ASCL2 is a key regulator of the proliferation–differentiation equilibrium in the esophageal epithelium

The esophagus is protected from the austere environment by a squamous epithelium (Rosekrans et al., 2015). In mice, its flat keratinized epithelium is composed of three layers (basal, suprabasal and superficial) and renews every 7 to 10 days. Under normal conditions, cell proliferation is limited to the basal layer (Squier and Kremer, 2001). As they migrate towards the lumen, basal cells differentiate into suprabasal cells and then into superficial cells. The Notch and BMP pathways are crucial in regulating this process (Ohashi et al., 2010; Jiang et al., 2015).

A rare population of stem cells ensures the renewal of the esophageal epithelium (Giroux et al., 2017; Busslinger et al., 2021). Adult stem cells are defined by their unique capacities of self-renewal and potency (Zakrzewski et al., 2019). Some stem cell populations are also radio-resistant and/or quiescent (Hua et al., 2012; Asfaha et al., 2015; Giroux et al., 2017, 2018). In the esophagus, heterogeneity has been observed in the basal layer and the presence of stem cells has been suggested. Indeed, Cluster of differentiation 71 (CD71) and integrin α6 (Itgα6) expression was used to describe a population with a higher stemness (Croagh et al., 2007). Also, a label-retaining population expressing cluster of differentiation 34 (CD34) was shown to contribute to epithelial regeneration following injury (Kalabis et al., 2008). Finally, heterogeneity was furthermore characterized using cluster of differentiation 73 (CD73), integrin β4 (Itgβ4) and Itgα6. This study identified a Itgα6/Itgβ4^highCD73⁺ population, which was less proliferative but had a higher clonogenic potential (DeWard et al., 2014).

The first identification of esophageal stem cells in mice was achieved using a model of lineage tracing, Krt15-CrePR1; Rosa26^{mTomato/mEGFP} mice. Keratin 15⁺ (Krt15⁺) cells are long-lived basal cells that display several characteristics of stem cells: self-renewal, regenerative potential and potency (Giroux et al., 2017). Recently, the presence of Krt15⁺ stem cells in the human esophageal epithelium was confirmed by single-cell RNA sequencing. Indeed, a population of quiescent stem cells was described in the basal layer: COL17A1^high KRT15^high cells (Busslinger et al., 2021). These cells would be at the head of the hierarchy as suggested by a lineage tree analysis. However, we still ignore the pathways or factors that influence the esophageal stem cell gene signature. Interestingly, an analysis of the Krt15⁺ mouse esophageal cells transcriptome showed that the expression of the transcription factor Achaete-scute complex homolog 2 (Ascl2) was highly upregulated in these cells in comparison to other basal cells (Giroux et al., 2017). Furthermore, a slight increase in ASCL2 transcript was also reported in COL17A1^high KRT15^high cells, though statistically non-significant (Busslinger et al., 2021). Nevertheless, these results suggest that Ascl2 is overexpressed in stem/progenitor esophageal cells and could, therefore, play an important role in stemness maintenance in the esophageal epithelium.

ASCL2 is a member of the basic helix-loop-helix (bHLH) family and a gene target of the Wnt/β-catenin and Notch pathways (Moriyama et al., 2008; van der Flier et al., 2009). ASCL2 functions are tissue dependent. It has been implicated in embryonic development, differentiation, proliferation and stemness (Liu et al., 2019). During myogenesis, ASCL2 inhibits differentiation by promotor competition. In this context, ASCL2 overexpression in myoblasts also leads to the inhibition of proliferation (Wang et al., 2017). In intestinal stem cells, ASCL2 is an essential regulator of stemness in leucine-rich repeat-containing G protein-coupled receptor 5⁺ (Lgr5⁺) cells. Thus, following the conditional deletion of Ascl2, loss of Lgr5⁺ cells is observed in the intestine (van der Flier et al., 2009). ASCL2 is crucial for regenerating Lgr5⁺ cells through the dedifferentiation of progenitor cells following Lgr5⁺ cell ablation. The deletion of Ascl2 also prevents dedifferentiation and thereby recovery following irradiation (Murata et al., 2020).

As the role of ASCL2 remains unknown in stratified epithelia, like the esophageal epithelium, we investigated the impact of ASCL2 overexpression on self-renewal and differentiation in esophageal organoids. Our study shows that ASCL2 strongly contributes to the maintenance of the esophageal epithelium through cell differentiation and cell cycle regulation. Furthermore, it confers resistance to radiation and chemotherapeutic agents. These results suggest that ASCL2 is a key modulator of stemness in esophageal organoids.

To investigate the role of ASCL2 in the esophageal epithelium, we generated organoids overexpressing ASCL2 (ASCL2^OE) by lentiviral infection (Fig. 1A,B). ASCL2^OE organoids were significantly smaller than controls (Fig. 1E) and appeared less differentiated (Fig. 1C,D). Indeed, we observed less differentiation/keratinization by phase-contrast microscopy (dark center) and histology (pink center). Differentiation was further evaluated on Hematoxylin and Eosin (H&E)-stained sections using a scoring scale (well-differentiated, intermediate and poorly differentiated) as described in Fig. S1. Control (Ctl) organoids showed a higher percentage of well-differentiated organoids (∼53%), which were almost absent in ASCL2^OE organoid population (∼2%). In contrast, ASCL2^OE organoids were mostly poorly differentiated (∼72% versus 8% in Ctl; Fig. 1F).

Differentiation was furthermore explored using p63 as a marker for basal cells and K13 for suprabasal (differentiated) cells. In ASCL2^OE organoids, immunofluorescence revealed that the number of p63⁺ cells was increased (Fig. 2A,B), which was confirmed by western blotting (Fig. 2E). Surprisingly, p63⁺ cells were also present outside of the basal layer, i.e. the periphery of the organoid (Fig. 2A, second panel). Interestingly, double-positive cells (K13⁺ p63⁺) were more prominent in ASCL2^OE organoids (Fig. 2C). Thus, cells seem to remain in an intermediate state of differentiation, as they expressed both basal (p63⁺) and suprabasal (K13⁺) markers. mRNA levels of known suprabasal markers, Krt13 and Wnt5a (Busslinger et al., 2021), were reduced in ASCL2^OE organoids, while levels of the basal markers Clu and Krt14 were increased (Fig. 2D). All these results demonstrate that the morphological changes in ASCL2^OE organoids result, in part, from impaired differentiation.

Epithelial homeostasis requires fine-tuning between differentiation and proliferation. As ASCL2 could be a key regulator of cell differentiation, we also investigated the impact of its overexpression on proliferation and clonogenicity. First, to compare the clonogenic potential of ASCL2^OE cells, we quantified the organoid formation rate, i.e. the proportion of cells able to form organoids. ASCL2^OE cells had a lower capacity to form organoids than Ctl cells (Fig. 3A). Moreover, WST-1 assays suggested a reduction of proliferation in ASCL2^OE cells (Fig. 3B). This result was confirmed by 5-ethynyl-2′-deoxyuridine (EdU) incorporation assay. There were ∼38% less EdU⁺ cells in ASCL2^OE organoids compared to those in Ctl organoids (Fig. 3C). In agreement with these results, mRNA levels of Top2a, a known proliferation marker (Busslinger et al., 2021), were significantly decreased in ASCL2^OE organoids (Fig. 2D). We validated that these observations were not attributed to changes in cell viability (Fig. S2). These results altogether suggest that ASCL2 could be involved in regulating the clonogenic and proliferative potential of esophageal epithelial cells.

Given that proliferation and clonogenicity of ASCL2^OE cells were reduced in organoids, we hypothesized that progression through the cell cycle might be impaired. A more quiescent phenotype could also explain a decrease in proliferation. Dissociated cells of organoids were stained with propidium iodide to assess cell cycle phases. Interestingly, there was an important decrease in the percentage of cells in G2/M (17.18% for Ctl cells versus 2.53% for ASCL2^OE cells, Fig. 3D). In return, there were more cells in G0/G1 in the ASCL2^OE cell population. As the baseline of Hoechst and propidium iodide was not the same in both of our organoid lines, we treated organoids 24 h prior harvesting with 100 ng/ml of nocodazole, an inhibitor of the cell cycle that causes arrest in G2/M, to confirm the gating strategy for each cell cycle phase (Fig. S3). To further explore cell cycle progression in our model, we stained organoid cells with the DNA stain Hoechst 33342 and the RNA stain Pyronin Y, which allow the discrimination of cells in G0, G1, S and G2/M. Our results revealed that there was an ∼10% increase in the number of cells in G0 for the ASCL2^OE population, suggesting a possible effect of ASCL2 on quiescence (Fig. 3E). Interestingly, mRNA expression of a known marker of quiescent cells, Clu, was considerably increased in ASCL2^OE organoids (Fig. 2D). Thus, reduced proliferation and clonogenicity in ASCL2^OE organoids could be the result of changes in cell cycle progression with an increase in the number of quiescent cells.

To determine the molecular changes occurring following ASCL2 overexpression, we performed data-independent acquisition (DIA) mass spectrometry (MS) on both organoid populations. We identified 37 upregulated and 57 downregulated proteins in ASCL2^OE organoids compared to Ctl organoids. Of note, keratinocyte differentiation-associated protein (KRTDAP), a protein responsible for the regulation of keratinocyte differentiation and maintenance of stratified epithelia (Oomizu et al., 2000), was the most downregulated protein in ASCL2^OE organoids (Fig. 4A). Gene Ontology (GO) analysis revealed that several biological processes related to differentiation (epidermal cell, keratinocyte and epithelial cell differentiation) and keratinization were over-represented in our dataset (Fig. 4B). Several proteins associated with differentiation were differentially expressed in ASCL2^OE organoids, such as small proline-rich protein 3 (SPRR3) and macroH2A.2 histone (H2AFY2; Fig. 4C). As ASCL2 is a transcription factor, we performed alignment of its motif on the promoter of the genes encoding for the ten most downregulated or upregulated proteins in our proteomic results. Three of these genes, Ly6g5c, Tmbim1 and Crabp2, showed putative binding sites for ASCL2 in their 500 bp promoter region (Fig. S5). On the other hand, proteins related to the cell cycle were also modulated in ASCL2^OE organoids (Fig. 4D). p16^INK4a (Cdkn2a), a protein implicated in cell cycle progression and capable of inducing cell cycle arrest (Medema et al., 1995), was the fourth most upregulated protein in our dataset (Fig. 4A). p16^INK4a and SMAD1 upregulation were confirmed by western blotting (Fig. S4). Interestingly, SMAD1 is involved in the BMP pathway, which is crucial for esophageal development and homeostasis. Some proteins associated with stem cell population maintenance were also significantly regulated in ASCL2^OE organoids, suggesting that ASCL2 could be involved in stemness as well (Fig. 4E). These molecular changes in ASCL2^OE organoids strongly correlate with the biological changes observed, confirming that ASCL2 plays an important role in differentiation and the cell cycle in esophageal organoids.

ASCL2 regulates several functions associated with stem cells, namely, cell proliferation and differentiation. Therefore, we investigated the contribution of ASCL2 to another key characteristic of stem cells, resistance to anti-cancer treatments. To do so, organoids were irradiated with 5 or 10 Gy on day 1 post-seeding, and the number of organoids formed was evaluated. As expected, both populations showed reduced organoid formation rate compared to that of their non-irradiated control (Fig. 5A). Particularly, ASCL2^OE organoids maintained a higher capacity to form organoids following 5 or 10 Gy irradiation than that of controls. We also tested chemo-resistance in our organoid model. Organoids were treated with cisplatin, camptothecin or 5-fluorouracil (5-FU) at day 1 and the number of organoids was counted at day 7. Interestingly, more organoids were formed in the ASCL2^OE organoid line following each treatment than in controls (Fig. 5B). In conclusion, these results showed that ASCL2 overexpression confers radio- and chemo-resistance capacities to esophageal epithelial cells.

Epithelial homeostasis is typically assured by stem cells. Renowned for their high self-renewal capacity and potency, this subpopulation of epithelial cells also contributes to tissue regeneration in response to injury. In the esophagus, the presence of a true stem cell population remained controversial until recently (Croagh et al., 2007, 2008; Doupé et al., 2012; DeWard et al., 2014). In 2017, Krt15⁺ esophageal stem cells were identified in mice (Giroux et al., 2017), and their presence was recently confirmed in humans as quiescent COL17A1^high KRT15^high cells (Busslinger et al., 2021). However, we still ignore the molecular players dictating their distinct features. Herein, we investigated the role of ASCL2, which was previously identified as the most upregulated transcription factor in Krt15⁺ stem cells (Giroux et al., 2017), using esophageal organoids. Organoid culture is the predilected model to study functions associated with stemness, such as clonogenicity, differentiation and cell proliferation, as it maintains stem cell growth.

Upon ASCL2 overexpression, differentiation was severely impaired in organoids. Keratinization, a sign of terminal differentiation in the mouse esophagus, was drastically reduced histologically. Proteomic data also supported dysregulation in biological processes related to keratinization and differentiation. ASCL2 can act as a transcriptional activator but also as an inhibitor. Sequence alignment suggests putative binding sites for ASCL2 on the promoter regions of the genes encoding for three downregulated proteins. One of those proteins, Crabp2, is implicated in the transport of retinoic acid to the nucleus and promotes differentiation (DeWard et al., 2014). These observations corroborate other studies reporting a role for ASCL2 in the differentiation of trophoblast cells, myoblasts and lymphocytes (Liu et al., 2014; Wang et al., 2017; Bogutz et al., 2018; Bao et al., 2020; Varberg et al., 2021). Interestingly, ASCL2 was also recently associated with the dedifferentiation of intestinal crypt cells in response to stem cell ablation to cope with the injury and induce regeneration (Murata et al., 2020). The differentiation changes observed in our study were confirmed by staining for specific markers of basal and suprabasal cells. An enhanced number of basal (p63⁺) cells was observed in ASCL2^OE organoids as well as an aberrant presence of cells expressing basal (p63) and suprabasal (K13) cell markers. This observation suggests an impairment in cell fate decisions in these organoids. Of note, ASCL2 was shown to regulate cell fate in the intestine (van der Flier et al., 2009; Giakountis et al., 2016) and in embryonic stem cells (Kinoshita et al., 2021). In the normal esophageal epithelium, cells migrate toward the lumen while undergoing differentiation from basal to suprabasal cells (Rosekrans et al., 2015). Herein, a large proportion of cells remains in an intermediate state, suggesting that terminal differentiation is dysregulated and that cell fate is compromised.

As alterations in cell fate decisions might impact clonogenic potential, we evaluated the organoid formation rate and it was indeed reduced in ASCL2^OE organoids. In accordance with decreased clonogenicity, ASCL2 overexpression reduced cell proliferation in esophageal organoids. ASCL2 has been previously associated with reduced cell proliferation in Schwann cells, myoblasts and neural stem cells (Küry et al., 2002; Wang et al., 2017; Liu et al., 2019). However, in the small intestine, ASCL2 supports a high proliferation (van der Flier et al., 2009; Schuijers et al., 2015). This difference could be due to the higher dependency of intestinal tissue to the Wnt/β-catenin pathway than that of esophageal tissue, amongst others. The reduced proliferation observed in ASCL2^OE organoids correlated with a lower number of cells in G2/M and an increase of cells in G1/G0, suggesting a blockage in cell cycle progression. Notably, proteomic data showed an increased expression of p16^INK4A (Cdkn2a) in ASCL2^OE organoids compared to that in controls. This cell cycle inhibitor leads to an arrest in the G1 phase of the cell cycle (Medema et al., 1995), which could explain our observations. Furthermore, Pyronin Y/Hoechst 33342 staining showed an increase in G0 cells, suggesting induction of quiescence following ASCL2 overexpression. Quantitative PCR (qPCR) results also showed induced expression of Clu, a gene enriched in quiescent basal cells in the human esophagus (Busslinger et al., 2021). High expression of Clu was also observed in a damage-induced quiescent cell population in the gut (Ayyaz et al., 2019). Overall, in response to ASCL2 overexpression, organoids are less clonogenic, proliferative, differentiated and potentially more quiescent, suggesting again a major impairment in cell fate.

Cell fate decision is mainly associated with stem cells. Of note, proteins associated with stem cell maintenance were regulated in ASCL2^OE organoids. Another distinct feature of stem cells is a higher capacity to cope with stressors such as radiation and chemotherapeutic drugs. In the esophagus, Krt15⁺ stem cells were shown to be resistant to high-dose radiation (Giroux et al., 2017), so we would expect higher resistance in ASCL2^OE organoids as well. Indeed, ASCL2 overexpression reduced sensitivity to radiation. Higher resistance to three chemotherapeutic agents (5-FU, cisplatin and camptothecin) was also observed with ASCL2^OE organoids. Aberrant upregulation of ASCL2 was previously associated with 5-FU resistance in gastric cancer (Kwon et al., 2013). In addition, ASCL2 was linked to drug resistance in colorectal cancer cells (Tanaka et al., 2019). More recently, high expression of ASCL2 in esophageal adenocarcinoma was proposed as a poor prognosis of survival (Shibahara et al., 2022). Interestingly, proteomic analysis of ASCL2^OE organoids showed altered expression of SPRR3, a protein that has been associated with resistance to DNA damage inducers in esophageal cancer cells (Luo et al., 2013).

Overall, our results demonstrate that ASCL2 contributes to epithelial homeostasis as a critical regulator of the differentiation-proliferation balance. Indeed, ASCL2^OE organoids displayed dysregulation in differentiation, cell proliferation, clonogenicity and the cell cycle. These various alterations suggest impaired cell fate decisions. These phenotypes correlated with proteomic changes in biological processes associated with these functions. Furthermore, ASCL2^OE organoids were also more resistant to radiation and chemotherapeutic drugs, suggesting a role for ASCL2 in maintaining homeostasis and in response to injury. Further studies will be required to define how ASCL2 mechanistically regulates these pathways. Inactive DNA-binding ASCL2 and ASCL2 derivatives with specific mutations could be used to mechanistically underline the phenotypes observed in ASCL2^OE organoids. This study highlights a new role for ASCL2 in the esophageal epithelium, contributing to our understanding of the molecular mechanisms involved in the fine-tuning of the homeostasis in this epithelium.

Organoids were established from the esophageal epithelium of wild-type 8-week-old C57BL/6J mice as described previously (Giroux et al., 2017). All animal procedures were approved by the Université de Sherbrooke Institutional Committee for the Use and Care of Animals. Briefly, the esophagus was open longitudinally and incubated in dispase (CB-40235, ThermoFisher Scientific). The epithelium was then peeled from the submucosa and incubated in trypsin to obtain a single-cell suspension. Isolated cells were then seeded in Matrigel™ (354230, Corning) and grown as organoids in organoid medium [Advanced DMEM/F12 (12634010, Gibco) supplemented with 1× GlutaMAX™ (35050061, ThermoFisher Scientific), 1× HEPES (330-050-EL, Wisent), 1× penicillin-streptomycin (450-201-EL, Wisent), 1× N2 supplement (17502048, ThermoFisher Scientific), 1× B27 supplement (17504044, ThermoFisher Scientific), 0.1 mM N-acetylcysteine (A7250-5G, Sigma-Aldrich), 50 ng/ml recombinant EGF (315-09, PeproTech), Noggin/R-Spondin homemade conditioned medium and 10 µM Y27632 (A11001-50, AdooQ Bioscience)]. The medium was changed every other day, and organoids were passaged after 7 days of culture. For harvesting or passaging, organoids were retrieved from Matrigel™ using dispase. At this step, organoids were fixed in zinc formalin for 4 h at 4°C for histology or used for protein or RNA extraction. For passaging, organoids were further dissociated using trypsin and filtered using a 40 µm cell strainer to obtain a single-cell suspension. Cells were counted and seeded again in Matrigel™. Organoids were regularly tested for mycoplasma contamination.

The Gateway method was used to generate lentiviral construction for protein overexpression (FLAG-ASCL2). The sequence was synthesized and cloned in pUC57 vectors (Bio Basic; Markham, Ontario, Canada). pDONR221 was used as a donor vector (graciously provided by Professor François-Michel Boisvert, Université de Sherbrooke) and pLX303 as a destination vector (graciously provided by Professor Marie-Josée Boucher, Université de Sherbrooke). The pUC57 vector containing the gene of interest was digested using the restriction enzymes BamH1 (R3136S, New England Biolabs) and Nde1 (R0111S, New England Biolabs). The purified insert [purification kit Monarch (T1020S, New England Biolabs), following the supplier's recommendations] was recombined in donor vectors by BP recombination. Afterward, the vector was transformed in Stellar™ competent bacteria (636766, Takara Bio). Bacteria were seeded on agar Petri dishes with kanamycin and, following amplification of positive colonies, the vector was extracted by miniprep using the GeneJetPlasmid Miniprep kit (K0503, ThermoFisher Scientific) following the supplier's recommendations. LR recombination was completed using pDONR221 vectors containing the gene of interest and the pLX303 destination vector. Those are already listed as gifted in the beginning of the paragraph. The vector was transformed in One Shot™ Stbl3™ competent bacteria (C737303, Invitrogen), positive colonies were amplified, and the vector was extracted using the QIAfilter Plasmid Midi kit (12243, Qiagen). Plasmids were validated by sequencing.

Overexpressing (ASCL2^OE) and Ctl (pLX303-empty) organoid lines were obtained by lentiviral infection. Viruses were produced in HEK293T cells by transfection (Lipofectamine™ 3000; L3000015, ThermoFisher Scientific), harvested after 48 h in the supernatant and filtered using a 0.45 µm filter. The viral solution was concentrated overnight using Lenti-X™ concentrator (one volume of Lenti-X™ for three volumes of supernatant; 631231, Takara Bio) according to the manufacturer's recommendations. 20,000 organoid cells (single-cell suspension) were incubated for 1 h at room temperature for Spinfection with concentrated virus and polybrene (H9268, Sigma-Aldrich). The supernatant was then removed and cells were seeded in Matrigel™ for 7 days of culture. At day 7, organoids were passaged and antibiotic selection with blasticidin (R21001, ThermoFisher Scientific) was performed for 7 days.

Organoids were lysed using RIPA buffer supplemented with protease and phosphatase inhibitors. Samples were flown through a 27G needle several times, incubated for 10 min on ice and then sonicated at 5 s intervals at 20% amplitude for 30 s. Following centrifugation, the supernatant was recovered. Protein extracts were quantified using the bicinchoninic acid (BCA) method, diluted in Laemmli buffer and then incubated for 5 min at 95°C. Proteins were separated on SDS-PAGE gels and then transferred to polyvinylidene fluoride (PVDF) membranes. Transfer and protein loading were confirmed by Ponceau Red staining. The membrane was blocked with milk [5% powdered milk in PBS/Tween 0.05% (PBS/T)] or 5% BSA in PBS/T for 1 h at room temperature, and then incubated overnight at 4°C in primary antibody diluted in blocking solution (see Table 1). The membrane was then incubated 1 h at room temperature with HRP-coupled secondary antibodies. The signal was visualized using the Enhanced Chemiluminescence kit (ECL; RPN2232, VWR or 1705062S Bio-Rad) and observed by autoradiography or a chemiluminescent imaging system (Azure 280, VWR). Uncropped images of western blots are shown in Fig. S6.

Following zinc formalin fixation, organoids were resuspended in a plug of 2.5% gelatine/2% agar. Following circulation and paraffin embedding, specimens were sectioned (4-5 µm) using a Histocore Multicut microtome (Leica). For histological analysis, H&E coloration was then performed. Slides were imaged using a slide scanner (Hamamatsu Nanozoomer 2.0-RS). The diameter of organoids was measured using NDP.view2 software (Hamamatsu). Differentiation was evaluated using the following criteria by two unbiased investigators: well-differentiated (normal differentiation with a keratinized center), intermediate (basal and suprabasal layers) and poorly differentiated (overall basal cell organoids) (Fig. S1). For immunofluorescence, slides were rehydrated and antigens were retrieved in citric acid. Slides were blocked using 5% BSA and incubated overnight at 4°C in primary antibodies diluted in 1% BSA solution (see Table 1). They were then incubated for 1 h at room temperature with secondary antibodies diluted in 1% BSA solution. Sections were mounted using VECTASHIELD^® (H-1200, Vector Laboratories) with DAPI. Slides were then imaged on a Zeiss Axioscope upright microscope or a confocal Zeiss LSM 880 two-photon microscope.

RNA was extracted using the RNeasy Plus Mini kit (74134, Qiagen) following the supplier's recommendations. 1 µg of RNA was then reverse transcribed using SuperScript III Reverse Transcriptase (18080-044, ThermoFisher Scientific) and oligo-dT_12-18 according to the manufacturer's recommendations. FastStart Essential DNA Green Master (06402712001, Roche) was used for qPCR analysis on a LightCycler^® 480 (Roche). Primers used can be found in Table 2. Results were calculated in normalized ratio using the reference gene Gapdh.

When organoids were passaged, 300 cells were seeded in 20 µl of Matrigel™ in a 48-well plate. Organoids were cultured as described above. Assays were done in quadruplicate for each condition and the number of organoids formed was counted 7 days post seeding. For the radio-resistance assay, organoids were irradiated at day 1 post seeding with 5 or 10 Gy and non-irradiated organoids were used as controls. For the chemo-resistance assay, organoids were incubated at day 1 post seeding with chemotherapeutic agents (30 μM cisplatin, 13119, Cayman Chemicals; 50 nM camptothecin, 11694, Cayman Chemicals; 10 μM 5-FU, 14416, Cayman Chemicals) or vehicle for 36 h.

For the WST-1 assay, when organoids were passaged, 1000 cells were seeded in 5 µl of Matrigel™ in a 96-well plate. Organoids were cultured as described above. Assays were done in quadruplicate for each condition and a negative control containing no cells was prepared. Proliferation was evaluated using the WST-1 reagent (5015944001, Sigma-Aldrich) according to the manufacturer's recommendations. The absorbance was measured using a plate reader. For EdU incorporation assay, 20,000 cells were seeded per well in a 48-well plate. At day 4 post seeding, organoids were incubated with 10 μM of EdU (C10419, ThermoFisher Scientific) for 16 h. Organoids were then harvested and dissociated to obtain a single-cell suspension. Cells were stained for EdU using the Click-iT™ EdU Alexa Fluor™ 647 Flow Cytometry Assay Kit (C10419, ThermoFisher Scientific) according to the manufacturer's instructions. EdU⁺ cells were quantified using a BD Fortessa flow cytometer and FlowJo software.

The cell cycle was assessed by RNA and DNA staining. Organoids were harvested at 7 days post seeding and dissociated to a single-cell suspension. Cells were fixed for least 2 h using 70% ethanol. Cells were then stained (50 μg/ml propidium iodide, P1304MP, ThermoFisher Scientific; 4 μg/ml Pyronin Y, 418630010, ThermoFisher Scientific; 2 μg/ml Hoechst 33342, H3570, ThermoFisher Scientific) and the cell cycle phase was assessed using a BD Fortessa flow cytometer and FlowJo software.

Ctl and ASCL2^OE organoids were extracted in a solution of 8 M urea, 1 M NH₄HCO₃ and 20 mM HEPES, pH 8.0. Then, the samples were sonicated on ice at 5 s intervals at 20% amplitude for 60 s and centrifuged at 16,000 g for 10 min at 4°C. 30 μg of proteins was reduced by adding dithiothreitol (final concentration of 5 mM). Samples were incubated for 2 min at 95°C, followed by incubation at 30 min at room temperature. Protein alkylation was completed by the addition of chloroacetamide (final concentration of 7.5 mM; Sigma-Aldrich) for 20 min in the dark at room temperature. Urea concentration was diluted (final concentration of 2 M) by adding 50 mM NH₄HCO₃ (Sigma-Aldrich). 1 μg of MS-grade trypsin (ThermoFisher Scientific) was added to each sample for protein digestion and incubated overnight at 30°C with agitation.

Digestion was stopped with trifluoroacetic acid (final concentration of 0.2%; Sigma-Aldrich). Digested peptides were purified with C18 tips (ThermoFisher Scientific Pierce) according to the manufacturer's recommendations. Using a centrifugal evaporator, samples were dried at 60°C for at least 3 h and resuspended in 1% formic acid (FA).

Samples were combined to create the DIA library, which was fractionated using the Pierce High pH Reversed-Phase Peptide Fractionation Kit (84868, ThermoFisher Scientific) according to the manufacturer's recommendations. Peptides were eluted in eight fractions and dried using a centrifugal evaporator at 60°C for at least 3 h. Samples were then resuspended in 1% FA.

All mass spectrometry experiments were performed at the Mass spectrometry and Proteomics platform at Université de Sherbrooke. From each fraction, 250 ng of peptides were injected into an HPLC (nanoElute, Bruker Daltonics) and loaded onto a trap column (constant flow of 4 µl/min; Acclaim PepMap100 C18 column, 0.3 mm internal diameter×5 mm, Dionex Corporation). An analytical C18 column (1.9 µm beads size, 75 µm×25 cm, PepSep) was used for elution for over a 2 h gradient of acetonitrile (5-37%) in 0.1% FA at 400 nl/min while being injected into a TimsTOF Pro trapped ion mobility mass spectrometer (TIMS) equipped with a Captive Spray nano electrospray source (Bruker Daltonics). Data were acquired using data-dependent auto-MS/MS with a 100-1700 m/z mass range, with parallel accumulation-serial fragmentation (PASEF) enabled and PASEF scans set at 10 (1.17 s duty cycle) and a dynamic exclusion of 0.4 min, m/z-dependent isolation window and collision energy of 42.0 eV. The intensity threshold was set at 2500, and the target intensity was set at 20,000.

Using the MaxQuant software (version 2.0.3.0; www.maxquant.org) and the UniProt mouse proteome database (February 2022, 55,286 entries), the raw files were analyzed to obtain the DIA library. The following settings were used for the MaxQuant analysis (with group-specific TIMS-DDA type parameters): one miscleavage was authorized; cysteine carbamidomethylation was a fixed modification; trypsin was the enzyme (lysine/arginine was not before a proline); methionine oxidation and protein N-terminal modification were included in the analysis; precursor and fragment ions had a mass tolerance of 20 ppm; ‘PSM FDR’, ‘protein FDR’ and ‘site decoy fraction’ identification values were set to 0.05; 1 was the minimum peptide count; ‘match between runs’ and ‘second peptides’ were permitted; and a transfer q-value of 0.3 was used to run MaxQuant. From this DDA analysis, the ‘peptides.txt’, ‘evidence.txt’ and ‘msms.txt’ files were used for DIA analysis.

Data were acquired with the diaPASEF mode and using the parameters described for DDA analysis. One mobility window with 27 mass steps (m/z between 114 and 1414 with a mass width of 50 Da) per cycle (1.27 s duty cycle) was used for each single TIMS (100 ms) in diaPASEF mode, which covers in the m/z-ion mobility plane the diagonal scan line for charged peptides +2 and +3.

MaxQuant software (version 2.0.3.0) and UniProt mouse proteome database (02/2022, 55,286 entries) were used to analyze the raw files. The settings were as described above, except for the following: Maxquant and TIMS MaxDIA were selected; three files (‘peptides.txt’, ‘evidence.txt’ and ‘msms.txt’) from the DIA library were used.

Single peptides and proteins positive for the ‘reverse’, ‘only identified by site’ or ‘potential contaminant’ categories were excluded. Proteomic data were then analyzed using the iPathwayGuide software (Advaita Bio) with a threshold of 1.5 for log₂ fold-change [calculated with flq (label-free quantification)-intensity] and 0.05 for P-value. GO terms for biological processes were corrected with a false discovery rate (FDR). Graphs were obtained using R software. Networks were obtained using the Cytoscape software. Proteomics data have been deposited to the ProteomeXchange with PRIDE repository partner (PXD039477).

Alignment and putative binding sites were obtained using the Contra v3 web tool (Kreft et al., 2017). The following parameters were used: visualization, organism Mus musculus, gene Ascl2 (NM_008554) and 500 bp promoter region. The stringency was as followed: core=0.95, similarity matrix=0.85. Alignment was performed on the ten most upregulated and downregulated proteins obtained by MS analysis.

Statistical analysis and graphs were obtained using GraphPad Prism 9 software. The following statistical tests were performed: two-tailed paired and unpaired Student's t-test, two-tailed unpaired Welch's t-test and two-way. Tests were corrected for variance, multiple comparisons and normality. Error bars shown on graphs correspond to the standard deviation or standard error. The significance of the P-value is represented as follows: *P≤0.05, **P≤0.01, ***P≤0.001 and ****P≤0.0001.

We sincerely thank all members of the Giroux Lab for scientific discussions. We also thank Dominique Lévesque, Alexis Gonneaud, Professor Marie-Josée Boucher, Professor François-Michel Boisvert and Professor Steve Jean for technical support and providing reagents. We are also thankful to these facility cores from the Faculty of Medicine and Health Sciences – Université de Sherbrooke for providing services and guidance: Histology and electron microscopy platform, Proteomic and mass spectrometry platform, and Light microscopy and flow cytometry platform.