https://orcid.org/0000-0002-4140-3717
Abstract
SAMP1/YitFcsJ (SAMP1) mice spontaneously develop terminal ileitis resembling human Crohn disease. SAMP1 mice have exhibited alteration of epithelial cell lineage distribution and an overall proliferation of the crypt cell population; however, it has not been evaluated whether epithelial differentiation is impaired because of dysfunction of intestinal stem cells (ISCs) or their niche factors.
Using the intestine of SAMP1 mice aged 10 to 14 weeks, morphometric alterations in the crypt-villus architecture, ISCs, crypt cells, and differentiated cells; organoid formation capacity of intestinal crypts; and niche signaling pathways were analyzed and compared with those of age-matched control AKR/J (AKR) mice.
The ileum of SAMP1 mice showed increased depth of intestinal crypts and decreased surface area of the villi compared with those in the ileum of AKR mice. The number of ISCs in the ileal crypts did not differ between SAMP1 and AKR mice; however, the number of Paneth cells decreased and the number of transient amplifying cells increased. The organoid formation rate of the ileal crypts of SAMP1 mice decreased significantly compared with that of AKR mice. The performance of RNA sequencing for intestinal crypts found that the expression of ISC niche factors, such as Wnt3, Dll1, and Dll4, was decreased significantly in the ileal crypts of SAMP1 mice compared with those of AKR mice. Among the ISC niche signals, the Notch signaling-related genes tended to be downregulated. In particular, immunocytochemistry revealed that the expression of Paneth cell–expressing Notch ligand Dll4 was significantly decreased in the intestinal tissue and organoids of SAMP1 mice compared with those of AKR mice.
Depletion of niche factors for ISCs contributes to the alteration of epithelial differentiation in SAMP1 mice.
Introduction
Crohn disease (CD) is characterized by chronic segmental inflammation of the gastrointestinal tract. The disease can occur anywhere from the mouth to the anus, but it most frequently involves the ileum.1,2 It is thought be developed from altered intestinal mucosal immunity to commensal microbiota and environmental factors in genetically susceptible individuals; however, the precise pathogenesis of CD remains unclear.3 The histological findings from endoscopic biopsy samples of patients with CD include the expansion of the proliferative zone in the intestinal crypts, increased crypt branching, and alteration in the pattern of epithelial cellular differentiation,4 suggesting that signaling pathways associated with epithelial cellular replication and differentiation may be amended in the ileum of patients with CD.5
SAMP1/YitFcsJ (SAMP1) mice develop spontaneous segmental ileitis that resembles human CD.6 These mice show a CD-like phenotype characterized by increased intestinal interferon-γ production by age 4 weeks, development of ileitis by age 10 weeks, and progress to intestinal strictures in nearly 50% of mice older than age 40 weeks.5,7-9 Therefore, experiments using SAMP1 mice are considered as a highly pertinent for elucidation of the pathological mechanisms of regional ileitis in patients with CD.9 Previous findings have shown that SAMP1 mice exhibit alteration of epithelial cell lineage distribution and an overall proliferation of the crypt cell population5,7; however, it has not been evaluated whether epithelial differentiation is impaired because of dysfunction of intestinal stem cells (ISCs) or their niche factors.
The spontaneous ileitis of SAMP1 mice develops by ages 10 to 12 weeks, which correlates with the early 20s in humans; the majority of patients with CD are diagnosed in their 20s. In this study, we evaluated structural alterations and cellular changes in the villi and crypts of SAMP1 mice aged 10 to 14 weeks. Furthermore, because cellular replication and epithelial differentiation occur in intestinal crypts, the gene expression pattern and replication potential of the ileal and jejunal crypts of SAMP1 mice were compared with those of age-matched control mice.
MATERIALS AND METHODS
Animals
SAMP1 (stock number 009355) and AKR/J (AKR; stock number 000648) mice were acquired from Jackson Laboratories (Bar Harbor, ME) and maintained under specific pathogen-free conditions. All were housed under a 12-hour light/dark cycle and were provided ad libitum. The duodenum (proximal third of the small intestine), jejunum (mid-third of the small intestine), ileum (distal third of the small intestine), proximal colon (proximal half of the large intestine), and distal colon (distal half of the large intestine) were harvested from SAMP1 and AKR mice aged 10 to 14 weeks. The Institutional Animal Care and Use Committee of the Samsung Biomedical Research Institute at the Samsung Medical Center approved the protocol (approval number 20150105005, approval date January 19, 2015).
Morphometric Index of the Crypt-Villus Structure
Hematoxylin and eosin (H&E)-stained sections were used to measure the number of intestinal crypts per 5 mm of the small and large intestine, the villus height (from the tip of the villus to the crypt-villus junction), the crypt depth (from the crypt-villus junction to the bottom of the crypt), the villus width (at the mid-portion of the villus), and the crypt width (at the mid-portion of the crypt). The villus surface area was calculated using the following formula: villus surface area = [villus width × villus length + (villus width / 2 + crypt width / 2)2 − (villus width / 2)2] / (villus width / 2 + crypt width / 2).10
Alcian Blue and Immunohistochemical Staining
After heat-induced epitope retrieval, staining was done with Alcian blue solution (Sigma-Aldrich, St. Louis, MO) and antibodies against chromogranin A (1:1000; Abcam, Cambridge, UK), lysozyme (1:3000; Abcam), olfactomedin 4 (OLFM4; 1:400; Cell Signaling Technology, Danvers, MA), proliferating cell nuclear antigen (1:100; Abcam), and delta-like ligand 4 (Dll4; 1:200; LSBio, Seattle, WA).
5-ethynyl-2’-deoxyuridine Assay
Intraperitoneal injection with 200 μg of 5-ethynyl-2’-deoxyuridine (EdU; Sigma-Aldrich) was performed 24 hours before the mice were killed. Incorporation of EdU into DNA was conducted with the Click-iT EdU Alexa Fluor 488 Imaging kit (Thermo Fisher Scientific, Waltham, MA). We performed OLFM4 co-staining to differentiate ISCs from transient amplifying (TA) cells.
Terminal Deoxynucleotidyl Transferase-Mediated Biotin-Deoxyuridine Triphosphate Nick-End Labeling Assay
Apoptosis-associated DNA fragmentation was detected using an in situ cell death detection kit (Sigma-Aldrich). Positive and negative control sections were incubated with 10 U/mL recombinant DNase I solution and with the absence of the terminal transferase enzyme.
Immunofluorescent Staining
Formalin-fixed, paraffin-embedded sections were stained with antibodies against Dll4 (1:200 dilution; LSBio). Fluorescent secondary antibodies used were Alexa Fluor 488 goat anti-mouse IgG2b (1:1000 molecular probes, Invitrogen, Carlsbad, CA), and nuclei were stained with 4’,6-diamidino-2-phenylindole (Sigma-Aldrich).
Mouse Intestinal Crypt Isolation
Murine small and large intestinal crypts were isolated as previously described.11,12 Briefly, AKR and SAMP1 mice aged 12 to 16 weeks were euthanized, and the duodenum, jejunum, ileum, and colon were harvested and emptied with ice-cold phosphate-buffered saline (PBS). Each segment of the intestine was incubated in PBS with 1 mM dithiothreitol (DTT, Sigma-Aldrich) and 10 mM ethylenediaminetetraacetic acid (EDTA, Sigma-Aldrich) at 4°C for 30 minutes. The samples were washed and vortexed with cold PBS. The washed solution was passed through a 70 μm pore cell strainer (BD Biosciences, Bedford, MA). The number of isolated crypts was counted with an inverted optical microscope.
Organoid Reconstitution Assay
Isolated crypts were cultured at a concentration of 100 crypts per 25 μL of 3-dimensional Matrigel matrix (BD Biosciences) in a culture medium consisting of advanced Dulbecco’s Modified Eagle Medium/Ham’s F-12 (Thermo Fisher Scientific) containing 1× antibiotic-antimycotic (Thermo Fisher Scientific), 2 mM glutamax (Invitrogen), 10 mM HEPES (Invitrogen), 1× N2 (Invitrogen), 1× B27 (Invitrogen), 50 ng/mL epidermal growth factor (Invitrogen), 1 mM N-acetylcysteine (Sigma-Aldrich), 100 ng/mL recombinant murine Noggin (Peprotech, Rocky Hill, NJ), and 500 ng/mL recombinant human R-spondin1 (R&D Systems, Minneapolis, MN). For the first 2 days, 2.5 μM glycogen synthase kinase 3 inhibitor (CHIR99021; Stemgent, Cambridge, MA) was added. The culture medium was replaced every 2 days. Micrographs of in vitro cultures were obtained 3 and 7 days after plating using an inverted optical microscope. The organoid formation rate was calculated as the percentage of viable organoids per 100 intestinal crypts.
RNA Sequencing of Jejunal and Ileal Crypts
Jejunal and ileal crypts were isolated from AKR (n = 3) and SAMP1 (n = 3) mice. Total RNA was extracted from intestinal crypts using the RNeasy Mini Kit (Qiagen, Valencia, CA). We performed RNA sequencing (RNA-seq) using total RNA samples with an RNA integrity number > 8 and a quantity > 10 μg. An RNA-seq library was implemented via complementary DNA amplification, end repair, 3’ end adenylation, adapter ligation, and amplification. Sequencing of the transcriptome library was performed using the 100-bp paired-end mode of the TruSeq Rapid PE Cluster Kit and the TruSeq Rapid SBS Kit (Illumina, San Diego, CA).
RNA-Seq Data Analysis
Reads from FASTQ format files were mapped against the Mus musculus reference genome (mm10) using TopHat version 2.0.6. The mapping results were analyzed using HTSeq version 0.6.1 to quantify transcript abundance. Read counts were normalized, and differentially expressed genes were identified using the edgeR package (version 3.28.1) and MeV software (version 4.9.0).
Quantitative Real-Time Reverse-Transcriptase Polymerase Chain Reaction
Quantitative polymerase chain reaction (qPCR) was performed with the following primers: Dll4 (F: 5’-GATGTGATGAGCAGCATGGA-3’, R: 5’-ACCTCGG TTCAGGCACTGTC-3’), Vil1 (F: 5’-GGTGGTTAGAGAAGTTG CTA-3’, R: 5’-TGGAAGAGTTGTTGGAAGAT-3’), Apoa1 (F: 5’-ACCGTTCAGGATGAAAACTGTAG-3’, R: 5’-GTGACTCAGGAGTTCTGGGATAAC-3’), Tff3 (F: 5’-AACCG GGGCTGCTGCTTTG-3’, R: 5’-GAGGTGCCTCAGAAGGT GC-3’), Muc13 (F: 5’-GCCAGTCCTCCCACCACGGTA-3’, R: 5’-CTGGGACCTGTGCTTCCACCG-3’), and GAPDH (F: 5’-ACCCAGTCCATGCCATCAC-3’, R: 5’-TCCACCA CCCTGTTGCTGTA -3’).
Statistical Analysis
Differences between experiment groups were analyzed using the Mann-Whitney U test, the Kruskal-Wallis test, and a linear regression model using the statistical software GraphPad Prism (GraphPad Software, La Jolla, CA). Measurements were displayed as a box plot with minimum-to-maximum whiskers and a bar plot with standard error. Statistical significance was set at a level of P < 0.05.
RESULTS
The crypt-villus structure of the small and large intestine of SAMP1 and AKR mice aged 10 to 14 weeks is presented in Fig. 1A. The SAMP1 mice exhibited spontaneous ileitis (Supplementary Fig. 1). The number of crypts in the small and large intestines did not significantly differ between the SAMP1 and AKR mice (Fig. 1B). However, the crypt depth in the SAMP1 ileum was significantly increased compared with that in the AKR ileum (P < 0.001; Fig. 1C). The surface area of the villi in the SAMP1 ileum was significantly decreased compared with that in the AKR ileum (P < 0.001; Fig. 1D). In the H&E-stained tissue, the number of intermediate cells containing eosinophilic granules was increased in the elongated ileal crypts of SAMP1 mice. Eosinophilic granules were prominent in cells located at the crypt base. In the ileum of AKR mice, eosinophilic granule-containing cells were restricted to positions 1 to 4 at the crypt base. Because morphometric differences between AKR and SAMP1 mice were limited to the ileum and human CD is characterized by ileitis, we compared the ileum of AKR and SAMP1 mice.
Crypt-villus structure of AKR and SAMP1 mice according to intestinal location. A, H&E staining. The scale black bar indicates 300 μm and the scale white bar indicates 100 μm. B, Number of crypts. C, Crypt depth. D, Villus surface area. Measurement was performed in 10 intact villi and 10 crypts selected from each intestinal segment from SAMP1 (n = 6) and AKR mice (n = 6).
The elongated crypts in the ileum of SAMP1 mice contained several cellular components, including ISCs, Paneth cells, and proliferating TA cells. The number of ileal ISCs did not differ between AKR and SAMP1 mice, although the OLFM4 staining intensity was reduced in the SAMP1 ileum (Fig. 2A). The expression of ISC markers—LGR5, SMOC2, MSI1, AXIN2, EPHB2, and OLFM4—did not show any significant differences in the ileal and jejunal crypts of SAMP mice compared with those in AKR mice (Supplementary Fig. 2). Compared with AKR mice, the number of lysozyme-stained Paneth cells was significantly decreased in the ileal crypts, but not in the jejunal crypts, of SAMP1 mice (Fig. 2B). The expression of the Paneth cell markers LYZ1 and DEFA5 was diminished significantly in the SAMP1 crypts compared with those in the AKR crypts (Supplementary Fig. 3). Moreover, the number of proliferating cell nuclear antigen–stained TA cells was significantly increased in the ileal crypts, but not in the jejunal crypts, of SAMP1 mice compared with those in age-matched control AKR mice (Fig. 2C). In the EdU assay, the number of replicating S phase cells per crypt in the ileum of SAMP1 mice was significantly increased compared with that of control mice (Fig. 2D).
Comparison of the cellular components of jejunal and ileal crypts between SAMP1 and AKR mice. A, OLFM 4 staining of ISCs. B, Lysozyme staining of Paneth cells. C, Proliferating cell nuclear antigen staining of TA cells. D, EdU assay. Immunohistochemically positive and EdU-positive cells were counted in 10 intact villi or crypts selected from each intestinal segment of SAMP1 (n = 3) and AKR mice (n = 3).
In addition, altered epithelial differentiation was noted in the ileum of SAMP1 mice. Compared with that of AKR mice, the number of Alcian blue–stained goblet cells was significantly amplified in the ileal villi, but not in the jejunal villi, of SAMP1 mice (Fig. 3A). The expression of the enterocyte markers VIL1 and APOA1 and the goblet cell markers TFF3 and MUC13 did not show any significant differences in the ileum and jejunum crypts of SAMP mice compared with those in AKR mice (Supplementary Fig. 4). The number of chromogranin-stained enteroendocrine cells in the ileal and jejunal villi did not significantly differ between AKR and SAMP1 mice (Fig. 3B). A terminal deoxynucleotidal transferase–mediated biotin-deoxyuridine triphosphate nick-end labeling assay was performed to evaluate apoptosis in intestinal crypt-villi structure. Terminal deoxynucleotidal transferase–mediated biotin-deoxyuridine triphosphate nick-end labeling (+) cells were significantly increased in the villi of SAMP1 mice compared with those of AKR mice (Fig. 3C).
Differentiated cells in the villi of ileum and jejunum in SAMP1 and AKR mice. A, Alcian blue staining for goblet cells (blue color). B, Immunohistochemistry for enteroendocrine cells (brown color). Immunohistochemically positive cells were counted in 10 intact villi or crypts in each intestinal segment of SAMP1 (n = 3) and AKR mice (n = 3).
In the organoid reconstitution assay, the organoid formation ability of intestinal crypts tended to be lower in SAMP1 mice than in AKR mice (Fig. 4A). The organoid formation rate of ileal crypts was significantly decreased in SAMP1 mice compared with that in AKR mice (day 3: 35.2% ± 6.6% vs 19.7% ± 9.0%; P < 0.01; day 7: 32.1% ± 7.8% vs 16.1% ± 7.3%; P < 0.01; Fig. 4B).
Organoid reconstitution assay used for AKR and SAMP1 mice. A, Bright-field micrographs of organoids (×40). B, Organoid formation rate on days 3 and 7 after intestinal crypts were embedded in Matrigel. Experiments was performed in triplicate using duodenal, jejunal, ileal, and colonic organoids derived from intestinal crypts of SAMP1 (n = 3) and AKR mice (n = 3).
The RNA-seq of jejunal and ileal crypts isolated from AKR and SAMP1 mice revealed 1434 differentially expressed genes. The gene expression pattern of the intestinal crypts was clustered between AKR and SAMP1 mice (Fig. 5A). The gene expression patterns of the ileal and jejunal crypts exhibited a clear distinction in principal component analysis and cluster dendrograms of AKR mice but not of SAMP1 mice (Fig. 5B). Gene ontology and kyoto encyclopedia of genes and genomes (KEGG) pathway analysis identified the enrichment of inflammatory response in the crypt of SAMP1 mice compared with that of AKR mice. In the crypt of SAMP1 mice, the gene expression of proinflammatory cytokines, such as interferon-γ, interleukin (IL)-1β, and IL-6; anti-inflammatory cytokines, such as IL-4 and IL-10; and prostaglandin-endoperoxide synthase (PTGS, known as cyclooxygenase)/prostaglandin E synthase (PTGES) pathway was upregulated, whereas the gene expression of JAK3, MAF, NFKB2, and TNF was upregulated in the crypt of AKR mice (Fig. 5C); however, the gene expression profile for pro- and anti-inflammatory genes was not separated clearly between the AKR and SAMP1 mice (Supplementary Fig. 5).
RNA-seq of jejunal and ileal crypts of AKR (n = 3) and SAMP1 mice (n = 3). A, Heatmap of DEGs. B, Principal component analysis. C, Top 5 gene ontology biological processes and KEGG pathwayes between AKR and SAMP1 mice crypts. D, Hierarchical clustering heatmap of the inflammation-related genes. E, Hierarchical clustering heatmap of the intestinal crypt cell lineage-specific genes. F, DEGs in the intestinal stem cell niche factors. G, DEGs in the Wnt signaling pathway. H, DEGs in Notch signaling pathway. DEGs indicate differentially expressed genes.
This study focused on intestinal epithelial differentiation regardless of inflammatory status, and the heatmap illustrating the expression profiles of crypt cell markers and the expression of Paneth cell markers was decreased in the crypts of SAMP1 mice, especially the ileal crypts (Fig. 5D). In analyzing ISC niche factors, we found that the differential expression of Dll1, Dll4, Tfgb3, and Wnt9A was identified between the AKR and SAMP1 mice (Fig. 5E and Supplementary Fig. 6). Although both Wnt and Notch signaling were altered, the alteration of the Notch signaling pathway was more prominent than Wnt signaling pathway (Figs. 5F and 5G). Interestingly, among the Paneth cell markers, Paneth cells presenting ISC niche factors, such as Dll1, Dll4, and Wnt3, were decreased significantly in SAMP1 crypts (Supplementary Fig. 7).
Among ISC niche factors, Dll4 was consistently decreased in SAMP1 crypts compared with AKR crypts. In the qPCR, immunofluorescent, and immunohistochemical staining, Dll4 expression was significantly decreased in the intestinal crypts and organoids of SAMP1 mice compared with those of AKR mice. The qPCR results confirmed that the expression of Dll4 was significantly decreased in the jejunal and ileal crypts of SAMP1 mice compared with those of AKR mice (Fig. 6A). Immunofluorescence showed that Dll4 was expressed around the +4 position in the jejunal and ileal crypts of AKR mice. However, Dll4 was not expressed in the jejunal and ileal crypts of SAMP1 mice (Figs. 6B and 6C). In addition, the Dll4-stained cells were significantly decreased in the SAMP1 mice–derived organoids compared with those of AKR mice (P < 0.001; Figs. 6D and 6E).
Altered expression of Dll4 in jejunal and ileal tissues and organoids of AKR and SAMP1 mice. A, qPCR in the jejunal and ileal tissues of AKR and SAMP1 mice. B, Immunofluorescent staining of Dll4 in the jejunal and ileal tissues of AKR and SAMP1 mice. The scale white bar indicates 30 μm. qPCR and immunofluorescent staining were conducted in triplicate for jejunal and ileal crypts obtained from AKR (n = 3) and SAMP1 mice (n = 3). C, Number of immunofluorescent Dll4 (+) cells in crypt. Measurement was performed in 10 intact crypts selected from jejunal and ileal tissues in AKR (n = 3) and SAMP1 mice (n = 3). D, Immunohistochemical staining of Dll4 in intestinal organoids of AKR and SAMP1 mice. E, Number of Dll4 (+) cells in organoids. Measurement was performed in 10 organoids derived from AKR (n = 3) and SAMP1 mice (n = 3).
Discussion
The present study provided additional details on the expansion of the crypt epithelial cell population and explored the signaling pathways associated with ISC self-renewal and differentiation. Our findings showed that the ISC population was not increased, the Paneth cells decreased, and the TA cells expanded significantly. Most of these TA cells were stained with Alcian blue, suggesting an aberrant increase in the production of secretory progenitors. In addition, a morphometric index of the crypt-villus structure clearly showed that the surface area of the villi was significantly reduced, which indicated impaired epithelial differentiation and diminished absorptive surface in SAMP1 mice. These results indicated that in the ileum of SAMP1 mice, the production of TA cells from ISCs was increased; however, the differentiation of these cells into mature cells were impaired. In the ex vivo organoid reconstitution assay, the organoid formation of ileal crypts in SAMP1 mice was significantly lower than that of AKR mice. Furthermore, we investigated the ISC/niche signaling using RNA-seq and found that the expression of niche factors for ISCs was significantly decreased, in particular the Notch signaling–related genes, in the ileal crypts of SAMP1 mice compared with those of AKR mice. Given that regeneration and differentiation are necessary to restore the epithelial cell lining, the altered differentiation and regeneration potential of ileal crypts in SAMP1 mice may contribute to the development of ileitis.
The small bowel epithelium is regenerated every 3 to 5 days.13 This process is fueled by ISCs that divide into daughter or progenitor cells, which subsequently differentiate into mature differentiated cells.14 Aberrant activation of the mucosal immune system and proinflammatory cytokine overproduction promote epithelial cell apoptosis.15 Therefore, CD is attributed to mucosal ulceration, followed by epithelial regeneration.5 In healthy intestines, the self-renewal and differentiation of ISCs are regulated along the crypt-villus axis spatially. In contrast, the ileal epithelium in CD shows the alteration of crypt-villus architecture, which mainly occurs in intestinal crypts rather than in villi.16 Intestinal crypts contain ISCs intermingled with Paneth cells; progenitor cells are also a major component of crypts. This phenomenon may suggest the impairment of replication and the differentiation of crypt cells as a contributing mechanism of the development and perpetuation of CD.
In this study, we evaluated whether compared with control AKR mice, SAMP1 mice, which develop segmental ileitis resembling human CD, exhibit alterations in the crypt-villus architecture, differentiation of epithelial cells, mini-gut intestinal organoid formation, and ISC niche signaling pathways. The crypt-villus structure and the pattern of epithelial differentiation of SAMP1 mice aged 10 to 14 weeks were altered only in the ileum and not in the proximal small bowel or colon. There was no difference in the number and gene expression of ISCs between SAMP1 and AKR mice. However, cellular differentiation was altered in the ileum of SAMP1 mice. The ileal crypts comprised a decreased number of Paneth cells and an increased number of secretory progenitor cells and exhibited proliferative zone expansion, elongation, and branching. In addition, a concomitant reduction in the villus surface area covered by absorptive enterocytes was observed. These data suggest that the transition from ISCs to progenitor and differentiated cells was regulated by a distinct mechanism in the ileal crypts of SAMP1 mice. Therefore, we concluded that the depletion of niche factors, not the dysfunction of ISCs themselves, induced the abnormal differentiation of ISCs in SAMP1 mice.
To functionally assess the intestinal crypts of SAMP1 mice, an ex vivo organoid reconstitution assay was performed. Intestinal crypt–derived organoids highly resemble the crypt-villus epithelial architecture of the small bowel17; therefore, intestinal organoids established from whole crypts facilitate the study of self-renewal and differentiation of ISC niche factors independent of the contribution of their individual components—the Lgr5-positive ISC and Paneth cells. In this study, the organoid reconstitution assay revealed that the regenerative capacity of the ileal crypts of SAMP1 mice was significantly lower than that of control mice.
The self-renewal, maintenance, proliferation, and differentiation of intestinal epithelial cells are preserved by ISCs and progenitor cells residing in intestinal crypts, which are under the control of ISC niche signaling pathways (mainly Notch and Wnt/β-catenin signaling).18 Cellular proliferation was promoted in the ileal crypts of SAMP1 mice, whereas differentiation was impaired. The number of secretory progenitors was increased and the number of reciprocally absorptive progenitors was decreased, resulting in a reduced epithelial absorptive surface. The RNA-seq enables a hypothesis-free approach to gene expression in intestinal crypts, allowing the identification of unique features of signaling pathways regulating the replication and differentiation of ISCs and progenitor cells.19 Bulk RNA-seq of intestinal mucosa incorporates various cell types, including immune cells, fibroblasts, and myofibroblasts. We isolated intestinal crypts and performed RNA-seq to analyze the gene expression patterns of crypt cells. Paneth cells presenting ISC niche factors were decreased in SAMP1 crypts compared with AKR crypts. Previous studies have suggested that Dll1, Dll4, and Wnt3 are expressed in Paneth cells.20,21 Dll4 and Dll1 act as ligands for Notch receptors, and the depletion of Dll1/4 may induce the impairment of Notch signaling, which regulates the proliferation and differentiation of ISCs.
In addition, the depletion of Wnt3, the canonical Wnt ligand, can affect ISC survival and proliferation. Wnt9A, formerly known as Wnt14, has been classified as both a canonical22 and a noncanonical ligand.22 In spite of the failure to reach statistical significance (q value < 0.10), Wnt9A was increased in the SAMP1 ileum and jejunum (P = 0.13). Although the function of Wnt9A has not been clear, Wnt9A has suppressed human colorectal cancer cell proliferation through a noncanonical Wnt pathway.23 The Dll1, Wnt3, and Wnt9A expression in this study also showed significant differences between the SAMP1 and AKR mice (Mann-Whitney P value); however, given our small sample size, statistical significance could not be reached in the false discovery rate statistics.
Research has shown that Dll4 is expressed in the crypt cells, including Paneth cells.24 Therefore, the decreased number of Paneth cells may have contributed to the decreased expression of Dll4. Notch receptor activation by Dll1, Dll2, and Dll4 and Jagged 1 and 2 expressed by neighboring cells leads in general to the cleavage of the Notch intracellular domain of the receptor. The Notch intracellular domain then translocates into the nucleus to activate downstream gene transcription via the Hes and Hey families of transcription factors. In the ileum of SAMP1 mice, the decreased expression of Dll4 induced alterations in Notch signaling, resulting in altered epithelial cell differentiation, impaired epithelial renewal, and ileitis development. When both Wnt and Notch signaling are activated, ISCs are preserved in a self-renewal state; however, when Notch signaling is inhibited, Wnt signaling promotes ISC differentiation toward a secretory progenitor lineage.25
This study has some limitations. First, the inflammatory status in the intestinal tissue of SAMP1 mice was not considered, although immune cells and cytokines in the lamina propria contribute to intestinal epithelial differentiation, regeneration, and wound repair.26,27 One previous study clearly showed that SAMP1 ileitis results from dysregulated mucosal immunity secondary to mucosal barrier dysfunction, not from a primary immunological defect.7 Therefore, we focused on intestinal epithelial differentiation and regeneration regardless of inflammatory status. Second, SAMP1 organoids unable to maintain. This finding may have been expected because the organoid formation ability of SAMP1 mice was decreased because of decreased Paneth cell differentiation and endogenous niche factor depletion that subsequently altered epithelial differentiation. If possible, long-term subcultured SAMP1 organoids may be helpful to display the characteristics of ISCs under inflammation-free conditions. Third, the crypt preparation method may have affected the gene expression pattern. However, the same preparation method was used in both AKR and SAMP1 mice and resulted in a similar effect in both groups. In addition, the crypt isolated using EDTA and DTT contained not only crypt cells but also subepithelial myofibroblast and nonmyofibroblast mesenchymal cells. However, the expression level of the subepithelial myofibroblast markers Acta2 and Myh11 and the nonmyofibroblast mesenchymal cells Foxl1 and CD34 was negligible and showed no significant difference between AKR and SAMP1 mice. Finally, microbial colonization and its metabolites contributed to the differentiation of the intestinal epithelium. This study did not exclude the effects of the intestinal microbiome and its metabolites.
Conclusions
The results indicate that characteristic ileitis in SAMP1 mice may be caused by a reduced number of Paneth cells; decreased expression of niche factors, especially Dll4-Notch signaling; altered epithelial differentiation, and impaired epithelial regeneration to restore epithelial barrier function, resulting in chronic inflammation of the mucosal layer (Supplementary Fig. 8). Future studies are required to ascertain how to restore the alteration of epithelial differentiation in SAMP1 mice, which will provide researchers a novel therapeutic molecule in CD therapeutics.
Supported by: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP) (NRF-2018R1D1A1B07048376, NRF-2019R1I1A1A01062205, and NRF-2019R1A2C2010404).
References
