Murine model of Hirschsprung-associated enterocolitis II: Surgical correction of aganglionosis does not eliminate enterocolitis

Abstract

Background

Hirschsprung disease (HD) results from aganglionosis of the colon, is linked to acute and chronic enterocolitis (known as Hirschsprung-associated enterocolitis) despite successful corrective surgery, and can lead to bacteremia and even death. The genetic and molecular mechanisms underlying these disorders are largely unknown.

Methods

We developed a microsurgical corrective pull-through procedure in mice, and applied that to Ednrb^−/− mice, which manifest aganglionic megacolon that is very similar to HD. Wild-type littermates (Ednrb^+/+) also underwent identical surgery. At prespecified time points postoperatively, mice were sacrificed, and histopathologic analyses of intestinal inflammation were performed. Mice of both genotypes were sacrificed after the postoperative recovery period to determine if corrective surgery itself caused inflammation. Stooling patterns were assessed as well to determine if intestinal function normalized after surgery.

Results:

There was no difference in histopathological enterocolitis scores after recovery from surgery. Stooling patterns in Ednrb^−/− and Ednrb^+/+ mice were similar postoperatively, suggesting normalization of intestinal function. However, with time, approximately 40% of Ednrb^−/− mice developed clinical illness consistent with enterocolitis. No control (Ednrb^+/+) mice developed clinical enterocolitis. Histopathological enterocolitis scores in the 40% of Ednrb^−/− mice that developed clinical enterocolitis postoperatively were significantly worse than those of healthy postoperative Ednrb^−/− mice. In contrast, none of the Ednrb^+/+ control mice exhibited postoperative long-term inflammation.

Conclusions

Microsurgical pull-through operation in Ednrb^−/− mice produces a mouse model that closely resembles key features of Hirschsprung-associated enterocolitis, enabling controlled study of genetic and molecular mechanisms in Ednrb^−/− mice and other genotypes that produce similar phenotypes.

Hirschsprung disease (HD) and similar disorders occur because of absent ganglia in the large intestine. Patients with these types of diseases suffer from enterocolitis that can lead to bacteremia, sepsis, and even death. Treatment consists of a surgical pullthrough operation that removes the aganglionic section of colon, and is curative in most cases. However, 20% to 38% of patients continue to suffer from enterocolitis despite technically successful surgery, a condition known as Hirschsprung-associated enterocolitis (HAEC). For these individuals, there are few or no treatment options available other than limited and temporary symptomatic relief.

The reasons why enterocolitis occurs, persists and becomes a chronic disease in some patients with HD are unknown. Studies of HAEC have been almost entirely observational and correlational, and little insight into molecular mechanisms has emerged. Two factors now make detailed study of HD and its associated postoperative enterocolitis feasible: (1) the existence of mouse and rat genotypes that are characterized by aganglionosis or hypoganglionosis of the large intestine, and (2) the development of a microsurgical pullthrough operation in our laboratory that allows afflicted animals with otherwise very limited life spans to survive indefinitely [1]. The rodent models can be used to study HD and its associated enterocolitis include the Ednrb-null mouse [2,3], piebaldlethal mouse [4,5], and the endothelin receptor B-deficient rat (spotting lethal rat) [6-8]. Additional mutant rodent models that can now be used to study HD include the Endothelin-3 ligand-deficient mouse [9], the Hoxb5 dominant negative conditional (Cre-Lox) transgenic genotype [10], erbB2/nestin-Cre conditional mutant mouse [11], the Dom spontaneous mutant mouse [12], a conditional β-1 integrin knockout mouse [13], trisomy 16 mice [14,15], and mice deficient in the c-ret proto-oncogene [16], and the fmc/ fmc (familial megacecum and colon) rat [17] that also display a phenotype that mimics HD.

We previously reported successful development of a microsurgical pullthrough procedure in Ednrb^−/− mice [1]. We next performed phenotypic characterization of these mice and created and validated a histopathologic scoring system to analyze enterocolitis in semiquantitative fashion [3]. Here we report that technically successful pullthrough surgery in Ednrb^−/− mice does not eradicate enterocolitis in a proportion of mice. Therefore, this model appears particularly well-suited to study genetic and molecular mechanisms of HAEC and similar diseases where pullthrough surgery does not eliminate enterocolitis.

1. Methods

1.1. Animals

Animal protocols were approved by the Institutional Animal Care and Use Committee at Cedars-Sinai Medical Center. Ednrb^tm1Ywa/J mice on a hybrid C57BL/6J-129Sv background; JAX stock # 003295) were purchased from Jackson Laboratory (Bar Harbor, Me) and housed at the animal facility of Cedars-Sinai Medical Center. The mice were maintained on a 12-hour light/dark cycle at 25°C and received regular mouse chow and water ad libitum. A breeding colony of the Ednrb^tm1Ywa/J heterozygous mice was established and maintained at Cedars-Sinai Medical Center. The Ednrb^tm1Ywa/J is a targeted null mutation of the endothelin receptor B gene that segregates in an autosomal recessive manner. The homozygotes (we refer to as Ednrb^−/− in this article) are easily identified by the white coat color and progressively enlarging abdomens owing to aganglionosis leading to megacolon [2]. The wild-type and heterozygote littermates (we refer to as Ednrb^+/+ and Ednrb^+/−, respectively) are phenotypically normal and therefore indistinguishable from one another and, hence, required genotyping using a polymerase chain reaction–based assay.

1.2. Study design

Two groups of mice were used in this study: 19 Ednrb^+/+ mice (Group 1) and 19 Ednrb^−/− mice (group 2). All mice underwent the pullthrough operation at 3 to 4 weeks of age (weight range group 1: 12-17 g; group 2: 11-14 g). Animals of each genotype that survived through the perioperative period of two weeks were then included into one of the groups. All study animals had health assessments daily including body weight monitored every 2 to 3 days. Ednrb ^+/+ mice were euthanized at predetermined time points of 5, 6, 8, 10, 12, and 16 weeks after surgery, whereas Ednrb^−/− mice were euthanized at 4, 6, or 8 weeks after surgery. Any animal appearing clinically ill was euthanized when it developed illness (clinical enterocolitis) defined as: reduced activity and weight loss and either new onset of loose stools or lack of stooling. At the time of euthanization for all mice, the colon and ileum was harvested for histopathological scoring [3].

1.3. Pullthrough procedure and perioperative management

The microsurgical procedure, pre- and postoperative care was performed as described by Zhao, et al [1]. The murine microsurgical pullthrough procedure was based on that described by Swenson and is shown in Fig. 1. The single-stage microsurgical pullthrough procedure resects approximately 40% of the total colon, owing to the lack of mesenteric mobility in the distal portion of the colon of both wild type and Ednrb^−/− mice; hence, the operation for each genotype is essentially identical. The length of aganglionosis is variable in the Ednrb^−/− mice in our colony with a mean length of 2.2 cm from the anus and a total colon length of 10 to 12 cm (data not shown). We typically resected 4 to 5 cm of colon and rectum, placing the resection margin well proximal to the aganglionic segment and transition zone. Furthermore, we performed histological evaluation of the proximal margin using hematoxylin and eosin staining to confirm the presence of ganglion cells after the operation was completed. We found no evidence of retained aganglionosis or hypoganglionosis in pullthrough segments of the study animals.

Fig. 1.

Murine colon pullthrough operation. A, Resection of dilated megacolon and aganglionic segment. B, Preparation of pullthrough segment. C, Completed pullthrough operation with coloanal anastomosis.

1.4. Stooling pattern assessment

We performed analyses of stooling patterns in 4 mice of each genotype at 4 weeks after pullthrough operation as described previously [1].

1.5. Histological preparation of tissues

The colon and intact pullthrough segment (including anus) and 2 cm of terminal ileum was excised, gently washed with cold phosphate-buffered saline and fixed with 10% neutral buffered formalin at room temperature overnight. The specimens were cut into 3 segments (1 cm in length) and recorded as anus and pullthrough anastomosis (segment 1), proximal colon (segment 2), and terminal ileum (segment 3) (Fig. 2). All specimens were embedded in paraffin and the anus and pullthrough anastomosis segment was sectioned in a sagittal orientation while all other segments were sectioned in a transverse orientation. Each segment was sampled at 3 regions and serially sectioned using 6-mm sections and stained with hematoxylin and eosin.

Fig. 2.

Pathologic segments evaluated after pullthrough operation. Segment 1 included the coloanal anastomosis and approximately 1 cm of distal pullthrough segment. Segment 2 included 1 cm of colon proximal to the ileum, and segment 3 included 1 cm of terminal ileum.

1.6. Histological grading for enterocolitis

One of the authors (DD) scored colon segment 2 using the validated murine enterocolitis scoring system described by Cheng, et al [3]. We chose segment 2 to evaluate because it was consistently in the ganglionic region of the colon of Ednrb^−/− mice and could also be compared to Ednrb^+/+ animals (control) without megacolon. In the Ednrb^+/+ group, animals were evaluated and scored at 6 weeks (4 animals) and 16 weeks (5 animals) after the pullthrough operation and were found to have essentially identical scores. Weeks 8, 10, and 12 were not scored, nor included in the analysis.

1.7. Statistical analysis

Data are presented as percent frequency or means ± standard deviation where appropriate. Frequency data was tested using a Fisher exact test. Comparisons between groups of enterocolitis score data were made using a two-sided nonparametric Wilcoxon rank sum test. Correlations between variables were calculated as a Spearman's rank correlation.

2. Results

2.1. Stooling pattern

We evaluated the stooling pattern of the post-pullthrough Ednrb^+/+ and Ednrb^−/− mice at 4 weeks after surgery (Fig. 3). The stooling pattern was similar in both groups and no statistically significant difference was found. The stool consistency was soft and formed in both groups.

Fig. 3.

Stooling frequency after pullthrough operation. Four Ednrb^+/+ and 4 Ednrb^−/− mice, 4 weeks after pullthrough operation had stooling frequency assessed over a 5-day period. There was no significant difference in numbers of stools per day using Wilcoxon rank sum test.

2.2. Enterocolitis scores after the pullthrough operation

We compared the histopathologic enterocolitis scores of all Ednrb^+/+ and Ednrb^−/− mice after pullthrough operation and found that inflammation, depth, and total scores were very similar. Fig. 4. The Ednrb^+/+ mice demonstrate low grade (score 2-3) inflammation in the proximal colon after the pullthrough operation ranging from 5 to 16 weeks, indicating this inflammation is not a transient condition only occurring early in the postoperative course. Of note, the SD of the Ednrb^−/− mice was significantly greater than the wild type group, indicating wide score variability within this group. Next we performed subgroup analyses of the Ednrb^−/− mice, comparing enterocolitis scores of the healthy subgroup (euthanized at 4 weeks) to the clinically ill subgroup (mean age of survival of 4.5 ± 2.9 weeks) (Fig. 5, Table 1). We found markedly increased enterocolitis scores in the clinically ill animals (5.4 ± 0.5) compared to the healthy animals (1.1 ± 1.5). Although all of the Ednrb^+/+ animals were clinically healthy after the pullthrough operation, about 40% of the Ednrb^−/− animals develop clinical illness that is likely enterocolitis and has features similar to HAEC in children (Fig. 6). Interestingly, the clinically healthy Ednrb^−/− subgroup has significantly lower enterocolitis scores than the healthy Ednrb^+/+ (3.0 ± 0.5, P < .05). The significance of this finding is unclear given that both groups are healthy, although it lends support to the argument that the pullthrough operation in the Ednrb^−/− mice is effective in removing the constipation from aganglionosis and restoring a similar stooling pattern to the post-pullthrough Ednrb^+/+ mice.

Fig. 4.

Enterocolitis scores of post-pullthrough mice. Enterocolitis scores (inflammation, depth and total) were compared between each group of Ednrb^+/+ and Ednrb^−/− mice. No significant difference between mean enterocolitis scores was found.

Fig. 5.

Enterocolitis scores of healthy and clinically ill postpullthrough Ednrb^−/− mice. Enterocolitis scores (inflammation, depth and total) were compared between healthy and clinically ill Ednrb^−/− mice. There was a statistically significant difference in the inflammation, depth, and total enterocolitis scores between the healthy and ill mice (Wilcoxon rank sum test P < .05).

Table 1.

Enterocolitis scores ofpost-pullthroughEdnrb^−/− mice separated by health status

	Healthy (n= 11)	Clinically ill (n = 8)
	Mean ± SD	Mean ± SD
Depth	0.7 ± 1.0	4.0 ± 0.0*
Inflammation	0.4 ± 0.5	1.4 ± 0.5*
Total	1.1 ± 1.5	5.4 ± 0.5*

Depth, inflammation, and total enterocolitis scores show a statistically significant difference between the healthy and clinically ill group (Wilcoxon rank sum test P < .05).

^*Significantly different from healthy group (P < .05).

Fig. 6.

Percent of the healthy animals after the pullthrough operation of each genotype. There was a statistically significant difference between (Fischer exact test, P < .05).

2.3. Correlation of time after pullthrough operation to enterocolitis score

To evaluate if duration of time after the pullthrough procedure correlated with total enterocolitis scores, Spearman's rank correlations were performed on all mice, each genotype as a group, and healthy and clinically ill Ednrb^−/− mice. There was no significant correlation between total enterocolitis scores and time at euthanasia, either across all mice, or within each group (data not shown).

3. Discussion

The pathologic mechanisms mediating HD and HAEC characterized by acute and chronic enterocolitis because of aganglionosis or hypoganglionosis are poorly understood. Until now, controlled experimental studies could not be performed because aganglionic and hypoganglionic genotypes of mice and rats suffer from acute constipation and intestinal inflammation that steadily and rapidly leads to failure to thrive or bacteremia, sepsis, and death. We developed a microsurgical pullthrough operation to prolong life in these mice [1] and developed a semiquantitative histopathologic scoring system to better evaluate enterocolitis [3]. Here, we report initial medium-term postoperative follow-up data in surgically corrected Ednrb^−/− mice.

Pullthrough surgery was technically successful, in that the aganglionic section of colon was removed, and stooling pattern normalized. Importantly, however, removal of the aganglionic colon did not prevent postoperative enterocolitis. This situation is very similar to that of the 20% to 38% of patients with HD who suffer from enterocolitis after pullthrough surgery (ie, HAEC). Approximately 40% of Ednrb^−/− mice developed postoperative enterocolitis, as evidenced by their significantly elevated histopathologic inflammation scores. Since we have a murine model that manifests aganglionic colon, a surgical corrective procedure that does not eliminate enterocolitis in a significant proportion of mice [1], and a semiquantitative histopathologic grading method to track the degree of intestinal inflammation [3], it is now possible to perform detailed, controlled, mechanistic studies.

We found that inflammation severity and depth in surgically corrected Ednrb^−/− mice were no different from those in postoperative Ednrb^+/+ mice after the postoperative recovery period (Fig. 4). There was no other evidence that corrective surgery itself contributed to intestinal inflammation beyond the postoperative recovery period. Although Ednrb^−/− mice had greater mortality postoperatively, this appears most likely owing to the fact that these mice were considerably smaller and less vigorous at the time of surgery than their Ednrb^+/+ counterparts [1]. In addition, we believe that a portion of the Ednrb^−/− mice likely developed enterocolitis preoperatively [3]. Hence, the Ednrb^−/− mice were poorer candidates for surgery, and some of them did not fare well after the operation, despite careful attention to perioperative and postoperative care.

Subtle genetic differences among strains of mice can contribute to susceptibility to postoperative enterocolitis, and may also impact the severity of enterocolitis as well. Furthermore, a number of other factors may impact enterocolitis, such as specifics of animal care and housing, whether or not antibiotic treatment is given, perioperative care, and probably other factors [1].

Several mouse and rat genotypes manifest aganglionosis or hypoganglionosis of portions of the small and large intestines [2,6-9,11-18]. This suggests that HD and associated diseases as well as HAEC are determined by multiple genetic mechanisms. Application of our approach to these genotypes may unravel complex genetic interactions and reveal molecular mechanisms and causes of HD and HAEC. Expanded understanding of these disorders should then enable development of more effective therapies targeted at the mechanisms that emerge.