Mapping and Characterization of Isolated Lymphoid Follicles in Patients with Hirschsprung-Associated Enterocolitis

Abstract

Hirschsprung-associated enterocolitis (HAEC) is a severe complication of Hirschsprung’s disease (HSCR), but its pathogenesis remains unclear. Gut-associated lymphoid tissue (GALT), including isolated lymphoid follicles (ILFs), plays a key role in mucosal immunity. This study investigated the presence, distribution, and phenotype of ILFs in colonic tissue from 35 HSCR patients and 12 controls. Histological and immunohistochemical analyses revealed mature ILFs in both ganglionic and aganglionic bowel, with no significant differences in densities between HSCR patients with and without HAEC. However, patients that developed HAEC showed signs of enhanced B-cell activation in ILFs. Patients with a stoma had significantly increased ILF density, suggesting that luminal content may influence ILF development. Despite the absence of enteric neurons, ILFs were structurally similar to those described in adults, indicating that ILF formation occurs independently of a functional enteric nervous system. These findings provide new insights into mucosal immunity in HSCR and suggest a role for activation of the mucosal B-cell compartment in HAEC pathogenesis.

Introduction

The gastrointestinal associated lymphoid tissue (GALT) plays a crucial role in initiating adaptive immunity within the gut. GALT encompasses multi-follicular Peyer’s patches located in the distal ileum, the appendix vermiformis, and numerous isolated lymphoid follicles (ILFs) dispersed along the entire length of the intestine. ¹ ILFs are highly organized lymphoid structures strategically positioned in close proximity to specialized epithelial microfold (M)-cells which facilitate sampling of luminal antigens. The distribution and organization of ILFs are extensively studied in adults, but characterization of ILFs in the pediatric population is limited.^2,3 While ILFs contribute to immune homeostasis, studies have also suggested their potential involvement in initiation and perpetuation of inflammation in conditions like Crohn’s disease and ulcerative colitis.^1,2

Hirschsprung’s disease (HSCR) is a congenital intestinal motility disorder with a reported incidence of 1 in 5000 and a higher prevalence in males. ⁴ HSCR is characterized by a deficiency of enteric nerve cells in the distal part of the colon, resulting in symptoms such as abdominal distension, vomiting, and delayed passage of meconium. ⁵ Treatment typically involves surgical resection of the affected colonic segment during the first few months of life.^6,7 Hirschsprung-associated enterocolitis (HAEC) is a recognized and potentially life-threatening complication of HSCR. Symptoms of HAEC encompass explosive diarrhea, abdominal distension, and fever.^8-10 Certain patient groups, including those with trisomy 21, long segment disease, and obstructive symptoms, are at higher risk of developing HAEC.^10-14 The etiology of HAEC remains largely unknown; however, human studies have indicated that dysregulated mucosal immunity plays a role in its development. Infants with trisomy 21 exhibit inherent deficiencies in both cellular and humoral immunity, which may contribute to their heightened risk for HAEC.^15,16 Furthermore, research has revealed that HSCR patients who develop HAEC show a significant increase in plasma cells within the lamina propria, yet exhibit reduced levels of luminal IgA. ¹⁷ Similar observations have been made in mouse models of HSCR. ¹⁸ Because IgA plays a crucial role in regulating the composition of the microbiota, diminished levels of luminal IgA may, at least in part, explain why dysbiosis is more frequently observed in patients with HAEC. ¹⁹

Several genes are important for controlling the morphogenesis and differentiation of the enteric nervous system (ENS), and mutations or deletions of such genes interfere with ENS development. Mutations of tyrosine kinase receptor gene RET are responsible for 50% of familiar HSCR and 15-20% of sporadic HSCR.^20,21 Importantly, RET is a key regulator of Peyer’s patch organogenesis, directly linking the pathogenesis of HSCR with the development of GALT. ²² Whether ENS is involved in the development of ILFs is not known.

On this background, we hypothesized that alterations of colonic ILFs might be implicated in the development of HAEC. We performed an extensive mapping and phenotyping of ILFs in colonic tissue from HSCR patients and age-matched control tissue to investigate whether there were differences in numbers and/or cellular composition of colonic ILFs between patients with and without HAEC, and between ganglionic and aganglionic bowel.

Materials and Methods

A total of 35 HSCR patients registered in the “Biobank for congenital pediatric surgical diseases” at Oslo University Hospital were included. Detailed clinical characteristics were obtained from the patients’ medical records. HAEC was defined based on a Pastor score ≥4. ⁹ Control samples were collected from patients who underwent sigmoidostomy closure following anorectoplasty for imperforate anus or full-thickness rectal biopsies due to severe idiopathic constipation.

All tissue samples were transmural, fixed in formalin, embedded in paraffin, and cut at 4 µm for further analysis. In all HSCR patients, at least 2 different sections from both ganglionic and aganglionic segments were stained with hematoxylin and eosin (H&E). The number of sections examined per patient varied depending on the length of the ganglionic and aganglionic segments. The tissue from the control samples were small colon samples from the area proximal to the stoma, and rectum biopsies from children with severe idiopathic constipation. In all control samples, at least 2 sections were examined. All tissue sections were digitally scanned at ×200 magnification using a Pannoramic MIDI II scanner (3DHistech) and examined in SlideViewer (3DHistech). The density of ILFs was calculated by dividing the number of ILFs by the mucosal surface length $\left(\frac{\mathrm{number}\mathrm{of}\mathrm{ILF}}{\mathrm{mm}\mathrm{surface}\mathrm{mucosa}}\right)$ . The length of the mucosa was measured using the length tool in SlideViewer.

To determine the 3-dimensional localization of ILFs, 6 tissue sections, 40 µm apart, were examined. The ILFs were registered as mucosal (M)- ILF if the follicle was confined above the muscularis mucosae or as submucosal (SM)-ILF if it crossed the muscularis mucosae. ²

For immunohistochemistry we employed a Ventana Discovery ULTRA automated slide stainer (Roche Diagnostics) for chromogenic detection and used the following primary monoclonal antibodies: CD20 (clone L26 1:200, Agilent Technologies), CD21 (clone 2G9, 1:10, Invitrogen), CD45RA (clone 4KB5, 1:1000, Abcam), CD38 (SPC32, 1:40, Novocastra Leica), and CD3 (polyclonal A 0452 1:75, Agilent Technologies) followed by Discovery OmniMap anti-Mouse HRP or OmniMap anti-Rabbit HRP (Roche Diagnostics) and Discovery ChromoMap DAB (Roche Diagnostics).

Peripheral Node Addressin (clone MECA-79, 1:6000, Novus Biologicals) was detected by using polyclonal Rabbit anti-Rat IgM (1:500, Jackson ImmunoResearch) followed by Discovery OmniMap anti-Rabbit HRP (Roche Diagnostics) and Discovery ChromoMap DAB (Roche Diagnostics, Pleasanton, Calif). Cell conditioning solution 1, pH 8.5 (Roche Diagnostics) for 48 minutes was used for antigen retrieval. Sections were counterstained in Shandon Instant Hematoxylin (Thermo Scientific) and finally mounted in Eukitt Quick-hardening mounting medium (Sigma-Aldrich). Quantification of CD38+ cells in ILFs was done with QuPath software (0.7.0-64×). We used the standard positive cell detection thresholds for “optical density sum.” The follicle border was annotated using the closed polygon annotation tool. The average number of cells counted within each ILF was 3160 cells.

Statistical analysis was conducted using STATA 17.0 MP parallel edition and Prism GraphPad 9.1.0. For data that did not follow a normal distribution, a log transformation was applied to achieve normality. Group comparisons were analyzed using Student’s t-test. A P-value of <.05 was considered statistically significant.

Results

Demographics of HSCR Patients and Controls

Thirty-five HSCR patients, except one HSCR patient, had aganglionosis confined to the rectosigmoid colon. Control patients were older than the HSCR patients, and the sex distribution was different. Seventeen (49%) HSCR patients experienced one or more incidents of HAEC. Interestingly, 14 patients experienced HAEC after the pull-through operation (Tables 1 and 2).

Table 1.

Demographics From All Patients With Hirschsprung’s Disease Included in the Study, n (%).

Patient characteristics
Male, n (%)	28 (80)
Down syndrome, n (%)	4 (11)
Transition zone
Rectosigmoid	34 (97)
Long segment	1 (3)
Pre pull-through stoma, n (%)	9 (25)
Ileostomy	7
Colostomy	2
Age at operation, months, median (min-max)	4 (2-82)
Patients with HAEC, n (%)	17 (49)
Pre pull through	3
Post pull through	10
Pre and post pull-through	4
Follow up time, months, median (min-max)	70 (9-168)

Table 2.

Demographic From Pediatric Control Patients, n (%).

	Pediatric control n = 12
Demographic profile	Anorectal malformations n = 6	Idiopathic constipation n = 6
Male, n (%)	1 (16)	1 (16)
Age at operation, years, median (min-max)	0.9 (1.1-1.6)	4.4 (3-11.5)

Colonic tissue of the Pediatric Population Contains Mature Gut-Associated Lymphoid Tissue

Histological evaluation of H&E-stained tissue sections from ganglionic and aganglionic colon of HSCR and age-matched controls revealed isolated lymphoid follicles (ILFs) in the mucosa and submucosa (Figure 1(A)-(D)). Serial sectioning of colonic tissue from both HSCR patients and controls showed that ILFs could be divided anatomically into 2 types: submucosal ILFs (SM-ILFs) that always penetrated muscularis mucosa and mucosal (M)-ILFs that were confined in the mucosa. ¹ Immunohistochemical stainings showed that the ILFs consisted of a central area with CD20+ B cells surrounded by CD3+ T cells (Figure 2). In both compartments many B cells and T cells expressed CD45RA, reflecting a naïve phenotype. Many ILFs contained germinal centers with CD21+ follicular dendritic cells and high numbers of Ki67-expressing cells, suggesting ongoing antigen-dependent B-cell activation. The ILFs also contained peripheral node addressin (PNAd)-expressing high endothelial venules in in the T-cell area. Together, this shows that colonic tissue from both aganglionic and ganglionic segments contains mature ILFs that were strategically situated close to the surface epithelium to facilitate sampling of luminal antigens.

Figure 1.

(A) H&E staining of colon tissue from a pediatric control, showing a M-ILF localized above muscularis mucosae. (B) H&E staining of healthy pediatric colon tissue showing a SM-ILF. (C) M-ILF in the ganglionic segment of the colon from a child with Hirschsprung’s disease. (D) SM-ILF in the aganglionic segment. All scale bars: 200 μm.

Figure 2.

Immunohistochemistry stainings with antibodies to CD45RA, CD3, CD20, Ki67, CD21, and PNAd of ILFs in pediatric controls, ganglionic, and aganglionic tissue from patients with Hirschsprung disease. CD45RA, CD3, CD20, Ki67, and PNAd scale bar: 200 μm. CD21 scale bar: 100 μm.

The Density of Colonic ILFs is not Different Between HSCR Patients that Develop and Those That Do Not Develop HAEC, But Patients With Stoma Have Higher Density of ILFs

To investigate whether the bowel tissue from patients who developed HAEC differed from those who did not, we stratified the patients based on whether they had an enterostomy before the pull-through operation, as it is well-documented that diversion colitis, characterized by an increased number of ILFs, can occur distal to the stoma. ²³ There was no significant difference in the density of ILFs between the ganglionic, aganglionic segment and controls, nor was the any difference between HSCR patients with and without HAEC (Figure 3). Moreover, whether HAEC developed before or after surgery had no impact on ILF density. However, those that had been treated with stoma before the pull-through operation had significantly higher numbers of ILFs, both in the ganglionic and aganglionic segment. Apart from increased numbers of ILFs, often with prominent germinal centers, there were no overt signs of mucosal inflammation in stoma tissue (Figure 4).

Figure 3.

Mean density of ILFs as ILF/mm surface mucosa of pediatric controls (red), ganglionic (black), and aganglionic (grey) segments of HSCR with and without stoma. Patients that developed HAEC before pull through operation are indicated (open circle).

Figure 4.

H&E staining of colon from a HSCR patient with (A) and without (B) diversion colitis. Arrows on ILFs. Scale bar: 2000 μm.

The relative number of CD38+ B cells is significantly higher in ILFs that develop HAEC.

Although the number of ILFs was similar in patients that developed HAEC with those that did not, we wanted to examine whether there were other signs of immune activation. In organized lymphoid tissue CD38 is highly expressed on activated germinal center-like B cells and B cells that differentiate towards plasma cells. To explore the activation status of the B-cell compartment we determined the number of CD38+ cells in ILFs of patients without stoma. Interestingly, the relative number of CD38+ cells ILFs was significantly higher in the ganglionic segment of HSCR patients that experienced HAEC (Figure 5), suggesting enhanced B-cell activation in the colonic mucosa of these patients (Figure 5).

Figure 5.

(A) Percentage of CD38+ cells of all cells in ILFs of ganglionic and aganglionic bowel segments comparing patients with and without HAEC. (B) Representative immunohistochemical image of an ILF showing CD38+ cells (brown). Scale bar: 100 μm.

Discussion

In this study, we found no significant differences in the density of ILFs in colon between HSCR patients who developed HAEC with those that did not experience HAEC. However, signs of B-cell activation, attested by a relative increase of CD38+ cells in ILFs, was observed in patients that developed HAEC.

Previous research has suggested that changes of the mucosal immune system could explain the development of HAEC, particularly noting an increased number of plasma cells within the lamina propria in HSCR patients developing HAEC. ¹⁷ It has been proposed that priming of B cells in colonic ILFs contributes significantly the accumulation of plasma cells to colonic mucosa.^1,2 Thus, our findings showing enhanced B-cell activation in ILFs of patients developing HAEC supported the occurrence of increased numbers of plasma cells in the colonic mucosa. ¹⁷ Of particular interest, we found that increased numbers of CD38+ cells in ILFs were more prominent in the ganglionic segment. This strongly suggests that the B-cell compartment is activated in the remaining colon after surgery, which accords with the fact that 14 out of 17 HAEC patients developed inflammation after surgery (Table 1).

Studies have demonstrated an intimate neuron-immune cross-talk in the gut. ²⁴ Interestingly, it has been shown that the RET gene, involved in HSCR pathogenesis, plays a crucial role in Peyer’s patch development. ²⁵ Based on this information, it was reasonable to hypothesize that the absence of enteric neurons would affect the organization of GALT in the colon. However, our findings suggest that colonic ILFs develop independent of a functioning ENS. Whether Peyer’s patches in HSCR patients are affected remains to be determined.

An important limitation of this study is the small sample size, which may underestimate differences between HSCR patients with or without HEAC. Inherent biological variability, particularly within the pediatric control group, may also be a confounding factor of this study.

Current knowledge about the organization of colonic ILFs in infants is limited. By comparing the phenotype and density of ILFs between HSCR patients and controls, we showed that that the density colonic ILFs does not seem to be significantly affected by HSCR. Furthermore, we found that both the organization, phenotype, and the anatomical localization of colonic ILFs are similar to that described in adults. ¹ Already 1 to 2 months after birth, colonic ILFs display prominent germinal centers, suggesting that these organized lymphoid structures play an important role for immune responses to luminal antigens in early infancy. Interestingly, in patients with a stoma the density of ILFs increased significantly, without overt signs of inflammation. This suggests that changes in the passage and composition of the luminal content in colon enhance the development of ILFs.

Conclusion

This study provides novel insights into the organization and development of isolated lymphoid follicles in the pediatric colon, particularly in the context of HSCR. We demonstrated that mature ILFs are present in both ganglionic and aganglionic bowel segments, with no significant differences in densities between HSCR patients with or without HAEC. However, we found signs of B-cell activation in ILFs in patients that develop HAEC, suggesting that enhanced priming of the mucosal B-cell compartment is involved in HAEC. ILF density was significantly increased in patients with a stoma, indicating that exposure to luminal content influences ILF development. Together, these results underscore the importance of ILFs in early-life mucosal immunity and suggest that local activation of B cells in ILFs play a role in HAEC pathogenesis.