Influence of microbiota on intestinal stem cell behaviour and differentiation in inflammatory bowel diseases.

Ulcerative colitis (UC) and Crohn’s disease (CD) are the two major forms of inflammatory bowel diseases (IBD), which are chronic and progressive diseases of the gastrointestinal tract. The exact cause is unknown, but it has been suggested that the diseases result from an abnormal immune response towards environmental factors, such as the microbiota, in a genetically predisposed person. Current therapies aim to stop this inflammatory cascade. The most widely used and successful therapy so far has been anti-tumour necrosis factor (TNF) antibodies, but also drugs with other modes of action are currently in use in the clinic (anti-adhesion molecules) or will soon become available (anti-interleukin (IL-) 12/23 and anti-Janus kinase inhibitors). Many patients are primary non-responders, and another large group of patients develop secondary loss of response over time to these antibodies. Discovering new possible therapeutic targets is therefore of great importance, which is still difficult due to insufficient knowledge of disease pathogenesis.

Microbiota not only play an important part in disease, but also in homeostasis. A dysbiosis in the microbiota has been described in IBD, with an increase in pro-inflammatory species and a lack of anti-inflammatory bacteria. Furthermore, patients with IBD may have decreased mucosal barrier function, resulting in an influx of bacteria towards and in the epithelium. The epithelium consists of crypts and villi, with absence of villi in the colon. All intestinal epithelial cells are generated by Leucine Rich Repeat Containing G Protein-Coupled Receptor 5 (LGR5)+ intestinal stem cells (ISCs), which are located at the bottom of the crypts. ISCs undergo differentiation when they move upward due to cellular drift away from the niche, which provides stem cell factors. In IBD, due to leakiness of the barrier, an interaction may take place between bacteria and ISCs, but little is known about the effect microbiota may have on ISC functioning.

Organoids are an excellent model to study the specific interaction between ISCs and microbiota. Intestinal organoids consist out of ISCs, and grow as self-organizing 3D structures in Matrigel, which constitutes the extracellular matrix. With factors as Wnt family member 3A (WNT3A), R-spondin-1, noggin, epidermal growth factor and others, cells undergo long-term expansion while remaining genetically stable.  In this project, we investigated if organoids may be used in IBD research, as a future model for organoid-microbiota interactions.

First, we established a library of intestinal organoids from patients with IBD and control patients without IBD, by isolating crypts from mucosal biopsies and growing them as organoids. We found no statistically significant differences in organoid forming capacity between controls and UC or CD patients. However, we observed a non-significant trend of decreased organoid forming efficiency of biopsies derived from macroscopically inflamed areas.

To see if organoids retain all biological functions, we cultured established organoids for at least two weeks, and then performed a differentiation by withdrawal of WNT3A, nicotinamide and p38-inhibitor. Under these conditions, ISCs differentiate in all the downstream cell lineages found in the intestinal epithelium. We found that in differentiated organoids, MUCIN2 mRNA levels (a goblet cell marker) were decreased in CD organoids compared to controls. Furthermore, Atonal bHLH Transcription Factor 1 (ATOH1), a transcription factor controlling differentiation of cells to the secretory lineage, was expressed at significantly lower levels in differentiated organoids from both UC and CD patients vs. controls. There was a significant increase in differentiated organoids from UC patients vs. controls for Tight junction protein 1 (TJP1), and a trend for increased expression in CD organoids. This increase in TJP1 is in direct contrast with the increased permeability seen in patients with IBD. It may be that there is a decrease of other tight junction proteins and perhaps these organoids are compensating by up-regulating TJP1. Lastly, when comparing ileum and colonic organoids we noticed, as reported before, that organoids retain their location specificity, as reflected by increased levels of villin-1, alkaline phosphatase (enterocyte marker), and chromogranin A (enteroendocrine cell marker) in ileal samples. Due to heterogeneity in differentiation, we were unable to corroborate the found differences on mRNA levels on the protein level.

Next, we aimed to investigate if organoids have an inflammatory response when challenged with inflammatory stimuli, such as TNFα, or the bacterial compounds muramyl dipeptide (MDP) and lipopolysaccharide (LPS). LPS nor MDP evoked a response in transcriptional activity after 24 hours exposure to three different concentrations of LPS or MDP (0.5 µg/ml, 5 µg/ml, and 10 µg/ml), while 100 ng/ml TNFα induced expression of various genes: C-X-C motif chemokine ligand- (CXCL) 8 and 3, TNFα, Nuclear factor kappa B (NFκB) and two NFκB inhibitors. Our results were in line with previously reported findings that gastric organoids also are unresponsive to LPS, but do express CXCL8 after exposure to flagellin, TNFα, or live Helicobacter pylori culture.

Patients with IBD experience disease flares followed by disease remission. How mRNA levels of inflammatory markers develop over time, from primary tissue to organoids, was unknown. We therefore included patients with active inflammation (macroscopically assed during endoscopy). We obtained normal and inflamed appearing tissue from the same patient, performed crypt isolation and organoid culture, and analysed mRNA expression levels in biopsies and organoids after 14 days in culture. As expected, LGR5 levels were enriched in organoids vs. biopsies. Markers of inflammation, IL1β and TNFα, were generally increased in UC and CD patients’ biopsies. The expression of these markers was decreased in organoids in all patient groups compared to the original biopsies. Strikingly, we observed an increase in CXCL8 for all groups in organoids vs. controls, suggesting that culture conditions induce its expression or it is required for organoid growth. CXCL3 expression levels were comparable to CXCL8 trends. Importantly, overall expression levels in organoids between the different patient groups were not significantly altered. We also performed this for ileal tissues from CD patients and controls, and found mostly similar patterns in de- or increases in the analysed genes. Thus, at the mRNA level, markers of inflammation reach baseline levels in organoids quickly, but it needs to be proven whether this is due to removal from the (inflammatory) environment, or that non-epithelial cells (e.g. immune cells) are responsible for the detected inflammation in biopsies.

Although the organoid model is very suitable to determine the effect of microbiota or its components on ISCs, it does not include other cell types such as those from the immune system. We therefore also performed in vivo experiments using a murine model of colitis, in which mice received 2.5% dextran sodium sulphate (DSS) in drinking water for 5 days, followed by 7 days of recovery with normal drinking water. It had been shown previously that colitis leads to the ablation of LGR5+ ISCs in a reporter mouse model (LGR5-LacZ) after 5 days of DSS, and upon withdrawal of DSS it takes 3-5 days before these cells reappear, indicating that there is regeneration through another source of cells. Our data were in line with what had been reported previously. Lgr5 expression levels were still at only 50% 7 days after withdrawal of DSS. Musashi RNA-binding protein 1 (Msi1), a marker of reserve ISCs, was equally affected as Lgr5. On the level of inflammation, we found that expression of TNFα peaked directly after 5 days of DSS, and it was not significantly altered compared to control levels 7 days after withdrawal of DSS. Interferon gamma (IFNγ) and IL1β on the other hand increased after 5 days of DSS with the highest increase 7 days after withdrawal of DSS, indicative of a primary trigger by TNFα, but a more TH-driven immune response later. Furthermore, IL1β may also have a function in epithelial recovery. This model may be used to study in vivo effects of microbiota on colitis and ISCs.

In conclusion, organoids derived from IBD patients show largely comparable transcriptional profiles to control patient organoids. Furthermore, organoids derived from normal and inflamed tissues quickly reach baseline mRNA levels, and unexpectedly, CXCL3 and CXCL8 were induced in organoids. We could confirm previously published findings of the effect of colitis on ISCs in a murine model, and this will facilitate studies on the influences of microbiota. We have established a large organoid library from controls and patients with IBD, which represent a reliable in vitro model for further studies and characterization of the response of organoids with different genetic backgrounds to microbial or other stimuli.