Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Jan 1.
Published in final edited form as: Gastroenterology. 2014 Oct 13;148(1):126–136.e6. doi: 10.1053/j.gastro.2014.09.042

In Vitro Expansion of Human Gastric Epithelial Stem Cells and Their Responses to Bacterial Infection

Sina Bartfeld 1, Tülay Bayram 1, Marc van de Wetering 1, Meritxell Huch 1, Harry Begthel 1, Pekka Kujala 2, Robert Vries 1, Peter J Peters 2, Hans Clevers 1
PMCID: PMC4274199  NIHMSID: NIHMS634587  PMID: 25307862

Abstract

Background & Aims

We previously established systems for long-term, 3-dimensional (3D) culture of organoids from mouse tissues (intestine, stomach, pancreas, and liver) and human intestine and pancreas. We describe conditions required for long-term 3D cultures of human gastric stem cells. The technology can be applied to study the epithelial response to infection with Helicobacter pylori.

Methods

We generated organoids from surgical samples of human gastric corpus. Culture conditions were developed based on those for the mouse gastric and human intestinal systems. We used microinjection to infect the organoids with H pylori. Epithelial responses were measured using microarray and quantitative PCR analyses.

Results

Human gastric cells were expanded indefinitely in 3D matrigel cultures. We cultured cells from healthy gastric tissues, single-sorted stem cells, or tumor tissues. Organoids maintained many characteristics of their respective tissues, based on their histology, expression of markers, and euploidy. Organoids from healthy tissue expressed markers of 4 lineages of the stomach and self-organized into gland and pit domains. They could be directed to specifically express either lineages of the gastric gland, or the gastric pit, by addition of nicotinamide and withdrawal of WNT. Whereas gastric pit lineages had only marginal reactions to bacterial infection, gastric gland lineages mounted a strong inflammatory response.

Conclusion

We developed a system to culture human gastric organoids. This system can be used to study H pylori infection and other gastric pathologies.

Keywords: stomach cancer, gastric epithelium, primary cells, tissue engineering

Introduction

The stomach is devided into three regions: The forestomach (mice) or cardia (human), the corpus and the pyloric antrum. The stomach lumen is lined with a monolayer of epithelial cells that is organized in flask-like invaginations, each of which consists of several glands that feed into a single luminal pit. The epithelium constantly renews itself and the stem cells fuelling this process reside in the gastric glands1–4. Similar as in the mouse intestine5,6, symmetric divisions and neutral drift leads to monoclonality of antral gastric units7. Mouse studies have proposed several markers for gastric stem cells4,8–11. Two markers, Lgr5 and Troy, have allowed identification of cells that have the capacity to self-renew and to generate the different lineages of the stomach in vivo4,11.

Research on human gastric stem cells is currently limited. The analysis of spontaneous mutations in the Cytochrome c oxidase gene has shown that some, but not all, human gastric units are monoclonal, allowing the conclusion that at one point in life, multipotent stem cells have resided in these units3. However, direct evidence for the presence of multipotent gastric stem cells into adulthood is lacking.

One of the major functions of the gastrointestinal epithelium is to shield the body from infections and maintain a peaceful coexistence with the gut commensals. Studies on host-pathogen or host-commensal interactions rely on the use of established model systems such as infection of animals or cancer cell lines12, but for many pathogens and commensals, such model systems have not been established.

The gastric pathogen Helicobacter pylori (H. pylori) is one of the most successful pathogens. It uses a range of biological strategies to ensure persistency which enables it to colonize the stomach of about half of the world’s population13. Chronic infection can cause gastric ulcers and gastric cancer13. Currently, in vivo experimental studies use rodent models to understand H. pylori-infection. While mouse studies are certainly useful, the clinical outcome of infection in mice is usually a mild gastritis that does not progress to ulceration or cancer. Alternatively, the Mongolian gerbil can develop cancer after H. pylori infection, but these animals are outbread and the study of host factors is therefore limited12. Other studies use gastric cancer cell lines which typically harbor oncogenic mutations. Human primary cells would represent the gastric epithelium much more closely, but current techniques are limited to isolation of differentiated (mostly mucous) cells that are not able to self-renew and thus can only be maintained for some days14–16. There exists no expanding primary gastric culture system that enables research of primary human gastric cells.

Here we present a gastric culture system that allows indefinite (> 1 year) expansion of human gastric cells. The cultures differentiate into the gastric lineages and can be used as tool to study stem cell biology as well as the response of the epithelium to infection.

Materials and Methods

Human tissue material

Human corpus tissue was obtained from 17 patients (12 men, 5 women, age range 41–87 years) that underwent partial or total gastrectomy at the University Medical Centre Utrecht. 10 patients were diagnosed with gastric cancer and 7 with esophageal cancer. This study was approved by the ethical committee of the University Medical Centre Utrecht. Samples were obtained with informed consent.

Organoid culture

A detailed protocol for gastric culture is provided in the supplement. Briefly, glands were extracted from 1 cm2 of human tissue using EDTA in cold chelation buffer17, seeded in Matrigel (BD Biosciences) and overlaid with medium containing Advanced Dulbecco’s modified Eagle medium/F12 supplemented with penicillin/streptomycin, 10 mmol/L HEPES, Glutamax, 1xB27 (all from Invitrogen), N-Acetylcysteine 1 mM (Sigma-Aldrich). To the basal medium, growth factors were added as indicated in the figures. Final human stomach culture medium contained essential components EGF 50 ng/mL (Invitrogen), Noggin conditioned medium 10%, R-spondin1 conditioned medium 10%, Wnt conditioned medium 50%, FGF10 200 ng/ml (Peprotech), Gastrin 1 nM (Tocris), TGFβi 2 µM (A-83-01, Tocris). Facultative component is Nicotinamide 10 mM (Sigma-Aldrich). After seeding RHOKi 10 µM (Y-27632, Sigma-Aldrich) was added. Additional tested components were: IGF 100 ng/mL (Peprotech), p38 inhibitor 10 µM (SB202190, Sigma-Aldrich), GSK3β inhibitor 3 µM (CHIR99021, Axon Medchem), PGE2 500 nM (Tocris).

Approximately 1 cm2 of cancer tissue was cut in small fragments and washed in cold chelation buffer until supernatant was clear. Fragments were subjected to enzymatic digestion by collagenase 1,5 mg/mL (Gibco) and hyaluronidase 20 µg/mL (Sigma) in 10 mL Advanced DMEM F12 (GIBCO) supplemented with antibiotics (Primocin, Invivogen) for 1 h at 37°C with shaking. Cells were washed twice in Advanced DMEM F12, seeded into Matrigel and overlayed with medium containing HEPES, Glutamax, Penicilline, Streptomycine, B27, n-Acetylcysteine, EGF, R-spondin1, Noggin, Wnt, FGF10, Gastrin, TGFβ-inhibitor and RHOK-inhibitor as above.

Bacterial culture and infection

Bacterial strains and culture conditions are specified in the supplement. For infection studies, organoids were seeded in 50 µL Matrigel in 4 well multidishes (Thermo Scientific). Antibiotics-free medium was refreshed every 2–3 days, with a minimum of 3 medium changes before infection to allow removal of antibiotics from the culture. Organoids were microinjected on day 10 after seeding with an approximate multiplicity of infection (MOI) of 50 unless otherwise stated. For calculation of MOI, organoids were disrupted into single cells by EDTA and cells counted (approximately 4000 cells per organoid). To achieve a final MOI of 50, bacteria were suspended in Advanced DMEM F12 at a density of 1×109/mL and organoids were injected with approximately 0.2 µL bacterial suspension using a micromanipulator and microinjector (M-152 and IM-5B, both Narishige) under a stereomicroscope (Leica MZ75) inside a sterile bench (CleanAir). For viability test, organoids with injected bacteria were picked and each organoid was lysed in 200 µL BHI medium containing 0.5% saponin for 15 minutes with repeated pipetting. 10 µL of 1:10 dilution rows were plated on horse serum agar plates. For heat inactivation, bacteria were subjected to 56°C for 1 h. To test inflammatory stimuli, organoids were incubated with medium containing the following substances in the final concentrations: LPS from E. coli (Invivogen, 1 µg/mL), recombinant human TNFα (BD Pharmingen, 10 ng/mL), recombinant human IL-1β (Sigma-Aldrich 100 ng/mL), CpG ODN 1668 (Enzo, 1 µg/mL), Flagellin from S. typhimurium (Invivogen, 100 ng/mL).

The reader is referred to supplementary methods for FACS, PCR and microarray, cell viability assay, karyotyping, histology and imaging.

Results

Establishment of human stomach cultures

To generate a culture system for human gastric epithelium, we isolated gastric glands from human gastric corpus tissue (figure 1A) and observed their growth under different culture conditions. We started from the conditions for mouse gastric epithelium4, containing EGF, Noggin, R-spondin1, Wnt, FGF10 and Gastrin (termed ENRWFG). Isolated glands from human donors could form organoids in these conditions with very low efficiency and with a limited life span in vitro.

Figure 1. Human gastric cultures expand in vitro.

Figure 1

Open in a new tab

(A) Scheme and image of gland isolation. (B) Nicotinamide (Nic) increases formation of organoids, while p38 inhibitor and TGFβ inhibitor do not. Bars represent average of triplicates of one culture with standard deviation. (C) Effect of several growth factors and inhibitors on the human gastric culture. TGFβi increases lifespan of organoids up to 30 weeks. (D) Removal of any of the factors EGF, noggin, R-spondin1, Wnt, FGF10, gastrin and TGFβi limits the growth of the culture. Removal of nicotinamide enables long term growth (> 1 year). In B and D, each bar represents a newly established culture. (E) Example of an organoid growing from a human gland. The insert shows a budding structure. (F) Karyogram after 3 months in culture. Scale bars 100 µm.

We then tested a panel of growth factors and inhibitors for organoid-forming efficiency, phenotype of the organoids, and longevity of the human gastric cultures. TGFβ inhibitor, p38 inhibitor, GSK2β inhibitor and PGE2 were chosen because of the relevance of the respective pathways in cancer. IGF is expressed in normal gastric tissue10. Nicotinamide suppresses Sirtuin activity19. Similar as for human intestine17, Nicotinamide increased the number of human gastric organoids formed (figure 1B and supplementary figure 1A). It was therefore included in the subsequent culture condition. IGF, p38 inhibitor, GSK3β inhibitor and TGFβ inhibitor all induced budding structures in a concentration-dependent manner (supplementary figure 1B) and had a positive effect on the lifespan of the organoids (figure 1C). PGE2 induced growth of large cysts and also prolonged the lifespan of the cultures. Addition of TGFβ inhibitor increased the life span to a maximum of half a year (figure 1C), while all other factors had no such effect. We therefore only added TGFβ inhibitor to the ENRWFG culture medium.

To analyze the importance of the single factors, we then withdrew each of the components from the medium. Without EGF, Noggin, R-spondin1 or Wnt, organoid formation was strongly reduced and cultures deteriorated within 1–3 weeks (figure 1D and supplementary figure 1C). Removal of FGF10, Gastrin or TGFβ inhibitor allowed growth for 10–20 weeks. Removal of Nicotinamide increased the lifespan of the cultures (figure 1D). Thus, addition of Nicotinamide promoted initial organoid formation, but limited the lifespan of the cultures. We therefore used it as a facultative culture component.

The cultures developed in a stereotypical manner: After seeding, glands sealed and formed small cysts which subsequently expanded. Many organoids initially stayed cystic. With expansion of the culture, organoids became more uniform and consisted of several buddings that surrounded a central lumen (Figure 1E). Cultures were grown for 1 year with bi-weekly splitting rates of 1:5 without losing any of the features described. After 3 months of culture, chromosomal metaphase spreads of 2 patients were obtained and either 15 or 6 karyograms were aligned. There was no indication of chromosomal aberrations (figure 1F). Organoids described there have all been generated from corpus tissue. However, organoids can also be generated from cardia or pyloric antrum and expand similarly under the here described culture conditions (3 months tested).

Human gastric organoids differentiate into gastric lineages and self-organize into gland and pit domains

We then analysed the cellular composition of the organoids in the culture condition for optimal longevity (ENRWFG_Ti). PCR indicated that the organoids expressed the stem cell marker LGR5 as well as the gastric epithelial markers mucin 5AC (MUC5AC), pepsinogen (PGC), somatostatin (SST), mucin 6 (MUC6), trefoil factor 1 (TFF1) and trefoil factor 2 (TFF2). As expected for gastric cultures, they did not express the intestinal markers mucin 2 (MUC2), caudal type homeobox (CDX) 1 and CDX2 (figure 2A). As expected for organoids derived from the corpus region of the stomach, the antral markers gastrin and PDX1 were not expressed according to microarray analysis comparing organoids to corpus and pyloric glands. Transcriptional profiling also indicated that markers of parietal cells and ECL cells, which are usually present in human corpus tissue, are not expressed in the organoids (microarray available online).

Figure 2. Human gastric organoids differentiate into four gastric lineages and self-organize into gland and pit domains.

Figure 2

Open in a new tab

(A) mRNA expression of gastric marker genes. LGR5 was used as marker for stem cells, MUC5AC for gastric pit mucous cells, PGC for chief cells, SST for enteroendocrine cells, MUC6 for gland mucous cells. Gastric trefoil factors 1 (TFF1) and TFF2 are also expressed. Housekeeping gene GAPDH was used as loading control. MUC2, CDX1 and CDX2 are markers of intestinal tissue and were used to control for tissue specificity. (B) Images of stained paraffin sections of gastric mucosa and organoids. Periodic acid-Schiff (PAS) staining also marks pit mucous cells. (C) Confocal images of wholemount organoids. (D) Proliferating cells are labeled by 5-ethynyl-2’-deoxyuridine (EdU) incorporation. Scale bar 100 µm. Culture condition: ENRWFG_Ti.

Histological staining of paraffin sections as well as immunofluorescence staining of whole organoids revealed remarkable organization. MUC5AC- and MUC6-positive mucous cells divided the organoids into gland -and pit-domains: while the budding structures consisted mostly of MUC6-positive mucous gland cells, the central lumen was lined with MUC5AC-positive mucous pit cells. PGC-positive chief cells and rare SST-positive enteroendocrine cells were scattered throughout the organoid (figure 2B and C). Staining for H-K-ATPase was negative, confirming the absence of parietal cells (figure 2B). EdU staining showed the presence of proliferative cells dispersed through the organoid (figure 2D).

Directed differentiation of human gastric organoids to gland or pit lineages

In the gastric mucosa, stem cells reside in the glands and produce progenitors that differentiate into pit cells as they migrate upwards to the pit4. In the mouse stomach, expression of Wnt target genes (such as Troy, Lgr5 and Axin2) occurs in a gradient with high expression in the gland bottom and no expression in the pit11. We therefore hypothesized, that it should be possible to generate organoids that only resemble the gland domains, while the progenitors can be directed towards the pit lineage by manipulating Wnt signal strength.

Comparing cultures with and without Nicotinamide, staining of paraffin section showed that the major effect of Nicotinamide was the prevention of differentiation into MUC5AC-positive pit cells (supplementary figure 2). Thus, the condition ENRWFGNiTi generated organoids that lack the pit-domain and only resemble the gland-domains.

To direct these gland-type organoids to the pit lineage, we used a two-step protocol: Organoids were grown for 10 days in the full medium (ENRWFGNiTi) and then Wnt was withdrawn from the medium for 4 days to allow differentiation. During the differentiation phase, organoids underwent a phenotypical change, in becoming more cystic with less pronounced glands (figure 3A). To globally assess the effect of Wnt withdrawal, we performed microarray analysis. As expected, Wnt was necessary for the expression of known stem cell markers such as LGR5 and TROY (figure 3B). Moreover, removal of Wnt led to a decrease in expression of the chief cell marker PGC and the mucous neck cell marker MUC6. In turn, expression of the mucous pit cell marker MUC5AC was upregulated (figure 3B). The regulation of known Wnt pathway targets (LGR5, TROY, AXIN2, CD44 (Ref11)) as well as the expression of PGC, MUC6 and MUC5AC was confirmed by quantitative PCR (figure 3C) and conventional PCR (figure 3D). Gastric TFF1 and 2 are also expressed (figure 3D). Markers of intestinal tissue (MUC2, CDX1, CDX2) were not expressed in organoids irrespective of the treatment (figure 3D).

Figure 3. Directed differentiation of organoids into pit or gland cell lineages.

Figure 3

Open in a new tab

(A) Examples of the phenotypical change upon Wnt withdrawal. Cultures were grown in ENRWFGNiTi for 10 days and subsequently either kept with Wnt (left panel) or without Wnt (right panel) for 4 days. Cultures under Wnt withdrawal loose budding structures and become large cysts. (B) Differential gene expression in the two growth conditions measured by genome-wide microarray of cultures from 3 donors. Removal of Wnt reduces expression of stem cell markers (LGR5, TROY), and increases expression of MUC5AC. (C) Quantitative PCR of known Wnt-response genes and stomach specific genes. mRNA was normalized to GAPDH housekeeping gene. Bars represent normalized average of triplicates with standard deviation. (D) PCR of gastric and intestinal markers. Intestinal markers MUC2, CDX1 and CDX2 are not expressed. (E) Images of stained paraffin sections show differential expression of MUC5AC and MUC6. PGC is expressed in both types of organoids. (F) Scheme of gastric gland and the two types of organoids. In all experiments organoids were grown as described in A. Scale bar 100 µm.

Staining of paraffin sections revealed two distinct types of organoids: With Wnt, organoids resembled glands with MUC6-positive mucous gland cells in the budding and high numbers of PGC-positive chief cells but virtually no MUC5AC positive pit cells (figure 3E left panel). Without Wnt, organoids had high numbers of MUC5AC pit-cells, fewer PGC-positive chief cells and only occasional MUC6 positive gland structures (figure 3E right panel). SST-positive enteroendocrine cells were very rare in all conditions. Quantifications of the 4 cell lines in the 3 conditions confirmed the changes in cellular compositions of the organoids (supplementary figure 3). Thus, human gastric organoids can be directed into gland- or pit-type organoids, suggesting a potential role for a Wnt-gradient in human gastric homeostasis (figure 3F).

In summary, we can generate 3 different types of organoids that mostly differ in the composition of mucous producing cells: (1) Ever-expanding cultures of organoids that comprise four gastric lineages organized into gland- and pit domains ("complete-type"), in ENRWFG_Ti medium (2) organoids with only gland domains ("gland-type") in ENRWFGNiTi medium and (3) organoids that consists of high numbers of pit cells ("pit-type") in ENR_FGNiTi medium.

Single gastric stem cells can generate organoids that differentiate into four lineages of the stomach

To analyze the differentiation capacity of putative human gastric stem cells, we sorted single cells from gastric mucosa. For this, we plotted the cells using forward scatter area versus forward scatter peak linear and gated on the single cell population (figure 4A). The quality of the sort was confirmed by microscopic analysis. To increase organoid forming efficiency and to avoid anoikis, we used Nicotinamide and Rho kinase inhibitor for this experiment. 0.1% of sorted cells formed organoids (figure 4B) that could be expanded at a 1:5 ratio on a bi-weekly basis (figure 4C). They expressed the gastric markers MUC5AC, PGC, SST, MUC6, TFF1, TFF2, but not intestinal markers (MUC2, CDX1, CDX2) as shown by PCR (figure 4D). The cellular composition of single-cell derived organoids was very similar to the one of gland-derived organoids as shown by IHC. In the presence of Wnt and Nicotinamide organoids contain PGC-positive chief cells, MUC6-positive mucous neck cells and very rare SST-positive enteroendocrine cells (“gland type organoids”). After 4 days of Wnt withdrawal, MUC5AC-positive mucous pit cells appear (“pit-type organoids”) (figure 4E). Quantifications corroborated these results (supplementary figure 3). In summary, we did not observe any differences between single-cell- and gland-derived organoids in terms of longevity, expansion rate, marker gene expression or composition of cell types. Cultures shown in figure 4E are 7 months old, demonstrating that the different cell lineages are maintained over time. We conclude that the single cells behaved as multipotent stem cells.

Figure 4. Single stem cells from human gastric mucosa can form gastric organoids.

Figure 4

Open in a new tab

(A) Scheme of isolation and FACS plot. FACS was used to generate single cells based on FSC. (B) Example of an organoid growing from a single cell. (C) Example of two single cell clones that were cultured for 1 year (here the first 3 passages = 6 weeks). (D) mRNA expression of gastric marker genes. Intestinal markers MUC2, CDX1 and CDX2 are not expressed and confirm tissue specificity. (E) Images of stained paraffin sections of organoids. MUC5AC expression increases after Wnt removal. Culture condition: ENRWFGNiTi (“Wnt”) or 4 days of Wnt withdrawal (“no Wnt”). Scale bar 100 µm.

In vitro expansion of human gastric cancer organoids

Treatment of gastric cancer patients depends on availability of tests for drug discovery and sensitivity. Currently, gastric cancer cell lines are available, but no system allows comparison of cancerous and normal cells from the same patient. Having established the culture condition for human normal organoids, we reasoned that human gastric tumors could also be expanded under the same conditions. To establish the culture, cells were isolated from the tumor using collagenase and hyaluronidase, seeded into Matrigel and embedded in ENRWFG_A medium. In parallel, organoids were established from normal tissue (figure 5A). Chromosomal metaphase spreads were obtained from the tumor organoids (figure 5B). 7 spreads were counted and showed aneuploidy with chromosome numbers between 70 and 160. Tumor cultures were reminiscent of the original tissue in terms of morphology shown by Hematoxylin and Eosin (H&E) staining and p53 accumulation as shown by p53 staining (figure 5C upper panel). To further analyze the possible mutation of the p53-pathway in this tumor, we used nutlin-3, which inhibits the interaction between p53 and MDM2 and thereby induces cell cycle arrest. Nutlin-3 requires functional p53 and MDM2 for its activity, thus cancer cells with mutated p53 are not affected by this compound20. As expected, the normal organoids were strongly inhibited in their growth by nutlin-3 as quantified by a luciferase-based assay. In contrast, tumor organoids were insensitive to nutlin-3 treatment, indicating mutated p53 pathway in this tumor (figure 5D).

Figure 5. Normal and tumor organoids.

Figure 5

Open in a new tab

(A) Brightfield image of tumor organoids (left) and normal organoids (right) established from the same patient. (B) Metaphase spread the tumor organoids shown in A. Counts of 7 spreads showed 70–160 chromosomes. (C) Images of stained paraffin sections of the original tumor tissue and organoids shown in A. (D) Organoids from tumor are insensitive to nutlin-3. Organoids were grown for 1 week with indicated concentrations of nutlin-3. Cell viability was assessed using luciferase assay. Bars represent average of triplicates with standard deviation. Scale bar 100 µm

Infection of human gastric cultures with Helicobacter pylori induces inflammatory response

In an infected individual, H. pylori colonizes the lumen of the stomach and has there contact with the apical side of the epithelium. In the organoids, the apical side of the polarized epithelium faces the lumen of the 3D structure. To enable bacteria to reach the natural side of infection, we established microinjection of the organoids (figure 6A left). Injection was confirmed by microscopy using GFP expressing H. pylori and E-cadherin as epithelial counterstain (figure 6A right). Plating of bacteria from organoids 2 h after injection verified that the bacteria are alive inside the organoids (supplementary figure 4 A and B). Electron microscopy showed that bacteria were engaged in very intimate contact with the epithelial cells (figure 6B).

Figure 6. Primary response of gastric organoids to H. pylori infection.

Figure 6

Open in a new tab

(A) Scheme and brightfield image of microinjection of H. pylori into an organoid (left). Confocal image of infected organoid (Lumen infected, LI) and non-infected organoid (NI) (right). Bacteria were visualized with using GFP-expression, organoids were counterstained with E-Cadherin (red) and DAPI (blue). (B) Electron microscopy demonstrates close interaction of bacteria with epithelial cells. Arrows indicate the binding site (C) Differential gene expression upon infection of human gastric organoids. Genome-wide microarray was performed on 3 independent experiments using organoids derived from 3 patients. (D) Confocal images of infected (LI) and non-infected organoids (NI). Staining for NF-κB subunit p65 (red) indicates nuclear p65 in the infected organoid. (E) qPCR for IL-8 mRNA after injection of H. pylori or medium (mock). Bars represent average of duplicates with standard deviation. (F) Organoids were differentiated into gland-type organoids (with Wnt) or pit-type organoids (4 days Wnt withdrawal) and subsequently microinjected with H. pylori. After indicated time, mRNA was extracted and expression of IL-8 was measured with quantitative real time PCR. Bars represent average of duplicates with standard deviation. LI = Lumen infected organoid, NI = Noninfected organoid. Scale bar 200 nm (C) or 100 µm (B, E). * = Matrigel.

To determine the global primary response of the infected epithelium, we used microarray analysis. After 2 h of infection, 25 genes were regulated 2-fold with a p-value < 0.5 (supplementary table 1). The highest upregulated gene was human chorionic gonadotropin beta (CGB), a gene that has been associated with gastric cancer21. Many other highly upregulated genes were targets of the NF-κB pathway (figure 6C), known to be activated in H. pylori infection18,22,23. To test whether indeed this pathway was activated, we stained the NF-κB subunit p65 and found that after one hour of infection, p65 was translocated to the nuclei of the cells in infected organoids. Of note, neighbouring organoids that did not contain bacteria did not show any p65 nuclear translocation (figure 6D). A well-known target of NF-κB is the chemokine IL-8 that attracts neutrophils and thereby promotes the inflammation24. Microarray analysis had already indicated that IL-8 was upregulated in the organoids. Quantitative PCR of IL-8 confirmed this. IL-8 was not upregulated in control organoids that were injected only with medium (mock) (figure 6E). Induction of IL-8 depended on the MOI (supplementary figure 4C). In human epithelial cell lines NF-κB activation depends on the cytotoxicity associated gene pathogenicity island (cagPAI) of the bacteria18,23. In human gastric organoids, IL-8 expression neither depended on the cagPAI, nor on bacterial viability (supplementary figure 4D). Three cagPAI-independent stimuli have been reported to activate NF-κB via toll like receptors (TLRs) in H. pylori infection: LPS, Flagellin, or bacterial DNA23. Human gastric organoids are inert to purified LPS or CpG oligodeoxynucleotides (ODN) (supplementary figure 4D), while these substances induce IL-8 in other cells (data not shown). In contrast, organoids mount a strong IL-8 response when incubated with purified flagellin or controls TNFα and IL-1β (supplementary figure 4D). Thus, generally, the organoids react to flagellin but are inert to LPS and CpG ODN. To analyse the importance of H. pylori flagellum, we microinjected an aflagellated bacterial mutant. This mutant still induced IL-8 expression indicating that the bacterial flagellum is not the major inducer of IL-8 in this system (supplementary figure 4E).

We then asked whether specific cells in the organoids respond to the bacteria and used our differentiation protocol to generate gland-type or pit-type organoids which we subsequently microinjected with H. pylori. IL-8 expression was substantially higher in gland-type organoids than in pit-type organoids (figure 6F).

Discussion

Here we present a long-term 3D organoid culture system for primary, untransformed human gastric epithelium as well as human gastric cancer. Using this culture, we provide direct evidence for the presence of stem cells in adult human gastric tissue. The cells can be directed to differentiate into specific lineages of the stomach. The organoids mount an NF-κB-driven inflammatory response to infection and the strength of this response depends on the differentiated cell types in the organoids.

The presence of stem cells in the human adult stomach is expected, yet has not been demonstrated previously. The organoids we present here can be grown from FACS-isolated single cells and generate four lineages of the stomach: Pit mucous cells, gland mucous cells, chief cells and enteroendocrine cells. Of the enteroendocrine cells, we identified SST-expressing cells, but not corpus-specific ECL cells. We could also not detect parietal cells. We assume the culture conditions are not optimal to allow differentiation into these cell types. Once clonal organoids are established, they expand without apparent limitation (> 1 year), defying the Hayflick limit. Thus, the isolated cells can self-renew, are long-lived and multipotent, fulfilling the classical criteria for stem cells.

In the intestine, the pathological activation of the Wnt pathway in cancer represents a deregulation of the controlled activation necessary for normal stem cell driven tissue homeostasis25. In the stomach, the role of the Wnt pathway is less clear. Up to 30% of gastric tumours are found to carry an activated Wnt pathway26,27, while mutations in the Wnt pathway drive tumorigenesis in the mouse4,28. Two of the known stem cell markers in the mouse stomach, Troy and Lgr5, are Wnt target genes4,11. Here we provide additional evidence for the importance of the Wnt pathway in human gastric epithelium. First, establishment and growth of human gastric organoids depends on Wnt and R-Spondin1. Second, upon withdrawal of Wnt, organoids differentiate into pit-lineage cultures. In the intestine, the Wnt-secreting Paneth cells provide the niche for the stem cells17 and competition for niche space determines the fate of the stem cell daughter cells5,6. It seems likely that there is a Wnt-source at the bottoms of gastric glands and that the migration of daughter cells upwards towards the gastric surface directs the differentiation into the pit-lineage.

We use human organoids to analyse the primary response of the human epithelium to H. pylori and find robust NF-κB activation. H. pylori has been shown to activate this transcription factor in various human and murine cell lines18,22,23,29. Also, a study using short-lived human primary mucous cells showed induction of IL-8 (Ref30). Results here indicate that in human organoids, IL-8 expression is independent of viability of the bacteria and independent of TLR4, 5 and 9 signalling. Further studies need to analyze the precise signalling pathways leading to NF-κB activation in this system. The human organoids allow us further to compare the NF-κB response in cells of the pit and the gland lineages. We find that the gland lineages respond with higher amounts of IL-8 than the pit lineage. This is in line with earlier studies that analysed the importance of bacterial chemotaxis in infection. These studies found that wild-type bacteria can colonize the gastric glands, but bacterial mutants with defects in chemotaxis were only able to colonize the surface mucus. After months of infection, the bacteria in the glands had induced a higher inflammation and T cell response than the bacteria in the surface layer31,32. Our finding is also in line with the general idea that the gastroepithelial lining protects itself from chronic inflammation by creating a certain “blindness” on the surface33. Two mechanisms are likely to underlie the relatively low response of the gastric surface cells observed here: (i) The surface cells promote physical separation from the bacteria by forming a thick mucus layer. (ii) The host restricts receptors initiating the NF-κB response to the deeper glands which should be less in contact with bacteria33,34. Future research has to determine whether one or both (or a now not anticipated mechanism) restricts the pit cell inflammatory response.

In summary, the organoids described here present a new model of self-renewing gastric epithelium grown from stem cells that can be directed into the different lineages of the stomach. It represents a model that is much closer to the gastric epithelium than currently used cell lines. Organoids can be grown from surgical resections as well as from biopsies and can be expanded without apparent growth limitation. This method also allows to grow parallel samples from normal as well as cancerous gastric cells from the same patient. This will enable their use for future patient-derived disease models, drug screens, gastric stem cell research and for the study of host pathogen interactions.

Supplementary Material

01

Acknowledgements

We are very thankful to the patients that allowed us to perform this study and to the Biobank of the UMC Utrecht for providing us with patient material. We thank Thomas F. Meyer for bacterial strains and for initial support with bacterial infection. We thank Karen Otteman for bacterial strains. We thank Jeroen Korving, Margriet Westerveld and the Hubrecht imaging Center for technical assistance.

Grant support: This work was supported by an EU Marie Curie Fellowship (EU/300686-InfO) to S.B.; EU Marie Curie Fellowship (EU/236954-ICSC-Lgr5) to M.H.; EU/232814-StemCellMark to M.v.d.W. and R.V., NIH/MIT Subaward 5710002735 to H.B.; Research Prize of the United European Gastroenterology Foundation to H.C..

Abbreviations

ENRWFGNiTi

EGF+R-spondin1+Noggin+Wnt+FGF10+Gastrin+Nicotinamide+TGFβ inhibitor culture condition

H. pylori

Helicobacter pylori

Footnotes

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 citable 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.

Disclosures: We disclose that H.C. and M.H. hold patents for organoid culture. Otherwise the authors disclose no conflict of interest.

Transcript profiling: Microarray data is available on the GEO database accession number GSE60557

Author contribution: S.B. study design; acquisition of data; analysis and interpretation; writing of the manuscript; T.B. acquisition of data; M.v.d.W. analysis of data; M.H. support during inital phase of the project; R.V. ethical approval; H.B. Histology; P.K. EM studies; P.J.P. supervision of EM studies; H.C. supervision and writing of the manuscript.

References

  • 1.Karam SM, Leblond CP. Dynamics of epithelial cells in the corpus of the mouse stomach. I. Identification of proliferative cell types and pinpointing of the stem cell. Anat. Rec. 1993;236:259–279. doi: 10.1002/ar.1092360202. [DOI] [PubMed] [Google Scholar]
  • 2.Bjerknes M, Cheng H. Multipotential stem cells in adult mouse gastric epithelium. Am. J. Physiol.- Gastrointest. Liver Physiol. 2002;283:G767–G777. doi: 10.1152/ajpgi.00415.2001. [DOI] [PubMed] [Google Scholar]
  • 3.McDonald SAC, Greaves LC, Gutierrez-Gonzalez L, et al. Mechanisms of Field Cancerization in the Human Stomach: The Expansion and Spread of Mutated Gastric Stem Cells. Gastroenterology. 2008;134:500–510. doi: 10.1053/j.gastro.2007.11.035. [DOI] [PubMed] [Google Scholar]
  • 4.Barker N, Huch M, Kujala P, et al. Lgr5+ve Stem Cells Drive Self-Renewal in the Stomach and Build Long-Lived Gastric Units In Vitro. Cell Stem Cell. 2010;6:25–36. doi: 10.1016/j.stem.2009.11.013. [DOI] [PubMed] [Google Scholar]
  • 5.Snippert HJ, Flier LG van der, Sato T, et al. Intestinal Crypt Homeostasis Results from Neutral Competition between Symmetrically Dividing Lgr5 Stem Cells. Cell. 2010;143:134–144. doi: 10.1016/j.cell.2010.09.016. [DOI] [PubMed] [Google Scholar]
  • 6.Ritsma L, Ellenbroek SIJ, Zomer A, et al. Intestinal crypt homeostasis revealed at single-stem-cell level by in vivo live imaging. [[Accessed March 12, 2014]];Nature. 2014 doi: 10.1038/nature12972. Available at: http://www.nature.com/doifinder/10.1038/nature12972. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Leushacke M, Ng A, Galle J, et al. Lgr5+ Gastric Stem Cells Divide Symmetrically to Effect Epithelial Homeostasis in the Pylorus. Cell Rep. 2013;5:349–356. doi: 10.1016/j.celrep.2013.09.025. [DOI] [PubMed] [Google Scholar]
  • 8.Qiao XT, Ziel JW, McKimpson W, et al. Prospective Identification of a Multilineage Progenitor in Murine Stomach Epithelium. Gastroenterology. 2007;133:1989–1998. doi: 10.1053/j.gastro.2007.09.031. .e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Arnold K, Sarkar A, Yram MA, et al. Sox2+ Adult Stem and Progenitor Cells Are Important for Tissue Regeneration and Survival of Mice. Cell Stem Cell. 2011;9:317–329. doi: 10.1016/j.stem.2011.09.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Mills JC, Andersson N, Hong CV, et al. Molecular characterization of mouse gastric epithelial progenitor cells. Proc. Natl. Acad. Sci. 2002;99:14819–14824. doi: 10.1073/pnas.192574799. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Stange DE, Koo B-K, Huch M, et al. Differentiated Troy+ Chief Cells Act as Reserve Stem Cells to Generate All Lineages of the Stomach Epithelium. Cell. 2013;155:357–368. doi: 10.1016/j.cell.2013.09.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.O’Rourke JL, Lee A. Animal models of Helicobacter pylori infection and disease. Microbes Infect. 2003;5:741–748. doi: 10.1016/s1286-4579(03)00123-0. [DOI] [PubMed] [Google Scholar]
  • 13.Salama NR, Hartung ML, Müller A. Life in the human stomach: persistence strategies of the bacterial pathogen Helicobacter pylori. Nat. Rev. Microbiol. 2013;11:385–399. doi: 10.1038/nrmicro3016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Takahashi M, Ogura K, Maeda S, et al. Promoters of epithelialization induce expression of vascular endothelial growth factor in human gastric epithelial cells in primary culture. FEBS Lett. 1997;418:115–118. doi: 10.1016/s0014-5793(97)01354-9. [DOI] [PubMed] [Google Scholar]
  • 15.Richter-Dahlfors A, Heczko U, Meloche RM, et al. Helicobacter pylori-infected human antral primary cell cultures: effect on gastrin cell function. Am. J. Physiol. 1998;275:G393–G401. doi: 10.1152/ajpgi.1998.275.3.G393. [DOI] [PubMed] [Google Scholar]
  • 16.Krueger S, Hundertmark T, Kuester D, et al. Helicobacter pylori alters the distribution of ZO-1 and p120ctn in primary human gastric epithelial cells. Pathol. - Res. Pract. 2007;203:433–444. doi: 10.1016/j.prp.2007.04.003. [DOI] [PubMed] [Google Scholar]
  • 17.Sato T, Stange DE, Ferrante M, et al. Long-term Expansion of Epithelial Organoids From Human Colon, Adenoma, Adenocarcinoma, and Barrett’s Epithelium. Gastroenterology. 2011;141:1762–1772. doi: 10.1053/j.gastro.2011.07.050. [DOI] [PubMed] [Google Scholar]
  • 18.Bartfeld S, Hess S, Bauer B, et al. High-throughput and single-cell imaging of NF-kappaB oscillations using monoclonal cell lines. BMC Cell Biol. 2010;11:21. doi: 10.1186/1471-2121-11-21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Denu JM. Vitamin B3 and sirtuin function. Trends Biochem. Sci. 2005;30:479–483. doi: 10.1016/j.tibs.2005.07.004. [DOI] [PubMed] [Google Scholar]
  • 20.Tovar C, Rosinski J, Filipovic Z, et al. Small-molecule MDM2 antagonists reveal aberrant p53 signaling in cancer: implications for therapy. Proc. Natl. Acad. Sci. U.S.A. 2006;103:1888–1893. doi: 10.1073/pnas.0507493103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Yakeishi Y, Mori M, Enjoji M. Distribution of beta-human chorionic gonadotropin-positive cells in noncancerous gastric mucosa and in malignant gastric tumors. Cancer. 1990;66:695–701. doi: 10.1002/1097-0142(19900815)66:4<695::aid-cncr2820660418>3.0.co;2-n. [DOI] [PubMed] [Google Scholar]
  • 22.Viala J, Chaput C, Boneca IG, et al. Nod1 responds to peptidoglycan delivered by the Helicobacter pylori cag pathogenicity island. Nat. Immunol. 2004;5:1166–1174. doi: 10.1038/ni1131. [DOI] [PubMed] [Google Scholar]
  • 23.Ferrero RL. Innate immune recognition of the extracellular mucosal pathogen, Helicobacter pylori. Mol. Immunol. 2005;42:879–885. doi: 10.1016/j.molimm.2004.12.001. [DOI] [PubMed] [Google Scholar]
  • 24.Shimoyama T, Crabtree J. Bacterial factors and immune pathogenesis in Helicobacter pylori infection. Gut. 1998;43:S2–S5. [PMC free article] [PubMed] [Google Scholar]
  • 25.Clevers H. The Intestinal Crypt, A Prototype Stem Cell Compartment. Cell. 2013;154:274–284. doi: 10.1016/j.cell.2013.07.004. [DOI] [PubMed] [Google Scholar]
  • 26.Horii A, Nakatsuru S, Miyoshi Y, et al. The APC gene, responsible for familial adenomatous polyposis, is mutated in human gastric cancer. Cancer Res. 1992;52:3231–3233. [PubMed] [Google Scholar]
  • 27.Clements WM, Wang J, Sarnaik A, et al. β-Catenin mutation is a frequent cause of Wnt pathway activation in gastric cancer. Cancer Res. 2002;62:3503–3506. [PubMed] [Google Scholar]
  • 28.Radulescu S, Ridgway RA, Cordero J, et al. Acute WNT signalling activation perturbs differentiation within the adult stomach and rapidly leads to tumour formation. Oncogene. 2013;32:2048–2057. doi: 10.1038/onc.2012.224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Bauer B, Moese S, Bartfeld S, et al. Analysis of Cell Type-Specific Responses Mediated by the Type IV Secretion System of Helicobacter pylori. Infect. Immun. 2005;73:4643–4652. doi: 10.1128/IAI.73.8.4643-4652.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Ogura K, Takahashi M, Maeda S, et al. Interleukin-8 Production in Primary Cultures of Human Gastric Epithelial Cells Induced by Helicobacter pylori. Dig. Dis. Sci. 1998;43:2738–2743. doi: 10.1023/a:1026671815512. [DOI] [PubMed] [Google Scholar]
  • 31.Howitt MR, Lee JY, Lertsethtakarn P, et al. ChePep Controls Helicobacter pylori Infection of the Gastric Glands and Chemotaxis in the Epsilonproteobacteria. mBio. 2011;2:e00098-11–e00098-11. doi: 10.1128/mBio.00098-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Rolig AS, Carter JE, Ottemann KM. Bacterial chemotaxis modulates host cell apoptosis to establish a T-helper cell, type 17 (Th17)-dominant immune response in Helicobacter pylori infection. Proc. Natl. Acad. Sci. 2011;108:19749–19754. doi: 10.1073/pnas.1104598108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Pedron T, Sansonetti P. Commensals, bacterial pathogens and intestinal inflammation: an intriguing menage a trois. Cell Host Microbe. 2008;3:344–347. doi: 10.1016/j.chom.2008.05.010. [DOI] [PubMed] [Google Scholar]
  • 34.Abreu MT, Fukata M, Arditi M. TLR signaling in the gut in health and disease. J. Immunol. Baltim. Md 1950. 2005;174:4453–4460. doi: 10.4049/jimmunol.174.8.4453. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

01
NIHMS634587-supplement-01.pdf (486.3KB, pdf)

RESOURCES