B-lymphocyte–intrinsic and –extrinsic defects in secretory immunoglobulin A production in the neural crest–conditional deletion of endothelin receptor B model of Hirschsprung-associated enterocolitis

Abstract

Hirschsprung disease (HSCR) is a common cause of intestinal obstruction in the newborn. Hirschsprung-associated enterocolitis (HAEC) is a significant and life-threatening complication of HSCR, affecting up to 60% of patients. Animal models of endothelin receptor B (EdnrB) mutation reliably model human HSCR and HAEC. We previously demonstrated intestinal dysbiosis and a gut-specific deficiency of B-lymphocyte–produced secretory IgA (sIgA), the primary effector molecule of mucosal immunity, in mice with homozygous neural crest cell–conditional deletion of EdnrB (EdnrB^NCC−/−). To determine mechanisms for sIgA deficiency, we examined intrinsic and extrinsic aspects of B-lymphocyte development and function. Expression of the endothelin axis components [endothelin-1 (ET-1), endothelin-3 (ET-3), endothelin receptor A (EdnrA), EdnrB] were determined over a developmental time course. B-lymphocyte survival and Ig production were assayed in vitro. Polymeric Ig receptor (pIgR)–mediated IgA transport into the intestinal lumen was interrogated. We found endothelin axis component (EdnrA, EdnrB, ET-1, ET-3) expression in developing extramedullary hematopoietic organs and that some splenic B lymphocytes express EdnrB. Splenic B lymphocytes from EdnrB^NCC−/− mice showed no intrinsic defect in survival vs. wild-type (WT) B lymphocytes. In vitro stimulation of splenic B lymphocytes demonstrated decreased IgA, IgG, and IgM production in EdnrB^NCC−/− vs. WT mice. Additionally, small intestinal pIgR was decreased ∼50% in EdnrB^NCC−/− mice. These results suggest an intrinsic B-lymphocyte defect in antibody production as well as an extrinsic defect in IgA transport in the EdnrB^NCC−/− model of HAEC. Our results are consistent with human HAEC observations of decreased luminal sIgA and mouse models of other inflammatory bowel diseases, in which decreased pIgR is seen in concert with a dysregulated microbiota. Finally, our results suggest targeting the dysbiotic microbiome and pIgR-mediated sIgA transport as potential therapeutic approaches in prevention and treatment of HAEC.—Medrano, G., Cailleux, F., Guan, P., Kuruvilla, K., Barlow-Anacker, A. J., Gosain, A. B-lymphocyte–intrinsic and –extrinsic defects in secretory immunoglobulinA production in the neural crest–conditional deletion of endothelin receptor B model of Hirschsprung-associated enterocolitis.

Hirschsprung disease (HSCR) is a common cause of intestinal obstruction in the newborn (1). HSCR is a polygenic disease characterized by an absence of enteric nervous system (ENS) ganglion cells in the distal hindgut, extending from the rectum to a variable distance proximally, and results from a failure of cranial-caudal neural crest cell (NCC) migration. Hirschsprung-associated enterocolitis (HAEC) is a significant and life-threatening complication of HSCR, impacting up to 60% of children with HAEC (2). HAEC has an incompletely defined, multifactorial etiology including dysmotility, microbiome dysbiosis, impaired epithelial barrier function, and impaired immune responses (3, 4).

Multiple genetic defects have been associated with HSCR, most commonly mutations of rearranged during transfection and of endothelin receptor B (EdnrB), both of which are involved in NCC migration and ENS development. In contrast to mouse models with deleted rearranged during transfection gene, in which an extreme phenotype of total intestinal aganglionosis is found, mouse models of EdnrB deletion closely mimic human HSCR, with aganglionosis confined to the distal hindgut. EdnrB and its ligand, endothelin-3 (ET-3), regulate enteric NCC proliferation, migration, and differentiation. Components of the endothelin axis [ligands: endothelin-1 (ET-1), ET-3; receptors: endothelin receptor A (EdnrA), EdnrB] are coexpressed in a wide variety of tissues and undergo coordinated changes in expression, suggesting a tightly regulated, autocrine-paracrine mechanism of action (5, 6). There is extensive literature demonstrating modulation of lymphocyte function by endothelins (6). Additionally, components of the endothelin axis are overexpressed in the serum and bowel wall of patients with inflammatory bowel disease and endothelin receptor antagonists have been shown to reduce inflammation in mouse models of chemically-mediated colitis (7–9).

The gut-associated lymphoid tissue is the largest lymphoid organ in the body and is responsible for providing protection against a variety of antigens that may gain access to the host, including food particles, commensal and pathogenic bacteria, and their toxins (10). Barrier function against pathogens and foreign particles, provided by a mucus gel layer composed of glycoproteins and secretory IgA (sIgA), appears to be impaired in HSCR (4, 11, 12). HSCR patients with HAEC demonstrate increased IgA-containing plasma cells in the bowel wall but decreased in luminal sIgA compared with patients without HAEC (11). In the mouse, mature B lymphocytes that can produce IgA originate primarily from the spleen, and reduced splenic size, abnormal splenic architecture, and reduced lymphocytes in the spleen have been described in EdnrB-knockout animals and in animals with homozygous NCC-conditional deletion of EdnrB (EdnrB^NCC−/−) (4, 13). Furthermore, EdnrB^NCC−/− mice display a gut-specific deficiency of sIgA (4).

Impaired mucosal immunity, abnormal microbiota, intestinal barrier dysfunction, and dysmotility all appear to contribute to the pathogenesis of HAEC (3). ENS dysfunction can result in microbiome dysbiosis through impaired motility (14). When this is followed by impaired intestinal barrier function and an abnormal immune response, HAEC develops. The goals of this study were to determine the temporal and spatial expression pattern of endothelin axis components in developing extramedullary hematopoietic organs, how these patterns of expression are altered in EdnrB^NCC−/− mice, and if EdnrB^NCC−/− B lymphocytes display intrinsic or extrinsic defects in IgA production.

MATERIALS AND METHODS

Animal care and use

All animal care procedures were approved by the Animal Care and Use Committees of the University of Wisconsin–Madison (Protocol M01394) and University of Tennessee Health Science Center (Protocols 16–021 and 16–051). Mice with NCC-conditional deletion of EdnrB were generated by mating animals with a floxed EdnrB allele (EdnrB^flex3/flex3) (003295, background: C57BL/6; The Jackson Laboratory, Bar Harbor, ME, USA) with mice expressing Cre recombinase under the control of a Wingless-related integration site-1 enhancer element, resulting in either wild-type (WT), heterozygous NCC-conditional deletion of EdnrB (EdnrB^NCC+/−), or EdnrB^NCC−/− (15). Conventional EdnrB-knockout mice were generated by insertion of bacterial β-galactosidase (lacZ) into the EdnrB gene locus [background: C57BL/6; provided by Jim Pickel, National Institutes of Health (NIH) National Institute of Mental Health Transgenics Core Facility (Bethesda, MD, USA)] and are referred to as EdnrB^lacZ/+ (heterozygote) and EdnrB^lacZ/lacZ (null) in this study (16). Animals of both genders were included, and gender was considered a biologic variable. No gender differences were noted. Mice were housed in a specific-pathogen–free environment and were allowed ad libitum access to standard rodent chow and water.

Tissue collection

Animals underwent timed matings, and the day the vaginal plug was identified was defined as embryonic d (E) 0.5. Embryos were isolated from pregnant dams that had been anesthetized with isoflurane and humanely euthanized by cervical dislocation. Postnatal animals were humanely euthanized using isoflurane and cervical dislocation. Tissues (liver, spleen, thymus, heart, colon) were collected, snap frozen in liquid nitrogen, and stored at −80°C. Three biologic replicates of each tissue type were collected at 5 developmental stages [E14.5, E16.5, and E18.5; postnatal d (P) 0 and P21] for analysis.

RNA isolation, quantification, purity, and integrity

Total RNA was isolated from frozen tissue (E14.5–P0) using silica-membrane–based columns (RNeasy Micro Kit; Qiagen, Germantown, MD, USA). P21 samples were isolated using an RNeasy Mini Kit (Qiagen). Total RNA from liver samples of all developmental stages was isolated using the RNeasy Mini Kit. 2-ME was added to the lysis buffer to prevent degradation by RNases. Fully automated RNA isolation (QiaCube, Qiagen) was used for samples at E18.5, P0, and P21, whereas E14.5 and E16.5 samples were processed manually. On-column DNAse digestion was performed for all samples. RNA concentration and purity (260:280 nm ratio) was determined using a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). RNA integrity number (RIN) was determined by capillary electrophoresis (Bioanalyzer 2100; Agilent Technologies, Santa Clara, CA, USA).

Synthesis of cDNA was carried out in the presence of a ProtoScript II First Strand cDNA synthesis kit (New England BioLabs, Ipswich, MA, USA) using 150 ng RNA, 1 μl of oligo d(T)23 VN (50 μM), and 1 μl of random primer mix (60 μM) in a final volume of 20 μl according to the manufacturer’s instructions. To preamplify small amounts of cDNA, we used the Taqman PreAmp Master Mix (Thermo Fisher Scientific). The PreAmp mix was performed using a primer pool consisting of 50 nM final concentration of each primer and cDNA at a final concentration of 22.5 ng in a final volume of 10 μl. The preamplification reaction was conducted as follows: 1 cycle of 95°C for 10 min and 14 cycles of 95°C for 15 s and 60°C for 4 min. After preamplification, samples were stored at −20°C.

Primer and probe design and real-time PCR

We used ProbeFinder [Universal ProbeLibrary (UPL) Assay Design Center; Roche, Basel, Switzerland] for mouse species to design primers and UPL probes for 6 reference genes (Table 1). Real-time PCR reactions were prepared in a final volume of 10 μl containing 0.1 μl of UPL probe (10 mM), 0.2 μl of left (10 mM) and right (10 mM) primers, 5 μl of 2× Kapa probe fast quantitative PCR (qPCR) master mix (Kapa Biosystems, Wilmington, MA, USA), 2 μl of cDNA, and 2.7 μl molecular grade water. Three biologic replicates for each sample were run in a 384-well white plate optimized for Roche 480 Light Cycler (Phenix Research Products, Candler, NC, USA). Nontemplate controls were run in triplicate. Interplate calibrators were also included in triplicates on each plate as a method for interrun variation compensation (Tataa Biocenter, Gothenburg, Sweden). All plates were run on a Roche LightCycler 480 machine using the Mono Color Hydrolysis Probe/UPL program (1 cycle of preincubation step at 95°C for 5 min, 45 cycles of denaturation step at 95°C for 10 s, 60°C for 30 s, and 70°C for 10 s) with the acquisition mode set to single at the last step of amplification.

TABLE 1.

Primers used for qPCR

Gene symbol	Gene name	NCBI reference #	Primer sequence, 5′–3′		UPL probe #
			Left	Right
ET-1	Mus musculus endothelin-1	NM_010104.3	GGACATCATCTGGGTCAACAC	TGGGAAGTAAGTCTTTCAAGGAA	85
ET-3	M. musculus endothelin-3	NM_007903.3	GCACCAGAGATGTCACCAGTT	AGTCTCCCGCATCTCTTCTG	108
EdnrA	M. musculus endothelin receptor type A	NM_010332.2	TCAGAAAACAGCCTTCATGC	ATGAGGCTTTTGGACTGGTG	34
EdnrB	M. musculus endothelin receptor type B	NM_007904.3	TCAGAAAACAGCCTTCATGC	GCGGCAAGCAGAAGTAGAA	83
HPRT1	Hypoxanthine guanine phosphoribosyl transferase 1	NM_013556.2	TCCTCCTCAGACCGCTTTT	CCTGGTTCATCATCGCTAATC	95
TBP	TATA-box binding protein	ENSMUST000 00014911.3	GGTCGCGTCATTTTCTCC	GGGTTATCTTCACACACCATGA	107

NCBI, National Center for Biotechnology Information.

Gene expression analysis

GenEx Pro software (v.6.1.0.757; MultiD Analyses, Gothenburg, Sweden) was used to analyze gene expression. Based previously published methods (17), we identified the combination of hypoxanthine phosphoribosyltransferase 1 (HPRT1) and TATA-box binding protein (TBP) as appropriate housekeeping genes for analysis of the 5 organs over embryonic and early postnatal time points. Heat maps for all genotypes (EdnrB^NCC+/−, EdnrB^NCC−/−, EdnrB^lacZ/+, EdnrB^lacZ/lacZ) were generated by normalizing each sample against its WT counterpart at the same time point and tissue type.

Protein expression of endothelin axis components

Protein expression of EdnrA and EdnrB in WT was confirmed by Western blot analysis in colon, heart, liver, spleen, and thymus at ages E14.5, E16.5, E18.5, P0, and P21. Frozen samples were homogenized using a disposable homogenizer inside a 1.5-ml tube. Once the sample was in suspension, T-PER tissue protein extraction reagent (Thermo Fisher Scientific) proceeded according to the manufacturer’s instructions. The proteins in the supernatant were separated on Any kD Mini-Protean TGX Precast Protein Gels (Bio-Rad, Hercules, CA, USA) under running conditions at 150 V for 50 min. Dual color (Bio-Rad) and SuperSignal molecular markers were separated along with the samples. The Western blot was transferred using the Trans-Blot turbo transfer apparatus and 0.45-μm nitrocellulose membranes. The membranes were blocked with 5% blotting-grade blocker (Bio-Rad) in PBS with 0.1% Tween 20 and incubated with one of the following primary antibodies: 1) anti-mouse glyceraldehyde-3-phosphate dehydrogenase (GAPDH) made in rabbit, 1:5000 dilution (G9545; MilliporeSigma, Burlington, MA, USA), 2) anti-mouse EdnrA made in rabbit, 1:500 dilution (PA3-065; Thermo Fisher Scientific), and 3) anti-mouse EdnrB (ETBR, M-74, pAb) made in rabbit, 1:500 dilution (sc-33538; Santa Cruz Biotechnology, Dallas, TX, USA), incubated overnight at 4°C. Secondary antibody used was anti-rabbit–horseradish peroxidase (1:2000) (7074S; Cell Signaling Technology, Danvers, MA, USA). All primary and secondary antibodies were diluted in blotting-grade blocker. Detection of immunoblots was carried out using SuperSignal Femto (Thermo Fisher Scientific). For relative quantification analysis, images were analyzed with Quantity One software (Bio-Rad) to obtain densitometric values. First, the signal values were determined for each band with the background signal subtracted from the signal of each individual band. The ratio of EdnrA or EdnrB to GAPDH for each sample was calculated.

A classic sandwich ELISA was employed to confirm protein expression of ET-1 and ET-3 in the same organs and time points used above. The ELISAs were performed and analyzed according to the manufacturer’s recommendations for ET-1 (LS-F35516; LSBio, Seattle, WA, USA) and for ET-3 (LS-F11351; LSBio).

Protein isolation and Western blot analysis of polymeric Ig receptor

Protein expression of polymeric Ig receptor (pIgR) was determined in postnatal intestinal samples by Western blot analysis. Briefly, protein from ileum tissue was isolated using T-PER tissue protein extraction reagent (78510; Thermo Fisher Scientific) according to the manufacturer’s protocols with added protease inhibitors. Subsequently, the proteins were quantified using Bradford assay, and equal quantities of denatured proteins were loaded on 7.5% Mini-Protean TGX Precast Gels (4561024; Bio-Rad). The separated proteins were then transferred onto nitrocellulose membranes. Membranes were incubated with goat anti-mouse pIgR pAb (AF2800, 0.2 µg/ml; R&D Systems, Minneapolis, MN, USA) and mouse anti-β actin monoclonal antibody (66009-1-Ig, 1:10,000; Proteintech, Rosemont, IL, USA) overnight at 4°C in Tris-buffered saline with 0.1% Tween 20 and 5% nonfat dry milk. Proteins were visualized using appropriate horseradish peroxidase–conjugated secondary antibodies (A16005; Thermo Fisher Scientific; sc-516102; Santa Cruz Biotechnology) and a chemiluminescent detection kit (34095, 34076; Thermo Fisher Scientific). Band intensity was quantified by digital densitometry using Quantity One software (Bio-Rad).

B-lymphocyte isolation and viability analysis

Spleens from EdnrB^NCC−/−, EdnrB^lacZ/lacZ, and WT animals were manually dispersed into single-cell suspension. B lymphocytes were isolated from spleen using positive selection with mouse CD19 microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany), counted, and plated at a concentration of 1 × 10⁶ cells per ml in 100 µl final volume of phenol-free Roswell Park Memorial Institute medium 1640 supplemented with 1× penicillin-streptomycin in a 96-well flat bottom plate and incubated at 37°C with 5% CO₂. Samples were analyzed over a time course: 0, 3, 5, and 7 h. Recombinant ATP standard was prepared (10 µM top dilution, 1:10 serial dilutions) for each time course, 100 µl per well of CellTiter Glo2.0 reagent (Promega, Madison, WI, USA) were added to the standard and the samples every hour, luminescence was measured using Gen5 software (BioTek Instruments, Winooski, VT, USA) using recombinant ATP standard, and the percentage viability was calculated.

Ig production by splenic B lymphocytes

Splenic B lymphocytes, isolated as above, were counted and plated at a concentration of 5 × 10⁶ cells per ml in 200 µl final volume of complete Roswell Park Memorial Institute medium 1640 supplemented with 1× penicillin-streptomycin and 10% FBS, and 10 µg/ml of LPS, in a 96-well flat bottom plate, incubated at 37°C with 5% CO₂. One nanogram per milliliter of TGF-β1 and 100 U/ml of IL-2 were added. Supernatants were collected at d 5 by transferring the whole sample to a V-bottom 96-well plate and centrifuging at 3214 g for 10 min. Supernatants were stored at −20°C for ELISA. Dilutions between 1:10 and 1:20 of supernatants were used in the ELISA for IgA and IgM to ensure that they would be in the linear range of each respective standard curve. A dilution of 1:200 was used for IgG. IgA, IgM, and IgG were determined by ELISA using an SBA Clonotyping System–horseradish peroxidase kit (5300-05; Southern Biotech, Birmingham, AL, USA) and C57BL/6 mouse Ig panel (5300-01B; Southern Biotech). All Ig standards had 3000 pg/ml in the top well, and 1:2 serial dilutions were performed to create linear standard curves.

Immunohistochemistry

Spleens from EdnrB^lacZ/+ were fixed in 4% paraformaldehyde and processed into 30% sucrose prior to embedding in optimal cutting temperature compound. Cryosections (16 μm) of spleen were stained with primary antibodies for CD3 (1:100, M7254, Agilent Technologies), B220 (1:50, 550286; BD Biosciences, San Jose, CA, USA), CD31 (1:100, 550274; BD Biosciences), and anti-lacZ (1:1000, Z3781; Promega). Secondary antibodies included Alexa Fluor 488 (1:500; Thermo Fisher Scientific), Alexa Fluor 568 (1:500; Thermo Fisher Scientific), and Alexa Fluor 647 (1:500; Thermo Fisher Scientific).

Small intestine from EdnrB^NCC−/− and WT mice was harvested and flushed with PBS. The intestinal segment was opened longitudinally along the mesenteric line, oriented as a Swiss roll, and embedded in Optimal Cutting Temperature Compound. Cryosections (10–20 μm) were stained with primary antibodies for IgA (1:50, ab9170; Abcam, Cambridge, MA, USA) and Alexa Fluor–conjugated CD19 (1:500, 115525; BioLegend, San Diego, CA, USA). Secondary antibody for IgA was Alexa Fluor 488 (1:200; Thermo Fisher Scientific). DAPI (300 nM, 422801; BioLegend) was used for nuclear staining.

Slides were imaged on a Nikon A1 confocal microscope and captured with Nikon Elements (Nikon, Tokyo, Japan) or imaged with a Zeiss 710 confocal microscope and captured with Zeiss Zen software (Carl Zeiss, Oberkochen, Germany). Image analysis was performed using ImageJ software (NIH; http://imagej.nih.gov/ij) with the Fiji (https://fiji.sc/) plugin package.

Statistical analysis

The data are expressed as means ± sem. Statistical significance was determined using Student’s t test, Mann-Whitney test, or ANOVA, as appropriate. Differences were considered to be statistically significant at P < 0.05. Statistical calculations were performed with GraphPad Prism v.7.0b (GraphPad Software, La Jolla, CA, USA).

RESULTS

Temporal and spatial expression of endothelin axis components in WT mice

To investigate the distribution of endothelin expression during development, we collected colon, heart, liver, spleen, and thymus from WT mice along a developmental time course. RNA was isolated and microcapillary electrophoretic RNA separation employed to estimate the RIN and the ratio of 28S:18S rRNA (18). RIN values ranged from 7.5 (liver) to 10 (colon, heart, spleen, and thymus). A summary of temporal expression patterns of ET-1, ET-3, EdnrA, and EdnrB in WT samples in major extramedullary hematopoietic organs is given in Fig. 1. Gene expression analysis in WT samples showed that ET-1 was stably expressed in heart, spleen, and thymus, with increased postnatal expression in liver and colon. ET-3 was highly expressed in the early embryonic colon and spleen. EdnrA was highly expressed in the embryonic colon, spleen, and thymus, as well as postnatal liver. EdnrB was highly expressed in the embryonic colon, spleen, and heart, as well as postnatal liver. Protein expression was confirmed with Western blot (EdnrA and EdnrB) and ELISA (ET-1 and ET-3) (Supplemental Fig. S1). High expression of ET-3 and EdnrB in colon during developmental stages is consistent with their known roles in ENS development. Expression of ET-3 and EdnrB in the spleen highlighted a potential role in lymphocyte development or function.

Figure 1.

Expression patterns of the endothelin axis components in WT hematopoietic organs. Gene expression analysis of ET-1 (A), ET-3 (B), EdnrA (C), and EdnrB (D) in colon, heart, liver, spleen, and thymus in WT samples. Data are shown as a heatmap in log₂ scale relative to the lowest expression for each target gene. Gene expression analysis was performed using the following steps: 1) efficiency correction, 2) normalization with reference genes (HPRT1 and TBP), 3) applied relative quantities to the lowest expressing target gene, 4) mean of 3 biologic replicates. Black indicates no expression, and white indicates high expression. Individual relative expression scales are presented in each panel. High expression of ET-3 and EdnrB in colon during developmental stages is consistent with their known roles in ENS development. Expression of ET-3 and EdnrB in the spleen highlighted a potential role in lymphocyte development or function.

Altered expression patterns of endothelin axis components in EdnrB^NCC−/− and EdnrB^lacZ/lacZ mice

Having established a baseline pattern of expression for endothelin axis components in WT mice, we next sought to understand how these expression patterns are altered in EdnrB^NCC−/− and EdnrB^lacZ/lacZ mice. Tissue samples were obtained from EdnrB^NCC+/−, EdnrB^NCC−/−, EdnrB^lacZ/+, and EdnrB^lacZ/lacZ mice along the same developmental time course, and qPCR was used to determine gene expression (Fig. 2). Values were normalized against WT samples at the same time point and tissue. Colon demonstrated the highest expression (green) at E14.5 compared with later developmental stages for all 4 genes analyzed (ET-1, ET-3, EdnrA, EdnrB). Interestingly, EdnrB^NCC+/− and EdnrB^NCC−/− mice demonstrated higher expression of EdnrB in all organs in all developmental stages compared with WT. This result is consistent with the original description of this model, in which higher EdnrB expression was observed in non-NCC of the gut during development (15). Likewise, EdnrB^lacZ/+ demonstrated higher EdnrB expression than WT except for in the spleen, in which no significant difference was observed. In contrast, EdnrB^lacZ/lacZ display consistently lower or no expression of EdnrB.

Figure 2.

Expression patterns of the endothelin axis components are altered in EdnrB^NCC−/− and EdnrB^−/− mice. Gene expression analysis of ET-1 (A), ET-3 (B), EdnrA (C), and EdnrB (D) in colon, heart, liver, spleen, and thymus in EdnrB^NCC+/−, EdnrB^NCC−/−, EdnrB^lacZ/+, and EdnrB^lacZ/lacZ samples compared with WT samples. Heatmaps were generated for all genotypes by normalizing the samples against their WT counterpart at the same time point and tissue (Fig. 1). Green indicates increased expression relative to WT, and red indicates decreased expression relative to WT. There is increased expression of ET-3 and EdnrB in the developing spleen of EdnrB^NCC−/− mice.

Splenic B lymphocytes express EdnrB

Because of the previously observed B-lymphocyte phenotype in EdnrB^NCC−/− mice, we next examined splenic B lymphocytes for EdnrB expression. Confocal microscopy of splenic cross-sections from postnatal EdnrB^lacZ/+ animals, which express the lacZ reporter at sites of EdnrB expression, demonstrates colocalization of lacZ (EdnrB) with B220 (B lymphocytes) but not CD3 (T lymphocytes) (Fig. 3A). Of note, EdnrB expression has also been described in NCC-derived perivascular cells that adopt an astrocyte-like, glial lineage in the spleen (19), indicating that multiple cell types express EdnrB in the spleen (Fig. 3B). We further validated B-lymphocyte expression of EdnrB in magnetic bead–sorted splenic B lymphocytes from EdnrB^NCC−/− mice (Fig. 3C).

Figure 3.

Identification of EdnrB-expressing cells in the spleen. A) Confocal microscopy of spleen from EdnrB^lacZ/+ animals demonstrates colocalization of lacZ (EdnrB) with B220 (B lymphocytes) but not CD3 (T lymphocytes). Scale bar, 200 μm. B) lacZ (EdnrB) also colocalizes with cells surrounding the vasculature (CD31) in the postnatal spleen, indicating that multiple cell types express EdnrB in the spleen. Scale bar, 50 μm. C) Real-time qPCR in magnetic bead–sorted B lymphocytes from WT and EdnrB^NCC−/− animals confirms expression of EdnrB. NS, not significant.

B-lymphocyte survival

To determine if neural crest specific deletion of EdnrB results in B-lymphocyte–intrinsic defects, we examined survival of splenic B lymphocytes in EdnrB^NCC−/− compared with WT mice (Fig. 4A). There were no significant differences in viability between EdnrB^NCC−/− or EdnrB^lacZ/lacZ B lymphocytes (Supplemental Fig. S2A) vs. WT, indicating that there is no intrinsic defect in B-lymphocyte survival in EdnrB-mutant models.

Figure 4.

EdnrB^NCC−/− B-lymphocyte viability and Ig production. A) Splenic B lymphocytes were isolated using CD19 microbeads, counted, and plated at a concentration of 1 × 10⁶ cells per ml in 100 μl final volume and incubated at 37°C with 5% CO₂. Recombinant ATP standard was prepared (10 µM top dilution, 1:10 serial dilutions) for each time course, and CellTiter Glo2.0 reagent (100 µl/well) was added to the standard and the samples every hour; luminescence was measured and calculated according the standard and transformed to percentage of viability. EdnrB^NCC−/− mice (n = 6) vs. WT mice (n = 9) showed no differences in survival (P = not significant). B–D) Splenic B-lymphocyte Ig production in vitro from EdnrB^NCC−/− mice (n = 6) vs. WT mice (n = 7) showed decreased IgA (B), decreased IgM (C), and decreased IgG measured as optical density at 450 nm (OD_450nm) (D). Bars represent means ± sem. *P < 0.05.

Ig production by B lymphocytes

In order to examine B-lymphocyte Ig production, we isolated splenic B lymphocytes from EdnrB^NCC−/− animals prior to HAEC onset, along with WT controls, and stimulated them in vitro with LPS, TGF-β1, and IL-2. After 5 d of incubation, supernatants were collected, and levels of IgA, IgM, and IgG were determined by ELISA (Fig. 4B–D). However, EdnrB^NCC−/− B lymphocytes demonstrated decreased IgA, IgM, and IgG compared with WT splenic B lymphocytes. These results indicate an intrinsic defect in B-lymphocyte function in EdnrB^NCC−/− animals.

pIgR expression and IgA transport

Evidence of inflammation during HAEC can be found in both the colon and small intestine (ganglionated and aganglionic bowel) and continues to occur in both humans and mice after the aganglionic segment is resected, suggesting that potential mucosal immune defects extend beyond the aganglionic segment (4). In mice, the majority of intestinal IgA production by B lymphocytes is induced by antigen presentation at the Peyer’s patches (PPs) (20), and we have previously demonstrated B lymphocytopenia of PPs and decreased luminal sIgA during HAEC in EdnrB^NCC−/− mice (4). B-lymphocyte–produced IgA dimerizes and is transported across the epithelial barrier by pIgR. As part of the transcytosis process, pIgR is cleaved by proteases and a secretory component is released as a complex with IgA, resulting in sIgA in the lumen. EdnrB^NCC−/− mice demonstrate decreased total pIgR and secretory component expression (Fig. 5). Additionally, based on immunohistochemistry, IgA appears to accumulate along the epithelial barrier of EdnrB^NCC−/− mice, consistent with failure to be transported into the lumen.

Figure 5.

pIgR expression and IgA transport. A) Representative Western blot showing expression of pIgR in ileum of WT and EdnrB^NCC−/− mice. B) Quantification of Western blots shows decreased total pIgR and secretory component (SC) expression in EdnrB^NCC−/− mice. *P < 0.05. C) Confocal microscopy of WT and EdnrB^NCC−/− ileum demonstrating high levels of IgA staining along the epithelial barrier of EdnrB^NCC−/− mice. Scale bar, 500 μm.

DISCUSSION

Our group and others have previously identified that EdnrB-mutant (conventional and neural crest–conditional) mice that model HSCR demonstrate splenic lymphopenia (4, 13). Additionally, we previously identified that EdnrB^NCC−/− mice demonstrate decreased luminal sIgA (4). These findings led us to interrogate the endothelin axis during extramedullary hematopoietic development and postnatal B-lymphocyte function in the EdnrB models of HSCR and HAEC. We identified expression of endothelin axis components (EdnrA, EdnrB, ET-1, ET-3) in developing hematopoietic organs, including high splenic expression of ET-3 in both models. We additionally noted that some splenic B lymphocytes express EdnrB. Splenic B lymphocytes from EdnrB^NCC−/− and EdnrB^lacZ/lacZ mice showed no intrinsic defect in survival vs. WT B lymphocytes. However, in vitro stimulation of splenic B lymphocytes demonstrated decreased IgA, IgG, and IgM production in EdnrB^NCC−/− vs. WT mice. Additionally, small intestinal lamina propria pIgR, which transports and secretes IgA into the lumen, was decreased ∼50% in EdnrB^NCC−/− mice, and IgA staining was increased along the epithelial border in these animals.

IgA is central to maintaining host-microbiota commensalism, and sIgA is the first line of antigen-specific immune defense in the gut (21). Human IgA deficiency is common (∼1:500 in the United States) and is associated with recurrent mucosal infections, increased incidence of autoimmune disease, and intestinal inflammation (including inflammatory bowel disease) (22, 23). Mouse models of impaired IgA secretion result in altered microbiota composition, increased susceptibility to colitis, and higher levels of bacterial translocation from the intestine to the mesenteric lymph nodes (24–26). A potential role for IgA in HAEC pathogenesis was first demonstrated by Imamura et al. (11). Using human surgical tissue samples, they noted an increase in IgA-containing plasma cells along the length of resected aganglionic bowel from patients with HAEC compared with HSCR patients without HAEC. They also found decreased luminal IgA (sIgA) in the same patients. Together, these results suggested decreased IgA production or transport in HSCR and HAEC.

The process of sIgA production and secretion into the small intestine lumen is complex (27). Antigen is presented across the epithelial barrier in the PPs. There, IgM⁺IgD^hi (mature) B lymphocytes are activated to produce IgA. In the mouse, these mature B lymphocytes derive primarily from the spleen. Activated mature B lymphocytes circulate through the mesenteric lymph, to the circulation, and home to the lamina propria as IgA-secreting plasma cells. IgA is then transported across the epithelial barrier by pIgR and secreted. We have previously demonstrated decreased, gut-specific luminal sIgA, a decreased mature B-lymphocyte population in PPs, and increased mature B lymphocytes in the spleen of EdnrB^NCC−/− mice, with abnormal location in the spleen (4). However, trafficking cues governing movement of B lymphocytes from the spleen to the PPs appeared intact. In the current study, we extended this finding by noting both decreased production of IgA (as well as other Ig) by splenic B lymphocytes in EdnrB^NCC−/− mice and decreased levels of pIgR (transport) in the same animals. Taken as a whole, our current and past results raise new questions about HAEC mechanisms. First, what are the precise roles for endothelin axis signaling in splenic lymphocyte development? Next, what are the mechanisms regulating intestinal pIgR expression in the setting of HSCR and HAEC? Furthermore, are these defects targetable for potential therapeutic benefit?

The primary components of the endothelin axis are the receptors, EdnrA and EdnrB, and their cognate ligands, ET-1 and ET-3. Components of this system have been identified throughout the body during both development and disease and participate in site- and disease-specific processes (5). For example, vascular endothelial cells produce ET-1, which exerts autocrine-paracrine functions by binding to EdnrA and EdnrB present on vascular endothelial cells and pericytes, thereby regulating endothelial cell proliferation and vascular tone. The endothelins also have known roles in promoting inflammation, both in the vascular system and through other cell types (28). ET-1 is synthesized by lymphocytes (29), is chemotactic for polymorphonuclear cells (30), induces thymocyte proliferation (31), and is produced by dendritic cells (along with increased expression of EdnrB) after stimulation with TLR agonists (32). In the current study, we noted diffuse expression of the endothelin axis components in the WT extramedullary hematopoietic system over a developmental time course. Specifically, we noted that splenic B lymphocytes express EdnrB, which is consistent with a recent report of EdnrA and EdnrB expression by circulating T and B lymphocytes in humans (33). Interestingly, we noted generally increased ET-3 expression in the developing spleen in both EdnrB^NCC−/− and EdnrB^lacZ/lacZ mice, as well as increased expression of EdnrB in the EdnrB^NCC−/− model. We additionally noted decreased Ig production by EdnrB^NCC−/− B lymphocytes (Fig. 4) but not by EdnrB^lacZ/lacZ B lymphocytes (Supplemental Fig. S2). This may reflect developmental differences in the B-lymphocyte populations in these 2 models caused by differential expression of the endothelin axis in the cells and the surrounding mesenchyme. Consistent with this hypothesis, another group has recently noted reduced EdnrB expression in B lymphocytes derived from trisomy 21 pluripotent stem cells and that knockdown of EdnrB in normal (disomic) cord blood cells resulted in defective B-cell lymphopoiesis (34). Together, these results support intrinsic defects in B-lymphocyte development and function in EdnrB-mutant animals.

Decreased epithelial expression of pIgR is a hallmark of multiple inflammatory bowel pathologies, including ulcerative colitis, Crohn’s disease, and experimental chemical colitis (25, 35–37). In these instances, pIgR is typically decreased in concert with a dysbiotic microbiota, dysfunctional TLR signaling, or both. We have previously identified microbiota dysbiosis in the EdnrB^NCC−/− model (38), and other investigators have made similar observations in global EdnrB-knockout animals (39). However, the time course of HAEC appears to differ between the EdnrB^−/− and EdnrB^NCC−/− models, making head-to-head comparisons difficult. For example, we observed a median survival of ∼28 d in EdnrB^NCC−/− mice (38), and other investigators have published ranges of survival from ∼3 wk of age to 5–6 wk of age for EdnrB^−/− animals, which suggests that, beyond model-specific differences, facility-specific microbiota changes contribute to disease severity and progression (40). Recent studies have identified that pIgR expression in intestinal epithelial cells is maintained through a complex interplay between TLR signaling and NF-κβ pathways in intestinal epithelial cells (41). Whether these pathways are perturbed in EdnrB-mutant mice or if they can be co-opted to therapeutic effect in HAEC has not yet been established. Multiple other environmental conditions can regulate pIgR expression as well (42), most notably diet (43). EdnrB^NCC−/− mice gain weight similarly to WT mice until ∼P15, at which point their oral intake decreases. Their weights are similar to WT mice until P16–18, at which point the weight curves diverge. In our previous analysis of B-lymphocyte populations in these animals, we accounted for weight in the comparisons of organ size and cellularity (4). However, the finding of decreased pIgR in EdnrB^NCC−/− mice may reflect effects from nutritional status in addition to the altered microbiota.

In summary, endothelin axis components (EdnrA, EdnrB, ET-1, ET-3) are widely expressed in developing extramedullary hematopoietic organs, and murine models of EdnrB mutation exhibit significant alterations in these patterns of expression. There appears to be an intrinsic B-lymphocyte defect in antibody production as well as an extrinsic defect in IgA transport in the EdnrB^NCC−/− model of HAEC. Our results are consistent with human HAEC observations of decreased luminal sIgA and mouse models of other inflammatory bowel diseases, in which decreased pIgR is seen in concert with a dysregulated microbiota. The current results and our prior investigations utilizing the EdnrB^NCC−/− model have refined our understanding of HAEC pathogenesis: ENS dysmotility (present from birth) results in microbiome dysbiosis (normal P16–18, dysbiotic by P21–24) (38). This precipitates intestinal barrier dysfunction (unpublished results) and an impaired immune response (P21–24: B-lymphocyte populations, sIgA, pIgR) (4), culminating in HAEC (P26–29) and death (P28–29) (38).

Currently, the only treatment for patients with HAEC is nonspecific and consists of systemic antibiotics, rectal irrigations, and bowel rest (2). Development of a targeted therapy for HAEC has the potential to decrease mortality. Our current results suggest that targeting the dysbiotic microbiome (e.g., through use of pre- or probiotic therapies, directed antimicrobials, or fecal microbiota transplant) and pIgR-mediated sIgA transport may be potential therapeutic approaches for the prevention and treatment of HAEC.

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.

Glossary

EdnrA

endothelin receptor A

EdnrB

endothelin receptor B

EdnrB^lacZ/+

conventional heterozygous EdnrB knockout

EdnrB^lacZ/lacZ

conventional null EdnrB knockout

EdnrB^NCC−/−

homozygous NCC-conditional deletion of EdnrB

EdnrB^NCC+/−

heterozygous NCC-conditional deletion of EdnrB

ENS

enteric nervous system

ET-1

endothelin-1

ET-3

endothelin-3

HAEC

Hirschsprung-associated enterocolitis

HPRT1

hypoxanthine phosphoribosyltransferase 1

HSCR

Hirschsprung disease

lacZ

β-galactosidase

NCC

neural crest cell

pIgR

polymeric Ig receptor

PP

Peyer’s patch

qPCR

quantitative PCR

RIN

RNA integrity number

sIgA

secretory IgA

TBP

TATA-box binding protein

UPL

Universal Prob eLibrary

WT

wild type

AUTHOR CONTRIBUTIONS

A. Gosain was responsible for study concept and design; G. Medrano, F. Cailleux, P. Guan, K. Kuruvilla, A. J. Barlow-Anacker, and A. Gosain were responsible for acquisition of data; G. Medrano, F. Cailleux, P. Guan, K. Kuruvilla, and A. Gosain were responsible for analysis and interpretation of data; G. Medrano and A. Gosain were responsible for drafting and revising of the manuscript; and all authors approved the final manuscript.