Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Mar 4.
Published in final edited form as: Cell Stem Cell. 2020 Dec 4;28(3):568–580.e4. doi: 10.1016/j.stem.2020.11.008

Mapping development of the human intestinal niche at single cell resolution

Emily M Holloway 1,*, Michael Czerwinski 2,*, Yu-Hwai Tsai 2,*, Joshua H Wu 2, Angeline Wu 2, Charlie J Childs 1, Katherine D Walton 1, Caden W Sweet 2, Qianhui Yu 3, Ian Glass 4, Barbara Treutlein 5, J Gray Camp 3,6, Jason R Spence 1,2,7,#
PMCID: PMC7935765  NIHMSID: NIHMS1648548  PMID: 33278341

SUMMARY

The human intestinal stem cell niche supports self-renewal and epithelial function, yet little is known about its development. We used single-cell mRNA sequencing with in situ validation approaches to interrogate human intestinal development from 7–21 weeks post conception, assigning molecular identities and spatial locations to cells and factors that comprise the niche. Smooth muscle cells of the muscularis mucosa, proximal to proliferative crypts, are a source of WNT and RSPONDIN ligands, while EGF is expressed far from crypts in the villus epithelium. Instead, an PDGFRAHI/F3HI/DLL1HI mesenchymal population lines the crypt-villus axis and is the source of the EGF family member NEUREGULIN1 (NRG1). In fetal intestinal enteroid cultures, NRG1, but not EGF, permitted increased cellular diversity via differentiation of secretory lineages. This work highlights complexities of intestinal EGF/ERBB signaling and delineates key niche cells and signals of the developing intestine.

eTOC Blurb

Spence and colleagues used scRNA-seq to characterize the cellular diversity of the developing human intestinal stem cell niche. Transcriptional and spatial profiling demonstrated that Neuregulin 1 (NRG1) is expressed by PGDFRAHI/F3+/DLL1HI subepithelial mesenchyme, and NRG1, but not EGF, permitted secretory lineage differentiation in enteroid culture.

Graphical Abstract

graphic file with name nihms-1648548-f0005.jpg

INTRODUCTION

The stem cell niche within a tissue is required to regulate stem cell maintenance, self-renewal and differentiation (Scadden, 2006). The niche is made up of both physical and chemical cues, including the extracellular matrix (ECM), cell-cell contacts, growth factors and other small molecules such as metabolites (Capeling et al., 2019; Cruz-Acuña et al., 2017; Gjorevski et al., 2016). Understanding the niche within various tissues has been central to understanding how tissues maintain homeostasis, and for understanding how disease may occur (Van de Wetering et al., 2002). Further, establishing proper in vitro niche conditions has allowed the growth and expansion of gastrointestinal tissue-derived stem cells in culture (Dedhia et al., 2016; Kretzschmar and Clevers, 2016). For example, through understanding that WNT signaling is important for maintaining intestinal stem cell (ISC) homeostasis (Muncan et al., 2006; Pinto et al., 2003; Sansom et al., 2004), blockade of BMP signaling by NOGGIN (NOG) promotes ectopic crypt formation (Haramis et al., 2004), and that EGF is a potent stimulator of proliferation (Goodlad et al., 1987; Ulshen et al., 1986), it was determined that WNTs, RSPONDINs (RSPOs), NOG and EGF can be utilized to expand and maintain ISCs in culture as 3-dimensional intestinal organoids (Ootani et al., 2009; Sato et al., 2009, 2011). This same information has been leveraged to expand and culture human pluripotent stem cell derived intestinal organoids in vitro (Capeling et al., 2020; Finkbeiner et al., 2015; Spence et al., 2010; Wells and Spence, 2014).

Despite significant progress over the past decade, it is also clear that current in vitro systems are still not optimized to most accurately reflect the in vivo environment. Ongoing efforts are aimed at improving the in vitro physical environment through biomimetic ECM (Capeling et al., 2019; Cruz-Acuña et al., 2017; Gjorevski et al., 2016), and by adjusting signaling cues to more accurately reflect the in vivo niche (Fujii et al., 2018). More recently, single cell technologies have started to reveal unprecedented amounts of information about the cellular heterogeneity of human intestinal tissue and the ISC niche during health and disease (Kinchen et al., 2018; Martin et al., 2019; Smillie et al., 2019), and will undoubtedly yield substantial information about cell types and niche cues that regulate ISCs in various contexts.

Here, we set out to better understand the cellular diversity and niche cues of the developing human intestine using single cell mRNA sequencing (scRNA-seq) to describe the transcriptional signatures and using fluorescent in situ hybridization (FISH) and immunofluorescence (IF) to define the location of cells and factors that make up the ISC niche. We sought to interrogate the cellular source of known stem cell niche factors and to identify new niche factors. We determined that a source of WNT and RSPO ligands that reside just below the proliferative crypt domains are ACTA2+/TAGLN+ smooth muscle cells of the muscularis mucosae. We further determined that EGF is not expressed in the mesenchyme, but is most abundant in the enterocytes of the villus epithelium, several cell diameters away from the proliferative region of the crypt. We identified a population of subepithelial cells (SECs) that lines the entire villus-crypt axis, marked by a F3+/PDGFRAHI/DLL1HI expression profile, and found that these cells express the EGF-family ligand NEUREGULIN1 (NRG1), which has recently been implicated as an important regenerative cue in the murine intestine (Jardé et al., 2020). Given that NRG1 is expressed in mesenchymal cells adjacent to the proliferative crypts, we tested the effect of both EGF and NRG1 on ISCs using fetal human duodenal enteroids. We observed that EGF potently stimulated proliferation, while NRG1 supported both enteroid and growth enhanced cell-type diversity. Together, these results suggest that NRG1 acts as a niche cue that when used in place of EGF, supports homeostasis in vitro, allowing both stem cell maintenance and differentiation human intestinal enteroids

RESULTS

Interrogating the developing human small intestine at single cell resolution.

Given that little is known about mesenchymal cell heterogeneity within the fetal human intestine, we aimed to better understand the mesenchymal cell populations that make up the developing human ISC niche. To do this, we obtained samples of human fetal duodenum, or duodenum with adjacent proximal small intestine (in the case of very young tissues) starting just after the onset of villus morphogenesis (47 days post conception; 47d) with samples interspersed up to the midpoint (132d) of typical full-term (280d) and performed histological and molecular analysis (Figure 1A–B). Major physical changes occur throughout this developmental window with rapid growth in length and girth, along with the formation of villi and crypt domains within the epithelium and increased organization and differentiation of smooth muscle layers within the mesenchyme (Chin et al., 2017) (Figure 1A).

Figure 1. Mesenchymal heterogeneity in the developing human duodenum.

Figure 1.

Open in a new tab

(A) H&E staining on human fetal intestine sections at a constant scale spanning from 54d to 130d (days post conception) (B) Timeline of specimens and corresponding number of cells profiled by scRNA-seq after filtering and ‘cleaning’ of ambient/background RNA (see STAR Methods) (n=8 biological specimens, n=1 technical replicate 47d, 59d, 72d, 80d, 101d, and 132d, n=2 technical replicates 122d and 127d). (C) UMAP visualization of each sample analyzed by scRNA-seq displayed by the age post conception. Clusters identities were assigned based on expression of canonical lineage markers (see also Figure S1). Mesenchymal (green), epithelial (blue), neuronal (yellow), immune (purple), and endothelial (pink) cell clusters were identified in all ages sequenced. (D) Following application of Harmony to mesenchymal cells from all time points, a force directed layout illustrates the relationship between timepoints. Cells are colored by sample identity (days post conception). (E) Feature plots of individual genes for various lineages are shown, including PDGFRA, F3, DLL1, NPY, GPX3, TAGLN, ANO1, and RGS5 plotted onto the force-directed layout presented in Figure 1C. (F) Representative images from FISH staining for F3 (pink) and immunofluorescent protein staining for SM22 (TAGLN protein product) with DAPI (grey) on the developing human intestine (n=1 biological replicate per timepoint) Scalebars represent 200μm. (G) Spatial characterization of PDGFRAHI/DLL1HI/F3HI and GPX3HI mesenchymal cells using FISH in the developing human intestine. Multiplexed FISH/IF for F3, PDGFRa, DLL1, NPY, GPX3 (pink or green) counterstained with DAPI (grey) and in one case ECAD (blue). Representative data is shown for n=1 biological replicate aged 132d for DLL1/F3, NPY/F3, and NPY/GPX3 and n=2 biological replicates for PDGFRa/F3, 120d specimen is shown. Scalebars represent 25 μm.

In order to capture the full complement of cell types that contribute to the developing human intestine, we dissociated full thickness intestinal tissue from 8 specimens ranging between 47 and132 days post conception and used scRNA-seq to sequence 2,830 – 3,197 cells per specimen after filtering and ambient RNA removal (see Methods). 24,783 total cells were used in the analysis after passing computational quality filtering (Figure 1B) Following dimensional reduction and visualization with UMAP (Becht et al., 2019; Wolf et al., 2018), we used canonical genes to annotate each sample individually by identifying major cell classes including epithelial, mesenchymal, endothelial, enteric nervous and immune cells (Figure 1C, Figure S1A). In order to focus our analysis on the mesenchymal niche populations found in each sample, we computationally extracted and re-clustered the mesenchyme (between 1,462–2,054 cells per specimen) and annotated a population of PDGFRAHI cells, which have also been described in mice (McCarthy et al., 2020), and ACTA2+/TAGLN+ smooth muscle cells (Figure 1E, Figure S1C). We also identified additional sub-clusters and gene expression patterns not previously described in mice, which we describe in greater detail below (Figure S1C). Given the dramatic morphological changes that take place across this development time (Figure 1A), we implemented Harmony (Nowotschin et al., 2019), an algorithm that allows interrogation of scRNA-seq data across discrete time points (Figure 1D–E, S1D).

Force directed layouts following Harmony implementation ordered cells broadly according to their developmental age (days post conception) (Figure 1D). The 47d cells were largely separate from other time points with the exception of an ACTA2HI/TAGLNHI/RGS5+ population of vascular smooth muscle cells (VSMC)(Muhl et al., 2020) (Figure 1D–E). Cells were then ordered according to developmental time, with cells from samples older than 101d (101d, 122d, 127d, 132d) clustering together. In addition to the VSMC population, this analysis suggested the emergence of several mesenchymal populations, including a PDGFRAHI/F3HI/DLL1HI population, a GPX3HI population, a TAGLNHI/RGS5- smooth muscle population and a prominent clusters of cells defined as fibroblasts based on their expression COLLAGEN genes (COL1A1, COL1A2) and DECORIN (DCN) (Figure 1E, Figure S1D) (Kinchen et al., 2018). F3 was recently shown to be expressed in a population of mesenchymal cells that is adjacent to the human colonic epithelium (Kinchen et al., 2018) and we identified that these cells were additionally characterized by their enrichment of NPY, DLL1, FRZB and SOX6 (Figure 1E, Figure S1B–D).

Mesenchymal cell lineages emerge across developmental time.

Force directed layouts following Harmony implementation suggested that different mesenchymal cell populations emerge across developmental time. For example, F3HI/PDGFRAHI cells emerged after approximately 59 days, while GPX3HI/F3LO/PDGFRALO and TAGLNHI/RGS5- smooth muscle cells emerge after approximately 80 days. To corroborate scRNA-seq analysis, we used a combinatorial staining approach, utilizing multiplexed fluorescent in situ hybridization (FISH) and immunofluorescence (IF) to examine F3 mRNA and SM22, the protein product of the TAGLN gene (Figure 1F). We found that F3 was not expressed at 59d but was clearly expressed in the villus mesenchyme by 78d, with expression becoming more restricted to the subepithelial cells as time progressed. Interestingly, we observed that F3 was expressed in the scRNA-seq at 59 days, however, it is possible that since fetal tissue staging is an approximation, samples may be slightly older or younger than their actual labeling, explaining slight discrepancies such as this. SM22 was expressed in the 59d intestine, but only in the outermost muscularis externa layer. SM22 expression in the muscularis mucosa, the layer closest to the intestinal epithelium and adjacent to the proliferative crypt domains, was first observed as poorly organized cells near the epithelium at 100d that became more organized after this time point. Single cell analysis and FISH/IF collectively suggest that the mesenchyme in the early fetal intestine is naïve and that mesenchymal cell emergence coincides with the formation of proliferative intervillus/crypt domains.

To understand how mesenchymal cell populations are spatially organized within the tissue after 100 days, we used FISH/IF and confirmed that F3HI cells co-express PDGFRA, DLL1 and NPY (Figure 1G). Interestingly, NPY marked a subset of F3HI subepithelial cells lining the villus, but was absent from villus subepithelial cells (Figure 1G; Figure S2D). GPX3HI/F3LO/PDGFRALO cells were most abundant within the core of intestinal villi and are observed sitting adjacent to NPYHI cells (Figure 1E).

Identifying putative human ISC niche factors in the developing gut.

It has been demonstrated that several niche factors allow adult and developing human and murine intestinal epithelium to be cultured ex vivo as organoids (Fordham et al., 2013; Fujii et al., 2018; Kraiczy et al., 2017; Sato et al., 2009, 2011; Tsai et al., 2018). These factors often include WNT and RSPO ligands, BMP/TGFβ antagonists and EGF, and are based on defined growth conditions that allow expansion of intestinal epithelium in vitro (Sato et al., 2009, 2011). Efforts have been made to determine more physiological niche factors for in vitro culture systems based on observed in vivo niche cues (Fujii et al., 2018), however niche factors have not been interrogated in the developing human gut using high resolution technologies such as scRNA-seq. To identify putative niche factors, we first determined which cells within the human fetal intestine expressed known niche factors. We observed that F3HI/PDGFRAHI subepithelial cells, and GPX3HI/F3LO/PDGFRALO villus core cells lacked robust expression of most known niche factors (Figure 2A–B), whereas the WNT pathway members with the highest expression were RSPO2, RSPO3 and WNT2B, which are expressed in the TAGLNHI/RGS5- smooth muscle cells and COL1A1HI fibroblasts, but are not expressed by the F3HI/PDGFRAHI subepithelial cells (Figure 2A–B). EGF is a critical driver of proliferation in murine enteroid culture (Basak et al., 2017); however, EGF expression was not observed in the mesenchyme, whereas the EGF family member NRG1 was abundant in the F3HI/PDGFRAHI cell population (Figure 2A–B). Of note, NRG1 was a robustly expressed EGF family member in the F3HI/PDGFRAHI cell population (Figure S3A). IF for SM22, combined with FISH for RSPO2, RSPO3, WNT2B, EGF and NRG1 revealed expression patterns that were consistent with scRNA-seq data (Figure 2C, Figure S2F).

Figure 2. Interrogating stem cell niche factors in the developing human intestine.

Figure 2.

Open in a new tab

(A) Summary schematic annotating the approximate expression domains of several mesenchymal subpopulations on the force directed layout as identified in Figure 1E and Figure S1C. (B) Feature plots of several individual ISC niche factors including EGF, NRG1, WNT2B, RSPO2, and RSPO3 in mesenchymal cells at all time points profiled. (C) Multiplexed FISH for niche factors EGF (green), NRG1 (green), WNT2B (pink), RSPO2 (pink), and RSPO3 (pink) coupled with immunofluorescent protein staining of SM22 (blue), DAPI (grey), and FISH for F3 (green) in developing human fetal crypts. Representative data are shown from n=2 biological replicates, 140d specimen shown for EGF and NRG1, while 132d specimen shown for RSPO2, RPSO3, and WNT2B. Lower magnification images for all panels are presented in Figure S2. (D) Following application of Harmony to epithelial cells from all time points, force directed layout illustrates the relationship among timepoints. Cells are colored by sample identity (days post conception). (E) Feature plots of EGFR, ERBB2, ERBB3, and ERBB4 plotted onto the force directed layout presented in Figure 2D. (F) FISH staining in developing human fetal crypts for EGFR (pink), ERBB2 (green), and ERBB3 (red) coupled with immunofluorescent staining for ECAD (blue) and DAPI (grey) Representative data are shown from n=1 132d biological specimen. Lower magnification images for all panels are presented in Figure S2. (G) Feature plots of EGF and the enterocyte marker FABP2 plotted onto the force-directed layout presented in Figure 2D. (H) Representative images of multiplexed FISH for EGF (pink) and NRG1 (green), coupled with immunofluorescent protein staining of MKI67 (blue) and DAPI (grey) (n=1 132d human fetal intestine). The white dotted line in images roughly defines the epithelial-mesenchymal boundary. Scalebars represent 25 μm.

Given the importance of EGF for in vitro enteroid culture, we further interrogated whether EGF and EGF receptors are expressed in the developing intestinal epithelium via scRNA-seq and FISH. All epithelial cells were extracted and re-clustered, and the data was visualized using a force directed layout following application of Harmony (Figure 2D). The ERBB receptors, including EGFR, ERBB2 and ERBB3 were broadly expressed throughout the epithelium, a finding that was confirmed by FISH, while ERBB4 was not expressed (Figure 2E–F). While EGF is not expressed in the intestinal mesenchyme (Figure 2A–C), we observed that EGF is expressed in a small subset of differentiated epithelial FABP2HI enterocytes (Figure 2G), a finding that was supported using co-FISH/IF and showed EGF expression is low/absent from the proliferative crypt domain, but is expressed several cell diameters above the MKI67+ crypt region and throughout the villus epithelium (Figure 2D). On the other hand NRG1 is expressed in F3HI /PDGFRAHI subepithelial cells adjacent to the crypt (Figure 2C,H; Figure S2B).

NRG1 does not support proliferation and growth in established enteroid cultures.

Based on the expression pattern of NRG1, we hypothesized that it may act as an ERBB niche signaling cue and may be physiologically relevant in vitro based on its localization and proximity to ISCs within the developing intestine in vivo. To interrogate the effects of NRG1 and EGF on the intestinal epithelium, we split established human fetal duodenum derived epithelium-only intestinal enteroids (established from a 142d specimen) in culture using standard growth conditions (WNT3A/RSPO3/NOG plus EGF, see Methods) into two groups. One group of enteroids was cultured in standard media with EGF (100ng/mL), the other was grown without EGF and was instead supplemented with NRG1 (100ng/mL) (Figure 3A). Following growth for 5 days in EGF or NRG1, enteroids did not appear phenotypically different (Figure 3B). Upon interrogation using immunofluorescence and FISH, we observed that EGF-grown cultures had both OLFM4+ and OLFM4- enteroids and that enteroids in this group were highly proliferative based on MKI67 staining (Figure 3C). NRG1 treated enteroids appeared to have more uniform OLFM4 expression but also had fewer KI67+ cells per field of view. To more closely interrogate these differences, each group was subjected to scRNA-seq to investigate transcriptional differences. To reduce any chances of batch effect, all processing for single cell sequencing for these groups was carried out at the same time in parallel, libraries were prepared in parallel, and samples were sequenced on the same lane (see Methods for detail). Despite varying only EGF or NRG1 in the culture, we observed a difference in gene expression between the two groups as visualized in UMAP plots illustrated by near complete independent clustering of cells by culture media composition (Figure 3D–E). The exception to this was cluster 4, which expressed proliferation markers (MKI67, TOP2A), and had a contribution from both samples (Figure 3E–F). Cluster 4 appeared to have a higher number of cells from the EGF grown enteroids, and proportionally ~4% (115/2,789) of cells from the EGF treated enteroids were in this cluster whereas <0.5% (10/2,720) of cells from the NRG1 treated enteroids were in this cluster (Figure 3G). These data support the MKI67 immunofluorescence staining suggesting that NRG1 had reduced proliferation. Enteroids from both groups broadly expressed OLFM4, although it appeared that expression levels were slightly higher when grown in NRG1, suggesting that these samples generally had little heterogeneity and were largely comprised of undifferentiated stem cells (Figure 3H). Consistent with this notion, differentially expressed genes associated with each cluster did not include genes canonically associated with differentiated cell types (Table S1). We did note that a subset of cells in the NRG1 clusters expressed enterocyte marker FAPB1 (Guilmeau et al., 2007) and a subset expressed secretory progenitor marker SPDEF (Gregorieff et al., 2009; Noah et al., 2010) (Figure 3H). These data suggest that some cells in the NRG1 cultures may be in the early stages of differentiation; however, expression patterns were not sufficiently different to form distinct clusters.

Figure 3. NRG1 does not support proliferation and growth of established enteroids lines.

Figure 3.

Open in a new tab

(A) Experimental schematic for data presented in 3A. (B) Representative stereomicroscope images after 5 days of growth in the presence of EGF (100ng/ml) or NRG1 (100ng/ml). Scalebars represent 500μm. Representative data shown from n = 2 biological replicates, 142d fetal sample shown(C) Representative images of FISH staining for OLFM4 (pink) or immunofluorescent protein staining for MKI67 (pink) coupled with ECAD (blue) and DAPI (grey) in enteroids grown in EGF (100ng/ml) or NRG1 (100ng/ml). Representative data shown from n = 2 biological replicates, 142d fetal sample shown. Scalebars represent 100μm. (D) UMAP embedding of enteroid scRNAseq data (5,509 cells total) demonstrating the 5 precited clusters (n = 1 biological sample sequenced, 142d fetal sample). (E) UMAP embedding of enteroid scRNAseq data colored by culture condition (EGF- 2,789 cells; NRG1– 2,720 cells). (F) Feature plot of MKI67 illustrating that most proliferating cells are within cluster 4. (F) Bar chart depicting the percentage of cells in cluster 4 from each treatment group. (H) Feature plots demonstrating the expression of the stem cell marker OLFM4, secretory progenitor marker SPDEF, and enterocyte marker FABP1. (I) Experimental schematic for enteroid forming assays (left). Stereoscope images of enteroids after single-cell passaging and 10 day growth without EGF-family ligands (control) or in the presence of EGF (100ng/ml), NRG1 (100ng/ml), or both EGF and NRG1 (100ng/ml each). (representative data shown from n = 2 biological replicates, 142d fetal sample shown) Scalebars represent 500μm.

To functionally evaluate the observation that proliferation was reduced in NRG1 treated enteroids, we bulk passaged enteroids and then allowed them to expand for 3 days in standard (EGF 100ng/mL) growth conditions. We then removed EGF for 24 hours, dissociated enteroids into a single-cell suspension and plated 5,000 cells per droplet of Matrigel. Immediately upon seeding single cells, we added standard growth media supplemented with no-EGF (control), with EGF (100ng/mL) only, with NRG1 (100ng/mL) only, or with NRG1 (100ng/mL) and EGF (100ng/mL) (Figure 3I). As expected, we observed robust re-establishment of enteroids after 10 days in the standard EGF condition. In contrast, we observed almost no enteroid recovery in the control and in the NRG1-only supplemented cultures, whereas this growth defect was rescued in the NRG1 plus EGF condition (Figure 3I). These functional results further suggest that EGF enhances proliferation relative to NRG1 in established enteroid cultures.

Long-term enteroid growth in NRG1 is associated with increased epithelial diversity in vitro.

The previous experiment was conducted with enteroids that had been established and expanded in long-term culture with EGF, and the experimental data (Figure 3) suggested that these cultures were highly dependent on EGF for proliferation. To determine the effects of different EGF-family members on establishment and long-term growth of enteroids, we cultured freshly isolated intestinal crypts in standard growth conditions (WNT3A/RSPO3/NOG, see Methods) supplemented with no EGF/NRG1 (control), EGF (100ng/mL), NRG1 (100ng/mL) or a combination of EGF and NRG1 (100ng/mL each) (Figure 4A). We used these cultures to carry out long-term passaging, imaging, quantitative enteroid forming assays and scRNA-seq (Figure 4A and S4). Enteroids were successfully established in all conditions (Figure 4B). All conditions successfully underwent serial passaging, with the exception of the controls (no EGF/NRG1), which failed to expand beyond initial plating (Passage 0; P0). To determine the effects of different growth conditions on enteroid forming ability, we performed a quantitative single cell passaging assay on surviving cultures at P2 (Figure S4B–C). To do this, we dissociated the three treatment groups into single cells and plated 1,000 single cells (Figure S4B) per droplet of Matrigel, allowed cultures to grow for 11 days, and quantified the number of recovered enteroids (Figure S4C). All groups re-established enteroids, albeit at a low efficiency, and there was no difference in enteroid forming efficiency between the EGF and NRG1 groups.

Figure 4. Establishment of new enteroid lines in NRG1 increases cell type diversity in vitro.

Figure 4.

Open in a new tab

(A) Experimental schematic. Enteroids were established from 132d (B-C) or 105d (D-J) human fetal specimen in the presence of EGF (100ng/ml), NRG1 (100ng/ml), both EGF (100ng/ml) and NRG1 (100ng/ml), or without any EGF or NRG1. Representative data shown from n = 4 biological replicates in total – 74d, 101d, 105d, 132d. (B) Representative stereoscope images of enteroids in each condition 8 days after placing isolated epithelium in Matrigel with growth factors. Scalebars represent 1mm (C) Representative images of FISH for OLFM4 (pink) or immunofluorescent protein staining for MKI67 (pink), MUC2 (pink), LYZ (pink) and ECAD (blue) and DAPI (grey) in enteroids after P0, 11 days of growth in the presence of EGF, NRG1, or both EGF and NRG1. Scalebars represent: 100μm. (D-J) scRNA-seq was performed on enteroids from EGF, NRG1, and dual EGF/NRG1 conditions after P1, 11 days of in vitro growth (n = 1 biological sample sequenced, 105d fetal sample). (D) UMAP embeddings of 13,205 enteroid cells colored by cluster identity. (E) UMAP embeddings of enteroid cells colored by sample identity (EGF- 3,262 cells, NRG1– 7,350 cells, and EGF/NRG1– 2,593 cells). (F) Bar charts depicting the cell type abundance (% of cells total sequenced) for each condition. Colors in graph correspond to Figure 4D. (G) Dotplots for the proliferation markers MKI67 and TOP2A. Both markers were enriched in clusters 4 and 8. (H) Bar chart depicting the proportion of cells sequenced that map to proliferative clusters (cluster 4 or 8) in each condition. (I) Feature plots for intestinal epithelial lineages include ISCs (OLFM4, clusters 0, 1, 2, 3, 4, 8), enterocytes (FABP2, SI - cluster 6), enteroendocrine cells (CHGA - cluster 12), and goblet cells (MUC2 - cluster 11). LYZ was broadly expressed across all enteroids conditions. (J) Bar chart depicting the proportion of cells sequenced for each condition are present in cluster 6 (enterocytes), cluster 12 (enteroendocrine cells), and cluster 11 (goblet cells). (K) Epithelial cells (828 cells) from primary intestine specimens (n= 4; 101d, 122d, 127d, 132d) were computationally extracted, re-clustered, and visualized using UMAP. (L) Cluster identities were assigned based on expression of canonical lineage markers (see Table S3 for differentially expressed genes in each cluster). (M) The Ingest function was used to map enteroids derived in the presence of EGF (100ng/ml), NRG1 (100ng/ml), or both EGF and NRG1 (100ng/ml) onto to primary intestinal epithelium reference presented in 4K. (N) The abundance of cells mapping to each of the 9 clusters identified in the in vivo intestinal epithelium was determined for the primary intestinal epithelium and for enteroids in each treatment group. Colors in graph correspond to Figure 4K.

Examining all three groups that remained after passaging (EGF, NRG1, EGF/NRG1) by FISH or immunofluorescence revealed that OLFM4 was expressed in all conditions, and MKI67 did not appear different per field of view (Figure 4C). The NRG1-only group appeared to have more MUC2 staining within the enteroid lumen, whereas both groups that included EGF (EGF-only, NRG1/EGF) had widespread LYZ expression within the epithelial cells (Figure 4C). Given the different IF staining patterns of MUC2 and LYZ observed when comparing treatment groups (Figure 4E), we interrogated the cellular makeup and molecular signatures of these enteroids using scRNA-seq. For each group we sequenced 3,262 cells grown in EGF (100ng/mL), 7,350 cells grown in NRG1 (100ng/ml) and 2,593 cells grown in both NRG1 and EGF (100ng/mL each). UMAP dimensional reduction showed that the NRG1 treated enteroids clustered distinctly from the EGF-only enteroids, and suggested that the NRG1/EGF enteroids shared a high degree of molecular similarity with EGF-only enteroids since these samples overlapped in the clustering (Figure 4E–F). Examining the cluster distribution for each sample, it was evident that EGF and EGF/NRG1 enteroids both contributed to the same clusters (clusters 1, 2, 8, 7), while NRG1 contributed to many distinct clusters (0, 3, 4, 11, 12) (Figure 4F). Upon interrogation of genes associated with various clusters (Table S2), proliferation genes were associated with two clusters - Cluster 4 (NRG1) and Cluster 8 (EGF and EGF/NRG1) (Figure 4G–H). Several clusters expressed the stem cell marker OLFM4 (EGF, EGF/NRG1 – clusters 1, 2, 8; NRG1 – clusters 0, 3, 4). Cluster 6 had a contribution from all 3 groups and expressed enterocyte genes (SI, FAPB2). Unique to the NRG1 grown enteroids, cluster 11 expressed genes associated with secretory and goblet cells (MUC2), and cluster 12 expressed genes associated with enteroendocrine cells (CHGA) (Figure 4I). We also interrogated LYZ expression given our immunofluorescence staining results. LYZ was expressed at higher levels in the EGF and NRG1/EGF enteroids, as suggested by IF; however, low level expression was also observed broadly in the NRG1 treatment group (Figure 4I). LYZ is canonically associated with Paneth cells; however, the fetal intestine does not possess Paneth cells until after 21 weeks post conception (Elmentaite et al., 2020; Finkbeiner et al., 2015). Given that the enteroids used here were generated from specimens earlier than 21 weeks (replicate experiments utilized 105d, 135d specimens), it is unlikely that LYZ expression is associated with Paneth cells. Taken together, these data shows that both EGF and NRG1 can promote long term survival of freshly established enteroids, but that they have a differing impact on gene expression and cellular diversity.

Mapping to an intestinal reference reveals cellular heterogeneity in enteroids

In order to further interrogate cellular heterogeneity in enteroids grown in EGF, NRG1 and EGF/NRG1, we used the human fetal epithelium as a high-dimensional search space to determine the potential correspondence enteroids and their in vivo counterparts. To do this, we implemented the Ingest function (Wolf et al., 2018), which uses an annotated reference dataset that captures the biological variability of interest, and projects new data onto the reference. We defined the major epithelial cell populations in the human fetal intestine using the four samples that were older than 100 days (101–132d). These samples were chosen based on the force directed layouts following Harmony augmentation, which suggested that major changes in development/differentiation were not taking place across these times (Figure 1–2). We defined 9 epithelial cell types based on published data (Table S3), including intestinal stem cells (ISCs, cluster 2 - LGR5, OLFM4), enterocytes (cluster 0 and 3 - FABP2, ALPI, RBP2)(Haber et al., 2017), BEST4+ enterocytes (cluster 5 - BEST4, SPIB)(Elmentaite et al., 2020; Parikh et al.), goblet cells and goblet cell precursors (cluster 4 - MUC2, SPDEF, DLL1)(Okamoto et al., 2008), tuft cells (cluster 7 – TRPM5, TAS1R3, SPIB)(Van Es et al., 2019; Howitt et al., 2020; Kaske et al., 2007), enteroendocrine cells (EECs, clusters 1 and 8 – CHGA, NEUROD1, PAX6, ARX, REG4)(Beucher et al., 2012; Du et al., 2012; Gehart et al., 2019; Haber et al., 2017) (Figure 4K–L). We then used Ingest to map enteroids grown in EGF, EGF/NRG1 or NRG1 onto the in vivo epithelium (Figure 4M) and determined the proportion of cells that mapped to each in vivo cell type and compared this with the distribution of cells seen in the primary intestine (Figure 4N). EGF and EGF/NRG1 samples shared similar distribution patterns, with the majority of cells from both conditions mapping to ISCs and enterocytes, with a minor population mapping to EECs (Figure 4M–N). NRG1 treated enteroids mapped to all cell types, including goblet cells (cluster 4), CHGA+ EECs (cluster 8), tuft cells (cluster 7), BEST4+ enterocytes (cluster 5), which were not present in EGF or EGF/NRG1 grown enteroids in this analysis (Figure 4M–N). These results further support that NRG1 grown enteroids have enhanced cellular differentiation relative to enteroids grown in EGF.

DISCUSSION

Subepithelial niche cells in the developing human intestine.

Based on scRNA-seq and confirmatory FISH, we observed that WNT2B, RSPO2 and RSPO3 are expressed in the muscularis mucosae in the human fetal intestine, as well as in cells more broadly defined as fibroblasts. However, the muscularis mucosae is a muscle layer that is spatially located closest to proliferative domain of the intestine. This is a unique finding when compared to recent single cell studies in the adult human colon, and compared to findings in the mouse. In the adult human colon, a source of WNT and RSPO external to the muscular mucosae has been identified as WNT2B/RSPO3+ fibroblasts (Smillie et al., 2019), whereas in mice WNT and RSPO ligands are expressed in cells that were recently coined ‘trophocytes’ and in PDGRFALO stroma cells in the small intestine(McCarthy et al., 2020). In addition, the identification of the Foxl1+ telocyte has represented a major advance in elucidating the cells and sources of many niche factors in the murine intestinal stem cell niche (Aoki et al., 2016; Shoshkes-Carmel et al., 2018). When Foxl1+ cells are genetically ablated, the crypt collapses; however recent single cell studies have shown that telocytes are not a major source WNT or RSPO ligands (Kim et al., 2020; McCarthy et al., 2020), and others have demonstrated that non-telocytes express RSPO3 (Greicius et al., 2018; Shoshkes-Carmel et al., 2018). Single cell studies in humans have initially focused on the colon (Kinchen et al., 2018; Smillie et al., 2019), so it remains to be seen if there is a unique expression pattern in the adult human small intestine and/or if there are changes in the cellular sources for WNT2B and RSPO2/3 as development progresses; however, as the gross anatomical structure of the intestine observed in the fetal stages starting at 100 days post conception and onward is maintained into adulthood (i.e. crypt-villus axis, muscle layers) it is possible that there are dramatic differences across species and regions of the gut for the major niche cells.

In the current work, we identify a subepithelial cell that lines the entire crypt-villus axis, marked by high levels of DLL1, F3 and PDGFRA expression. These cells can be further sub-divided using the marker NPY, which is expressed in the villus (NPY/DLL1/F3) but not the crypt. Within the F3 transcriptional signature, we also observed robust expression of FRZB and SOX6, which have previously been described in the human colon as a subepithelial cell population that expresses several WNT family members (WNT5A, WNT5B) and several BMP family members (BMP2, BMP5) (Kinchen et al., 2018). Thus, while the focus of the current manuscript is on EGF-family members, it is likely that the niche signaling role for the DLL1HI/F3HI is more complex and may involve secretion of activators and inhibitors of several other signaling pathways.

Establishing signaling gradients along the crypt-villus axis.

While difficult or impossible to test in human tissue, one can speculate that the robust levels of EGF expressed in the villus epithelium coupled with NRG1 expression in subepithelial cells along the crypt-villus axis help to establish, in effect, a gradient of EGF/NRG1 signaling by differential receptor binding/dimerization in different domains. For example, high NRG1 is expressed in the crypt-associated subepithelial cells, with low/no EGF being expressed in the in the crypt domain, whereas both NRG1 and lower levels of EGF are present in the putative TA zone, Finally, high NRG1 and high EGF is likely present in the villus, based on FISH data, but this area would also have the lowest levels of RSPO and WNT, given our localization data showing that RSPO and WNT ligands are low/absent in the villus core. Given these spatial differences, it is interesting to speculate that EGF normally acts as a differentiation factor in the absence of WNT signaling. However, the interplay between WNT signaling and EGF signaling is likely to intersect with other signaling pathways as well, given that gradients of BMP signaling are also established in along the crypt-villus axis (McCarthy et al., 2020).

Future work may focus on how dose-dependent gradients from multiple signaling pathways intersect with one-another to finely tune human stem cell maintenance and differentiation, and it is clear that there is still much to be understood about the human niche. In mice, a minimal set of niche factors are able to maintain intestinal enteroids in a homeostatic state with stem cells and differentiated cells taking on proper crypt-villus organization (Sato et al., 2009). This organization is dependent on Paneth cell differentiation and secretion of WNT ligands, allowing crypts to bud into their own spatial domains in enteroids (Sato et al., 2010; Serra et al., 2019). However, such a homeostatic state that allows both stem cell maintenance and differentiation has not been achieved in human enteroid cultures, and to date, human enteroids are largely maintained in a stem cell state, are switched to differentiation media to allow cytodifferentiation, or required engineered scaffolds and growth factor gradients to establish crypt-villus domains (Dame et al., 2018; Sato et al., 2011; Vandussen et al., 2014; Wang et al., 2017). Here we show that EGF signaling members may play a role in such a homeostatic state, given our findings that NRG1 permits cytodifferentiation while maintaining stem cells in vitro; however, it was clear that neither EGF nor NRG1 grown enteroids possessed crypt-villus architecture.

Taken together, our data reveals that the human fetal intestinal stem cell niche is composed of multiple cellular sources, and highlights a unique role for different ligands from the EGF family. The resources we provide here lay the groundwork to further interrogate cellular relationships in the human fetal intestine, provide an important benchmark for in vitro experiments, and will inform additional methods to generate more robust and physiologic culture conditions.

Limitations of the Study

Shortly after the initial submission of this manuscript, the world was gripped by a global pandemic caused by the SARS-CoV-2 virus. This forced massive shutdowns to biomedical research across the U.S. and across the world for many months, significantly impacting our ability to carry out new experiments. Fortunately, much of the data for this manuscript was on hand and allowed the authors to perform additional computational analysis and follow up studies also relied on archival tissue present in the laboratory prior to the shutdown. Nonetheless, there were some limitations imposed by the pandemic – access to new tissue for this study was effectively stopped, and prevented plans to isolate different mesenchymal cell populations (i.e. F3+/PDGFRAHI) in order to directly test the function of these cells to support enteroids in vitro. Another clear limitation in this study (irrespective of the global pandemic) is the limited ability to carry out experiments in human tissue. Our result in Figure 1 based on force directed layouts following the implementation of Harmony suggest that there are lineage trajectories of emerging mesenchymal cell populations over time. For example, our data suggest that cell trajectories change dramatically and diversify over time (Figure 1D–E) however, without the ability to carry out bona fide lineage tracing, inferred cell trajectories should be interpreted with caution. Similarly, the temporal resolution of our data, with data sets being one week or more apart, prohibits certain types of analysis, such as RNA velocity (La Manno et al., 2018), which would require much higher temporal resolution to draw meaningful inferences between the time points.

STAR METHODS

RESOURCE AVAILABILITY

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Jason R. Spence (spencejr@umich.edu).

Materials Availability

All unique/stable reagents generated in this study are available from the Lead Contact with a completed Materials Transfer Agreement

Data Code and Availability

Sequencing data generated and used by this study is deposited at EMBL-EBI ArrayExpress. Data sets for human fetal intestine (ArrayExpress: E-MTAB-9489) and human fetal intestinal enteroids (ArrayExpress: E-MTAB-9720) have been deposited. Accession numbers for deposited data are also provided in the Key Resources Table. Code used to process raw data can be found at https://github.com/jason-spence-lab/Holloway_Czerwinski_Tsai_2020.

KEY RESOURCES TABLE

REAGENT or RESOURCE SOURCE IDENTIFIER
Antibodies
Mouse anti E-Cadherin (1:500) BD Biosciences Cat#610181; RRID:AB_397580
Goat anti E-Cadherin (1:100) R&D systems Cat#AF748; RRID:AB_355568
Rabbit anti-SM22 (1:500) Abcam Cat#ab14106; RRID:AB_443021
Rabbit anti-Ki67 (1:500) Lab Vision Cat# RM-9106-S1; RRID:AB_149792
Rabbit anti-LYSOZYME (1:800) DAKO Cat# A0099; RRID:AB_2341230
Mouse anti-MUC2 (1:50) Santa Cruz Biotechnology Cat# SC-7314; RRID:AB_627970
Donkey anti-mouse 488 (1:500) Jackson ImmunoResearch Cat# 715–545-150; RRID:AB_2340846
Donkey anti-rabbit 488 (1:500) Jackson ImmunoResearch Cat# 711–545-152; RRID:AB_2313584
Donkey anti-goat 488 (1:500) Jackson ImmunoResearch RRID:AB_2336933; Cat# 705–545-147
Donkey anti-mouse 647 (1:500) Jackson ImmunoResearch RRID:AB_2340862; Cat# 715–605-150
Donkey anti-rabbit 647 (1:500) Jackson ImmunoResearch Cat# 711–605-152; RRID:AB_2492288
Biological Samples
human fetal intestine University of Washington Laboratory of Developmental Biology N/A
Chemicals, Peptides, and Recombinant Proteins
EGF R&D 236-EG
NRG1 R&D 5898-NR-050
N-acetylcysteine Sigma A9165–25G
Nicotinamide Sigma N0636–061
Critical Commercial Assays
Neural Tissue Dissociation Kit (P) Miltenyi 130–092-628
RNAscope® Multiplex Fluorescent Reagent Kit v2 ACD 323100
Deposited Data
Raw scRNA-seq data (human fetal duodenum) This study ArrayExpress: E-MTAB-9489
Raw scRNA-seq data (human intestinal enteroids) This study ArrayExpress: E-MTAB-9720
Experimental Models: Cell Lines
L-WRN Cells ATCC CRL-3276
Human fetal duodenal enteroids (HT-071, 142d) This study N/A
Human fetal duodenal enteroids (HT-323, 105d) This study N/A
Human fetal duodenal enteroids (HT-335 135d) This study N/A
Software and Algorithms
Scanpy, Ingest (Wolf et al., 2018) https://github.com/theislab/scanpy
Cellranger 10x Genomics https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/what-is-cell-ranger
UMAP (McInnes et al., 2018) https://github.com/lmcinnes/umap
Harmony (Nowotschin et al., 2019) https://github.com/dpeerlab/Harmony
Prism 8.3.0 Graphpad https://www.graphpad.com/scientific-software/prism/
Other
Matrigel Corning 354234
RNAscope Probe Hs-DLL1 ACD 532631
RNAscope Probe Hs-PDGFRA-C2 ACD 604481-C2
RNAscope Probe Hs-F3-C2 ACD 407611-C2
RNAscope Probe Hs-F3 ACD 407611
RNAscope Probe Hs-WNT2B ACD 453361
RNAscope Probe Hs-EGF ACD 605771
RNAscope Probe Hs-EGFR ACD 310061
RNAscope Probe Hs-ERBB2-C2 ACD 310081-C2
RNAscope Probe Hs-ERBB3 ACD 311941
RNAscope Probe Hs-NRG1-C2 ACD 311181-C2
RNAscope Probe Hs-RSPO2 ACD 423991
RNAscope Probe Hs-RSPO3–01 ACD 429851
RNAscope Probe Hs-NPY ACD 416671
RNAscope Probe Hs-GPX3-C2 ACD 470591-C2
RNAscope Probe Hs-NRG1-C2 ACD 311181-C2
RNAscope Probe Hs-LGR5 ACD 311021
RNAscope Probe Hs-OLFM4 ACD 311041
Open in a new tab

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Isolating, establishing and maintaining human fetal enteroids:

Fresh human fetal epithelium was isolated and maintained exactly as previously described (Tsai et al., 2018). Briefly, duodenal tissue was minced into small pieces using a scalpel under sterile conditions. To separate the epithelium from the underlying cell layers, minced biopsy specimens then were incubated in dispase (STEMCELL Technologies, 07923) for 30 minutes on ice in a petri dish. Dispase then was removed and replaced with 100% fetal bovine serum for 15 minutes on ice. To mechanically separate the tissue layers, a volume of Advanced DMEM/F12 (Gibco, 12634–028) equal to the initial volume of fetal bovine serum was added to the biopsy tissue before vigorously pipetting the mixture several times. Epithelial fragments then settled to the bottom of the petri dish where they were collected manually on a stereoscope by pipet. The epithelium then was washed with ice-cold Advanced DMEM/F12 and allowed to settle to the bottom of a 1.5 mL tube. The media then was withdrawn from the loose tissue pellet and replaced with Matrigel. The Matrigel containing the isolated epithelium then was gently mixed to suspend the cells evenly before being pipetted into individual 50 μL droplets in a 24-well plate. The plate containing the droplets then was incubated at 37°C for 15 minutes to allow the Matrigel to solidify before adding media.

Once enteroids were established, Matrigel droplets were dislodged from the culture plate and were pipetted in a petri dish. Healthy cystic enteroids were manually selected from the petri dish under a stereoscope to be bulk-passaged through a 30G needle and embedded in Matrigel (Corning, 354234). For single-cell passaging, healthy cystic enteroids were manually selected under a stereoscope and dissociated with TrypLE Express (Gibco, 12605–010) at 37°C before filtering through 40μm cell strainers. Cells were then counted using a hemocytometer (ThermoFisher) and embedded in Matrigel. For some experiments, enteroid lines were derived in the presence of NRG1 (100ng/ml) or NRG1 (100 ng/ml) and EGF (100 ng/ml) in addition to the standard EGF (100ng/ml) conditions in complete LWRN media (see media composition).

Media composition:

Culture media consisted of LWRN conditioned media generated as previously described (Miyoshi and Stappenbeck, 2013; Tsai et al., 2018) and human basal media [Advanced DMEM/F12 (Gibco, 12634–028); Glutamax 4 mM (Gibco, 35050–061); HEPES 20 mM (Gibco, 15630–080); N2 Supplement (2X) (Gibco, 17502–048), B27 Supplement (2X) (17504–044), Penicillin-Streptomycin (2X) (Gibco, 15140–122), N-acetylcysteine (2 mM) (Sigma, A9165–25G), Nicotinamide (20 mM) (Sigma, N0636–061)]. Complete media was comprised of 25% LWRN and 75% human basal media to which rhEGF (R&D, 236-EG, 100ng/ml) and/or rhNRG1 (R&D, 5898-NR-050, 100ng/ml) was added. For some experiments complete media was used without rhEGF or rhNRG1 as a control.

Human subjects:

Normal, de-identified human fetal intestinal tissue was obtained from the University of Washington Laboratory of Developmental Biology. All human tissue used in this work was de-identified and was conducted with approval from the University of Michigan IRB.

METHOD DETAILS

Experimental design of enteroid cultures:

The enteroid experiments in Figures 3 and 4 were carefully conducted to reduce batch effects in scRNA-seq data. All experiments comparing different treatment groups (i.e. EGF, NRG1, etc.) were carried out in parallel, with experiments and treatments being conducted at the same time. Cells were harvested and dissociated into single cell suspensions in parallel (see below). Since the 10X Chromium system allows parallel processing of multiple samples at a time, cells were captured (Gel bead-in-Emulsion - GEMS) and processed (i.e. library prep) in parallel. All samples were sequenced across the same lane(s) on a Novaseq 6000.

Single cell dissociation:

To dissociate human fetal tissue to single cells, fetal duodenum was first dissected using forceps and a scalpel in a petri dish filled with ice-cold 1X HBSS (with Mg2+, Ca2+). Whole thickness intestine was cut into small pieces and transferred to a 15 mL conical tube with 1% BSA in HBSS. Dissociation enzymes and reagents from the Neural Tissue Dissociation Kit (Miltenyi, 130–092-628) were used, and all incubation steps were carried out in a refrigerated centrifuge pre-chilled to 10°C unless otherwise stated. All tubes and pipette tips used to handle cell suspensions were pre-washed with 1% BSA in 1X HBSS to prevent adhesion of cells to the plastic. Tissue was treated for 15 minutes at 10°C with Mix 1 and then incubated for 10 minute increments at 10°C with Mix 2 interrupted by agitation by pipetting with a P1000 pipette until fully dissociated. Cells were filtered through a 70μm filter coated with 1% BSA in 1X HBSS, spun down at 500g for 5 minutes at 10°C and resuspended in 500μl 1X HBSS (with Mg2+, Ca2+). 1 mL Red Blood Cell Lysis buffer was then added to the tube and the cell mixture was placed on a rocker for 15 minutes in the cold room (4°C). Cells were spun down (500g for 5 minutes at 10°C) and washed twice by suspension in 2 mL of HBSS + 1% BSA, followed by centrifugation. Cells were counted using a hemocytometer, then spun down and resuspended to reach a concentration of 1000 cells/μL and kept on ice. Single cell libraries were immediately prepared on the 10x Chromium by the University of Michigan Advanced Genomics Core facility with a target of 5000 cells. The same protocol was used for single cell dissociation of healthy cystic enteroids manually collected under a stereoscope. A full, detailed protocol of tissue dissociation for single cell RNA sequencing can be found at www.jasonspencelab.com/protocols.

Single cell library preparation and transcriptome alignment

All single-cell RNA-seq sample libraries were prepared with the 10x Chromium Controller using either the v2 or v3 chemistry. Sequencing was performed on an Illumina HiSeq 4000 or NovaSeq 6000 with targeted depth of 100,000 reads per cell. Default alignment parameters were used to align reads to the pre-prepared hg19 human reference genome provided by the 10X Cell Ranger pipeline. Initial cell demultiplexing and gene quantification were also performed with the default 10x Cell Ranger pipeline.

Primary tissue collection, fixation and paraffin processing

Human fetal intestine tissue samples were collected as ~0.5 cm fragments and fixed for 24 hours at room temperature in 10% Neutral Buffered Formalin (NBF), and washed with UltraPure Distilled Water (Invitrogen, 10977–015) for 3 changes for a total of 2 hours. Tissue was dehydrated by an alcohol series diluted in UltraPure Distilled Water (Invitrogen, 10977–015). Tissue was incubated for 60 minutes each solution: 25% Methanol, 50% Methanol, 75% Methanol, 100% Methanol. Tissue was stored long-term in 100% Methanol at 4°C. Prior to paraffin embedding, tissue was equilibrated in 100% Ethanol for an hour, and then 70% Ethanol. Tissue was processed into paraffin blocks in an automated tissue processor (Leica ASP300) with 1 hour changes overnight.

Enteroid collection, fixation and paraffin processing

Enteroids were allowed to grow in Matrigel for several days following passaging. Once established, Enteroids in Matrigel were transferred gently with a cut pipette P1000 tip into a petri dish filled with cold DMEM/F12. Enteroids are then manually dissected from Matrigel under a dissecting stereomicroscope using fine forceps and transferred to a microcentrifuge tube. Enteroids are left upright for several minutes until they gravity sediment to at the bottom of the tube, at which time as much media as possible is gently withdrawn. Histogel (Thermo-scientific, HG-4000–012) is slowly added to cover the enteroids following the manufacturers protocol. Once Histogel has solidified, Histogel-embedded enteroids are transferred to a 5 mL conical tube and fixed in 10% NBF overnight at room temperature. Once fixed, they are processed into paraffin as described above, sectioned and staining for FISH and IF described below.

Multiplex Fluorescent In Situ Hybridization (FISH) and immunofluorescence (IF)

Paraffin blocks were sectioned to generate 5 μm-thick sections within a week prior to performing in situ hybridization. All materials, including the microtome and blade, were sprayed with RNase-away solution prior to use. Slides were baked for 1 hour in a 60°C dry oven the night before and stored at room temperature in a slide box with a silicone desiccator packet, and with seams sealed using parafilm. The fluorescent in situ hybridization protocol was performed according to the manufacturer’s instructions (ACD; RNAscope multiplex fluorescent manual protocol, 323100-USM) under standard antigen retrieval conditions and 30 minute protease treatment. Immediately following the HRP blocking for the C2 channel of the FISH, slides were washed three times for 5 minutes in PBS, then transferred to blocking solution (5% Normal Donkey Serum in PBS with 0.1%Tween-20) for 1 hour at room temperature. Slides were then incubated in primary antibodies overnight at 4°C in a humidity chamber. The following day, excess primary antibodies were rinsed off through a series of PBS washes. Secondary antibodies and DAPI (1 μg/ml) were added and slides were incubated at room temperature for 1 hour. Excess secondary antibodies were rinsed off through a series of PBS washes, and slides were mounted in ProLong Gold (TermoFisher, P36930). A list of antibodies and concentrations can be found in the Key Resources Table. All imaging was done using a NIKON A1 confocal and images were assembled using Photoshop CC. Z-stack series were captured and compiled into maximum intensity projections using NIS-Elements (Nikon). Imaging parameters were kept consistent for all images within the same experiment and any post-imaging manipulations were performed equally on all images from a single experiment.

QUANTIFICATION AND STATISTICAL ANALYSIS

Single-cell in silico analysis

All in silico analyses downstream of gene quantification were done using Scanpy with the 10x Cell Ranger derived gene by cell matrices (Wolf et al., 2018). For primary human tissue sample analysis in Figure 1 and 2, all samples were filtered to remove cells with less than 1000 or greater than 9000 genes, less than 3500 or greater than 25000 unique molecular identifier (UMI) counts per cell. Ambient/background signal was removed from the data using CellBender. “Remove-background” was used at 200 epochs to remove ambient RNA counts from all fetal intestine samples, and the de-noised data matrix was used for subsequent analysis (Fleming et al., 2019). De-noised data matrix read counts per gene were log normalized prior to analysis. After log normalization, 2000–3000 highly variable genes were identified and extracted. The normalized expression levels then underwent linear regression to remove effects of total reads per cell and cell cycle genes, followed by a z-transformation. Dimension reduction was performed using principal component analysis (PCA) and then uniform manifold approximation and projection (UMAP) on the top 9 principal components (PCs) and 30 nearest neighbors for visualization on 2 dimensions (McInnes et al., 2018; Polański et al., 2019). Clusters of cells within the data were calculated using the Louvain algorithm within Scanpy with a resolution of 0.6. For Figure 1 and 2, combined time series data for mesenchymal and epithelial cells were integrated using Harmony to generate augmented affinity matrices and plotted as force-directed layouts with ForceAtlas2 (Jacomy et al., 2014; Nowotschin et al., 2019).

For Figures 3 and 4, all samples were filtered to remove cells with too few or too many genes (Figure 3 - <2000, >9000; Figure 4 <250, >8000; Figure 5 <500, >3000) or with high unique molecular identifier (UMI) counts per cell (Figure 3 - 100000; Figure 4 – 10000), and a fraction of mitochondrial genes greater than 0.1–0.25. Data matrix read counts per gene were log normalized prior to analysis. After log normalization, 2000–3000 highly variable genes were identified and extracted. For Figure 3, the normalized expression levels then underwent linear regression to remove effects of total reads per cell and mitochondrial transcript fraction. Data was then scaled by z-transformation. Dimension reduction was performed using principal component analysis (PCA) and then uniform manifold approximation and projection (UMAP) on the top 11–20 principal components (PCs) and 15–30 nearest neighbors for visualization on 2 dimensions (McInnes et al., 2018; Polański et al., 2019). Clusters of cells within the data were calculated using the Louvain algorithm within Scanpy with a resolution of 0.2–0.4. Following initial PCA dimension reduction and UMAP visualization, further de-noising was not carried out for this analysis given the distinct cell clusters. In some cases, accompanying graphs were generated from summarizing the scRNA-seq data using Prism 8 software (Graphpad).

Scanpy’s Ingest functionality was used to map enteroids onto primary human fetal epithelial cells. Epithelial cells were identified (as in Figure S1) and extracted from a data matrix to include 828 intestinal epithelial cells from ages 101, 122, 127, and 132 days (ArrayExpress: E-MTAB-9489). Epithelial cells were annotated using canonical genes (Table S3). The extracted epithelial cell matrix then again underwent log normalization, variable gene extraction, z-transformation and dimension reduction to obtain a reference embedding. Ingest was then run to project each of the individual enteroid datasets onto the epithelial reference map.

Supplementary Material

Supplementary material
Table S1

Supplemental Table 1. Related to Figure 3. Differentially expressed genes in established enteroid cultures grown for 5 days in EGF or NRG1.

Table S2

Supplemental Table 2. Related to Figure 4. Differentially expressed genes in enteroid cultures derived in EGF and/or NRG1 conditions.

Table S3

Supplemental Table 3. Related to Figure 4. Differentially expressed genes among primary fetal intestinal epithelial cell clusters.

Highlights.

  • Cell diversity in the developing human intestine was interrogated using scRNA-seq

  • PGDFRAHI/F3HI/DLL1HI mesenchyme lines the epithelium and expresses NRG1

  • EGF, a common in vitro niche factor, is not abundant in the crypt domain.

  • Compared to EGF, NRG1 increases cellular diversity in enteroid culture

Acknowledgements:

We thank Judy Opp and the University of Michigan Advanced Genomics Core for their expertise operating the 10X Chromium single cell capture platform and sequencing expertise. We would also like to thank the University of Michigan Microscopy Core for providing access to confocal microscopes and image analysis software, and The University of Washington Laboratory of Developmental Biology staff.

Financial Support:

This work was supported by the Intestinal Stem Cell Consortium (U01DK103141 to J.R.S.), a collaborative research project funded by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) and the National Institute of Allergy and Infectious Diseases (NIAID). This work was also supported by the NIAID Novel Alternative Model Systems for Enteric Diseases (NAMSED) consortium ( U19AI116482 to J.R.S.), by a Chan Zuckerberg Initiative Seed Network grant (to J.R.S, B.T. and J.G.C), and by the University of Michigan Center for Gastrointestinal Research (UMCGR) (NIDDK 5P30DK034933). KW is supported by NIDDK R01KD121166. The University of Washington Laboratory of Developmental Biology was supported by NIH award number 5R24HD000836 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD). MC was supported by the Training Program in Organogenesis (NIH-NICHD T32 HD007505). EMH was supported by the Training in Basic and Translational Digestive Sciences Training Grant (NIH-NIDDK 5T32DK094775), the Cellular Biotechnology Training Program Training Grant (NIH-NIGMS 2T32GM008353), and the Ruth L. Kirschstein Predoctoral Individual National Research Service Award (NIH-NHLBI F31HL146162).

Footnotes

Competing interests

The authors have no competing interests.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Aoki R, Shoshkes-Carmel M, Gao N, Shin S, May CL, Golson ML, Zahm AM, Ray M, Wiser CL, Wright CVE, et al. (2016). Foxl1-expressing mesenchymal cells constitute the intestinal stem cell niche. Cell. Mol. Gastroenterol. Hepatol. 2, 175–188. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Basak O, Beumer J, Wiebrands K, Seno H, van Oudenaarden A, and Clevers H. (2017). Induced Quiescence of Lgr5+ Stem Cells in Intestinal Organoids Enables Differentiation of Hormone-Producing Enteroendocrine Cells. Cell Stem Cell 20, 177–190.e4. [DOI] [PubMed] [Google Scholar]
  3. Becht E, McInnes L, Healy J, Dutertre CA, Kwok IWH, Ng LG, Ginhoux F, and Newell EW (2019). Dimensionality reduction for visualizing single-cell data using UMAP. Nat. Biotechnol. 37, 38–47. [DOI] [PubMed] [Google Scholar]
  4. Beucher A, Gjernes E, Collin C, Courtney M, Meunier A, Collombat P, and Gradwohl G. (2012). The Homeodomain-Containing Transcription Factors Arx and Pax4 Control Enteroendocrine Subtype Specification in Mice. PLoS One 7, e36449. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Capeling M, Huang S, Mulero-Russe A, Cieza R, Tsai YH, Garcia A, and Hill DR (2020). Generation of small intestinal organoids for experimental intestinal physiology. In Methods in Cell Biology, (Academic Press Inc.), p. [DOI] [PubMed] [Google Scholar]
  6. Capeling MM, Czerwinski M, Huang S, Tsai Y-H, Wu A, Nagy MS, Juliar B, Sundaram N, Song Y, Han WM, et al. (2019). Nonadhesive Alginate Hydrogels Support Growth of Pluripotent Stem Cell-Derived Intestinal Organoids. Stem Cell Reports. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chin AM, Hill DR, Aurora M, and Spence JR (2017). Morphogenesis and maturation of the embryonic and postnatal intestine. Semin. Cell Dev. Biol. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cruz-Acuña R, Quirós M, Farkas AE, Dedhia PH, Huang S, Siuda D, García-Hernández V, Miller AJ, Spence JR, Nusrat A, et al. (2017). Synthetic hydrogels for human intestinal organoid generation and colonic wound repair. Nat. Cell Biol. 19, 1326–1335. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Dame MK, Attili D, McClintock SD, Dedhia PH, Ouillette P, Hardt O, Chin AM, Xue X, Laliberte J, Katz EL, et al. (2018). Identification, isolation and characterization of human LGR5-positive colon adenoma cells. Development 145, dev153049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dedhia PH, Bertaux-Skeirik N, Zavros Y, and Spence JR (2016). Organoid Models of Human Gastrointestinal Development and Disease. Gastroenterology 150, 1098–1112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Du A, McCracken KW, Walp E, Terry NA, Klein TJ, Han A, Wells JM, and May CL (2012). Arx is required for normal enteroendocrine cell development in mice and humans. Dev. Biol [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Elmentaite R, Ross A, James K, Ortmann D, Gomes T, Roberts K, Nayak K, Tuck L, Bayraktar OA, Heuschkel R, et al. (2020). Single-cell sequencing of developing human gut reveals transcriptional links to childhood Crohn’s disease. BioRxiv 2020.02.06.937110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Van Es JH, Wiebrands K, López-Iglesias C, Van De Wetering M, Zeinstra L, Van Den Born M, Korving J, Sasaki N, Peters PJ, Van Oudenaarden A, et al. (2019). Enteroendocrine and tuft cells support Lgr5 stem cells on Paneth cell depletion. Proc. Natl. Acad. Sci. U. S. A. 116, 26599–26605. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Finkbeiner SR, Hill DR, Altheim CH, Dedhia PH, Taylor MJ, Tsai Y-H, Chin AM, Mahe MM, Watson CL, Freeman JJ, et al. (2015). Transcriptome-wide Analysis Reveals Hallmarks of Human Intestine Development and Maturation In Vitro and In Vivo. Stem Cell Reports 4, 1140–1155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Fleming S, Marioni J, and Babadi M. (2019). CellBender remove-background: a deep generative model for unsupervised removal of background noise from scRNA-seq datasets. BioRxiv 791699. [Google Scholar]
  16. Fordham RP, Yui S, Hannan NRF, Soendergaard C, Madgwick A, Schweiger PJ, Nielsen OH, Vallier L, Pedersen RA, Nakamura T, et al. (2013). Transplantation of Expanded Fetal Intestinal Progenitors Contributes to Colon Regeneration after Injury. Cell Stem Cell. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Fujii M, Matano M, Toshimitsu K, Takano A, Mikami Y, Nishikori S, Sugimoto S, and Sato T. (2018). Human Intestinal Organoids Maintain Self-Renewal Capacity and Cellular Diversity in Niche-Inspired Culture Condition. Cell Stem Cell 23, 787–793.e6. [DOI] [PubMed] [Google Scholar]
  18. Gehart H, Van Es JH, Dekkers JF, Rios A, and Clevers Correspondence H. (2019). Identification of Enteroendocrine Regulators by Real-Time Single-Cell Differentiation Mapping Data Resources GSE113561 Gehart et al. Cell 176. [DOI] [PubMed] [Google Scholar]
  19. Gjorevski N, Sachs N, Manfrin A, Giger S, Bragina ME, Ordóñez-Morán P, Clevers H, and Lutolf MP (2016). Designer matrices for intestinal stem cell and organoid culture. Nature 539, 560–564. [DOI] [PubMed] [Google Scholar]
  20. Goodlad RA, Wilson TJ, Lenton W, Gregory H, McCullagh KG, and Wright NA (1987). Proliferative effects of urogastrone-EGF on the intestinal epithelium. Gut 28 Suppl, 37–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Gregorieff A, Stange DE, Kujala P, Begthel H, van den Born M, Korving J, Peters PJ, and Clevers H. (2009). The ets-domain transcription factor Spdef promotes maturation of goblet and paneth cells in the intestinal epithelium. Gastroenterology 137, 1333. [DOI] [PubMed] [Google Scholar]
  22. Greicius G, Kabiri Z, Sigmundsson K, Liang C, Bunte R, Singh MK, and Virshup DM (2018). PDGFRα + pericryptal stromal cells are the critical source of Wnts and RSPO3 for murine intestinal stem cells in vivo. Proc. Natl. Acad. Sci. U. S. A. 115, E3173–E3181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Guilmeau S, Niot I, Laigneau JP, Devaud H, Petit V, Brousse N, Bouvier R, Ferkdadji L, Besmond C, Aggerbeck LP, et al. (2007). Decreased expression of Intestinal I- and L-FABP levels in rare human genetic lipid malabsorption syndromes. Histochem. Cell Biol. 128, 115–123. [DOI] [PubMed] [Google Scholar]
  24. Haber AL, Biton M, Rogel N, Herbst RH, Shekhar K, Smillie C, Burgin G, Delorey TM, Howitt MR, Katz Y, et al. (2017). A single-cell survey of the small intestinal epithelium. Nat. Publ. Gr. 551, 333–339. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Haramis APG, Begthel H, Van Den Born M, Van Es J, Jonkheer S, Offerhaus GJA, and Clevers H. (2004). De Novo Crypt Formation and Juvenile Polyposis on BMP Inhibition in Mouse Intestine. Science (80-. ). 303, 1684–1686. [DOI] [PubMed] [Google Scholar]
  26. Howitt MR, Cao YG, Gologorsky MB, Li JA, Haber AL, Biton M, Lang J, Michaud M, Regev A, and Garrett WS (2020). The Taste Receptor TAS1R3 Regulates Small Intestinal Tuft Cell Homeostasis. ImmunoHorizons 4, 23–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Jacomy M, Venturini T, Heymann S, and Bastian M. (2014). ForceAtlas2, a continuous graph layout algorithm for handy network visualization designed for the Gephi software. PLoS One 9, e98679. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Jardé T, Chan WH, Rossello FJ, Kaur Kahlon T, Theocharous M, Kurian Arackal T, Flores T, Giraud M, Richards E, Chan E, et al. (2020). Mesenchymal Niche-Derived Neuregulin-1 Drives Intestinal Stem Cell Proliferation and Regeneration of Damaged Epithelium. Cell Stem Cell. [DOI] [PubMed] [Google Scholar]
  29. Kaske S, Krasteva G, König P, Kummer W, Hofmann T, Gudermann T, and Chubanov V. (2007). TRPM5, a taste-signaling transient receptor potential ion-channel, is a ubiquitous signaling component in chemosensory cells. BMC Neurosci. 8, 49. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kim JE, Fei L, Yin WC, Coquenlorge S, Rao-Bhatia A, Zhang X, Shi SSW, Lee JH, Hahn NA, Rizvi W, et al. (2020). Single cell and genetic analyses reveal conserved populations and signaling mechanisms of gastrointestinal stromal niches. Nat. Commun. 11, 1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kinchen J, Chen HH, Parikh K, Gervais F, Koohy H, and Simmons Correspondence A. (2018). Structural Remodeling of the Human Colonic Mesenchyme in Inflammatory Bowel Disease. Cell 175, 372–386. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Kraiczy J, Nayak KM, Howell KJ, Ross A, Forbester J, Salvestrini C, Mustata R, Perkins S, Andersson-Rolf A, Leenen E, et al. (2017). DNA methylation defines regional identity of human intestinal epithelial organoids and undergoes dynamic changes during development. Gut gutjnl-2017–314817. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kretzschmar K, and Clevers H. (2016). Organoids: Modeling Development and the Stem Cell Niche in a Dish. Dev. Cell 38, 590–600. [DOI] [PubMed] [Google Scholar]
  34. La Manno G, Soldatov R, Zeisel A, Braun E, Hochgerner H, Petukhov V, Lidschreiber K, Kastriti ME, Lönnerberg P, Furlan A, et al. (2018). RNA velocity of single cells. Nature 560, 494–498. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Martin JC, Chang C, Boschetti G, Merad M, Cho JH, and Correspondence EK (2019). Single-Cell Analysis of Crohn’s Disease Lesions Identifies a Pathogenic Cellular Module Associated with Resistance to Anti-TNF Therapy. Cell 178. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. McCarthy N, Manieri E, Storm EE, Saadatpour A, Luoma AM, Kapoor VN, Madha S, Gaynor LT, Cox C, Keerthivasan S, et al. (2020). Distinct Mesenchymal Cell Populations Generate the Essential Intestinal BMP Signaling Gradient. Cell Stem Cell 26, 391–402.e5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. McInnes L, Healy J, Saul N, and Großberger L. (2018). UMAP: Uniform Manifold Approximation and Projection. J. Open Source Softw. 3, 861. [Google Scholar]
  38. Miyoshi H, and Stappenbeck TS (2013). In vitro expansion and genetic modification of gastrointestinal stem cells in spheroid culture. Nat. Protoc. 8, 2471–2482. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Muhl L, Genové G, Leptidis S, Liu J, He L, Mocci G, Sun Y, Gustafsson S, Buyandelger B, Chivukula IV, et al. (2020). Single-cell analysis uncovers fibroblast heterogeneity and criteria for fibroblast and mural cell identification and discrimination. Nat. Commun. 11, 3953. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Muncan V, Sansom OJ, Tertoolen L, Phesse TJ, Begthel H, Sancho E, Cole AM, Gregorieff A, de Alboran IM, Clevers H, et al. (2006). Rapid loss of intestinal crypts upon conditional deletion of the Wnt/Tcf-4 target gene c-Myc. Mol Cell Biol 26, 8418–8426. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Noah TK, Kazanjian A, Whitsett J, and Shroyer NF (2010). SAM pointed domain ETS factor (SPDEF) regulates terminal differentiation and maturation of intestinal goblet cells. Exp. Cell Res. 316, 452–465. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Nowotschin S, Setty M, Kuo YY, Liu V, Garg V, Sharma R, Simon CS, Saiz N, Gardner R, Boutet SC, et al. (2019). The emergent landscape of the mouse gut endoderm at single-cell resolution. Nature 569, 361–367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Okamoto R, Tsuchiya K, Nemoto Y, Akiyama J, Nakamura T, Kanai T, and Watanabe M. (2008). Requirement of Notch activation during regeneration of the intestinal epithelia. AJP Gastrointest. Liver Physiol. 296, G23–G35. [DOI] [PubMed] [Google Scholar]
  44. Ootani A, Li X, Sangiorgi E, Ho QT, Ueno H, Toda S, Sugihara H, Fujimoto K, Weissman IL, Capecchi MR, et al. (2009). Sustained in vitro intestinal epithelial culture within a Wnt-dependent stem cell niche. Nat. Med. 15, 701–706. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Parikh K, Antanaviciute A, Fawkner-corbett D, Jagielowicz M, Aulicino A, lagerholm christoffer, Davis S, Kinchen J, chen HH, Khalid Alham N, et al. Colonic epithelial cell diversity in health and inflammatory bowel disease. [DOI] [PubMed] [Google Scholar]
  46. Pinto D, and Gregorieff A. (2003). Canonical Wnt signals are essential for homeostasis of the intestinal epithelium. Genes … 6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Pinto D, Gregorieff A, Begthel H, and Clevers H. (2003). Canonical Wnt signals are essential for homeostasis of the intestinal epithelium. Genes Dev. 17, 1709–1713. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Polański K, Young MD, Miao Z, Meyer KB, Teichmann SA, and Park J-E (2019). BBKNN: fast batch alignment of single cell transcriptomes. Bioinformatics. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Sansom OJ, Reed KR, Hayes AJ, Ireland H, Brinkmann H, Newton IP, Batlle E, Simon-Assmann P, Clevers H, Nathke IS, et al. (2004). Loss of Apc in vivo immediately perturbs Wnt signaling, differentiation, and migration. Genes Dev. 18, 1385–1390. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Sato T, Vries RG, Snippert HJ, van de Wetering M, Barker N, Stange DE, van Es JH, Abo A, Kujala P, Peters PJ, et al. (2009). Single Lgr5 stem cells build crypt-villus structures in vitro without a mesenchymal niche. Nature 459, 262–265. [DOI] [PubMed] [Google Scholar]
  51. Sato T, van Es JH, Snippert HJ, Stange DE, Vries RG, van den Born M, Barker N, Shroyer NF, van de Wetering M, and Clevers H. (2010). Paneth cells constitute the niche for Lgr5 stem cells in intestinal crypts. Nature 1–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Sato T, Stange DE, Ferrante M, Vries RGJ, van Es JH, van den Brink S, van Houdt WJ, Pronk A, van Gorp J, Siersema PD, et al. (2011). Long-term Expansion of Epithelial Organoids From Human Colon, Adenoma, Adenocarcinoma, and Barrett’s Epithelium. Gastroenterology 141, 1762–1772. [DOI] [PubMed] [Google Scholar]
  53. Scadden DT (2006). The stem-cell niche as an entity of action. Nature 441, 1075–1079. [DOI] [PubMed] [Google Scholar]
  54. Serra D, Mayr U, Boni A, Lukonin I, Rempfler M, Challet Meylan L, Stadler MB, Strnad P, Papasaikas P, Vischi D, et al. (2019). Self-organization and symmetry breaking in intestinal organoid development. Nature 569, 66–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Shoshkes-Carmel M, Wang YJ, Wangensteen KJ, Tóth B, Kondo A, Massasa EE, Itzkovitz S, and Kaestner KH (2018). Subepithelial telocytes are an important source of Wnts that supports intestinal crypts. Nature 557, 242–246. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Smillie CS, Biton M, Ordovas-Montanes J, Sullivan KM, Burgin G, Graham DB, Herbst RH, Rogel N, Slyper M, Waldman J, et al. (2019). Intra- and Inter-cellular Rewiring of the Human Colon during Ulcerative Colitis. Cell 178, 714–730.e22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Spence JR, Mayhew CN, Rankin SA, Kuhar MF, Vallance JE, Tolle K, Hoskins EE, Kalinichenko VV, Wells SI, Zorn AM, et al. (2010). Directed differentiation of human pluripotent stem cells into intestinal tissue in vitro. Nature 470, 105–109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Tsai Y-H., Czerwinski M, Wu A, Dame MK, Attili D, Hill E, Colacino JA, Nowacki LM, Shroyer NF, Higgins PDR, et al. (2018). A Method for Cryogenic Preservation of Human Biopsy Specimens and&nbsp;Subsequent Organoid Culture. Cell. Mol. Gastroenterol. Hepatol. 1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Ulshen MH, Lyn-Cook LE, and Raasch RH (1986). Effects of intraluminal epidermal growth factor on mucosal proliferation in the small intestine of adult rats. Gastroenterology 91, 1134–1140. [DOI] [PubMed] [Google Scholar]
  60. Vandussen KL, Marinshaw JM, Shaikh N, Miyoshi H, Moon C, Tarr PI, Ciorba MA, and Stappenbeck TS (2014). Development of an enhanced human gastrointestinal epithelial culture system to facilitate patient-based assays. Gut. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Wang Y, Gunasekara DB, Reed MI, DiSalvo M, Bultman SJ, Sims CE, Magness ST, and Allbritton NL (2017). A microengineered collagen scaffold for generating a polarized crypt-villus architecture of human small intestinal epithelium. Biomaterials 128, 44–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Wells JM, and Spence JR (2014). How to make an intestine. Development 141, 752–760. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Van de Wetering M, Sancho E, Verweij C, De Lau W, Oving I, Hurlstone A, Van der Horn K, Batlle E, Coudreuse D, Haramis AP, et al. (2002). The β-catenin/TCF-4 complex imposes a crypt progenitor phenotype on colorectal cancer cells. Cell 111, 241–250. [DOI] [PubMed] [Google Scholar]
  64. Wolf FA, Angerer P, and Theis FJ (2018). SCANPY: large-scale single-cell gene expression data analysis. Genome Biol. 19, 15. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary material
NIHMS1648548-supplement-Supplementary_material.pdf (13MB, pdf)
Table S1

Supplemental Table 1. Related to Figure 3. Differentially expressed genes in established enteroid cultures grown for 5 days in EGF or NRG1.

NIHMS1648548-supplement-Table_S1.xlsx (78.1KB, xlsx)
Table S2

Supplemental Table 2. Related to Figure 4. Differentially expressed genes in enteroid cultures derived in EGF and/or NRG1 conditions.

NIHMS1648548-supplement-Table_S2.xlsx (180.6KB, xlsx)
Table S3

Supplemental Table 3. Related to Figure 4. Differentially expressed genes among primary fetal intestinal epithelial cell clusters.

NIHMS1648548-supplement-Table_S3.xlsx (134.9KB, xlsx)

RESOURCES