The Notch and Wnt pathways regulate stemness and differentiation in human fallopian tube organoids

Abstract

The epithelial lining of the fallopian tube is of critical importance for human reproduction and has been implicated as a site of origin of high-grade serous ovarian cancer. Here we report on the establishment of long-term, stable 3D organoid cultures from human fallopian tubes, indicative of the presence of adult stem cells. We show that single epithelial stem cells in vitro can give rise to differentiated organoids containing ciliated and secretory cells. Continuous growth and differentiation of organoids depend on both Wnt and Notch paracrine signalling. Microarray analysis reveals that inhibition of Notch signalling causes downregulation of stem cell-associated genes in parallel with decreased proliferation and increased numbers of ciliated cells and that organoids also respond to oestradiol and progesterone treatment in a physiological manner. Thus, our organoid model provides a much-needed basis for future investigations of signalling routes involved in health and disease of the fallopian tube.

Figure 1. Single fallopian epithelial cells give rise to organoids that can be maintained long term.

(a) Representative confocal image of fallopian tube tissue labelled for acetylated tubulin, a marker of ciliated cells (red), the proliferation marker Ki67 (green) and DNA (DRAQ5, blue). Ki67-positive nuclei are dispersed along epithelial folds, with occasional clustering (asterisk). (b) Phase contrast images of spheroid formation and growth. Small spheres are already visible 1 day after seeding and expand to reach a diameter of over 100 μm within 7 days. The same culture shows no apparent morphological differences after 2 weeks and after 10 months in vitro. (c) Phase contrast microscopy images of focal planes inside mature organoids showing the epithelial invaginations and folding (arrows), which are routinely present but not visible in most images focused on the organoid surface. (d) Representative image of organoids generated with independently labelled GFP and mCherry epithelial isolates showing almost exclusively single-colour organoids. Quantification of the organoids from two independent experiments and donors (table) reveals that the percentage of hybrids is on average below 1%. (e) Time course of a monoclonal organoid generated from a single EpCAM+ cell sorted and embedded in Matrigel in a 96-well format. Monoclonal organoid cultures were successfully generated from four different donor isolates, confirming the presence of stem cells in the fallopian tube epithelium. (f) Confocal microscopy image of a monoclonal organoid generated from a single fallopian tube stem cell after 2 months of 3D culture (P3), labelled for b-Cat (red) and tubulin, showing ciliated and non-ciliated cells. (g) Confocal images of early organoid cultures at different time points labelled for b-Cat (red) and ac tubulin (green), revealing that cilia develop after 2 weeks in 3D culture.

Figure 2. Organoids match the in vivo epithelium in structure and distribution of markers.

Confocal images of organoid sections show that epithelium contains both ciliated cells positive for detyrosinated tubulin (green (a)) and secretory cells positive for PAX8 (green (b)) confirming differentiation. Terminally differentiated ciliated cells are PAX8-negative (c). (d) Epithelial cells form tight junctions visible by labelling for occludin (green, left). Organoids also contain occasional domains of epithelial cells ubiquitously expressing CA125 (green, middle) as well as basolateral vimentin (green, right). (e) Electron transmission microscopy images of the fallopian tube epithelium (upper panel) and organoid monolayer after 2 months in vitro (middle panel), showing high levels of similarity. Monolayer of the columnar epithelium with two cell types (ciliated and secretory), with clear apicobasal polarization closing off the empty lumen. Higher magnifications of the middle panel (square) reveal completely formed junctional complexes, which ensure separation of the apical from basolateral side (arrow, lower panel, right). The apical surface is covered by abundant microvilli (*) and fully developed cilia (***). Ruffling of the apical membrane on the non-ciliated cells (left panel) is highly suggestive of active release and secretion into the lumen (double arrow).

Figure 3. Organoid morphology and phenotype remains stable during long-term culture.

(a) Comparative confocal analysis of the organoids after 1 and 4 months in vitro reveals stability of the phenotype, with unchanged morphology and staining pattern of the adhesion markers. (b) Dynamics of organoid growth remains stable over time as determined by quantification of the relative number of proliferative Ki-67-positive cells in confocal sections at 1- and 8-month culture. (c) The proportion of ciliated cells also remains unchanged as determined by quantification of ciliated cells. Error bars represent±s.d. from a minimum of five independent fields of view.

Figure 4. Active Wnt signalling is required for expression of stemness factors in the organoids and supports their growth.

(a) Formation efficiency of the organoids determined by quantification in four different media with three independent donors (basic (B), FGF10 (F), noggin (N), EGF (E), RSPO1 (R)). Addition of EGF to the medium almost doubles sphere formation. Addition of RSPO1 strongly increases the size of the organoids, with a much higher number growing to above 300 μm diameter (dark grey fractions). Data are normalized to control conditions (B) ± cumulative s.d. of big and small organoids in duplicate wells. (b) During long-term expansion only organoids growing in complete medium maintain stable growth as seen in phase contrast images at 3 months in culture. (c) Active Wnt signalling is detected in the cells of the growing organoids by the TCF4-GFP reporter (green), which was introduced by lentiviral transduction. All cells containing reporter are labelled red (mCherry). Images are representative stills from a live-cell imaging experiment (Supplementary Movie 3). (d) qPCR analysis shows that addition of RSPO1 and Wnt3A to the medium for 7 days leads to a strong increase in the expression of hTERT and the proliferation marker Ki67 relative to GAPDH (ΔΔC_t). Increased expression of the Wnt targets AXIN2 and cMYC confirms pathway activation. Data are representative of two independent donor samples, and are represented as mean±s.d. of technical replicates. (e) Addition of 25% Wnt3A conditioned medium to 25% RSPO1 conditioned medium increases expression of Olfactomedin4 and Notch1, compared with control organoids grown in RSPO1 conditioned medium alone, as shown using qPCR. Data are representative of two independent patient samples and represented as mean±s.d. of technical replicates. (f) Phase contrast images of organoids 7 days post seeding in media supplemented with different Wnt pathway ligands. Pictures are representative of three independent patient samples. Notably, addition of Wnt3A on its own failed to improve sphere formation and growth, while RSPO1 greatly increased both number and size of spheres.

Figure 5. Inhibition of Notch signalling leads to downregulation of stemness-related genes.

(a) Inhibition of the Notch pathway by addition of DBZ (1 μM; lower panel) changes the differentiation pattern and structure of the organoids, leading to distinct changes in morphology, evident by increased folding (asterisk). (b) qPCR validation of selected candidate genes that were found to be downregulated in the microarray. The stemness marker Olfactomedin4, Wnt signalling components AXIN2 and LGR6, as well as the Notch target gene HES1 were all confirmed to be downregulated in independent patient samples. Data represent mean±s.d. of three independent donors. (c) Gene Set Enrichment Analysis of the regulated genes after DBZ treatment identified by microarray, compared with a set of stem cell signature genes from mouse intestine. The correlation plot reveals a significant enrichment in the set of downregulated genes (negative t-score). (d) Venn diagram showing that 78 of the 274 ‘stem cell signature genes' were also found to be significantly downregulated in the fallopian tube organoids on inhibition of Notch. (e) qPCR validation of selected candidate genes functionally related to cillia found to be upregulated in the microarray. ARMC4, DNAI1, FOXJ1 and LRRC were found to be consistently upregulated in four independent biological replicates. Data are presented as mean±s.d.

Figure 6. Notch signalling is required for maintaining a secretory phenotype.

Representative confocal images of control (a) and DBZ-treated (1 μM) organoids (b), showing an increase in the number of ciliated cells positive for acetylated tubulin (green), as well as a decrease in the number of proliferating cells positive for Ki67 (red) when Notch is inhibited. (c,d) Quantification of ciliated and actively proliferating cells in control versus DBZ-treated organoid sections from three different donors. Error bars represent±s.d. from a minimum of seven independent fields of view. (e) Organoids respond to oestradiol (500 pmol l⁻¹) and progesterone (50 ng ml⁻¹) stimulation as determined by gene expression profiling by microarray analysis of three independent donors. The colour matrix of the heat map depicts the log10(Ratio) of individual unstimulated (C) versus hormone-stimulated (estradiol, E or progesterone, P) comparisons while the hierarchical structures are illustrated as dendrograms at gene and patient levels.