Olfactomedin 4 orchestrates TGF-β/CEACAM6 axis, promoting cell-autonomous epithelial to mesenchymal transition during gallbladder epithelium carcinogenesis

Abstract

Gallbladder cancer (GBC) is a highly lethal disease which is usually diagnosed at advanced stage owing to unavailable screening tools and effective therapies. Persistent inflammation induced by various risk factors is the main cause of GBC, however, the molecular program remains elusive. Unveiling the molecular trajectory and events during gallbladder epithelium malignant transformation contribute to prevention and drug discovery for GBC. Single cell RNA sequencing was performed by using 4 gallbladder adenoma and cancer samples. Pseudotime trajectory analyses were employed to reconstruct epithelium transformation track in order to identify crucial genes which promoted GBC development. Functional and mechanism studies were performed to validate the regulatory network and cell behavior in vitro and in vivo. Three clusters of gallbladder epithelium were identified among GBC microenvironment which were characterized by distinct epithelial mesenchymal transition (EMT) and inflammation states. Cell adhesion molecular binding was the most significant GO term between EMT high and low states, in which OLFM4 was differentially expressed gene participated. Further studies implicated that OLFM4 could promote GBC metastasis and activate EMT through CEACAM6/AKT signaling cascade, and the CEACAM6 expression was regulated by TGF-β/Smad3 pathway. Interestingly, we disclosed that TGFβR1 was the functional receptor of OLFM4, through which the tumorigenic signaling of OLFM4 was transduced. These findings suggest the oncogenic role of OLFM4 during GBC carcinogenesis which can be a candidate biomarker of GBC, and OLFM4, TGFβR1 and downstream signaling elements are promising therapeutic targets for GBC.

Introduction

Gallbladder cancer is the most common malignant biliary tract tumor, which is generally diagnosed at the advanced stages owing to its atypical symptoms and no reliable screening biomarker [1, 2]. The lethal disease is characterized by aggressive behavior and poor prognosis with a 5-year survival rate of 5–20% [3]. It is incidentally discovered during laparotomy or laparoscopy performed on patients with symptomatic gallbladder diseases [4]. Although surgical resection is a potentially curative therapy for GBC, only 10–30% of patients can be considered for surgery, and limited therapeutics are optional for advanced GBC cases [5]. Currently, the gallstone is a major risk factor for GBC based on the evidence that gallstones are found in 65–90% of GBC patients [6, 7]. Several other factors, including polypoid gallbladder lesion, bile duct abnormalities, exposure to carcinogens, infection and genetics, are important factors contributing to GBC [8]. Chronic inflammation raised by either lithiasic or non-lithiasic origin is the dominant process facilitating the development of GBC [9]. It is clear that most GBC progressions followed two histogenic processes, including metaplasia-dysplasia-carcinoma and adenoma-carcinoma sequences [1]. Long-standing inflammation promotes sequential histological transformations of GBCs from normal gallbladder epithelium to invasive carcinoma accompanying with growing genetic alterations [10, 11].

Sustained inflammation induces DNA damage among cancer-driver genes through increased level of reactive oxygen and nitrogen species leading to the initiation and progression of GBC [12]. Employing the next-generation sequencing and omics strategy, a number of genetic aberrations were identified in GBCs, including DNA repair genes (TP53, ATM, and BRCA2), proto-oncogenes (ERBB2, SMAD4, and KRAS) and tumor suppressors (FHIT, RB1, and DCC) [8, 11, 13, 14]. Additionally, components of oncogenic signaling pathways are abnormally upregulated in the inflamed gallbladder epithelium, which contribute to the pathogenesis of GBC. Some studies demonstrated that PI3K-AKT-mTOR, MAPK-ERK, Hedgehog, Wnt, and Notch signaling pathways were dysregulated in GBC; however, potent therapeutics against these targets show limited improvements for GBC patients [1, 15–17]. Obviously, malignant transformation of gallbladder epithelium involves a broad range of oncogenes and their related pathways in an accumulative manner supported by the fact that GBC progresses from dysplasia to invasive carcinoma over 15 years [18]. The microenvironment evolved by GBCs is fundamentally distinct from that of their benign origins, characterized by altered cellular crosstalk, reprogrammed metabolic and signaling pathways, and a disrupted immune balance [19, 20]. The carcinogenesis routes of GBCs and related gene expression dynamics can be restored by single-cell transcriptomic analysis. Investigation of transformation process of GBCs may systemically disclose dysregulated genes and related signaling pathways shedding lights on therapeutic strategies.

In this study, we identify that OLFM4, an inflammatory related glycoprotein is upregulated during GBCs development by single-cell sequencing analysis [21]. Importantly, OLFM4 potently elicits epithelial to mesenchymal transition of gallbladder epithelium by TGFβ/Smad3-CEACAM6/AKT extended signaling pathway. The carcinogenic function of OLFM4 mainly relies on its ability of direct binding to TGF-β receptor type-1 (TGFβR1), triggering the release of TGF-β in GBC tumor microenvironment. Interestingly, suppression of either OLFM4 or CEACAM6 dramatically attenuates the metastasis of GBC in mice model. This investigation intends to elucidate the underlying mechanism promoting gallbladder from inflamed epithelium to invasive cancer and presents promising therapeutic targets for GBCs.

Materials and methods

Patient cohorts

Four gallbladder carcinoma samples subjected to scRNA-seq were obtained from patients undergoing surgical resection with gallbladder diseases at Yan’an Affiliated Hospital of Kunming Medical University and The Third Affiliated Hospital of Kunming Medical University. All clinical data were sourced from electronic medical records under informed consent, and sample collection was approved by the Ethical Committee of Yan’an Affiliated Hospital of Kunming Medical University (No. 2023-024-01), which was conducted in accordance with Declaration of Helsinki. Fresh surgical specimens were divided under sterile conditions: one portion was directly processed for scRNA-seq, and the other underwent formalin fixation for histopathological analysis. All diagnoses were independently confirmed by three pathologists through hematoxylin-eosin-stained section review. The clinical samples collected for scRNA-seq should meet the following criteria: inclusion criteria: 1) Diagnosis of gallbladder adenoma or carcinoma confirmed by histological methods; 2) Availability of complete clinical and pathological data, including TNM stage, treatment records information; 3) Availability of sufficient tumor tissue samples for analysis; 4) Received a standard surgery regimen relevant to the study objectives. Exclusion criteria: 1) Carcinomas originating in bile duct; 2) Lack of pathological diagnosis; 3) Incomplete key clinical data or follow-up records; 4) A history of other active malignant tumors within the past 5 years; 5) Received prior anticancer therapy that could interfere with the study’s endpoints.

Single-cell RNA-sequencing

Tissue samples were minced in a cell culture dish and enzymatically digested in centrifuge tubes using a mixture containing Collagenase I (Sigma), Collagenase IV (Sigma), DNase I (Sigma), and Dispase (Corning) at 37 °C for 30 min. The digested cell suspension was filtered through a 40-μm cell strainer (BD Falcon) to obtain a single-cell suspension. For viability assessment, aliquots of the suspension were mixed with 0.4% trypan blue at a 9:1 ratio and analyzed using the Countess® II Automated Cell Counter, ensuring ≥90% viability. Cell concentrations were adjusted to ≥1000 cells/μL prior to processing. Cell suspensions were processed through the 10X Genomics GemCode platform using microfluidic partitioning technology to generate Gel Bead-In-EMlusions (GEMs). Libraries were generated and sequenced from the cDNAs with the 10X Genomics 5’ platform with Chromium Next GEM Single Cell 5’ Reagent Kits v3.1. Upon dissolution of Gel Beads in GEMs, the encapsulated primers comprising Illumina® R1 sequences, 16-nucleotide 10x Genomics barcodes, 10-nucleotide unique molecular identifiers (UMIs), and poly-dT sequences were liberated into the cellular lysate. Subsequently, barcode-labeled and full-length cDNA were synthesized via reverse transcription using polyadenylated mRNA transcripts as templates.

Post-GEM reactions were purified using silane magnetic beads to remove residual reagents. Amplification of cDNAs was conducted via PCR to achieve sufficient material for library preparation. The R1 primer was integrated during GEM incubation, while P5/P7 sample index, and R2 primers were introduced in library construction through end repair, A-tailing, adapter ligation, and PCR. Final Illumina-ready libraries contained paired-end constructs flanked by P5/P7 primers, with Read 1 encoding 10x Barcodes and UMIs, and Read 2 capturing cDNA fragments. The generated cDNA library was sequenced using an Illumina NovaSeq X Plus.

Single-cell RNA-seq data processing and data quality control

Raw sequencing data (BCL files) were processed using 10X Genomics Cell Ranger (version 3.1.0) for FASTQ file conversion, read alignment, and gene expression quantification. Following quality control filtering of low-confidence barcodes and UMIs, sequencing reads were aligned to the GRCh38.100 human reference genome. Post-alignment data from individual samples were processed through Seurat (version 3.1.1) with the following quality thresholds: exclusion of cells exhibiting extreme UMI counts (<500 or >4000 genes detected), elevated mitochondrial content (>10%), or excessive sequencing depth (≥8000 UMIs) [22]. Potential doublets were computationally identified and removed using DoubletFinder (version 2.0.3) through artificial nearest-neighbor analysis (pANN scoring) calibrated to the experimental cell density [23]. To minimize the effects of batch effect and behavioral conditions on clustering, we performed dataset integration using Harmony R package to aggregate all samples, ultimately resolving 28,548 high-quality cells expressing 39,974 detectable genes from 4 samples [24]. Following log-normalization, integrated datasets were scaled and subjected to principal component analysis (PCA) for dimensionality reduction prior to downstream clustering analyses [25]. We annotated the cell types, and six clusters were identified based on expression of the following marker genes: CD2, CD3D, and CD3E for lymphocyte cells; EPCAM and KRT19 for epithelial cells; DCN and COL1A1 for fibroblast cells; VWF, RAMP2, and PECAM1 for endothelium cells; CD14 and CD68 for myeloid cells; IGHA1, IGHG1, and JCHAIN for plasma cells.

Trajectory analysis

To investigate dynamic biological processes, such as interconversion and evolutionary trajectories of different cell types, we applied the Monocle (Version 2.36.0) algorithm [26]. Signature genes expressed in at least 10% cells of the dataset and with a p < 0.01 calculated using the differentialGeneTest function were included to define the trajectory progress. The Ggridges package was used to analyze the frequency of distributed cells in different groups along the pseudotime axis. Cellular differentiation trajectories were reconstructed using Monocle (version 2.36.0) to investigate dynamic processes including cell state transitions and lineage progression. Ordering genes were selected based on two criteria: (1) detection in ≥10% of cells, and (2) significant differential expression (q-value < 0.01) determined by Monocle’s differentialGeneTest function with Benjamini-Hochberg correction. The ggridges package (version 0.5.3) was used to analyze the frequency of distributed cells in different groups along the pseudotime axis.

Enrichment analysis

Pseudotime-dependent genes were further subjected to GO and KEGG enrichment analysis using the clusterProfiler package (Version 3.11.0) with default settings. A nonparametric and unsupervised algorithm from the gene set variation analysis (GSVA) package (Version 1.42.0) was selected to assess the EMT scoring of different states generated with Monocle [27]. The signature genes of EMT and other related biological process were obtained from fifty hallmark gene sets in the MSigDB database (https://www.gsea-msigdb.org/gsea/msigdb).

Immunohistochemical (IHC) staining and grading

Gallbladder tissue samples were obtained from 108 chronic cholecystitis and 72 gallbladder carcinoma cases through surgical resection, fixation in 10% neutral buffered formalin and paraffin embedding. Serial 4 μm sections were subjected to de-paraffinized with xylene and graded ethanol hydration. Microwave-based antigen retrieval was performed in 0.01 M citrate buffer (pH 6.0), with subsequent endogenous peroxidase blockade using 3% H₂O₂ for 10 min and nonspecific binding inhibition with 5% horse serum for 1 h at room temperature. Sections were incubated overnight at 4 °C with anti-OLFM4 primary antibody (Proteintech 28432-1-AP, 1:100 in antibody diluent), followed by HRP-conjugated secondary antibody development and hematoxylin counterstaining for 3 min. The staining intensity was calculated based on the intensity of the positive staining (0 = negative; 1 = weak; 2 = moderate; 3 = strong) multiplied by the percent of positive cells.

Clinical tissue microarrays and Kaplan-Meier survival analysis

Two paraffin-embedded cancer tissue microarrays (TMA) containing 220 tumor tissues and 24 adjacent non-tumor tissues were obtained from the National Engineering Center for Biochip (Outdo Biotech, China), with ethical approval granted by the Institutional Review Board of Outdo Biotech. Patients were stratified into high and low expression groups based on OLFM4 expressions. Kaplan-Meier survival curves were generated to evaluate overall survival (OS) differences between cohorts, and the significance was assessed via log-rank tests.

Cell culture

GBC-SD and NOZ cell lines were purchased from the Cell Bank of Type Culture Collection of Chinese Academy of Sciences (Shanghai, China). All cell lines were propagated and cultured in accordance with instructions from the providing institution and utilized with 15 or fewer passages. GBC-SD were cultured in RPMI-1640 medium (Biological Industries) supplemented with 10% FBS (Biological Industries), and 100 U/mL penicillin and 100 μg/mL streptomycin (Biological Industries). NOZ were cultured in DMEM/F12 medium (Gibco) supplemented with 10% FBS (Biological Industries), and 100 U/mL penicillin and 100 μg/mL streptomycin (Biological Industries). Short Tandem Repeat profiling was used to authenticate cell lines before use in experiments. Cultured cells were maintained at 37 °C with 5% CO₂ in a humidified incubator.

Vector construction and transduction

OLFM4, CEACAM6 or TGFβR1 knockdown in GBC-SD and NOZ cells was performed using the hU6-MCS-CBh-gcGFP-IRES-puromycin and hU6-MCS-CMV-Neomycin vector. OLFM4 and CEACAM6 overexpression vectors were generated using the Ubi-MCS-SV40-EGFP-IRES-puromycin and Ubi-MCS-3FLAG-SV40-Neomycin vectors. Both vector and recombinant lentiviruses were designed and synthesized by GeneChem (Shanghai, China).

The interference sequences are shOLFM4-1:

“5-CCGGAGTGCAGAGCATTAACTATAACTCGAGTTATAGTTAATGCTCTGCACTTTTTTG-3“;

shOLFM4-2:

“5-CCGGCCCTAATGCTGCCTATAATAACTCGAGTTATTATAGGCAGCATTAGGGTTTTTG-3”;

shCEACAM6-1:

“5-CCGGGGTTTATCAATGGGACGTTCCCTCGAGGGAACGTCCCATTGATAAACCTTTTTG-3”;

shCEACAM6-2:

“5-CCGGGGAACGATGCAGGATCCTATGCTCGAGCATAGGATCCTGCATCGTTCCTTTTTG-3”;

shTGFBR1-1: ‘5-CCGGGCCTTGAGAGTAATGGCTAAACTCGAGTTTAGCCATTACTCTCAAGGCTTTTTG-3”;

shTGFBR1-2:

“5-CCGGGCCACAGATACCATTGATATTCTCGAGAATATCAATGGTATCTGTGGCTTTTTG-3”;

shControl (Scramble):

“5-CCGGAGTTCTCCGAACGTGTCACGTCTCGAGACGTGACACGTTCGGAGAATTTTTG-3”.

The OLFM4 and CEACAM6 overexpression lentiviral vector was built using NCBI Reference Sequence: NM_006418 (OLFM4) and NM_002483.7 (CEACAM6). GBC-SD and NOZ cells were co-cultured with 4 μg/mL puromycin for 7 days or 300 μg/mL G418 for 14 days for screening, and immunoblotting was performed to validate the level of protein expression.

Western blot

Protein samples (20–40 μg) were separated on 10% SDS-PAGE gels and subsequently transferred to PVDF membranes (Millipore). Following 1 h blocking with 5% BSA/TBST at room temperature, the membranes were incubated with primary antibodies (diluted in 5% BSA/TBST) at 4 °C overnight. The primary antibodies used in the Western blotting assay are summarized in Supplementary Table S1. After three 10-min TBST washes, membranes were probed with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 hour at room temperature. Following three additional TBST washes (10 min/wash), protein bands were visualized using enhanced chemiluminescence detection reagent (ECL, ProteinTech) with a chemiluminescence imaging. In TGF-β/Smad3 signaling blockage assay, cells were treated with TGF-β type I receptor kinase inhibitor IN1130 (Cat No. HY-18758, MCE) and phosphor-Smad3 inhibitor E-SIS3 (Cat No. HY-13013, MCE). For rescue experiments, OLFM4-knockdown GBC cells were incubated with recombinant human TGF-β protein (Cat No. HZ-1011, ProteinTech) with different doses and time series indicated. To explore the role of p-AKT in CEACAM6/EMT cascade, p-AKT inhibitor (Capivasertib, Cat No. HY-15431, MCE) was employed to treat GBC cells at different concentrations and immunoblotting was performed to examine the expression of EMT markers.

Cell invasion and migration assays

Two strains of GBC cells (GBC-SD and NOZ) overexpressing and knocking down by OLFM4 or CEACAM6 were employed to investigate their invasion and migration promoting function. For cell invasion assays, 24-well plates with Transwell chambers (8 μm pore, Corning) were coated with 100 μL (400 μg/mL) Matrigel before use (Matrigel Basement Membrane Matrix High Concentration, Corning). Cell migration assays were conducted without Matrigel coating using the same procedures described above. GBC cells transfected with different lentivirus or treated with Capivasertib were trypsinized and washed twice with PBS before being resuspended in serum-free DMEM/F12 medium (Gibco) and seeded into the upper chamber. During p-AKT inhibition assay, the culture medium was supplemented with Capivasertib at different concentrations indicated. The lower chamber was filled with 500 μL of culture medium containing 20% FBS. Cells in upper chambers were wiped out with cotton swabs after 24 h of incubation at 37 °C, 5% CO₂. Cells were fixed in methanol and stained with 1% crystal violet solution, and the chambers were photographed with a microscope (Eclipse 50i POL, Nikon) equipped with a camera. Then, cells were quantified using the ImageJ software.

In vivo tumor metastasis models

For tumor metastasis models, six-week-old male nude mice were purchased from SiPeiFu Biotechnology Co., Ltd. (Beijing, China). After acclimatization for one week, the mice were randomly assigned to either the experimental or control group. Approximate 1 × 10⁶ GBC-SD or NOZ cells stably expressing either OLFM4 or CEACAM6 were injected into mice by tail vein injection. Correspondingly, GBC cells transfected with shRNA vector against OLFM4 or CEACAM6 were inoculated with same cell mounts. Cells transfected with negative control or scramble RNA vector were taken as control, and all of viral vector were marked with firefly luciferase. In vivo imaging and quantification of bioluminescence were performed on a BLT AniView600 Living Image system. D-luciferin, a substrate for firefly luciferase, was injected intraperitoneally (150 μl D-luciferin at 30 mg/mL, Cat No. E1605, Promega) 5 min before luminescence imaging. After injection, mice were placed in a Sealed Optical Imaging Tray, ventilated with 1.5% isoflurane. All animal experiments were approved by the Animal Ethics and Welfare Committee (AEWC) of Yan’an Affiliated Hospital of Kunming Medical University (No. 2023025).

Bulk RNA sequencing and data analysis

Total RNA was isolated from GBC-SD cells stably transfected with scramble or shOLFM4 construct using TRIzol (Invitrogen). Following RNA quality assessment, mRNA was enriched using Oligo(dT) magnetic beads. The RNA sequencing libraries were constructed with the KAPA Stranded RNA-Seq Library Preparation Kit (Illumina) according to the manufacturer’s protocol. Libraries were subjected to 150 bp paired-end sequencing on an Illumina HiSeq2000 platform with three independent biological replicates per group. Raw sequencing reads were aligned to the human reference genome GRCh38.87 using STAR aligner (Version 2.7.10). Gene expression levels were quantified as fragments per kilobase million (FPKM) using Cufflinks (Version 2.2.1). Differential gene expression analysis between Scramble and shOLFM4 groups was performed using the limma R package (v3.54.2) with thresholds of |log2FC | >1, p-value < 0.01 and q-value < 0.05. Protein-protein interaction networks were reconstructed using the STRING database (Version 11.5) and visualized by Cytoscape (Version 3.9.1).

Immunofluorescence and confocal microscopy

GBC-SD and NOZ cells were plated on coverslips at 50,000 cells per well in a 24-well plate. After three times washing with PBS, the cells were fixed in 4% paraformaldehyde in PBS for 10 min. Permeabilization was performed using 0.1% Triton X-100 in PBS for 10 min and blocked with 5% horse serum for 1 h at room temperature, which was followed by primary antibodies staining. For GBC tissue samples, sections were obtained from gallbladder carcinoma cases, fixated in 10% neutral buffered formalin and embedded in paraffin. The 4 μm sections were subjected to de-paraffinized with xylene and graded ethanol hydration. Microwave-based antigen retrieval was performed in 0.01 M citrate buffer (pH 6.0), and the sections were blocked with 5% horse serum for 1 h at room temperature. Primary antibodies of OLFM4, CEACAM6 or TGFBR1 (summarized in Supplementary Table S1) was added and incubated overnight at 4 °C. After three times washing, corresponding secondary antibodies were added at 1:500 (Alexa Fluor 488, Alexa Fluor 594, Abcam). After three times washing, the coverslips were mounted using antifade mounting medium with DAPI (Cat No. P0131, Beyotime Biotechnology). Stained cells were imaged using a confocal microscope (LSM980, ZEISS) or Fluorescence microscope (Axio Scope5, ZEISS). For analysis of colocalization, two independent experiments were quantified using ImageJ and the Coloc 2 plug-in.

ELISA

For measurement of TGF-β in the medium, GBC-SD cells, which were knocked down or overexpressed by OLFM4, were cultured in 12-well plates and starved in the serum-free medium for 12 h. The supernatants were collected and centrifuged at 300 g for 5 min at 4 °C. TGF-β concentrations of the medium were determined by ELISA using the Human TGF-β ELISA kit (Cat No. EHC107b, Neobioscience) according to the manufacturer’s instructions.

Co-immunoprecipitation

Approximate 1 × 10⁶ GBC-SD or NOZ cells were harvested and washed twice with PBS. The cells were lysed in 1 ml lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, Millipore) supplemented with cOmplete™ EDTA-free Protease Inhibitor Cocktail (Roche) and PhosSTOP Phosphatase Inhibitor Cocktail Tablets (Roche) for 30 min on ice. The cell lysates were centrifuged at 4 °C for 5 min at 15,000 × g. The protein concentrations were determined by BCA-based method. For inputs, 500 µg of cell lysate was mixed with 5× Loading Buffer (Solarbio) and heated for 10 min at 100 °C. Immunoprecipitations were performed by employing Catch and Release kit v2.0 (Cat. #17-500, Millipore). Briefly, total 500 µl of reaction mass, including 500 µg of cell lysates, 4 µg of antibody (anti-OLFM4, anti-TGFβRI or normal IgG antibody), 10 µl of antibody capture affinity ligand and appropriate amount of wash buffer, was incubated in a spin column on a rotating wheel at 4 °C for 12 h. Then, the column was centrifuged and washed three times with 400 µl wash buffer. The elutes were collected for immunoblotting by adding denaturing elution buffer containing β-ME.

His pull-down assay

For pull-down assay, anti-His-tag mAb-Magnetic Beads (Cat No. D291-11, MBL) were blocked with 5% BSA for 1 h, then incubated with OLFM4 Protein (Cat No. HY-P71179, MedChemExpress) and TGFBR1 Protein (Cat No. T07-13G, SignalChem) at 4 °C for 2 h on a rotating wheel. In competitive binding assays, TGF-β was added at concentrations of 0.5 and 0.25 μg/mL (Cat No. HZ-1011, Proteintech). After incubation, the supernatants were completely discarded using a magnetic stand, and the protein-bound beads were washed five times with wash buffer. Proteins were eluted by adding loading buffer and heated at 100 °C for immunoblotting.

Results

Sequential transformation of three epithelial cell subpopulations in GBC tumor microenvironment

In order to explore the altered biological processes during the development of GBC, scRNA-seq analyses were performed on one gallbladder adenoma and three gallbladder adenocarcinomas without any treatment. The clinicopathological information of all the cases included was presented in Fig. S1. After rigorous data quality control and filtering, we obtained scRNA-seq profiles for 28,548 cells, which were categorized into six distinct cell populations based on known lineage markers through graph-based clustering of uniform manifold approximation and projection (Fig. 1A, B). These cell subsets included epithelium, fibroblast, endothelium, lymphocyte, plasma and myeloid. Among these cell subpopulations, we captured three epithelium cell clusters (cluster 4, 8, and 13) with distinct gene expression profile (Fig. 1C). To further study the heterogeneity of the epithelial cell subtypes, we performed trajectory analysis in pseudotime manner and formulated a transition model of cell distribution using Monocle 2. All the single cells in these three epithelial subsets were deconstructed and reassembled into three branches of cells based on their transcriptional similarity. By defining the origin of cell progression, we found that the gallbladder epithelia polarized following two different evolutionary trajectories (Fig. 1D). The cell cluster 13 was the pre-branch of the trajectories, and the epithelial cells transformed towards cluster 8 and cluster 4 respectively (Fig. 1E). We defined the three branches of cells as three epithelium states (state 1, 2, and 3) by taking the bifurcation as tipping point, which did not significantly influence the constitutions of original cell clusters (Fig. 1F–H). The EMT score of state2 and 3 were significantly higher than state1, suggesting that epithelial cells dedifferentiated along these two paths (p < 0.001) (Fig. 1J). Then, we conducted gene ontology (GO) analysis of differentially expressed genes (DGEs) in state1, state2 and state3 to disclose the regulon that activated the EMT of gallbladder epithelium. Interestingly, cell adhesion molecular binding was the most significant GO term in either state2 vs. state1 or state3 vs. state1 (Fig. 2A, B). The two groups of DGEs annotated with cell adhesion molecular binding were extracted and ranked by fold change. The top 15 DGEs were shown and visualized by percentage of target protein expressed and average expression scale (Fig. 2C). Importantly, OLFM4 involved in modulation of inflammatory process was upregulated in both state2 and state3, and the expression was progressively increased along with the malignant transformation process of gallbladder epithelium (Fig. 2D–F). Immunohistochemistry staining confirmed the enhanced expression of OLFM4 in gallbladder cancer compared to chronic cholecystitis (Fig. 2G; Table 1). Kaplan-Meier survival analysis of TCGA GBC patients suggested that high expression of OLFM4 significantly correlated with worse prognosis (p = 0.02) (Fig. 2H).

Fig. 1. Cellular atlas and epithelium trajectory analyses of gallbladder adenoma and carcinomas using scRNA-seq.

A Uniform Manifold Approximation and Projection (UMAP) plots visualizing 6 cell types in gallbladder samples included. B Violin plots showing marker genes expressions for each cell type. C UMAP plots indicating 3 subclusters of gallbladder epithelium (cluster4, 8 and 13) screened by EPCAM. D Pseudotime trajectory analysis of epithelium extracted from all samples by Monocle2. E Trajectory of gallbladder epithelium labeled with different epithelium clusters. F Malignant transformation trajectory of epithelium divided into three states (State1/2/3) based on bifurcation. G Histography implicating the components of cluster 4, 8, and 13 in each state. H UMAP plots showing cell grouping of all single cells in different states. I Comparison of EMT scores of epitheliums in State1/2/3. t-tests were performed to assess the significance of EMT scores between different states (****p < 0.0001).

Fig. 2. Functional and differential gene expression profiles analyses of epitheliums in State 1, 2 and 3.

A, B Gene ontology analyses of differentially expressed genes between states indicated. C Bubble plots showing the gene signatures of “cell adhesion molecule binding” GO term in different states. D and E Volcanic plots of differentially expressed genes between different states indicated. F Heatmap of gene expressions in gallbladder carcinogenesis program. Left: trajectory from state1 to state2 with 3994 DGEs grouped in three clusters; Right: trajectory from state1 to state3 with 6109 DGEs grouped in 7 clusters. G Representative gallbladder cancer samples illustrating differential expression of OLFM4 between specimens from cholecystitis and gallbladder cancer. H The overall survival rate of GBC patients with high or low OLFM4 expression (p = 0.02). Data were obtained from TCGA for analysis. The median OLFM4 expression in each cohort was used as the cutoff, and a log-rank test was performed for significance.

Table 1.

Correlation analysis between OLFM4 expression and clinicopathologic characteristics of GBC patients.

Clinicopathologic characteristics		OLFM4 expression in gallbladder tissue		Total	P value
		Low expression n (%)	High expression n (%)
	Chronic cholecystitis	45(41.67)	63(58.33)	108	0.000
	Gallbladder cancer	58(80.56)	14(19.44)	72
Chronic cholecystitis
Gender	Male	22(66.67)	11(33.33)	33	0.267
	Female	41(55.26)	34(44.74)	75
Age	≤50	37(64.91)	20(35.07)	57	0.065
	>50	21(46.67)	24(53.33)	45
Clinical diagnosis	Chronic cholecystitis	13(56.52)	10(43.48)	23	0.843
	Chronic cholecystitis with a gallstone	50(58.82)	35(41.18)	85
Total		45(41.67)	63(58.33)	108a
Gallbladder cancer
Gender	Male	11(73.33)	4(26.67)	15	0.471
	Female	47(82.46)	10(17.54)	57
Age	≤60	24(80.00)	6(20.00)	30	0.920
	>60	34(80.95)	8(17.54)	42
TNM stage	T1 + T2	3(50.00)	3(50.00)	6	0.048
	T3 + T4	55(83.33)	11(16.67)	66
Clinical stage	Ⅰ/Ⅱ	12(54.55)	10(45.45)	22	0.494
	Ⅲ/Ⅳ	46(92.00)	4(8.00)	50
Lymphatic metastasis	N0	42(77.78)	12(22.22)	54	0.001
	N1 + N2	16(88.89)	2(11.11)	18
Total		58(80.56)	14(19.44)	72

^aAge information was available for 103 of the 108 total cases.

Elevated OLFM4 expression in malignant epithelial cells drives EMT in GBC

By labeling the expression profiles of OLFM4 and mesenchymal markers on epithelial cell clusters, we found that OLFM4, N-cadherin, ZEB-1, Vimentin, MMP9 and Snail were upregulated in cluster4 and cluster8, contrarily, the expression of MMP2 was enhanced in cluster13, and no significance was observed in Slug expression (Fig. 3A). To further investigate that OLFM4 was the regulon promoting EMT, we overexpressed and knocked down OLFM4 in two GBC cell line (GBC-SD and NOZ) in vitro, and measured the expression of EMT related genes. As shown in Fig. 3B, OLFM4 increased the mesenchymal markers, including ZEB1, N-cadherin, MMP9, MMP2, Vimentin, Slug and snail, and OLFM4 knockdown dramatically downregulated the expressions of corresponding markers. In accordance with observations in gene expression pattern, OLFM4 overexpression significantly promoted cell migration and invasion in GBC-SD and NOZ, depletion of OLFM4 dramatically weakened cell migration and invasion (Figs. 3C, D, S2). To explore the role of OLFM4 on tumor metastasis in vivo, we generated GBC mouse model by inoculating GBC-SD and NOZ via tail-vein injection (Fig. 3E). Upregulation of OLFM4 in both GBC-SD and NOZ promoted lung metastasis, and OLFM4 knockdown significantly decreased GBC cells spreading to lung, as evidenced by signals changes of bioluminescence (Figs. 3F, G and S3).

Fig. 3. OLFM4 activated EMT and facilitated GBC metastasis.

A UMAP plots showing expression levels of OLFM4 and EMT-related genes among three gallbladder epithelium clusters. B Western blot for expression of EMT markers by OLFM4 knockdown and overexpression, respectively with both GBC-SD and NOZ. scramble, scramble shRNA; shOLFM4#1 and shOLFM4#2 represent two independent shRNAs targeting OLFM4; a-Tubulin as loading control. C, D Transwell migration and invasion assays of GBC-SD and NOZ with either OLFM4 knockdown and overexpression. E Experiment timeline of GBC cells inoculation and bioluminescence detection. F Left: bioluminescence images for mice injected with GBC-SD of scramble, shOLFM4, Vector and OLFM4 overexpression at day 0 and day 28. Middle: bioluminescence curve of mice injected with GBC-SD in which quantitation of bioluminescence was recorded in each assessment day. Right: histography analyses of bioluminescence quantities of mice in all GBC-SD groups on the 28th day. t-test were performed to evaluate significance (***p < 0.001. **p < 0.01). G Left: bioluminescence images for mice injected with NOZ, which were grouped by treatments indicated. Middle: Bioluminescence curve of mice inoculated with NOZ. Right: histography analyses of bioluminescence quantities of mice treated with NOZ at 28th day (*p < 0.05).

CEACAM6/PI3K-AKT is downstream pathway of OLFM4 activating EMT

To ascertain the EMT signaling pathways triggered by OLFM4, bulk RNA-seq was performed on GBC-SD cells treated with shRNA against OLFM4 to analyze the genes potentially regulated by OLFM4. Approximate 2197 DGEs (p < 0.01, FDR < 0.01 foldchange>2 or <0.5) were identified in GBC-SD depleted with OLFM4, including 1004 upregulated and 1175 downregulated genes. All the DGEs were input into the STRING database to predict the protein-protein interactions among DGEs. By picking up the close associations with OLFM4, we obtained 14 DGEs directly interacting with OLFM4 (Fig. 4A). By screening the interacting gene nodes with degree score, we selected CEACAM6 as the functional gene potentially regulated by OLFM4 for further investigation. Interestingly, CEACAM6 is a glycoprotein expressed on cell surface playing a role in cell adhesion and tumor progression, and the protein was co-mentioned with OLFM4 according to STRING database [28, 29]. However, the regulatory relationship has not been established yet. By examining the expression profiles on scRNA-seq data, we found that OLFM4 and CEACAM6 were specifically expressed on epithelium cluster 4 and 8, which were characterized with high EMT score (Fig. 4B, C). Next, immunofluorescence assay was performed on GBC tissue specimen to validate the expression correlation between OLFM4 and CEACAM6. It was evident that OLFM4 and CEACAM6 showed similar expression pattern in both OLFM4 high and low expression GBC cases (Fig. 4D). CEACAM6 expression was increased by OLFM4, and OLFM4 knockdown resulted in reduction of CEACAM6 expression in GBC-SD and NOZ (Fig. 4E). To understand the role of CEACAM6 in EMT, we assessed the expressions of EMT related genes following CEACAM6 overexpression and knockdown. It was apparent that CEACAM6 significantly triggered the expression of mesenchymal genes in GBC cells, and depletion of CEACAM6 repressed their expressions (Fig. 4F). Consistent with OLFM4, forced expression of CEACAM6 significantly promoted invasion and migration of GBC cells in vitro, and CEACAM6 knockdown inhibited the processes above (Fig. 4G, H and Fig. S4). In tail-vein injection mouse model, CEACAM6 overexpression caused augmented lung accumulations of GBC cells, and depletion of CEACAM6 impeded lung metastasis (Fig. 4I and Fig. S5). It was previously reported that CEACAM6 induced EMT of gastric cancer via the PI3K/AKT signaling pathway. In order to ascertain the molecular basis of GBC EMT, immunoblotting was performed to examine whether the AKT phosphorylation of GBC cells was affected by either OLFM4 or CEACAM6. Interestingly, the AKT phosphorylation of GBC cells was repressed by OLFM4 or CEACAM6 knockdown, and the mesenchymal markers expressions was inhibited simultaneously (Fig. 5A–D). In both OLFM4 and CEACAM6 overexpressed GBC cells, a significant increase in p-AKT and EMT key regulators expression suggesting PI3K/AKT triggering EMT signaling transduction. (Fig. 5A–D). In order to further support the probability that OLFM4 initiated EMT via CEACAM6/PI3K-AKT pathway, we performed rescue assay on CEACAM6 knockdown GBC cells. As shown in Fig. 5E, F, OLFM4 dramatically increased CEACAM6 expression, phosphorylation of AKT and mesenchymal markers expression in both GBC-SD and NOZ cells; however, depletion of CEACAM6 expression in GBC cells by shRNA could significantly reverse PI3K-AKT signaling pathway activation and mesenchymal markers upregulations triggered by OLFM4. To further investigate the role of p-AKT in CEACAM6 enhanced EMT, AKT phosphorylation inhibitor (Capivasertib) was employed to reverse the EMT promoted by CEACAM6. Interestingly, we found that the expressions of EMT markers were significantly repressed by Capivasertib in CEACAM6 overexpressed GBC cells (Fig. 5G). By treating with Capivasertib, the migration and invasion of CEACAM6 overexpressed GBC cells have been significantly attenuated in dose-dependent manner (Fig. 5G, H and Fig. S6). All these data indicated that OLFM4 potently promoted the EMT process of GBC via CEACAM6/PI3K-AKT pathway.

Fig. 4. OLFM4 initiates EMT through upregulation of CEACAM6.

A Interactive network of genes potentially downregulated by OLFM4. The correlations among genes were constructed based on STRING database. B Left: UMAP plots visualizing all cell clusters among GBC TMEs. Right: UMAP plots showing the co-expression and cell specificities of OLFM4 and CEACAM6. C Bubble plots quantifying the expression scale and percentage of OLFM4 and CECAM6 in all cell clusters. D Immunofluorescence staining of OLFM4 and CEACAM6 in GBC specimens labeled with Alex488 and Alex594 respectively. Representative images captured by fluorescence microscopy suggesting co-expression of OLFM4 and CEACAM6. E Validation of CEACAM6 expression regulated by OLFM4 using Western blot through OLFM4 knockdown and overexpression in both GBC-SD and NOZ. F Measurement of EMT markers expressions after CEACAM6 knockdown and overexpression in GBC-SD and NOZ. G, H Left: comparisons of invasion and migration abilities of GBC cells between each paired groups indicated by transwell assays. Right: histography analyses of each paired groups interfered by CEACAM6 expression. t-tests were performed for each paired group. (**p < 0.01, ***p < 0.001 and ****p < 0.0001). I Left: bioluminescence images of mice models inoculated with GBC cells (GBC-SD and NOZ), which were overexpressed or knockdown with CEACAM6. Middle: bioluminescence curves of mice in different groups indicated. Right: histography analyses of bioluminescence quantities between groups with corresponding treatments. Scramble: scramble vector, shCEACAM6: shRNA targeting CEACAM6, Vector: negative control vector, CEACAM6: CEACAM6 overexpression. T-test was performed to assess significance (*p < 0.05, **p < 0.01).

Fig. 5. CEACAM6/AKT is the downstream signaling of OLFM4.

A, B Measurement of AKT phosphorylation and EMT markers expression following OLFM4 knockdown and overexpression in both GBC-SD and NOZ is indicated. C, D Western blot quantifying AKT phosphorylation and EMT markers expression following CEACAM6 knockdown and overexpression in both GBC-SD and NOZ. E, F Rescue assay implicating that OLFM4 promoted EMT through CEACAM6/AKT in GBC-SD and NOZ: western blot showing AKT phosphorylation and EMT markers expressions in various OLFM4 overexpression and CEACAM6 knockdown combinations. G Western blot analysis of p-AKT and EMT marker expression in CEACAM6 overexpressing GBC cells (GBC-SD and NOZ), which were treated with 5 μM p-AKT phosphorylation inhibition (Capivasertib) for 48 h. H Left: Effects of AKT phosphorylation inhibition by Capivasertib on GBC cell migration and invasion. GBC-SD and NOZ cells were treated with Capivasertib at indicated concentrations, respectively, for 24 h and invasion was assessed by Transwell assay. Right: Histography analyses of each paired group treated by Capivasertib. t-tests were performed for each paired group. (*p < 0.05, ***p < 0.001 and ****p < 0.0001).

OLFM4 upregulates CEACAM6 via TGF-β/Smad3 signaling pathway

To identify the mechanism which was responsible for regulation of CEACAM6 by OLFM4, gene expression dynamics during gallbladder epithelium carcinogenesis were analyzed to screen DGEs whose expression trend resembled these two genes. There were 5225 DGEs categorized into five major gene clusters in characterized patterns (Fig. S7). Genes in cluster 1 were gradually increased from State1, peaked at State2 and slightly decreased at State3, containing both OLFM4 and CEACAM6 (Fig. 6A). In order to gain insight into the biological process of this gene set, gene set enrichment analysis was performed among all DGEs in cluster 3. Consistently, EMT signature was positively correlated with OLFM4 expression, and TGF-β was listed in the leading-edge subset, which might response to OLFM4-promoting EMT (Fig. 6A and Fig. S7). Besides, the scRNA-seq analysis indicated that the expression profiles of OLFM4, TGF-β, and CEACAM6 were correlated with each other along with the GBC malignant transformation trajectory (Fig. 6B). To investigate whether TGF-β/Smad3 signaling worked as a pathway between them, we quantified TGF-β concentrations of condition mediums harvested from controls, OLFM4 knockdown and overexpression groups. It was evident that OLFM4 knockdown significantly decreased TGF-β secretion of GBC cells in vitro, in contrast, forced OLFM4 expression in GBC cells elevated TGF-β level in cell culture mediums (Fig. 6C). To further validate that TGF-β/Smad3 activation was required for modulation of CEACAM6 expression by OLFM4 in GBC cells, we examined the Smad3 phosphorylation (p-Smad3) and expression changes of corresponding signaling pathway apparatus following treating GBC cells with OLFM4 shRNA, overexpression lentivirus and signaling inhibitors. Downregulation of OLFM4 in either GBC-SD or NOZ cells significantly decreased TGF-β expression and phosphorylation of Smad3, whereas OLFM4 overexpression activated TGF-β/Smad3 signaling pathway in GBC cells (Fig. 6D). Blockage of TGF-β signaling transduction dramatically reversed Smad3 phosphorylation and CEACAM6 expression promoted by OLFM4 by treating GBC cells with TGF-β receptor I kinase inhibitor (IN-1130) in both time- and dose- dependent manners (Fig. 6E, F). Additionally, impairment of TGF-β/Smad3 pathway using Smad3 inhibitor (SIS-3) suppressed the expression of CEACAM6 in the GBC cells expressing OLFM4 as IN-1130 (Fig. 6G, H). Interestingly, neither IN-1130 or SIS-3 disturbed TGF-β expressions in OLFM4 overexpressed GBC cells, which implicated that the regulation of TGF-β expression by OLFM4 in GBC was independent of Smad3 phosphorylation (Fig. 6E–H). Furthermore, we treated OLFM4 null GBC cells with recombinant TGF-β, and found that exogenous delivery of TGF-β significantly increased p-Smad3 and CEACAM6 in dose- and time-dependent way (Fig. 6I, J).

Fig. 6. OLFM4 regulates CEACAM6 expression via TGF-β.

A Upper: series test of cluster analysis of pseudotime trajectory DEGs in state1/2/3 based on Mufuzz method. Cluster I (1030 DEGs) was picked for further analysis for the reason that OLFM4 and CEACAM6 were included. Lower: gene set enrichment analysis of DEGs in cluster I in which OLFM4 expression was significantly increased. B Two dimensional and trajectory analyses showing the dynamic expression of OLFM4, CEACAM6 and TGF-β along the evolutionary path. C Comparisons of TGF-β secretion by ELISA using supernatants of GBC-SD treated either by shRNA targeting OLFM4 or OLFM4 overexpression vector. D Western blot assays for examining TGF-β expressions and phosphorylation of SMAD3 in either GBC-SD or NOZ by repressing or increasing OLFM4 expression. E, F Examinations of TGF-β, CEACAM6 expressions and SMAD3 signaling in OLFM4 overexpressed GBC cells (GBC-SD and NOZ) which were treated with IN-1130 (TGF-β receptor I kinase inhibitor) in time and dosage dependent manners. G, H Measurements of TGF-β, CEACAM6 expressions and SMAD3 signaling in OLFM4 overexpressed GBC cells (GBC-SD and NOZ) treated with SIS-3 (SMAD3 phosphorylation inhibitor) in time and dosage dependent manners. I, J Evaluations of CEACAM6 expressions and SMAD3 signaling in GBC cells (GBC-SD and NOZ), which were treated with shRNA targeting OLFM4 and stimulated with recombinant TGF-β in time and dosage dependent manners.

TGF-βRI is a functional receptor of OLFM4 mediating TGF-β release

Numerous studies have reported that TGF-β signaling was multifunctional during tumorigenesis, including cell cycle progression, adhesion and differentiation, and it could prolong its tumor-promoting effect through autoinduction. Importantly, most of the signaling pathways in cancer are interplayed and mutated. We speculated that OLFM4 stimulated the production of TGF-β in gallbladder cancer by employing TGFβ-Smad3 signaling. To test the hypothesis, co-immunoprecipitations were performed to study the interaction between OLFM4 and TGFβR1. It was shown that both TGFβR1 and OLFM4 could be co-precipitated with each other in GBD-SD and NOZ (Fig. 7A). By his-tag pull down assay, it was evident that OLFM4 was able to bind to TGFβR1 directly without any mediator in vitro (Fig. 7B). However, it was elusive whether OLFM4 binding competed with its canonical TGF-β ligand–receptor interaction. Therefore, we performed his pull-down assays by adding two different doses of recombinant TGF-β protein to the OLFM4–TGFβRI binding system. The presence of TGF-β significantly weakened the binding between OLFM4 and TGFβRI in a dose dependent manner, indicating that OLFM4 and TGF-β compete for binding to TGFβRI (Fig. 7C). We further validate the direct interaction between OLFM4 and TGFβR1 by examining their colocalization on cell compartment. By confocal microscopy, we found that the expressions of OLFM4 and TGFβR1 were colocalized in GBC-SD and NOZ cells which was evidenced by the overlapping of the two fluorescence signals from the two proteins on the images (Fig. 7D). The colocalization of OLFM4 and TGFβR1 were validated in gallbladder cancer specimens from patients (Fig. 7D). For functional assay, it was shown that the expressions of both TGF-β and CEACAM6 were significantly repressed by TGFβR1 knockdown in both GBC-SD and NOZ cells (Fig. 7E). Next, the functional binding of OLFM4 to TGFβR1 were further investigated by “rescue” assays. Exogenous addition of recombinant OLFM4 could significantly increase expressions of TGF-β and CEACAM6 in GBC cells, however, the effects could be reversed by TGFβR1 knockdown (Fig. 7F). Finally, the TGFβR1 expressions in OLFM4 overexpressed GBC cells were repressed to examine whether the TGF-β/CEACAM6 axis signaling was transduced through it. As shown in Fig. 7G, OLFM4 overexpression dramatically increased TGF-β production, Smad3 phosphorylation and CEACAM6 expression in GBC cells, whereas, OLFM4 signaling transduction has been impeded by TGFβR1 knockdown.

Fig. 7. TGFβR1 is functional receptor of OLFM4 transducing oncogenic signaling.

A Co-immunoprecipitation assays demonstrating interactions between OLFM4 and TGFβR1 in both GBC-SD and NOZ. Antibodies used for immunoprecipitations and immunoblotting were from different animal species. IP: immunoprecipitations; IB: immunoblotting. B His tag pull down assay illustrating direct interaction between OLFM4 and TGFβR1. C His tag pull-down assay exploring competitive binding of OLFM4 and TGF-β to TGFβR1. The binding between OLFM4 and TGFβR1 was inhibited by TGF-β in a dose-dependent manner. D Immunofluorescence assays indicating colocalization of OLFM4 and TGFβR1 in GBC-SD, NOZ and GBC tissues from two representative cases. Images were captured by confocal laser scanning microscopy. OLFM4 and TGFβR1 were labeled by Alex488 and Alex594, respectively. The fluorescence intensities of green and red channels in GBC-SD and NOZ were quantified by ImageJ software (line charts). E Western blot showing expression of TGF-β, CEACAM6 and OLFM4 in GBC-SD and NOZ repressed with TGFβR1 expression. TGFβR1 knockdown was performed with two independent shRNAs. F Immunoblotting showing rescue assays for recombinant OLFM4 promoting TGF-β and CEACAM6 expressions. The TGFβR1 expressions were repressed in both GBC-SD and NOZ for rescuing OLFM4 function. G Immunoblotting showing rescue assays for endogenous OLFM4 activating TGF-β/smad3/CEACAM6 signaling in GBC-SD and NOZ cells.

High expression of OLFM4 in GBC is associated with poor prognosis

In order to examine the associations between expression of OLFM4 and survival, GBC tissue microarrays analysis was performed based on immunohistochemistry. It was shown that varied expression profiles of OLFM4 and CEACAM6 were found in either tumor and stroma (Fig. 8A). Compared to matched paracancerous tissues of GBC samples included, OLFM4 expression was significantly increased in GBC tissues (****p < 0.0001), and CEACAM6 expression was increased in GBC as well (*p < 0.05) (Fig. 8B, C). By employing Kaplan-Meier survival analysis, we found that high OLFM4 expression in GBC was significantly correlated with the reduced overall survival of GBC patients (Fig. 8D).

Fig. 8. High OLFM4 expression in GBC patients indicated poor prognosis and summary of OLFM4 impelled gallbladder carcinogenesis in this study.

A Immunohistochemistry staining of OLFM4 expression in GBC clinical samples on tissue microarray. Images were representative samples demonstrating high and low expression of OLFM4 and CEACAM6, respectively. B, C Violin plots showing differential expressions of OLFM4 and CEACAM6 between tumor and adjacent tissues (****p < 0.0001, *p < 0.05). D The overall survival rate of GBC patients in cohorts of high and low OLFM4 expressions. log-rank test was performed for significance. E Graphic outline of the mechanism that OLFM4 regulated TGF-β/CEACAM6 axis promoting EMT in GBC cells.

DISCUSSION

Carcinoma of the gallbladder is a highly lethal disease with poor prognosis. It is frequently diagnosed incidentally during laparotomy or laparoscopy for gallstone. Reliable screening biomarker and effective therapeutic agent remain a challenge because GBC pathogenesis was poorly understood. Persistent inflammation was recognized as the dominant factor which led to gallbladder epithelium carcinogenesis. It can impair genomic stability of gallbladder epithelial cells by oxidative stress accumulation of DNA damage and induce EMT, ultimately promoting malignant transformation. The fundamental question in GBC etiology is what molecular process and regulators contributing to chronic inflammation-driven malignant transformation. To answer this question, we employed high-throughput single-cell RNA-seq to dissect the heterogeneity of GBC, identified the epithelium cell trajectory, and revealed the regulatory network promoting this oncogenic process. Interestingly, we found three subtypes of epithelium based on transcriptomic divergence exhibiting varied inflammatory phenotypes that were closely associated with EMT score (Fig. S8). Elevated inflammations were identified in epithelium of state 2 and 3, in which other cancer hallmarks gene signatures scoring including angiogenesis, hypoxia, metastasis and invasion were significantly increased suggesting that the malignant characteristics of GBC were related with inflammatory reaction (Fig. S8). Coincided with molecular subtypes of GBC patients, GBC patients could be stratified into three subtypes based on a 95-gene transcriptional signature. Two subtypes of GBC patients with poor prognosis showed adverse clinicopathological features, and the other good survival subtype has opposite feature [30]. These findings implicated that transcriptomic heterogeneity among GBC epithelium and patients making it more complicated for treatment.

Remarkably, our study disclosed that cell adhesion molecule binding was the dominant GO term that impelled GBC progression. Further analysis indicated that OLFM4 annotated with this GO term above was significantly upregulated in both of epithelium state 2 and 3. Subsequent investigations revealed that OLFM4 could bind to its receptor TGFβR1 to promote expression and secretion of TGF-β in GBC, which was evidenced by protein-protein interaction and “rescue” assays. Among GBC tumor microenvironment, TGF-β could upregulate CEACAM6 expression depending on phosphorylation of SAMD3. Importantly, we experimentally verified that CEACAM6 could activate EMT through PI3K/AKT pathway and enhance migration and invasion of GBC both in vitro and in vivo. OLFM4 is a glycoprotein which is strongly expressed in gastrointestinal tract with multifunction including immunity modulation and tumor promotion or suppression depending on different organs [21]. OLFM4 expression is upregulated in inflamed colon tissue and H. pylori-infected gastric mucosa and may function as a gastrointestinal inflammation regulator [31–33]. In ulcerative colitis, for example, TGF-β and OLFM4 are co-upregulated in both inflamed and endoscopically uninflamed regions, where their expression is accompanied by EMT markers such as suppressed E-cadherin and induced vimentin and α-SMA [32]. This co-expression pattern suggests that TGF-β, a master regulator of inflammation and tissue remodeling, may be regulated by OLFM4 under inflammatory stimuli. According to previous studies, aberrant OLFM4 expressions were found in GBC, which were consistent with our findings [34, 35]. However, few studies have explored the tumor-promoting role of OLFM4 during gallbladder carcinogenesis, especially corelated with its inflammation modulation function. Exploration of interactions among GBC TME implicated that OLFM4 inhibited T cells immune response by upregulating the expression of PD-L1 on gallbladder epithelium, which corelated OLFM4 anti-inflammatory effect with tumor-promoting function [34]. Interestingly, we found that a variety of proinflammatory cytokines were dramatically upregulated in epithelium of state 2 and 3, and OLFM4 expression was positively associated with epithelial inflammation intensity (Fig. S9). The increased expression of OLFM4 could promote migration and invasion in gallbladder epithelium via EMT in a cell autonomous manner, and comparable function was observed in keratinocytes in which OLFM4 directly remodeled the actin cytoskeleton and regulated cell-cycle progression [36]. CEACAM6, participating in cell adhesion and intracellular signaling, was upregulated in various types of cancers. It was reported that the CEACAM6 expression was regulated by TGF-β, and could induce EMT through PI3K/AKT signaling in gastric cancer, which were consistent with our findings in gallbladder cancer [37, 38]. Moreover, studies have shown that CEACAM6 was highly expressed in cholangiocarcinoma tissues, and its expression was negatively correlated with the degree of tumor differentiation. Silencing CEACAM6 could reduce the phosphorylation levels of key molecules in the SRC/PI3K/AKT signaling pathway [39]. Further studies have confirmed that the differential expression of CEACAM6 regulates AKT activity in a c-Src-dependent manner [40].

Conventionally, it was commonly known that TGFβR1 activated EMT through either SMAD2/3 phosphorylation or TRAF6 related ubiquitination pathways [41, 42]. Based on our study, OLFM4 acted as a noncanonical ligand of TGFβR1 triggering production of TGF-β independent of SMAD3, which was supported by the fact that TGF-β expressions in GBC-SD and NOZ were undisturbed by treatment of inhibitors against either ALK5 (IN1130) or SMAD3 phosphorylation (SIS-3). Interestingly, upon TGF-β stimulation, we observed a biphasic pattern of Smad3 phosphorylation that was different from the conventional rapid and transient kinetics reported in some studies [43, 44]. In our time-course experiments, p-Smad3 levels peaked at 1 h, declined at 2 h, but subsequently rose again after 4 h, reaching asecond peak at 6 h (Fig. 6I). While a previous study demonstrated that the Smad3/4 protein complex binds to the CEACAM6 promoter upon TGF-β1 stimulation, and no significant amounts of CEACAM6 promoter DNA fragment was observed between 1 h and 24 h after TGF-β1 stimulation [45]. The results indicated that there was no significant difference in p-Smad3 level between 1 h and 24 h. This finding was consistent with ours indirectly.We speculated that the chronic inflammation in GBC has evolved alternative signaling pathways which sustained Smad3 phosphorylation, although the precise mechanisms underlying this prolonged activation remain to be elucidated. The limitation of current study provided an important direction for future investigation. However, it was reported that OLFM4 was upregulated in mice-small intestine when TGFβR1 was genetically inactivated [46]. In both colorectal cancer and leukemic cells, it seemed that activation of TGF-β signaling can downregulate OLFM4 expression, which implying OLFM4 was modulated by TGF-β [47]. In our study, the expression of OLFM4 was not affected by TGF-β signaling on the basis that no apparent OLFM4 expression was observed in GBC cells by stimulation of recombinant TGF-β (Fig. 6I, J). Again, the expression of OLFM4 was affected by treating GBC cells with TGF-β signaling inhibitors (Fig. 6E–H). In brief, the relationship between OLFM4 and TGF-β may be organ-specific, and the regulatory network of these oncogenic molecules may be established during GBC tumorigenesis. In this novel pathway, TGF-β was the downstream signaling mediator that was able to initiate EMT in cancer cells. Interestingly, CEACAM6 was upregulated in response to TGF-β, and further strengthened EMT in GBC cells via PI3K/AKT pathway, by which OLFM4 could promote GBC EMT in a cascading manner. Furthermore, it was evident that OLFM4 has employed TGFβR1 as functional receptor to transduce signals, and triggered EMT by TGF-β, suggesting that crosstalk among oncogenic pathways occurred in GBC. It was probably that the signal leakage contributed to malignancy and drug-resistant properties of GBC. Similarly, EMT genes were overexpressed in GBC patients with poor prognosis and higher recurrence, and mutations of genes related with TGF-β were enriched [48]. These data suggested that TGF-β signaling and downstream EMT could act as dominators in GBC TME which may be hot spots for therapeutic targets.

We delineated the mechanism that OLFM4 promoted EMT by TGF-β/CEACAM6 axis in response to cholecystitis (Fig. 8E). However, further investigation should be performed due to the limitations. We postulate that OLFM4 may upregulate TGF-β in gallbladder epithelium via ubiquitination of TGFβR1 caused by TRAF6 [49]. Among GBC TME, a great amount of TGF-β from immune cells should be considered, and it would further strengthen the expression of CEACAM6 and EMT in epithelium. Moreover, overexpression of OLFM4 in GBC has not enhanced cell proliferation, however, OLFM4 knockdown could inhibit cell growth both in vitro and in vivo. It was possible that the basal OLFM4 expression in gallbladder cancer cells was enough to promote cell proliferation, and increasing expression of OLFM4 could not significantly enhance cell replication. Gallbladder epithelium with normal phenotype should be adopted in OLFM4 promoting cell expansion, migration and invasion assays. The direct interaction between OLFM4 and TGFβR1 was confirmed by CoIP and pull-down assay. The docking analysis should be performed to find out the interactive residuals of both OLFM4 and TGFβR1, and CoIP assays are necessary for validation with key residues mutated, which will be informative for drug development. In summary, our study has provided a molecular cell trajectory of gallbladder epithelium carcinogenesis, which was inflammation-related. The exploration of the mechanism offers insights into key molecular events and signal transducers, which will accelerate development of diagnostic tools and targeted therapeutics for GBC.

Author contributions

SY and WW conceived, designed the study and wrote the manuscript. SY, WH, QZ, SL, XC and ZW conducted experiments and analyzed the data. ZL, HW, RW and QY contributed to the collection of clinical specimens. WH and LW provided clinical pathology evaluations. XW, MM and LL provided technical support and commented on the study. RL and JT supervised the study and revised the manuscript. RL, JT and WW are responsible for the overall content as the guarantor.

Data availability

The data underlying this article will be shared on reasonable request upon contact with the corresponding author. The raw sequence data have been deposited in the GSA-Human database under accession number HRA017631 and are publicly available at https://ngdc.cncb.ac.cn/gsa-human. All codes and scripts used for association studies are available on request.

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participant

All animal experiments were approved by Animal Ethics and Welfare Committee (AEWC) of Yan’an Affiliated Hospital of Kunming Medical University (No. 2023025). All clinical data were sourced from electronic medical records under informed consent, and sample collection were approved by Yan’an Affiliated Hospital of Kunming Medical University (No. 2023-024-01).

Supplementary information

The online version contains supplementary material available at 10.1038/s41388-026-03808-7.