Abstract

   Barrett’s esophagus (BE) is a precancerous condition closely linked to
   chronic gastroesophageal reflux disease, characterized by the abnormal
   transformation of esophageal squamous mucosa into specialized
   intestinal-type epithelium, significantly elevating the risk of
   esophageal adenocarcinoma (EAC). Recurrent acidic bile reflux promotes
   epithelial-mesenchymal transition (EMT), a critical event driving
   malignant progression. However, the underlying molecular mechanisms
   remain incompletely understood. Here, we identify the long non-coding
   RNA small nucleolar RNA host gene 1 (SNHG1) as a novel regulator of EMT
   in BE, mediating its effects through the UNC-52-like kinase 1
   (ULK1)-Notch1 signaling axis and autophagy modulation. Using BAR-T and
   CP-C cell models, we demonstrate that SNHG1 expression is elevated
   following acidic bile salt exposure, enhancing EMT characteristics by
   promoting the phosphorylation of ULK1 and activating Notch1 signaling.
   Pharmacological interventions targeting autophagy (Rapamycin) and Notch
   signaling (DAPT) further confirmed that SNHG1’s effects on EMT are
   mediated via modulation of the autophagy-Notch1 interplay. We further
   validated our results in vivo using the previously described
   IL1β-induced Lgr5-CreERT2; p16^flox/flox/Kras^G12D mouse model, which
   reliably reproduces the histological progression of Barrett’s-like
   dysplasia in the squamocolumnar junction (SCJ), confirming SNHG1’s
   critical role in regulating EMT and BE progression. Additionally, SNHG1
   expression was significantly elevated in patients who progressed to
   low- or high-grade dysplasia, as confirmed by diagnostic endoscopic
   biopsies. Collectively, our study uncovers SNHG1 as a central molecular
   mediator linking acidic bile-induced EMT and autophagy regulation via
   the ULK1-Notch1 axis, highlighting its potential as a therapeutic
   target for preventing BE recurrence and progression to EAC.

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s43556-025-00285-4.

   Keywords: Barrett’s esophagus, Small nucleolar RNA host gene 1 (SNHG1),
   UNC-52-like kinase 1 (ULK1), Notch1, Epithelial-mesenchymal transition
   (EMT)

Introduction

   Barrett’s esophagus (BE) is a premalignant condition in which the
   normal stratified squamous epithelium lining the distal esophagus is
   replaced by specialized intestinal-type columnar epithelium, primarily
   as a result of chronic exposure to gastric acid and bile reflux from
   long-standing gastroesophageal reflux disease (GERD). BE is estimated
   to affect up to 1.6% of the general population globally [[40]1], with a
   higher prevalence in Western countries—particularly among middle-aged
   to older white males with chronic GERD symptoms [[41]2]. In the United
   States, for example, the prevalence of BE among individuals undergoing
   endoscopy for reflux symptoms may reach as high as 10–15%. While the
   annual incidence of progression from non-dysplastic BE to esophageal
   adenocarcinoma (EAC) is relatively low—approximately 0.1% to 0.5% per
   year—the lifetime risk remains substantial, especially in patients with
   additional risk factors such as male sex, smoking history, central
   obesity, and confirmed dysplasia [[42]3, [43]4]. EAC is a highly
   aggressive cancer with poor prognosis, and BE remains the only
   well-established precursor lesion for its development. Accordingly,
   clinical guidelines recommend that patients with confirmed BE undergo
   regular endoscopic surveillance to monitor dysplasia and early
   neoplastic transformation. Once dysplasia is detected, endoscopic
   eradication therapies—most notably radiofrequency ablation (RFA)—are
   considered the gold standard for preventing progression to invasive
   cancer due to their minimally invasive nature and favorable safety
   profile [[44]5]. However, even after successful initial treatment,
   recurrent intestinal metaplasia is observed in approximately 10% of
   patients per year [[45]6].

   The precise mechanisms underlying the recurrence of Barrett’s
   metaplasia after RFA remain elusive; however, sub-squamous intestinal
   metaplasia (SSIM) has been proposed as a potential cause [[46]7,
   [47]8]. SSIM refers to intestinal metaplastic glands situated beneath
   the squamous epithelium within the esophageal lamina propria, often
   evading detection by conventional endoscopic methods [[48]9]. RFA
   treatment is designed to induce controlled thermal injury, effectively
   disrupting the superficial metaplastic epithelium while preserving the
   submucosa to minimize the risk of esophageal strictures [[49]10].
   Consequently, this technique may inadequately target SSIM glands that
   lie deeper beneath the protective squamous epithelial layer,
   potentially contributing to disease recurrence [[50]11]. Nevertheless,
   the pathogenesis and molecular regulation of SSIM remain poorly
   understood.

   A critical pathogenic mechanism proposed for SSIM involves the
   migration of Barrett’s epithelial cells into the squamous lamina
   propria [[51]12]. In this context, Barrett’s epithelial cells acquire
   mesenchymal properties through epithelial-mesenchymal transition (EMT),
   enabling penetration into the underlying tissue before potentially
   reverting to an epithelial state, thereby initiating SSIM formation.
   EMT in Barrett’s cells is frequently induced by chronic exposure to
   acidic bile salts, as observed in both cell-culture experiments and
   animal models of reflux esophagitis induced by esophagojejunostomy
   [[52]11, [53]13]. These conditions stimulate EMT via activation of
   several key pathways, notably the vascular endothelial growth factor
   (VEGF) signaling pathway, which involves the zinc finger E-box-binding
   homeobox proteins (ZEB1/2) and transcription factors such as Snail and
   Slug, which suppress epithelial markers like E-cadherin (CDH1)
   [[54]13–[55]15]. Despite these observations, the specific molecular
   events underlying acidic bile salt-induced EMT and its implications for
   SSIM development require further elucidation. The recurrence of BE
   following treatment presents not only a major challenge in clinical
   management but also underscores the urgent need to deepen our
   understanding of the cellular and molecular mechanisms underlying BE
   persistence and relapse. In particular, the roles of EMT and SSIM as
   potential"hidden reservoirs"warrant focused investigation. Elucidating
   these processes may offer novel therapeutic targets and strategies for
   precision treatment and long-term management of BE patients.

   In the present study, we investigated the effects of prolonged exposure
   to acidic bile salts on EMT in Barrett’s epithelial cells, with
   particular attention to the underlying molecular mechanisms. Our
   findings highlight the Notch signaling pathway as a crucial mediator of
   EMT regulation in Barrett’s epithelial cells, notably influencing
   autophagy levels through modulation of UNC-52-like kinase 1 (ULK1)
   phosphorylation. Furthermore, we identified the long non-coding RNA
   (lncRNA) small nucleolar RNA host gene 1 (SNHG1) as a central
   regulatory molecule involved in modulating the ULK1-Notch1 signaling
   axis, thereby orchestrating the EMT process. These results provide
   significant new insights into the molecular mechanisms driving
   Barrett’s esophagus pathogenesis and progression toward malignancy.

Results

Acidic bile salts induce epithelial-mesenchymal transition in Barrett’s
esophagus through the Notch1 signaling pathway

   EMT is recognized as a critical process in the progression from BE to
   EAC, yet the precise upstream mechanisms driving EMT in BE remain
   poorly defined [[56]16]. While previous studies have demonstrated that
   acidic bile salts (ABS) can induce EMT-like features in BE, primarily
   through inflammatory and other signaling pathways, the specific
   molecular regulators underlying this process are not well-characterized
   [[57]13, [58]17]. Notably, although Notch1 signaling has emerged as a
   prominent regulator of EMT in various cancers and fibrotic diseases,
   its role in mediating EMT within Barrett’s esophagus has not been
   clearly established, representing a significant gap in current
   research.

   To address this gap, we first verified the induction of EMT by treating
   Barrett’s epithelial cells with acidic bile salt medium for 30 minutes
   a day for 14 days (AB14D). Subsequent assays confirmed significant
   increases in migration and invasion capabilities (Fig. [59]1a-d),
   coupled with hallmark EMT marker changes: reduced E-cadherin and
   enhanced expression of N-cadherin, Vimentin, MMP-2, MMP-9, and nuclear
   Snail (Fig. [60]1e-g). These findings reinforced that acidic bile salts
   robustly promote EMT phenotypes in Barrett’s epithelial cells. We then
   explored Notch1 signaling, considering its known regulatory function in
   EMT across other contexts. Analysis of publicly available data
   ([61]GSE26886) revealed Notch1 is significantly downregulated in BE
   compared to normal esophageal squamous epithelium (ESE) but becomes
   reactivated in EAC tissues (Fig. [62]1h, i), highlighting its potential
   dual role as a biomarker and a regulator during disease progression. To
   further substantiate the rationale for selecting the Notch signaling
   axis as our research focus, we conducted pathway enrichment analysis
   based on differentially expressed genes (DEGs) from the [63]GSE26886
   dataset, comparing BE tissues with EAC. As shown in Fig. [64]1j, the
   Notch signaling pathway ranked highest among all significantly enriched
   KEGG pathways, exhibiting the greatest fold enrichment and statistical
   significance (adjusted P < 0.05). While other EMT-related pathways such
   as TNF, MAPK, and thyroid hormone signaling also showed modest
   enrichment, their association was less robust compared to Notch
   signaling. This finding highlights the Notch pathway as a dominant
   upstream regulatory mechanism in BE progression. Our experiments
   demonstrated significant activation of Jagged1/Notch1 signaling upon
   acidic bile salt treatment (Fig. [65]1k). Crucially, silencing Notch1
   effectively reversed ABS-induced EMT, confirming Notch1’s essential
   regulatory role in driving EMT phenotypes and invasiveness in Barrett’s
   esophagus cells (Fig. [66]1l).

Fig. 1.

   [67]Fig. 1
   [68]Open in a new tab

   The regulation of EMT features in Barrett’s epithelial cells exposed to
   acidic bile salt is attributed to Notch1. a,b The scratch method was
   used to evaluate the capacity of cells to migrate. Scale Bar, 200 µm.
   c,d The number of migratory and invasive cells was measured by
   trans-well analysis. Scale Bar, 100 µm. e,f N-cadherin’s (shown in red)
   fluorescence intensity was evaluated by immunofluorescence analysis,
   which was then quantitatively analyzed. DAPI labeling was utilized to
   make the cell nuclei visible (blue). Scale Bar, 10 µm. g Using β-actin
   as a reference, Western blot analysis was used to assess the protein
   expression levels of N-cadherin, Vimentin, MMP2, snail, and MMP9. h, i
   Publicly available data ([69]GSE26886) demonstrate that Notch1
   expression is downregulated in BE compared to normal ESE but
   reactivated in esophageal adenocarcinoma (EAC) tissues. j KEGG
   enrichment analysis of Publicly available data ([70]GSE26886) between
   BE and EAC. k Using β-actin as a control, Western blot analysis
   evaluated the protein expression of Jagged1 and Notch1. l Silencing
   Notch1 reverses ABS-induced EMT, confirming the critical role of Notch1
   in driving EMT and invasiveness in Barrett’s esophagus cells. The
   control (Con) conditions refer to the absence of AB14D treatment. We
   used one-way ANOVA with Tukey’s post hoc test for multiple group
   comparisons, and Student’s t-test or Mann-Whitney U test for two-group
   comparisons. The values are the means ± standard error of at least
   three separate studies

   Collectively, these results establish the novel involvement of Notch1
   signaling in mediating EMT induced by acidic bile salts in BE,
   addressing an important gap in understanding Barrett’s esophagus
   progression and providing potential new avenues for therapeutic
   intervention.

Autophagy modulates acidic bile salt-induced EMT via Notch1 signaling in
Barrett’s epithelial cells

   Autophagy has previously been implicated in regulating EMT processes
   via modulation of the Notch1 signaling pathway [[71]18]. Given this
   potential regulatory role, we hypothesized a connection between
   autophagy suppression and Notch1 signaling during ABS-induced EMT in
   Barrett’s epithelial cells.

   To test this hypothesis, we evaluated autophagy activity in Barrett’s
   epithelial cells exposed to acidic bile salts for either a short (30
   minutes, AB30M) or prolonged period (AB14D). Western blot analysis
   revealed that short-term exposure (AB30M) significantly reduced p62
   levels and modestly increased LC3-II expression compared to control
   cells, suggesting enhanced autophagic flux under acute acidic bile
   stimulation. However, under chronic exposure (AB14D), LC3-II levels
   were significantly reduced, while p62 levels returned to near-baseline,
   indicating suppressed autophagy under chronic ABS exposure (Fig.
   [72]2a). Electron microscopy and LC3 autophagic flux assays further
   confirmed reduced autophagic vesicle formation in AB14D-treated BAR-T
   cells (Fig. [73]2b, c). Subsequently, we investigated whether reversing
   this suppression by activating autophagy could mitigate the EMT
   phenotype. Western blot analyses demonstrated that Rapamycin (RAPA), an
   autophagy activator, effectively restored autophagy, significantly
   elevating LC3-II levels and reducing SQSTM1/p62 expression compared to
   untreated controls (Fig. [74]2d). Functionally, RAPA-treated BAR-T and
   CP-C cells exhibited significantly lower migration rates in scratch
   wound-healing assays (Fig. [75]2e) and decreased invasion capacities in
   transwell assays (Fig. [76]2f) compared to cells exposed solely to
   acidic bile salts. Immunofluorescence assays showed notably reduced
   N-cadherin expression after RAPA treatment (Fig. [77]2g). Further
   Western blot analysis confirmed that, under the conditions of EMT
   induced by AB14D, RAPA activated autophagy, upregulating the epithelial
   marker E-cadherin and significantly downregulating the expression of
   mesenchymal markers N-cadherin, MMP2, MMP9, and Vimentin (Fig. [78]2h).
   Additionally, under the conditions of EMT induced by AB14D,
   RAPA-mediated autophagy activation significantly inhibited Notch
   signaling, as evidenced by reduced levels of Notch1 and Snail proteins
   (Fig. [79]2h).

Fig. 2.

   [80]Fig. 2
   [81]Open in a new tab

   Autophagy inhibits EMT in acidic bile salt treatment Barrett’s
   epithelial cells by Notch1 signaling. a Western blot analysis using
   β-actin as a control was used to evaluate the expression of the p62 and
   LC3 proteins. b,c Electron microscopy and LC3 autophagic flux assays
   further confirm reduced autophagic vesicle formation in AB14D-treated
   cells. d Using β-actin as a reference, Western blot analysis revealed
   the expression of p62 and LC3 proteins after RAPA therapy. e Following
   RAPA treatment, the scratch technique was used to assess the cell
   migratory ability. Scale Bar, 200 µm. f After receiving RAPA therapy,
   the cells’ ability to migrate and invade was ascertained using
   transwell examination. Scale Bar, 100 µm. g After RAPA therapy,
   immunofluorescence analysis evaluated the fluorescence intensity of
   N-cadherin (shown in red) and performed a quantitative analysis on it.
   The blue-colored DAPI labeling made the cell nuclei visible. Scale Bar,
   10 µm. h Using β-actin as a reference, Western blot analysis assessed
   the protein expression of Vimentin, MMP-2, MMP-9, N-cadherin, Notch1,
   Snail, and E-cadherin after RAPA therapy. The control (Con) conditions
   refer to the absence of AB14D treatment. We used one-way ANOVA with
   Tukey’s post hoc test for multiple group comparisons, and Student’s
   t-test or Mann-Whitney U test for two-group comparisons. Values
   represent mean values ± SEM from at least three independent experiments

   Collectively, these findings indicate that suppression of autophagy is
   necessary for ABS-induced EMT, and reactivation of autophagy
   effectively inhibits this process by modulating the Notch1 signaling
   pathway. Thus, targeting autophagy presents a promising therapeutic
   strategy for preventing EMT and its associated progression in Barrett’s
   esophagus.

ULK1-Notch1 axis participates in Barrett’s epithelial cell EMT stimulated by
acidic bile salts

   ULK1 plays a crucial role in autophagy initiation, serving as a primary
   target of mTOR-mediated phosphorylation. RAPA, a potent autophagy
   activator, inhibits mTOR and consequently regulates ULK1
   phosphorylation, thereby facilitating autophagy induction [[82]19]. In
   our experiments, Western blot analyses revealed a significant decrease
   in p-ULK1 (Ser757) expression following RAPA treatment compared to
   untreated control cells (Fig. [83]3a), confirming effective autophagy
   activation.

Fig 3.

   [84]Fig 3
   [85]Open in a new tab

   DAPT controls autophagy by means of the ULK1-Notch1 pathway. a Western
   blot analysis using β-actin as a control was used to evaluate the
   protein expression of ULK1, p-ULK1(S757), p62, and LC3 after RAPA
   therapy. b Protein-protein docking analyses reveal a strong interaction
   between ULK1 and Notch1, with a docking score of 99.059. Visualizations
   show hydrogen bonds between specific residues of Notch1 and ULK1,
   indicating a potential physical interaction. c Following stimulation
   with acidic bile salt, Notch1 and ULK1 in Barrett’s epithelial cells
   were shown to physically interact, as demonstrated by Co-IP tests. d
   Using β-actin as a reference, Western blot analysis assessed the
   protein expression of ULK1, p-ULK1(S757), p62, and LC3 after treatment
   with RAPA or the γ-secretase inhibitor DAPT. We used one-way ANOVA with
   Tukey’s post hoc test for multiple group comparisons, and Student’s
   t-test or Mann-Whitney U test for two-group comparisons. The values are
   the means ± standard error of at least three separate studies

   To further investigate the potential link between autophagy and Notch1
   signaling through ULK1, we conducted protein-protein docking analyses
   using Discovery Studio 4.0 software with the ZDOCK module. The
   calculated docking score between NOTCH1 and ULK1 proteins was 99.059,
   indicative of a strong interaction. Docking visualizations identified
   hydrogen bonds formed between NOTCH1 residues (His363, Ser385, Gly2865,
   Val1867, and Cys1871) and ULK1 residues (Asp40, Lys13, Glu83, Met84,
   and Lys50) (Fig. [86]3b). Co-immunoprecipitation (Co-IP) assays further
   validated the physical interaction between Notch1 and ULK1 proteins in
   Barrett’s epithelial cells (Fig. [87]3c). To assess the functional role
   of Notch1 signaling in regulating ULK1 phosphorylation and autophagy in
   Barrett’s epithelial cells, we used DAPT, a γ-secretase inhibitor known
   to effectively suppress Notch activation by blocking its cleavage
   [[88]20]. Western blot analysis demonstrated that treatment with either
   RAPA or DAPT alone significantly elevated LC3 expression while reducing
   Notch1, p-ULK1, and p62 levels in acidic bile salt (AB14D)-treated
   Barrett’s epithelial cells (Fig. [89]3d). Notably, the combined
   administration of RAPA and DAPT amplified these effects, indicating
   even more pronounced reductions in Notch1, p-ULK1, and increased LC3
   compared to single treatments (Fig. [90]3d). These results suggest that
   DAPT-mediated Notch1 inhibition blocks acidic bile salt-induced ULK1
   phosphorylation, thus promoting autophagy activity.

   In summary, our findings demonstrate that the ULK1-Notch1 axis plays a
   central role in mediating acidic bile salt-induced EMT in Barrett’s
   epithelial cells by modulating autophagy. Specifically, ULK1 directly
   interacts with Notch1, and their phosphorylation interplay governs
   autophagy suppression and EMT promotion. Pharmacological inhibition of
   Notch signaling (DAPT) or mTOR signaling (RAPA) not only disrupts ULK1
   phosphorylation but also enhances autophagic activity and mitigates EMT
   features. These results highlight the ULK1-Notch1 signaling axis as a
   critical regulatory hub in the pathogenesis of Barrett’s esophagus.

SNHG1 clinically correlates with Barrett’s progression and mediates the
interaction between ULK1 and Notch1

   LncRNAs function as regulators of gene expression, exerting diverse
   effects on numerous crucial physiological processes. These effects
   include acting as chromatin modification factors, X chromosome
   inactivation factors, enhancers, transcriptional regulatory factors,
   and post-transcriptional regulatory factors, among others [[91]21,
   [92]22]. Certain lncRNAs are known to interact with ULK1, potentially
   influencing its modification and function [[93]23]. Therefore, we
   performed weighted correlation network analysis (WGCNA) to identify
   critical eigengenes associated with the initiation and progression of
   BE and to construct a regulatory network (Fig. [94]4a).

Fig. 4.

   [95]Fig. 4
   [96]Open in a new tab

   Construction of the core regulatory network for Barrett’s esophagus
   patients. a Workflow illustrating the identification of lncRNAs
   associated with BE progression through WGCNA and RIP-seq integration. b
   Scale independence and mean connectivity under each power. c The
   cluster dendrogram of each module. d The relationships of module-trait
   (correlation greater than 0.3). e, f Correlation between yellow and
   turquoise module membership and gene significance for Barrett. g The
   lncRNA subnetwork that regulated by ULK1 and Notch1. h Clinical
   validation of SNHG1 expression in BE patient biopsies (n = 20).
   Patients were stratified based on follow-up histological progression
   into indolent (non-progressors), LGD, and HGD groups. SNHG1 was
   significantly upregulated in patients progressing to LGD or HGD (P <
   0.05), as assessed by qPCR

   For network construction, we selected a soft threshold of 4 to ensure a
   scale-free distribution (Fig. [97]4b). The WGCNA analysis yielded 49
   distinct gene modules (Fig. [98]4c). ULK1 and Notch1 were found within
   the turquoise module, which showed a strong correlation with BE traits
   (correlation = −0.94, P = 5×10^−19; Fig. [99]4c, d). To explore lncRNAs
   regulated by ULK1 and Notch1 and their potential roles in BE
   progression, we constructed a regulatory network using genes from the
   turquoise and yellow modules, both demonstrating high module membership
   and gene significance for BE (Fig. [100]4e, f). From this regulatory
   network, we extracted the lncRNA subnetwork regulated by both ULK1 and
   Notch1, identifying 14 lncRNAs highly correlated with BE progression
   (Fig. [101]4g). We further intersected these 14 candidate lncRNAs with
   822 lncRNAs identified from RNA immunoprecipitation (RIP) sequencing
   targeting ULK1. This intersection identified six candidate lncRNAs:
   SNHG1, LINC00621, ARMCX5-GPRASP2, DLEU1, PCAT6, and NNT-AS1. However,
   qPCR validation in acidic bile salt-treated Barrett’s epithelial cells
   revealed that only SNHG1 exhibited significantly altered expression
   compared to control cells. These findings led us to hypothesize that
   SNHG1 mediates the interaction between ULK1 and Notch1, potentially
   influencing BE progression (Supplementary Fig. 1). To further
   substantiate the clinical relevance of SNHG1 in Barrett’s esophagus
   progression, we analyzed SNHG1 expression in a well-characterized
   cohort of BE patients (n = 20) who underwent longitudinal endoscopic
   surveillance (Fig. [102]4h). Based on histological evaluation and
   clinical follow-up, patients were classified into three categories:
   indolent (non-progressors), progressed to low-grade dysplasia (LGD),
   and progressed to high-grade dysplasia (HGD). Importantly, quantitative
   PCR analysis of endoscopic biopsy samples revealed that SNHG1
   expression was significantly upregulated in patients who progressed to
   LGD or HGD compared to the indolent group. These findings further
   validate the association between elevated SNHG1 expression and BE
   progression, reinforcing its potential utility as a biomarker for risk
   stratification in clinical settings.

   In summary, this section identifies SNHG1 as a clinically and
   mechanistically relevant lncRNA that modulates the interaction between
   ULK1 and Notch1, and is significantly upregulated in BE patients who
   later developed dysplasia. These findings support SNHG1 as a critical
   molecular of disease progression in Barrett’s esophagus.

SNHG1 regulates ULK1-Notch1 axis-mediated autophagy inhibition and EMT
activation in Barrett’s epithelial Cells

   We observed increased transcription levels of SNHG1 in Barrett’s
   epithelial cells upon exposure to acidic bile salts (Supplementary Fig.
   1). Weighted correlation network analysis (WGCNA) suggested that SNHG1
   participates in the ULK1-Notch1 signaling axis. To validate the
   physical association between SNHG1 and the ULK1-Notch1 axis, we
   performed Co-IP assays, which demonstrated that SNHG1 modulates the
   interaction between Notch1 and ULK1 (Fig. [103]5a, b). Specifically,
   SNHG1 overexpression enhanced the interaction, while its knockdown
   weakened the binding affinity between ULK1 and Notch1.

Fig. 5.

   [104]Fig. 5
   [105]Open in a new tab

   SNHG1 inhibits autophagy and activates EMT through modulation of the
   ULK1-Notch1 axis in Barrett’s epithelial cells. a The overexpression
   and silence efficiency of SNHG1. b Following SNHG1 overexpression or
   knockdown, Co-IP tests verified the physical connection between Notch1
   and ULK1 in Barrett’s epithelial cells. c The scratch experiment
   evaluated the ability of cells to migrate when SNHG1 was overexpressed
   or knocked down. Scale Bar, 200 µm. d After overexpressing or knocking
   down SNHG1, immunofluorescence analysis detected the fluorescence
   intensity of N-cadherin (shown in red) and quantitatively analyzed it.
   The blue-colored DAPI labeling made the cell nuclei visible. Scale Bar,
   10 µm. e Western blot analysis used β-actin as a reference to assess
   the protein expression of Notch1, E-cadherin, N-cadherin, MMP2, MMP9,
   and Vimentin after SNHG1 was overexpressed or knocked down. f Western
   blot analysis used β-actin as a reference to assess the protein
   expression of p-ULK1, ULK1, p62, and LC3 after SNHG1 was overexpressed
   or knocked down. g LC3 autophagic flux assays further support these
   findings. These results suggest that SNHG1 inhibits autophagy in BAR-T
   cells. h The migratory and invasive potential of cells after SNHG1
   overexpression can be reversed by the DAPT or RAPA when assessed by
   transwell analysis. Scale Bar, 100 µm. The empty vector (EV) serves as
   the overexpression control. Normal control (NC) refers to normal
   control, which is the siNotch1 control. We used one-way ANOVA with
   Tukey’s post hoc test for multiple group comparisons, and Student’s
   t-test or Mann-Whitney U test for two-group comparisons. The values are
   the means ± standard error of at least three separate studies

   Based on these findings, we hypothesized that SNHG1 might regulate EMT
   characteristics in Barrett’s epithelial cells. To test this hypothesis,
   we evaluated cell migration using scratch wound-healing assays.
   Enhanced SNHG1 expression significantly increased cell migration
   compared to controls, whereas SNHG1 knockdown via siRNA markedly
   inhibited cell migration (Fig. [106]5c). These results indicate that
   SNHG1 promotes EMT in Barrett’s epithelial cells. Immunofluorescence
   assays further supported this conclusion, showing substantially
   decreased N-cadherin expression after SNHG1 knockdown and elevated
   N-cadherin levels following SNHG1 overexpression (Fig. [107]5d).

   We next assessed EMT-related protein expression by Western blot.
   Overexpression of SNHG1 led to decreased E-cadherin and increased
   levels of Notch1, N-cadherin, MMP2, MMP9, and Vimentin compared to
   control groups (Fig. [108]5e). Conversely, SNHG1 silencing
   significantly increased E-cadherin expression and decreased Notch1,
   N-cadherin, MMP2, MMP9 and Vimentin (Fig. [109]5e). Moreover, Western
   blot analyses revealed that SNHG1 overexpression significantly
   increased p-ULK1 and p62 while decreasing LC3 levels. The opposite
   pattern was observed upon SNHG1 knockdown (Fig. [110]5f). LC3
   autophagic flux experiments provided further validation of these
   findings (Fig. [111]5g). Importantly, re-expression of SNHG1 in
   siSNHG1-transfected cells restored EMT and autophagy phenotypes,
   confirming the specificity of the knockdown and excluding off-target
   effects (Supplementary Fig. 2). Collectively, these results demonstrate
   that SNHG1 inhibits autophagy and promotes EMT by modulating the
   ULK1-Notch1 axis in Barrett’s epithelial cells. To further validate the
   regulatory relationship between SNHG1 and ULK1-Notch1 expression, we
   conducted rescue experiments. We found that the EMT enhancement induced
   by SNHG1 overexpression was reversed upon RAPA-mediated autophagy
   activation or DAPT-induced Notch1 inhibition. This indicates that the
   regulatory effect of SNHG1 on EMT is dependent on the ULK1-Notch1
   signaling pathway (Fig. [112]5h).

   Thus, these findings establish SNHG1 as a critical modulator of EMT in
   Barrett’s epithelial cells through inhibition of autophagy and
   enhancement of the ULK1-Notch1 signaling interaction. The reversal of
   EMT phenotypes by RAPA and DAPT further confirms that the effects of
   SNHG1 are mediated via the ULK1-Notch1 axis, highlighting its potential
   as a therapeutic target in Barrett’s esophagus.

SNHG1 modulates EMT in a Barrett’s esophagus mouse model
(p16^flox/flox/Kras^G12D)

   To validate our in vitro findings, we performed in vivo experiments
   using genetically engineered mice that spontaneously developed
   Barrett’s esophagus. Figure [113]6a illustrates the experimental
   design, in which p16^flox/flox/Kras^G12D mice underwent
   tamoxifen-mediated gene activation starting at three months of age and
   were continuously exposed to 0.2% bile acid (BA) in their drinking
   water until euthanasia at nine months. Tissue samples were then
   collected for histological analysis.

Fig. 6.

   [114]Fig. 6
   [115]Open in a new tab

   SNHG1 Modulates EMT in Squamocolumnar Junction (SCJ) Tissues of
   p16^flox/flox/Kras^G12D Mice a Experimental design of the
   p16^flox/flox/Kras^G12D genetically engineered mice. From 3 months of
   age, mice were treated with tamoxifen for gene regulation and exposed
   to 0.2% bile acid (BA) in drinking water until euthanasia at 9 months
   for tissue collection and analysis. b Histological analysis of
   inflammation and metaplasia in wild-type and p16^flox/flox/Kras^G12D
   mice. Quantitative analysis revealed significantly higher inflammation
   and metaplasia scores in p16^flox/flox/Kras^G12^D mice compared to
   wild-type controls (p < 0.05 and p < 0.01). Scale Bar, 100 µm. c
   Immunohistochemical staining of EMT markers in SCJ organoids from
   p16^flox/flox/Kras^G12D mice. Overexpression of SNHG1 (OE-SNHG1)
   resulted in reduced E-cadherin expression and increased N-cadherin
   levels, indicating enhanced EMT. Knockdown of SNHG1 (sh-SNHG1) reversed
   these effects, with increased E-cadherin and decreased N-cadherin
   expression. Scale Bar, 10 µm. The EV conditions refer to the empty
   vector, which serves as the overexpression control. NC refers to the
   normal control, which is shNotch1 control. We used one-way ANOVA with
   Tukey’s post hoc test for multiple group comparisons, and Student’s
   t-test or Mann-Whitney U test for two-group comparisons. The values are
   the means ± standard error of at least three separate studies

   Histological examinations showed that wild-type control mice did not
   display significant inflammation or metaplasia. In contrast,
   p16^flox/flox/Kras^G12D mice exhibited notable tissue damage,
   characterized by pronounced inflammation and evident metaplastic
   changes. Quantitative scoring confirmed significantly elevated
   inflammation (p < 0.05) and metaplasia (p < 0.01) in the
   p16^flox/flox/Kras^G12D group compared to wild-type controls (Fig.
   [116]6b).

   To further investigate SNHG1’s regulatory role in EMT within Barrett’s
   metaplastic cells at the mouse SCJ, we generated organoids from mouse
   SCJ tissues and conducted targeted gene manipulation of SNHG1.
   Immunohistochemical analysis revealed that SNHG1 overexpression
   (OE-SNHG1) resulted in decreased E-cadherin and increased N-cadherin
   expression, indicative of EMT enhancement. Conversely, SNHG1 knockdown
   (sh-SNHG1) reversed these effects, significantly elevating E-cadherin
   levels while reducing N-cadherin expression (Fig. [117]6c).

   These findings collectively underscore SNHG1’s critical role in
   modulating EMT processes and driving the progression of Barrett’s
   metaplasia within the p16^flox/flox/Kras^G12D mouse model.

Discussion

   BE is a condition characterized by abnormal remodeling of esophageal
   mucosa, closely associated with chronic gastroesophageal reflux disease
   (GERD). Specialized intestinal-type Barrett’s epithelial cells commonly
   replace normal esophageal mucosa, significantly increasing the risk of
   esophageal adenocarcinoma [[118]24]. Despite effective treatments such
   as radiofrequency ablation (RFA), recurrence remains frequent due to
   subsquamous intestinal metaplasia (SSIM), closely linked with EMT
   driven by chronic exposure to acidic bile salts [[119]6, [120]25,
   [121]26]. Acidic bile salts influence several critical signaling
   pathways, including VEGF and Notch1, facilitating EMT and malignant
   transformation [[122]13, [123]17]. Dysfunctional Notch signaling is
   widely recognized to accelerate cancer progression by influencing cell
   differentiation, proliferation, and invasiveness [[124]27–[125]29]. Our
   results specifically highlight the activation of Notch signaling as a
   crucial driver of EMT in Barrett’s epithelial cells exposed to acidic
   bile salts, reinforcing its significant role in BE progression toward
   adenocarcinoma.

   Autophagy, an essential cellular degradation process, plays a central
   role in modulating EMT through interactions with the Notch1 signaling
   pathway [[126]18, [127]30] Activation of autophagy facilitates the
   degradation of the Notch intracellular domain (NICD), a key
   transcriptional regulator of EMT and cancer metastasis [[128]18]. Our
   study further elucidates this interaction by demonstrating that NICD
   directly interacts with and phosphorylates ULK1, thereby suppressing
   autophagy initiation. RAPA, an mTOR inhibitor known to activate
   autophagy, was shown to reduce NICD levels and suppress EMT markers
   effectively. Additionally, using DAPT, a γ-secretase inhibitor that
   inhibits Notch signaling, we validated the critical role of Notch
   signaling in modulating autophagy via ULK1 phosphorylation. Notably,
   the combined application of DAPT and RAPA significantly enhanced
   autophagy and effectively inhibited EMT compared to either treatment
   alone. Co-immunoprecipitation (Co-IP) assays confirmed the direct
   interaction between Notch1 and ULK1, underscoring the importance of the
   ULK1-Notch1 axis in regulating EMT and autophagy in BE cells.

   Long non-coding RNAs (lncRNAs), defined as transcripts longer than 200
   nucleotides, play essential roles in regulating gene expression across
   diverse biological contexts [[129]31, [130]32]. Among them, SNHG1 has
   been previously implicated in tumor proliferation and metastasis,
   particularly in esophageal squamous cell carcinoma and gastric cancer,
   where it exerts oncogenic effects through pathways such as PI3 K/AKT or
   Wnt/β-catenin [[131]27]. However, its role in BE has remained
   unexplored. Our study identifies SNHG1 as a novel driver of BE
   progression by promoting EMT through modulation of the ULK1-Notch1
   axis. Functional experiments confirmed that altering SNHG1 expression
   significantly impacts EMT markers, cell migration, and autophagy
   status, both in vitro and in vivo.

   In contrast to previous studies that independently examined autophagy
   and Notch1 signaling in BE or other cancers [[132]33, [133]34], our
   findings mechanistically connect these two processes through ULK1, a
   key autophagy initiator. We demonstrate that ULK1 physically interacts
   with Notch1, and that this interaction is modulated by SNHG1, forming a
   unified regulatory circuit. This SNHG1–ULK1–Notch1 axis coordinates the
   suppression of autophagy and activation of EMT, a mechanism not
   previously reported in BE. Thus, our work provides a new dimension to
   the functional repertoire of SNHG1 and adds mechanistic depth to our
   understanding of EMT regulation under chronic acidic bile salt
   exposure.

   The implications of these findings are twofold. First, they suggest
   that SNHG1 could serve as a predictive biomarker for BE progression,
   particularly in identifying patients at risk for dysplasia following
   RFA treatment. Second, the observed reversibility of SNHG1-induced EMT
   via RAPA or DAPT highlights potential therapeutic strategies for
   halting or reversing BE evolution. Targeting SNHG1 directly, or its
   downstream signaling components ULK1 and Notch1, could offer new
   avenues for chemopreventive intervention and improve the long-term
   management of BE. These insights support future development of
   lncRNA-based diagnostics and therapeutics tailored to the molecular
   subtype of Barrett’s epithelium.

   This study has several limitations. First, our findings rely heavily on
   the BAR-T and CP-C cell lines, immortalized Barrett’s epithelial
   models, which may not fully capture the complexity and heterogeneity of
   human BE. Using additional models, such as patient-derived organoids,
   could enhance translational relevance. Second, although the
   p16^flox/flox/Kras^G12D mouse model provides valuable insights, it
   cannot entirely replicate the multifactorial and heterogeneous nature
   of BE in humans, particularly regarding environmental exposures and
   genetic diversity. Third, clinical validation of the SNHG1-ULK1-Notch1
   axis in patient samples is currently lacking, limiting direct clinical
   applicability. Further studies utilizing human tissues are necessary to
   confirm clinical relevance. Methodologically, Co-IP assays, while
   valuable, may not definitively capture transient or direct protein
   interactions, and additional validation methods are needed to confirm
   specificity.

   Collectively, our findings uncover novel molecular pathways underlying
   BE pathogenesis, emphasizing the significance of the SNHG1-ULK1-Notch1
   axis as a promising therapeutic target. This research provides
   foundational insights for developing personalized therapeutic
   strategies aimed at preventing recurrence and progression in patients
   with Barrett’s esophagus.

Materials and methods

Ethical approval of clinical studies and animal studies

   The study was approved by the Ethics Committee of West China Hospital,
   Sichuan University, and all participants signed informed consent forms
   (2020(1094)). All animal experimental protocols were approved by the
   Ethics Committee of West China Hospital, Sichuan University (Approval
   number: No.20220107005). All animal experiments were in accordance with
   the guide for the care and use of laboratory animals established by the
   United States National Institutes of Health.

Clinical specimen collection and classification

   Endoscopic biopsy specimens were obtained from patients with
   histologically confirmed BE at West China Hospital, Sichuan University.
   Written informed consent was obtained from all participants prior to
   tissue collection. For each patient, biopsy samples were obtained from
   the squamocolumnar junction during routine surveillance endoscopy and
   stored in RNAlater (Thermo Fisher Scientific, USA) at −80 °C until RNA
   extraction. Based on longitudinal histopathologic evaluation and
   clinical follow-up over a minimum of 24 months, patients were
   categorized into three groups: (1) indolent/non-progressors, (2)
   progressed to low-grade dysplasia (LGD), and (3) progressed to
   high-grade dysplasia (HGD). Only initial diagnostic biopsy specimens
   were used for molecular analysis. SNHG1 expression was quantified via
   quantitative real-time PCR (qPCR), and its association with progression
   status was statistically evaluated.

Cell culture and treatment

   Barrett’s epithelial cells (BAR-T and CPC cells) were purchased from
   ATCC (Virginia, USA). These cells were grown in 1640 medium from
   Hyclone (Logan, UT, USA), supplemented with 10% fetal bovine serum, and
   were maintained in an incubator at 37 °C in an atmosphere containing 5%
   CO[2].

Cell transfection

   The SNHG1 overexpression plasmids and siRNA were synthesized by
   Genechem (shanghai, China). Barrett’s epithelial cells cultured in DMEM
   were transfected with Lipofectamine 3000 (Invitrogen, Carlsbad, CA,
   USA) once they reached 70%–80% confluence, and the transfection was
   allowed to proceed for 36 hours.

Scratch wound assay

   Initially, 6-well plates were used to seed Barrett’s epithelial cells,
   which were then incubated for 24 hours. Next, using a sterile 200 μL
   pipette tip, each well’s cells were gently scraped. After three PBS
   rinses, cells were cultured for a further twenty-four hours in DMEM
   with different treatments according to their groups. Two times were
   recorded for the images: the first time at 0 hours and the second time
   24 hours later. Images were captured using a inverted microscope (Leica
   DMi8, Germany). Using ImageJ software, the wound area was defined and
   measured [[134]35].

Trans-well assay

   A Millipore trans-well chamber (Burlington, MA, USA) was used to hold a
   cell suspension at a concentration of 1 × 10^5 cells/mL (200 μL) in the
   upper compartment, and 800 μL of specialized media were placed in the
   lower compartment for every experimental group. After a 24-hour
   incubation period, the upper chamber cells were extracted, and the
   chambers were fixed with 4% paraformaldehyde for 30 minutes. Following
   that, staining was done for 20 minutes using a 0.1% crystal violet
   solution. Finally, the stained cells in five randomly selected sections
   of each well were quantified using an inverted microscope (Leica DMi8,
   Germany) [[135]36].

Immunofluorescence staining

   BAR-T and CPC cells were seeded onto sterilized glass coverslips in
   24-well plates and allowed to adhere overnight. Following treatment,
   cells were washed twice with PBS and fixed with 4% paraformaldehyde for
   15 minutes at room temperature. Cells were then permeabilized using
   0.1% Triton X-100 for 10 minutes and blocked with 5% bovine serum
   albumin (BSA) for 1 hour. Subsequently, cells were incubated with
   primary antibodies overnight at 4 °C. After washing with PBS, cells
   were incubated with secondary antibodies for 1 hour at room temperature
   in the dark. Nuclei were counterstained with DAPI (1 μg/mL) for 5
   minutes. Coverslips were mounted with anti-fade mounting medium and
   observed using a fluorescence microscope (Olympus IX71, Japan).
   Fluorescence intensity was quantified using ImageJ software.

Western blotting

   RIPA buffer (Beyotime, Shanghai, China) was utilized to lyse the
   Barrett’s epithelial cells. To quantify the protein levels, we employed
   the BCA protein assay kit (Beyotime, Shanghai, China). The separation
   of target proteins by their molecular weights was achieved through
   sodium dodecyl sulfate-polyacrylamide gel electrophoresis, followed by
   their transfer onto PVDF membranes (Millipore, Burlington, MA, USA).
   After that, a 2-hour room-temperature blocking stage was completed.
   After applying primary antibodies, secondary antibodies coupled with
   horseradish peroxidase were added to identify proteins. Enhanced
   chemiluminescence was used as the HRP substrate (4A Biotech, Beijing,
   China) to see the proteins. We employed ImageJ software (National
   Institutes of Health, USA) to do quantitative protein analysis
   [[136]37].

Antibodies

   The following antibodies were used: E-cadherin (20874-1-AP,
   proteintech, Wuhan, China), N-cadherin (A19083, ABclonal, Wuhan,
   China), MMP9 (10375-2-AP, proteintech, Wuhan, China), MMP2 (66366-1-Ig,
   proteintech, Wuhan, China), Vimentin (A19607, ABclonal, Wuhan, China),
   Snail (3879S, Cell Signaling Technology, Danvers, MA, USA), Jagged1
   (70109S, Cell Signaling Technology, Danvers, MA, USA), Notch1(3608S,
   Cell Signaling Technology, Danvers, MA, USA; ab52627, Abcam, Cambridge,
   UK), p62 (CST88588S, Cell Signaling Technology, Danvers, MA, USA), LC3
   (ab192890, Abcam, Cambridge, UK), p-ULK1 (14202S, Cell Signaling
   Technology, Danvers, MA, USA), ULK1 (8054S, Cell Signaling Technology,
   Danvers, MA, USA; ab167139, Abcam, Cambridge, UK), β-actin (66009-1-Ig,
   proteintech, Wuhan, China).

Quantitative Real-Time PCR

   Total RNA was extracted from Barrett’s epithelial cells using TRIzol
   reagent (Invitrogen, USA) according to the manufacturer’s instructions.
   RNA concentration and purity were measured using a NanoDrop 2000
   spectrophotometer (Thermo Fisher Scientific, USA). Complementary DNA
   (cDNA) was synthesized using the PrimeScript™ RT Reagent Kit with gDNA
   Eraser (Takara, Japan). Quantitative real-time PCR was performed using
   TB Green™ Premix Ex Taq™ II (Takara, Japan) on a CFX96 Real-Time PCR
   Detection System (Bio-Rad, USA). Relative gene expression was
   calculated using the 2^-ΔΔCt method, with β-actin as the internal
   control. Each reaction was conducted in triplicate. The primer
   sequences were listed in supplementary materials.

Transmission Electron Microscopy (TEM)

   To visualize autophagosomes, BAR-T cells were fixed in 2.5%
   glutaraldehyde (Sigma-Aldrich, USA) in 0.1 M phosphate buffer (pH 7.4)
   at 4 °C overnight. Samples were post-fixed with 1% osmium tetroxide for
   2 hours, dehydrated through a graded ethanol series, and embedded in
   epoxy resin. Ultrathin sections (70 nm) were obtained using an
   ultramicrotome (Leica EM UC7, Germany) and mounted on copper grids.
   Sections were stained with uranyl acetate and lead citrate, then
   examined under a transmission electron microscope (Hitachi HT7700,
   Japan). Autophagic vacuoles were identified and quantified in randomly
   selected fields at 20,000× magnification.

LC3 autophagic flux detection via dual fluorescent reporter

   To assess autophagic flux, BAR-T cells were transfected with an
   mRFP-GFP-LC3 tandem fluorescent-tagged plasmid (Genechem, Shanghai,
   China) using Lipofectamine 3000 (Invitrogen, USA) according to the
   manufacturer’s instructions. After 24 hours of transfection, cells were
   treated under indicated conditions (e.g., acidic bile salts). The
   mRFP-GFP-LC3 reporter enables discrimination between autophagosomes
   (GFP+/RFP+, yellow puncta) and autolysosomes (GFP−/RFP+, red puncta),
   as the GFP signal is quenched in the acidic lysosomal environment.
   Cells were imaged using a confocal laser scanning microscope (Leica TCS
   SP8, Germany). Quantification of LC3 puncta was performed using ImageJ.

Protein-protein docking analysis

   To predict the potential physical interaction between ULK1 and Notch1,
   protein–protein docking analysis was performed using Discovery Studio
   4.0 software (BIOVIA, San Diego, CA, USA). The three-dimensional
   structure of ULK1 was obtained from the Protein Data Bank (PDB ID:
   4WNO), which represents the crystal structure of the ULK1 kinase domain
   in complex with a small-molecule inhibitor. As a full-length
   experimentally resolved structure of Notch1 is currently unavailable, a
   high-confidence predicted model from the AlphaFold database
   (ID：AF-[137]P46531-F1) was employed for docking analysis. Rigid-body
   docking was conducted using the ZDOCK module, which evaluates candidate
   complexes based on shape complementarity, electrostatic interactions,
   and desolvation energy. The top-scoring complex was visualized in three
   dimensions, and the predicted binding interface was analyzed for
   hydrogen bond formation and spatial compatibility to infer potential
   interaction sites and binding modes between ULK1 and Notch1.

Co-IP assay

   Barrett’s epithelial cells were cultured for 1 hour in IP lysis buffer
   (Beyotime, Shanghai, China). The supernatant was produced by
   centrifuging cell lysates at 12,000 RPM for 10 minutes to separate
   them. A 10 μL portion of the supernatant was set aside for measuring
   the protein content, while the rest was gently mixed with specific
   antibodies or normal IgG from the same species and incubated at 4 °C
   for the entire night. The samples were incubated for 4 hours with the
   addition of Protein A/G magnetic beads (Thermo Fisher Scientific,
   Waltham, USA) and a wash with IP lysis solution. Subsequently, the
   beads were washed three times with PBS and twice with IP lysis
   solution. 2× SDS loading buffer (Beyotime, Shanghai, China), which was
   used to elute and denature the precipitated proteins for ten minutes at
   100 °C. Ultimately, Western blotting was used to examine the proteins
   [[138]38, [139]39].

WGCNA analysis

   To better understand the pathogenesis of Barrett’s esophagus patients,
   we downloaded the [140]GSE26886 dataset from the GEO database
   ([141]https://www.ncbi.nlm.nih.gov/geo/). In the downstream analysis,
   we included the expression data of Barrett’s esophagus patients (n=20)
   and Control samples (biopsies from patients with normal esophageal
   squamous epithelium, n=19). R package WGCNA (version 1.70–3) was used
   to explore the eigengenes closely related to the occurrence and
   development of Barrett’s esophagus and the critical interaction between
   eigengenes. Firstly, we use the pickSoftThreshold function to filter
   the optimal soft threshold for weighted correlation algorithm
   calculations from 1–20. Subsequently, using the optimal soft threshold
   of 4 as the power value, the correlation coefficients between genes
   were multiplied to obtain the adjacency matrix of all genes. The Tree
   method with deepSplit was used to identify and divide each gene module,
   and the minimum height of the merge module was set to 0.25, while the
   minimum module size for module detection was set to 30. The threshold
   for building the network is set to the default value of 0.02, and the
   resulting interactions are considered to have regulatory relationships.
   Cytoscape software (version 3.9.1) was used to construct the core
   regulatory network of Barrett’s esophagus patients.

Animal model and treatment protocol

   Genetically engineered Lgr5-CreERT2, IL1B^tg; p16^flox/flox;
   Kras^LSL-G12D; mice (C57BL/6 background) were obtained from
   GemPharmatech Co., Ltd. (Nanjing, China) and bred in a pathogen-free
   animal facility at West China Hospital. Mice were genotyped via PCR
   prior to study initiation. At 3 months of age, gene recombination was
   induced via intraperitoneal injection of tamoxifen (75 mg/kg/day,
   Sigma-Aldrich) for five consecutive days to activate Kras^G12D
   expression and knock out p16 expression.

   Following recombination, mice were supplemented with 0.2% deoxycholic
   acid (DCA, Sigma-Aldrich) in drinking water, for six months to induce
   chronic bile reflux conditions. Mice were monitored weekly for weight,
   behavior, and general health. At the end of the exposure period (9
   months of age), mice were euthanized, and the SCJ tissues were
   harvested for H&E staining, immunohistochemistry, and organoid culture.
   Histological scoring of inflammation and metaplasia was performed by
   two blinded gastrointestinal pathologists according to established
   criteria.

Mouse SCJ-derived organoid culture and SNHG1 manipulation

   Organoids were established from SCJ tissues as previously described
   with modifications. Briefly, harvested tissues were washed in cold PBS
   and incubated with 10 mM EDTA for 30 minutes at 4 °C to release
   epithelial crypts. After mechanical dissociation and filtration, crypts
   were embedded in Matrigel (Corning, 356231) and cultured in Advanced
   DMEM/F12 medium supplemented with: B27 supplement (1×, Gibco), N2
   supplement (1×, Gibco), 50 ng/mL EGF (PeproTech), 100 ng/mL Noggin
   (PeproTech), 100 ng/mL R-spondin1 (PeproTech), 10 mM Nicotinamide, 1.25
   mM N-acetylcysteine, 10 μM Y-27632 (for initial plating only).
   Lentiviral constructs for SNHG1 overexpression or knockdown (shRNA)
   were packaged using 293 T cells and transduced into organoids at
   passage 2. Organoids were harvested for immunohistochemical analysis of
   EMT markers (E-cadherin, N-cadherin), and images were quantified using
   ImageJ software.

Statistical analysis

   For the statistical analysis in this study, SPSS 24.0 and GraphPad
   Prism 7.0 were used. Comparisons were performed using the
   Kruskal-Wallis H test with Bonferroni’s adjustment for ranked data. The
   measurement data were presented as means ± standard deviations, and
   Tukey’s post hoc test was used after one-way analysis of variance
   (ANOVA) for multiple group comparisons. P < 0.05 was deemed as the
   statistically significant threshold. we define significance levels as
   follows: *** p < 0.001, ** p < 0.01, and * p < 0.05.

Supplementary Information

   [142]Supplementary Material 1.^ (444.1KB, pdf)

Acknowledgements