharp and transient increase in systemic TNF-α levels. However, we were not able to detect differing TNF-α levels with burn injury at the 12-h time point. This leads to the assertion that only early neutralization might be effective and, secondly, that TNF-α may only act as a transient trigger of apoptosis in the small bowel by activating a cascade of events, which in turn, consolidates the tissue response.
On the other hand Ogle et al. (11) and Wu et al. (20) showed increased local production of TNF-α in enterocytes and hepatocytes of guinea pigs after thermal injury. This increase in local expression of TNF-α in small bowel may also induce apoptotic changes. These changes in local TNF-α levels could be due to increased transcription or increased translation. However, we were not able to detect specific differences in local TNF-α mRNA expression. Here again, TNF-α may only act as a trigger mechanism for induction of apoptosis at an early time point after burn and not be expressed at elevated levels at the 12-h time point. The reported findings do exclude either of these mechanisms but also are not able to support either of these hypotheses. However, the presence of increased TNF-α levels may be irrelevant, as effects may also be realized with increased activity below the sensitivity of our methods. To elucidate the effective source of the TNF-α trigger requires further investigation.
The activation phase of cell death includes a variety of transduction pathways with signals implicating FasL-receptor interactions (FasL-Fas, TNF-α-TNFR) and the additional proteins (FADD, RIP, TRADD) (15,18). These proteins, acting as intermediaries with “death domains,” in turn activate procaspase-8 and consequent caspase-3 and -6. Although we were not able to show differences in caspase-8 activation with anti-TNF-α treatment, this does not preclude a TNF-α-mediated mechanism for increased gut epithelial apoptosis. Because the absolute number of cells undergoing apoptosis at a certain time point is 3–8 cell per 1,000 mucosal cells, differences on the protein expression levels would be predicted to be small, if at all detectable. However, the effects at a physiological level, here as mucosal atrophy, are clearly demonstrated and suggest that TNF-α is an effector, at some level, of diminished gut mucosal integrity.
It is speculated that the administration of anti-TNF-α in the early phase of injury binds to systemic serum TNF-α or locally abundant TNF-α. The early neutralization of circulating soluble and transmembrane forms of TNF-α thus removes the initiating factor for the apoptotic cascade in gut mucosa. A prevention of TNF-α/TNFR interaction would reduce effective signaling with caspase activation. An alternative mechanism might be the inhibition of signals activating the mitogen-activated protein kinase pathway (9).
From this study, we conclude that changes of gut mucosal homeostasis seen after severe burn are associated with activation of apoptosis by TNF-α-TNFR interaction. The effect of TNF-α neutralization on gut mucosal homeostasis partially reversed mucosal atrophy after burns in this study. It remains open for discussion whether this is, in fact, beneficial on the gut or systemic level. Anti-TNF-α strategies were indeed successful to improve survival and outcomes in injured animals (12). However, similar use in patients was unsuccessful. Only in defined studies in other animal models and perhaps in patients will it be possible to determine whether TNF-α neutralization may be of benefit. Whether these findings can be used to clinical advantage will require further study to elucidate the source of TNF-α involved in the response and, secondly, to determine whether inhibiting mucosal atrophy after burn is of clinical benefit.
This study was supported by Shriners Hospitals for Children Grant 8580.
This study was presented, in part, at the Surgical Forum of the Annual Meeting of the American College of Surgeons, Chicago, IL, October 22–28, 2000.
Address for reprint requests and other correspondence: Steven E. Wolf, Shriners Hospitals for Children, 815 Market St., Galveston, Texas 77550 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
March 28, 2002;10.1152/ajpgi.00149.2001
- Copyright © 2002 the American Physiological Society