p38 MAPK regulates Bax activity and apoptosis in enterocytes at baseline and after intestinal resection

Abstract

Increased apoptosis in crypt enterocytes is a key feature of intestinal adaptation following massive small bowel resection (SBR). Expression of the proapoptotic factor Bax has been shown to be required for resection-induced apoptosis. It has also been demonstrated that p38-α MAPK (p38) is necessary for Bax activation and apoptosis in vitro. The present studies were designed to test the hypothesis that p38 is a key regulator of Bax activation during adaptation after SBR in vivo. Enterocyte expression of p38 was deleted by tamoxifen administration to activate villin-Cre in adult mice with a floxed Mapk14 (p38-α) gene. Proximal 50% SBR or sham operations were performed on wild-type (WT) and p38 intestinal knockout (p38-IKO) mice under isoflurane anesthesia. Mice were killed 3 or 7 days after operation, and adaptation was analyzed by measuring intestinal morphology, proliferation, and apoptosis. Bax activity was quantified by immunoprecipitation, followed by Western blotting. After SBR, p38-IKO mice had deeper crypts, longer villi, and accelerated proliferation compared with WT controls. Rates of crypt apoptosis were significantly lower in p38-IKO mice, both at baseline and after SBR. Levels of activated Bax were twofold higher in WT mice after SBR relative to sham. In contrast, activated Bax levels were reduced by 67% in mice after p38 MAPK deletion. Deleted p38 expression within the intestinal epithelium leads to enhanced adaptation and reduced levels of enterocyte apoptosis after massive intestinal resection. p38-regulated Bax activation appears to be an important mechanism underlying resection-induced apoptosis.

short bowel syndrome (sbs) results after massive intestinal resection when the amount of remnant bowel is insufficient to absorb enteral nutrition. As a result, patients with SBS require supplemental nutrition by vein. Unfortunately, parenteral nutrition is laden with complications, including high cost, need for central venous access, catheter sepsis, liver failure, and death. Intestinal adaptation is a compensatory response in which the gut undergoes functional and morphological changes to improve nutrient absorption (29). If adaptation is complete, patients with SBS are able to achieve autonomy from parenteral nutrition.

Crypt and villus elongation are well-described morphological adaptive changes that occur after massive small bowel resection (SBR) and serve to increase mucosal surface area (45, 46). Not surprisingly, crypt cell proliferation is accelerated to generate more cells for mucosal growth (11, 12). Somewhat paradoxically, apoptosis, or programmed cell death, also increases (14, 17, 40). While a majority of scientific research has emphasized understanding the proliferative response after SBR, less is known about the mechanisms underlying resection-induced apoptosis. The importance of apoptosis is suggested by several studies, which have found that inhibition of apoptosis, coupled with stimulation of proliferation, may result in enhanced intestinal adaptation responses (5, 21, 35, 39, 44).

Apoptosis is triggered by two major pathways. The extrinsic or death receptor pathway involves extracellular signals that are transduced by specific receptors to initiate cell death. Alternatively, the intrinsic or mitochondrial pathway is operative when the internal cellular milieu prompts cell suicide (9). Using tumor necrosis factor receptor-1 and FAS-knockout mouse models, our laboratory has previously demonstrated that these two components of the extrinsic pathway are not necessary for resection-induced apoptosis (25). Based on these findings, we have, therefore, focused on elucidating how the intrinsic pathway mediates gut apoptosis during adaptation. Following SBR, increased expression of the proapoptotic protein Bax has been recorded in enterocytes, which correlates with augmented rates of apoptosis (4, 13, 27, 36, 38). Furthermore, using Bax-null mice, we have established that Bax expression is required for resection-induced apoptosis (22, 37).

Under normal conditions, Bax exists in the cytosol in an inactive form (47). During the process of apoptosis induction, Bax undergoes a conformational change (7), permitting Bax to translocate to the mitochondrion, oligomerize, and form pores in the outer mitochondrial membrane, thereby allowing cytochrome c and other pro-apoptotic molecules to escape into the cytoplasm (1, 20, 23, 24, 26, 47). In the cytosol, these factors activate caspases, which then execute the apoptotic cascade (24, 49).

Epidermal growth factor receptor (EGFR) signaling is important to the normal adaptive response after massive SBR (15, 16). Inhibition of the EGFR results in impaired adaptation and augments resection-induced apoptosis, while EGFR stimulation enhances intestinal adaptation and reduces gut apoptosis (15, 16). Recently, our laboratory has found that EGFR inhibition in vitro induces enterocyte apoptosis. In the dying enterocytes, the activated form of Bax was found to be localized to the mitochondria (34). Moreover, we discovered that, within cultured intestinal epithelial cells, p38-α MAPK (p38) was necessary for both Bax activation and apoptosis in the context of EGFR inhibition. Given the involvement of EGFR signaling during adaptation and apoptosis, we, therefore, sought to test the hypothesis that p38-α MAPK would regulate Bax activity during resection-induced apoptosis in vivo.

In the first set of experiments, we characterized the phenotype of an inducible, intestinal epithelial-specific p38-α MAPK-knockout (p38-IKO) mouse line. While a constitutive, intestinal epithelial p38 knockout mouse model has been recently described (32), this is the first time that p38 has been acutely deleted in enterocytes of adult mice. Next, we tested the effects of p38 deletion on intestinal adaptation and Bax activity after massive SBR.

METHODS

Animals

The protocol for this study was approved by the Washington University Institutional Animal Care and Use Committee (Protocol nos. 20070145 and 20100103; Washington University School of Medicine, St. Louis, MO). All mice were bred and housed in a standard vivarium with a 12:12-h day-night cycle.

Inducible, intestinal epithelial-specific p38-α-knockout mice were generated by crossing mice with a tamoxifen (TAM)-inducible Cre-fusion protein under control of the villin promoter (8) [vil-Cre ERT2; obtained via generous donation from Sylvie Robine, Curie Institute, Paris, France] with mice in which the floxed p38-α MAPK gene had been tagged for recombination (31) [p38(f/f); acquired via generous donation from Huiping Jiang, Boehringer Ingelheim Pharmaceuticals, Ridgefield, CT]. Mice harboring both the inducible Cre fusion protein and floxed p38 alleles [vil-Cre ERT2(+), p38(f/f)] were identified as p38-IKO mice. Littermates lacking Cre [vil-Cre ERT2(−), p38(f/f)] were used as wild-type (WT) controls. Unless noted otherwise, p38 expression was disrupted in 7-wk-old male mice by giving intraperitoneal TAM injections (0.5 mg/day for 2 consecutive days).

Genotyping

Polymerase chain reaction (PCR) was used to determine mouse genotype from tail clippings in the standard fashion using previously published protocols for vil-Cre ERT2 (8) and p38(f/f) (31).

Experimental Design

Baseline studies.

One week after the first TAM injection, mice were weighed and killed. Small bowel length was measured and collected for further analyses. We studied the effects of 2 and 4 wk of p38 deletion on small bowel length by injecting mice with TAM at 6 and 4 wk of age, respectively.

SBR.

One week after initial TAM injection, mice were randomly assigned to either 50% proximal SBR or sham operation. We compared parameters of intestinal adaptation (villus height, crypt depth, proliferation, and apoptosis) between WT and p38-IKO mice at postoperative day 3 (POD3) and postoperative day 7 (POD7).

Operative procedure.

Specific details of the 50% SBR and sham procedures have been described previously (18). Briefly, under isoflurane anesthesia, the 50% SBR was performed by transecting the small intestine 12 cm proximal to the cecum. A second transection was made proximally, 1–2 cm distal to the ligament of Treitz; the appropriate mesenteric blood vessels were ligated with 6–0 nylon suture; and intervening bowel was removed. An anastomosis of the remaining small intestine was then constructed using 9–0 nylon suture. The sham operation was completed by transecting and immediately reanastomosing the small bowel 12 cm proximal to the cecum.

Two days before surgery, mice were placed on a liquid rodent diet (Micro-Stabilized Rodent Liquid Diet LD101; Purina Mills, St. Louis, MO). For the first 24 h following operation, mice were allowed only water ad libitum. Thereafter, mice were fed liquid rodent diet ad libitum until death. Mice that died, had evidence of intestinal obstruction at the time of death, or appeared ill were excluded from further analyses.

Small bowel harvest.

At death, mice were anesthetized with a solution containing ketamine (100 mg/kg), xylazine (20 mg/kg), and acepromazine (3 mg/kg). The abdominal cavity was opened, and the bowel was immediately flushed with ice-cold phosphate-buffered saline with protease inhibitors and phosphatase inhibitors in situ. The intestine was then excised and measured after stretching with a 5-g weight. A 2-cm segment of small intestine was removed starting 12 cm proximal to the cecum. This piece of intestine was fixed in 10% formalin for histological analyses. The remaining remnant distal bowel was stored in ice-cold phosphate-buffered saline with inhibitors, and enterocytes were isolated, as our laboratory has previously described (10). This results in crypt and villi specimens with minimal contamination from the submucosal or muscular layers.

Histology.

Formalin-fixed specimens of intestine were embedded in paraffin, cut into 5-μm-thick longitudinal sections, and mounted on glass slides. Hematoxylin and eosin (H&E) stained sections were used to measure crypt depth and villus length with a video-assisted computer program (Metamorph, UIC, Downingtown, PA). A minimum of 20 well-oriented crypts and villi were measured per subject by investigators who were blinded to the experimental group. Crypt and villus densities were similarly measured by counting the number of crypts and villi in a 2-mm section of small bowel.

Crypt cell proliferation.

Ninety minutes before death, mice were given an intraperitoneal injection of 5-bromodeoxyuridine (BrdU; 0.1 ml/10 gm body wt; Zymed Laboratories, San Francisco, CA). Formalin-fixed tissue sections were immunostained for BrdU, as previously described (10). The number of positively-staining cells and the total number of cells per crypt were counted from at least 20 well-oriented crypts by blinded scoring. A proliferative index was calculated from the ratio of these measurements.

Crypt cell apoptosis.

Crypt apoptosis was measured from H&E and cleaved caspase-3 (CC3) stained slides, as previously described (43). Apoptotic bodies were determined by the presence of pyknotic and fragmented nuclei. One hundred well-oriented crypts were counted for each mouse by blinded scoring, and the apoptotic index was defined as the number of apoptotic bodies per 100 crypts. For CC3 staining, the same procedure was used as for BrdU staining, except the antibody was replaced with anti-CC3 antibody (1:15,000, Cell Signaling, Danvers, MA).

Crypt fission.

Crypt fission was counted from H&E stained slides by noting the number of crypts with branching from the crypt base from 100 well-oriented crypts by blinded scoring.

RNA extraction, quantitative real-time PCR, and Illumina array.

RNA was prepared from <100 mg of isolated material. Crypts, villi, or the muscle remnant were homogenized in lysis buffer (RNAqueous kit, Ambion, Austin, TX), and RNA was extracted according to kit instructions. RNA was stored at −80°C until use. Total RNA concentration was determined using a NanoDrop Spectrophotometer (ND-1000, NanoDrop Technologies, Wilmington, DE). RNA quality was evaluated using an Experion System with an RNA StdSens Chip and reagents (Bio-Rad Laboratories, Richmond, CA). For some RNA samples, an Illumina array was done on a MouseRef chip by the University Microarray Core Facility (Washington University, St. Louis, MO). Complementary DNA (cDNA) was prepared from high-quality RNA using an RT2 First Strand Kit (SABioscience, Fredrick, MD). cDNA was quantified using Quanti-iT OliGreen ssDNA Assay kit (Invitrogen, Carlsbad, CA). Equivalent amounts of cDNA were used in all reactions. β-Actin was used as the endogenous control, and a standard whole bowel sample was used as the calibrator. Mapk14 (p38-α MAPK) and Bax gene expression were determined using primers and reagents from SABioscience (Frederick, MD) and using an Applied Biosystems 7500 Fast Real-Time PCR system (Foster City, CA).

Western blotting.

We confirmed the phenotype (presence or absence of p38 protein) of every mouse by Western blot. Isolated enterocytes were stored at −80°C, reconstituted with Tris buffer, lysed with SDS sample buffer, heated at 100°C for 5 min, and sonicated. Protein concentrations were measured using an RC DC kit from Bio-Rad. Equal amounts of protein were subjected to Western blotting assay.

Bax activity.

To detect the activated Bax isoform, isolated crypts were lysed with 1% 3-([3-cholamidopropyl]dimethylammonio)-1-propanesulfonate (CHAPS) in Tris buffer for 90 min at 4°C and spun down at 14,000 rpm for 30 min; the supernatant was saved as CHAPS extracts; and protein concentrations in extracts were measured as above. Protein (2 mg) was precleared with 10 μl protein G plus A beads for 1 h at 4°C. The specific Bax 6A7 antibody, which recognizes the activated form of Bax, was used to immunoprecipitate the activated Bax at 4°C overnight. The beads were washed, and bound Bax was lysed with ×1 SDS sample buffer. Bax protein was resolved in 18% SDS-PAGE gel and analyzed with Western blot using an antibody against the total Bax.

Cyotchrome c quantification.

Mitochondrial and cytosolic fractions were separated from isolated crypts according to the specifications from Pierce's Mitochondrial Isolation Kit for Tissue (Thermo Scientific, Rockford, IL). Cytochrome c was detected by Western blotting.

Antibodies.

anti-p38-α MAPK, anti-p38-γ MAPK, anti-total p38 MAPK, anti-total Bax, and anti-actin antibodies are from Cell Signalling (Danvers, MA); anti-cytochrome c and anti-Bax 6A7 antibodies are from BD Pharmingen (San Diego, CA).

Statistics.

Statistics were performed with SigmaStat software (SPSS, Chicago, IL), and all results are presented as the mean ± SE. Single comparisons for normally distributed data were performed with Student's t-tests. When normality failed, a Mann-Whitney rank sum test was used. Two-way ANOVA was done for multiple comparisons. P values <0.05 were considered significant.

RESULTS

p38-α MAPK Was Successfully Deleted in Enterocytes After Administration of TAM

Protein levels of both p38-α and total p38 MAPK were reduced by nearly 95% after TAM administration, suggesting that p38-α is the primary isoform of p38 MAPK expressed in intestinal enterocytes (Fig. 1A). We confirmed that p38-α was the only isoform of p38 affected, since levels of p38-γ were unaffected in the knockout mice (data not shown). mRNA levels of the Mapk14 gene were reduced by 80% in both the crypts and villi of the p38-IKO mice (Fig. 1B). Furthermore, p38-α expression was not altered in the muscle layer of the bowel at both the mRNA (Fig. 1B) and protein level (data not shown), confirming that the gene deletion was successfully targeted to enterocytes.

Fig. 1.

p38-α MAPK is successfully deleted in enterocytes. Intestinal crypts (C) and villi (V) were isolated from wild-type (WT) and inducible, enterocyte-specific p38-α MAPK-null [p38-IKO or knockout (KO)] mice 1 wk after tamoxifen injection (0.5 mg/day × 2 doses). The underlying intestinal smooth muscle was also collected. A: Western blot analysis of total p38 MAPK and the p38-α isoform. Actin was used as the loading control. B: quantitative real-time PCR analysis of Mapk14 (the gene encoding p38-α MAPK) gene expression in crypts, villi, and underlying smooth muscle. Values are means ± SE; n = 4 in WT and KO groups. #P < 0.001.

Effects of p38 MAPK Deletion on the Small Intestine

One week after TAM injection, p38-IKO mice had a slightly, but statistically significant longer small intestine than WT controls [38.9 ± 0.5 cm (n = 39) vs. 36.6 ± 0.5 cm (n = 32), P = 0.001; Fig. 2A]. This longer small bowel effect persisted at 2 and 4 wk after gene deletion [2 wk: 42.2 ± 0.6 cm (n = 13) vs. 40.2 ± 0.7 cm (n = 17), P = 0.04; 4 wk: 39.6 ± 0.7 cm (n = 10) vs. 36.8 ± 0.7 cm (n = 10), P = 0.01]. There were no differences between groups in animal weights at initial TAM injection at any time. While mouse weight was not different between WT and p38-IKO mice both at 1 and 2 wk after TAM, mice in the knockout group were heavier 4 wk after p38 deletion (23.1 ± 0.5 vs. 21.5 ±0.6 g, P = 0.04).

Fig. 2.

p38 deletion causes small bowel hyperplasia. WT and p38-IKO (KO) mice were killed 1 wk after tamoxifen injection (0.5 mg/day × 2 doses). A: small bowel length as measured with a standardized 5-g weight. The middle line represents the mean, the edges of the box represent the 75th percentile, and the error bars show the 95th percentile. B: a proliferative index was calculated from a ratio of 5-bromodeoxyuridine (BrdU+) cells to total crypt cells. An apoptotic index was calculated both from the number of apoptotic bodies per 100 crypts on routine hematoxylin and eosin sections (C), or the number of crypt cells positively staining for cleaved caspase 3 (CC3; D). E: crypt fission. A crypt fission index was calculated from the number of crypts undergoing fission per 100 crypts. Values are means ± SE. *P < 0.05, %P < 0.01, &P < 0.005, #P < 0.001.

Rates of crypt cell proliferation were augmented by p38 deficiency, as evidenced by the higher rate of BrdU-labeled cells in the p38-IKO group compared with WT [9.5 ± 0.4 (n = 12) vs. 7.8 ± 0.5 BrdU-staining cells/crypt (n = 8), P = 0.02]. Because there was no difference in the total number of cells per crypt, the proliferative index (BrdU + cells/total number of crypt cells) was greater in the p38-IKO mice [24.03 ± 0.93% (n = 12) vs. 19.81 ± 1.27% (n = 8), P = 0.01; Fig. 2B]. The rates of apoptosis in crypt enterocytes was reduced by roughly one-half in the p38-IKO mice [H&E: 4.9 ± 0.6 (n = 11) vs. 8.9 ± 1.3 apoptotic bodies/100 crypts (n = 10), P = 0.008; CC3: 4.3 ± 0.7 (n = 11) vs. 9.3 ± 1.0 apoptotic bodies/100 crypts (n = 10), P < 0.001; Fig. 2, C and D].

Despite the changes in cell kinetics in the p38-IKO intestine, mucosal morphology was not altered at baseline. Neither crypt depth nor villus length was significantly different after intestinal epithelial p38 deletion compared with WT values at any time point. However, p38 deficiency did result in accelerated rates of crypt fission compared with WT values [5.4 ± 0.4 (n = 16) vs. 3.3 ± 0.5 fission events/100 crypts (n = 11), P = 0.004; Fig. 2E]. Increased fission as a consequence of accelerated enterocyte proliferation and reduced enterocyte apoptosis, but without villus or crypt growth, could have possibly accounted for the increased intestinal length in the p38-IKO mice. To test whether p38 deletion resulted in expansion of the intestinal stem cell (ISC) population, we measured expression of several ISC markers via microarray (3, 48). Expression of Lgr5, Msi1, Prom1, Bmi1, Ascl2, and Olfm4 was not significantly affected by intestinal epithelial p38 deficiency (data not shown).

Deletion of p38 MAPK Results in Enhanced Morphological Adaptation After Massive SBR

Disrupted p38 MAPK expression in the intestinal epithelium was confirmed in p38-IKO mice at 3 and 7 days after operation (Fig. 3). Both WT and p38-IKO mice lost significantly more weight after SBR compared with mice undergoing sham operation (P < 0.001; Table 1). However, weight loss was comparable between WT and p38-IKO mice 3 and 7 days after operation. The remnant bowel length distal to the anastomosis was not significantly different after SBR between WT and p38-IKO mice (Table 2).

Fig. 3.

p38 is absent in enterocytes for the duration of the postoperative period. Levels of p38-α were detected by Western immunoblotting in crypt and villus enterocyte protein isolated from WT and p38-IKO (KO) mouse intestine collected 3 (POD3) and 7 days after operation (POD7). Actin was used as the loading control.

Table 1.

Mouse body weight

	WT Sham	WT SBR	KO Sham	KO SBR
Postoperative day 3
n	18	20	18	18
Preoperative	23.9 ± 0.5	25.1 ± 0.5	23.7 ± 0.4	23.5 ± 0.5
Postoperative	22.7 ± 0.4	22.3 ± 0.4	22.6 ± 0.4	21.0 ± 0.5*
Absolute change	−1.2 ± 0.2	−2.8 ± 0.2‡	−1.1 ± 0.2	−2.5 ± 0.2‡
Percent change	−4.8 ± 0.7	−10.9 ± 0.7‡	−4.5 ± 0.8	−10.5 ± 0.9‡
Postoperative day 7
n	15	20	11	20
Preoperative	22.8 ± 0.5	23.6 ± 0.4	23.4 ± 0.8	23.5 ± 0.5
Postoperative	23.2 ± 0.6	22.0 ± 0.5	24.3 ± 0.9	21.5 ± 0.5†
Absolute change	+4.7 ± 0.4	−1.6 ± 0.4‡	+0.9 ± 0.3	−2.0 ± 0.4‡
Percent change	+2.2 ± 2.1	−6.9 ± 1.6‡	+3.9 ± 1.4	−8.3 ± 1.5‡

Values are means ± SE in g, except for percent change, which is shown in %; n, no. of animals. Preoperative weight was measured on the day of operation. Postoperative weight was taken on the day of death. Absolute change was calculated from the difference of pre- and postoperative weights. The percent change was calculated with reference to the preoperative weight. WT, wild type; SBR, small bowel resection; KO, knockout.

^*P < 0.05,

^†P < 0.005,

^‡P < 0.001.

Table 2.

Bowel lengths in operated mice

	WT Sham	WT SBR	KO Sham	KO SBR
Postoperative day 3
n	18	20	18	18
Distal remnant	12.9 ± 0.5	12.8 ± 0.6	12.0 ± 0.5	12.3 ± 0.7
Postoperative day 7
n	16	20	11	20
Distal remnant	13.5 ± 0.5	14.0 ± 0.3	11.7 ± 0.6*†	14.0 ± 0.5

Values are means ± SE in cm; n, no. of animals. The distal remnant was the length of small intestine distal to the anastomosis extending to the ileocecal valve.

^*P < 0.05,

^†P < 0.001.

As expected, intestinal crypts and villi grew in WT mice at both 3 and 7 days after SBR compared with sham. Crypt growth after SBR was more robust in the p38-IKO mice [POD3: 113.8 ± 3.8 μm (n = 12) vs. 101.2 ± 4.0 μm (n = 13), P = 0.01; POD7: 94.9 ±4.3 μm (n = 20) vs. 82.9 ± 2.3 μm (n = 19), P = 0.005; Fig. 4, A and C]. While intestinal villi in the p38-IKO group were not significantly longer 3 days after SBR compared with WT controls, villi were longer 7 days after SBR [309.9 ± 9.3 μm (n = 20) vs. 277.6 ± 6.6 μm (n = 19), P = 0.003; Fig. 4, B and D]. Crypt and villus densities were not significantly affected by either resection status or p38 deletion 3 days after operation (data not shown).

Fig. 4.

p38 deficiency results in enhanced intestinal adaptation after small bowel resection (SBR). WT and p38-IKO mice (KO) underwent either 50% proximal SBR or sham operation (transection and reanastomosis of the bowel alone). Crypt depth and villus length (means ± SE) were measured from H&E-stained sections of ileum. Crypt depth (A) and villus length (B) from ileum harvested on POD3 are shown. Crypt depth (C) and villus length (D) from ileum taken on POD7 are shown. *P < 0.05, %P < 0.01, &P < 0.005, #P < 0.001.

Deletion of p38 MAPK Alters Cellular Kinetics After SBR

Both WT and p38-IKO mice demonstrated accelerated crypt proliferation rates after SBR compared with sham-operated mice (Fig. 5A). However, p38-IKO mice had even greater rates of crypt proliferation compared with WT. While p38-IKO animals had more total cells per crypt after SBR than WT mice [55.4 ± 1.4 (n = 12) vs. 48.3 ± 1.3 total cells/crypt (n = 13), P < 0.001], they also had more crypt cells incorporating BrdU [18.2 ± 0.8 (n = 12) vs. 14.1 ± 0.5 BrdU+ cells/crypt (n = 13), P < 0.001], resulting in enhanced crypt proliferation rates [32.9 ± 1.0 (n = 12) vs. 29.4 ± 0.5% of cells/crypt incorporating BrdU (n = 13), P < 0.001]. The differences in crypt fission index between the WT and p38-IKO animals were no longer evident after SBR in the remnant bowel (Fig. 5B).

Fig. 5.

p38 deficiency alters enterocyte kinetics after SBR. Enterocyte kinetics were measured in WT and p38-IKO (KO) mice on POD3 following SBR and sham operation. A: a proliferative index was calculated from a ratio of BrdU+ cells to total crypts cells. B: the crypt fission index was calculated from the number of crypt fission events per 100 well-oriented crypts. An apoptotic index was calculated from the number of apoptotic bodies per 100 crypts from H&E (C) or CC3 immunostained slides (D). Values are means ± SE. *P < 0.05, %P < 0.01, #P < 0.001.

Deletion of p38 reduced crypt apoptosis following bowel resection. Crypt cell apoptosis increased in WT mice after bowel resection as expected (Fig. 5, C and D). The p38-IKO mice had reduced levels of crypt apoptosis compared with WT mice after SBR [H&E: 6.5 ± 0.6 (n = 12) vs. 10.8 ± 0.6 apoptotic bodies/100 crypts (n = 13), P < 0.001; CC3: 5.4 ± 0.6 (n = 12) vs. 7.4 ± 0.5 apoptotic bodies/100 crypts (n = 13), P = 0.01].

p38 MAPK Regulates Bax Activity in Small Intestine Enterocytes

Previously our laboratory found that Bax activity is regulated by p38 MAPK in rat intestinal epithelial cells in vitro (34). To determine whether Bax is also regulated by p38 in the more complicated in vivo state, we harvested small intestine crypts from nonoperated WT and p38-IKO mice and confirmed that Bax activity was diminished in the absence of p38 (P = 0.04; Fig. 6, A and B). Bax expression was unaltered by p38 deletion (data not shown). In addition to activated Bax, we also detected lower levels of cytochrome c in the cytoplasmic fraction of intestinal crypt cells in p38-IKO mice (P = 0.004; Fig. 6, C and D). Therefore, mice with p38 deletion have impaired ability to activate Bax, thereby preventing Bax from polymerizing, forming pores in the mitochondrial membrane, and allowing cytochrome c to escape into the cytosol.

Fig. 6.

Bax activity is augmented after SBR and is regulated by p38 MAPK. A: Bax activity was measured in WT and p38-IKO (KO) mice by immunoprecipitating the active form of Bax with an isoform-specific antibody (6A7), followed by Western blotting for total Bax. The amount of total Bax used as input is shown for control. B: the amount of active Bax relative to input was quantified and is shown in the graph adjacent; n = 3 in each group. C: cytochrome c was measured by Western blotting in cytoplasm fractions of isolated crypts from WT and KO mice. D: quantification of the cytochrome c level against actin; n = 8 for each group. E: C57/B6 mice underwent SBR or sham operation and were killed on POD3. F: the amount of active Bax was quantified against total Bax and is shown; n = 6 in each group. G: p38-IKO mice underwent either sham or SBR and were killed on POD3. H: quantification of active Bax against the input total Bax; n = 3 in each group. Values are means ± SE. *P < 0.05, &P < 0.005.

Bax Is Activated During Resection-induced Apoptosis

To better understand the mechanism by which p38 deletion leads to reduced enterocyte apoptosis after resection, we performed 50% SBR and sham operations on C57/B6 mice. Mice were killed on POD3, the distal remnant bowel was harvested, and crypts were isolated. Using immunoprecipitation to pull down the activated isoform of Bax, we found that significantly more Bax exists in its activated state after SBR compared with sham levels (P = 0.04; Fig. 6, E and F). This resection-induced increase in Bax activity was prevented in the p38-IKO mice (Fig. 6, G and H). These findings support our hypothesis that p38 MAPK regulates resection-induced enterocyte apoptosis by regulating Bax activity.

DISCUSSION

Herein, we demonstrate for the first time that the proapoptotic protein Bax is activated during intestinal adaptation after massive SBR. Coupled with increased Bax expression (4, 13, 27, 36, 38) this finding suggests that the activity of endogenous Bax is also a contributor toward the elevated rates of enterocyte apoptosis observed after SBR. Furthermore, we provide evidence supporting our hypothesis that Bax activation is mediated by p38 MAPK in vivo. Finally, we provide data suggesting that enterocyte-targeted p38 deletion reduces crypt apoptosis in mice by reducing Bax activity.

Perturbed expression of p38 in the intestinal epithelium results in enhanced intestinal adaptation after massive SBR. Our data suggest that this may be due to a combination of attenuated crypt apoptosis and accelerated proliferation. Inhibition of enterocyte apoptosis alone has variably affected resection-induced adaptation responses. Although we have not observed enhanced adaptation in Bax-null mice subjected to SBR (22), Tang et al. demonstrated enhanced adaptation after SBR in the same Bax-deficient background (39). We presently cannot account for the disparate results between the two studies, but it might be due to the different postoperative time intervals (1 mo vs. 1 wk) or regions of the bowel (distal ileum vs. jejunum) investigated. Notwithstanding, adaptation has been observed to be boosted in association with experimental conditions in which enterocyte proliferation is stimulated, while apoptosis is attenuated (5, 21, 35). Further investigation into manipulation of resection-induced apoptosis responses would, therefore, seem to have therapeutic benefit as a means to generate greater mucosal surface area in response to massive intestinal loss.

While p38 deletion augmented morphological adaptation after SBR, animal weights were not altered in our experiments. Given the greater magnitude of adaptation in p38-IKO mice, we would have expected less weight loss or quicker weight recovery after SBR. In these experiments, we only followed mice to POD7. Perhaps animal weights between groups would diverge at later time points after resection. Alternatively, it must be considered that functional adaptation after removing 50% of the small bowel may be sufficient for both groups of mice, and there would be no need to absorb or digest beyond minimal adaptive changes. Perhaps subjecting these mice to greater amounts of intestinal resection may reveal a functional benefit to the greater mucosal surface area. Finally, it is possible that the enterocytes in the p38-IKO mice are less mature, but compensated by the expanded mucosal surface area. Our laboratory has recently demonstrated adaptive mucosal growth in human infants (30). This observation validates the use of our animal model for understanding the regulation of structural features of adaptation after intestinal resection. Correlations between structural and functional adaptation, such as body composition, energy expenditure, and nutrient absorption, are important and are areas of current investigation.

Intestinal epithelial p38 deletion had marked effects on cellular kinetics at baseline. Our findings are in line with other investigators who have demonstrated that constitutive, enterocyte-specific knockout of p38 leads to increased crypt proliferation (32). We opted to use an inducible, enterocyte-specific p38-α knockout model for the present experiments because the development and function of other organ systems are less likely to be affected by this genetic manipulation. Acute inactivation of p38 MAPK expression led to increased crypt proliferation and decreased crypt apoptosis. Somewhat surprisingly, shifting the balance toward proliferation did not generate taller villi and deeper crypts at baseline. In addition, we found no evidence that these changes were being driven by an increase in the ISC population. Rather, we would propose that disrupting p38 expression contributed to crypt fission events, contributing to longitudinal growth of the bowel. We presently do not understand the mechanism for elevated rates of crypt fission in the p38-IKO mice at baseline.

We were surprised to note that differences between WT and p38-IKO mice in terms of bowel length and crypt fission rates were no longer apparent after bowel resection. We transected and reanastomosed the intestine at the same distance from the cecum (12 cm) in both groups of mice. Thus the lack of change after resection was verified in both groups by studying the same segment of bowel from the anastomosis to the cecum. These findings suggest that the effect of intestinal epithelial p38 deficiency has more of an effect on intestinal lengthening under normal conditions, but lost during the stimulus of intestinal resection. Similarly, p38 expression must have a greater effect on crypt fission in the baseline, nonperturbed state.

It is important to note that, while resection-induced apoptosis was reduced in p38-IKO mice, rates of apoptosis in the crypts were elevated after SBR. It is possible that stresses such as massive SBR deploy several redundant apoptotic pathways besides Bax that contribute to the increased apoptosis after SBR.

Therapies aimed at altering p38 activity would ideally be aimed at the specific target organ, in this case the gut. p38 MAPK is expressed ubiquitously and has many cellular functions; it is activated by environmental and genetic stresses and plays a role in inflammation and tissue homeostasis by regulating cell proliferation, differentiation, and survival (6, 33, 42). Consequently, blocking p38 function globally may be fraught with potential adverse effects. Indeed, in mouse models, global p38 deletion is embryonic lethal (19, 31). Inhibition of p38 might be expected to cause immune suppression, potentially making the recipient more prone to infections, which could compromise recovery (2, 28). In addition, p38 deletion in animal models has been show to enhance tumorigenesis (19, 41). Therefore, limiting the field effect of disrupted p38 expression to the intestinal epithelium would be sensible.

The timing of p38 inactivation may be important to maximally enhance adaptation. In the present experiments, we induced p38 deletion 1 wk before performing SBR to be sure that p38 was absent before any adaptation response had occurred. It would also be informative to explore the adaptation response after systemic inhibition of p38 MAPK. This may provide a novel means to enhance adaptation responses in patients who require a massive intestinal resection.

Conclusions

This work represents the first time that activation of Bax has been found to be an important mechanistic step during resection-induced apoptosis. Furthermore, these findings support our hypothesis that p38 is a key regulator of Bax activation. We propose that bowel resection induces a milieu of cellular stress within enterocytes, leading to p38 MAPK activation, and subsequent activation of Bax. Activated Bax causes cytochrome c release into the cytosol, which triggers cell death via the caspase cascade.

GRANTS

This work was supported by National Institutes of Health Grants R01 DK 059288 (B. W. Warner), T32 CA009621 (D. Wakeman, and J. A. Leinicke), U54 HG003079 (J. A. Santos), P30DK52574, Morphology and Murine Models Cores of the Digestive Diseases Research Core Center of the Washington University School of Medicine, and the St. Louis Children's Hospital Foundation, Children's Surgical Sciences Institute.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

D.W., J.G., C.R.E., and B.W.W. conception and design of research; D.W., J.G., J.A.S., W.S.W., J.E.S., M.E.M., and J.A.L. performed experiments; D.W. analyzed data; D.W., J.G., J.A.S., W.S.W., J.E.S., C.R.E., and B.W.W. interpreted results of experiments; D.W. prepared figures; D.W. drafted manuscript; D.W., J.G., C.R.E., and B.W.W. edited and revised manuscript; D.W., J.G., J.A.S., W.S.W., J.E.S., M.E.M., J.A.L., C.R.E., and B.W.W. approved final version of manuscript.