COLEC10 inhibits the stemness of hepatocellular carcinoma by suppressing the activity of β-catenin signaling

Abstract

Background

Liver cancer stem cells (CSCs) contribute to tumor initiation, progression, and recurrence in hepatocellular carcinoma (HCC). The Wnt/β-catenin pathway plays a crucial role in liver cancer stemness, progression, metastasis, and drug resistance, but no clinically approved drugs have targeted this pathway efficiently so far. We aimed to elucidate the role of COLEC10 in HCC stemness.

Methods

The Cancer Genome Atlas (TCGA) and the Clinical Proteomic Tumor Analysis Consortium (CPTAC) databases were employed to search for the association between COLEC10 expression and HCC stemness. Colony formation, sphere formation, side population, and limiting dilution tumor initiation assays were used to identify the regulatory role of COLEC10 overexpression in the stemness of HCC cell lines. Wnt/β-catenin reporter assay and immunoprecipitation were performed to explore the underlying mechanism.

Results

COLEC10 level was negatively correlated with HCC stemness. Elevated COLEC10 led to decreased expressions of EpCAM and AFP (alpha-fetoprotein), two common markers of liver CSCs. Overexpression of COLEC10 inhibited HCC cells from forming colonies and spheres, and reduced the side population numbers in vitro, as well as the tumorigenic capacity in vivo. Mechanically, we demonstrated that overexpression of COLEC10 suppressed the activity of Wnt/β-catenin signaling by upregulating Wnt inhibitory factor WIF1 and reducing the level of cytoplasmic β-catenin. COLEC10 overexpression promoted the interaction of β-catenin with the component of destruction complex CK1α. In addition, KLHL22 (Kelch Like Family Member 22), a reported E3 ligase adaptor predicted to interact with CK1α, could facilitate COLEC10 monoubiquitination and degradation.

Conclusion

COLEC10 inhibits HCC stemness by downregulating the Wnt/β-catenin pathway, which is a promising target for liver CSC therapy.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13402-024-00972-4.

Introduction

Liver cancer remains the sixth most common cancer and the third leading cause of cancer mortality [1]. Hepatocellular carcinoma (HCC) accounts for approximately 80% of the primary liver cancer. HCC is a biologically complex and highly heterogeneous disease, with its heterogeneity primarily driven by liver cancer stem cells (CSCs), also referred to as tumor-initiating cells (TICs) [2]. CSCs possess stem cell-like capacity to self-renew and resistance to chemotherapy, including greater colony-forming efficiency, higher proliferative output, and most importantly, greater ability to form tumors in vivo. The growth of HCC is fuelled by liver CSCs (LCSCs) that are also recognized as the root of tumor recurrence and therapy resistance [3]. Currently reported markers for LCSCs include CD13, CD24, CD44, CD47, CD90, CD133, ICAM1, EpCAM, LGR5, OV6 and et al. [2]. Targeting LCSC surface markers directly can effectively eliminate LCSCs in preclinical animal models [4], but some of the markers are also expressed in normal tissues, which will lead to undesirable side effects [5]. Instead, there are other potential directions to target HCC stemness specifically, like reversing the dysregulated biological processes [6, 7], regulating the microenvironment of LCSCs [8], and LCSC-directed immunotherapy [9].—However, efficient treatments targeting LCSCs have not yet been established in clinical practice. Therefore, exploring the regulatory mechanisms of HCC stemness and identifying preferable targets remains essential.

Convincing evidence suggests that approximately 20–35% of HCC cases exhibit activation of the Wnt/β-catenin pathway [10]. This pathway serves as a significant signaling cascade for the proliferation of LCSCs, and overexpression of β-catenin can enhance the self-renewal and in vivo tumorigenicity of these cells. Regulation of β-catenin stability is one of the main mechanisms by which Wnt/β-catenin activity is tuned. Normally, the β-catenin destruction complex, including adenomatosis polyposis coli (APC), casein kinase 1α (CK1α), glycogen synthase kinase 3β (GSK3β), and Axin, negatively controls β-catenin, which induces β-catenin phosphorylation and proteolysis [11]. When in the presence of Wnt ligands or mutations of key genes in this pathway, β-catenin accumulates, translocates to the nucleus, and induces the transcription of genes promoting cell proliferation and stemness by combining with T-cell factor/lymphoid enhancer factor (TCF/TEF) transcription factors.

KLHL22, a known E3 ligase adaptor, usually forms a functional cullin-RING E3 ubiquitin ligase complex with the scaffold protein CUL3 and the ring-finger protein RBX1 [12]. KLHL22 is reported to be involved in mitosis, mTORC1, and Wnt/β-catenin signaling pathways [13, 14]. We previously reported that COLEC10 was downregulated and associated with poor prognosis in HCC patients, and one of the mechanisms for COLEC10 inhibiting HCC was revealed [15]. In this study, we investigate the role of COLEC10 in HCC stemness. Through analysis of human HCC data, as well as in vitro and in vivo functional experiments, we identified COLEC10 as a critical suppressor of HCC stemness and progression. Mechanistically, COLEC10 inhibits the β-catenin pathway by promoting its degradation, thereby suppressing HCC stemness. Additionally, we explored the mechanisms underlying COLEC10 downregulation in HCC. Immunoprecipitation and ubiquitination assays indicated that KLHL22 interacts with COLEC10 and facilitates its monoubiquitination and degradation. Our findings reveal a novel function of COLEC10 in suppressing HCC development, making it a potential target for LCSC therapy.

Materials and methods

Bioinformatic analysis

One-class logistic regression (OCLR) was utilized to calculate the mRNA stem index (RNAsi) and DNA stem index (DNAsi) of the TCGA-HCC dataset [16]. Spearman analysis was used to calculating the correlation between COLEC10 mRNA level and mRNAsi or DNAsi. The Cancer Genome Atlas (TCGA), Gene Expression Omnibus (GEO), and the Clinical Proteomic Tumor Analysis Consortium (CPTAC) databases were utilized to analyze the relationship between COLEC10 and HCC stemness. The Contaminant Repository for Affinity Purification (CRAPome) database was used to analyze the specificity of the candidates in the mass spectrum (MS) results by uploading them to the database.

Patient samples

Primary human HCC samples were obtained from 40 patients undergoing hepatectomy at the Sun Yat-Sen University Cancer Centre (Guangzhou, China). The baseline characteristics of the 40 patients are detailed in Supplementary Table 1. The study was approved by the Institute Research Ethics Committee at the Third Affiliated Hospital of Sun Yat-sen University. Written informed consent was obtained from each patient.

Cell culture

HEK293T cells and HCC cell lines (PLC/PRF/5 and HepG2) were cultured in Dulbecco Modified Eagle Medium (DMEM; Invitrogen) containing 10% fetal bovine serum (Gibco) at 37 °C and 5% CO₂. All cell lines were acquired from Cellcook Biotech (Guangzhou, China), and cell line authentication was demonstrated by short tandem repeat profiling.

Plasmids, RNA oligonucleotides, stable cell line establishment, and transfection

The plasmid construction method of COLEC10 and four serial COLEC10 truncation mutants, and the establishment of COLEC10 stable cell lines have been described previously [15]. The complementary DNA (cDNA) encoding the full-length human KLHL22 gene was amplified using PCR from HEK293T cell cDNA library and subcloned into the pcDNA3.1(+) vector (Invitrogen). During PCR, a sequence encoding a hemagglutinin (HA) tag (TACCCATACGATGTTCCAGATTACGCT) was added to the C-terminus of KLHL22. The β-catenin-Flag plasmid was purchased from RiboBio Company (Guangzhou, P.R. China).

7TGP plasmid was a gift from Roel Nusse (Addgene plasmid # 24305; http://n2t.net/addgene:24305; RRID: Addgene_24305). The β-catenin activity reporter cells were established as previously described [15]. Briefly, the lentiviruses were produced by co-transfecting the 7TGP plasmid and the packaging plasmids psPAX2 coupled with pMD2.G (Addgene) into HEK293T cells using jetPRIME (101000046; Polyplus) for 48 h. Then cells were infected by the lentiviruses and selected by puromycin (Merck). The fluorescence intensity of GFP was determined by flow cytometry (Beckman-Coulter CytoFLEX LX Flow Cytometer) and the live-cell analysis system (IncuCyte, SX5).

The siRNAs were obtained from RiboBio Company (Guangzhou, China) or IGE Biotechnology Company (Guangzhou, China). The target sequences are listed in Supplementary Table 2. Cells were transfected with 100 nM siRNA using Lipofectamine RNAiMAX according to the manufacturer’s protocol (Invitrogen, USA). The effect of RNA silence was confirmed by western blotting.

Sphere formation assay

Cell spheres were cultured in DMEM/F12 medium (Gibco) supplemented with 20 ng/mL human recombinant EGF(236-EG-200; R&D), 20 ng/mL human recombinant basic FGF (233-FB-025; R&D), and 2% B27 supplement (12587010; Gibco) at 37 °C and 5% CO₂. HCC cells were seeded into 6-well low-attachment plates at a density of 1000 per well. Cells were replenished with supplemented medium every two days and the number of spheres (more than 50 μm) was calculated after two weeks.

Side population assay

Cultured cells were detached from culture dishes using 0.25% Trypsin-EDTA (Gibco), and 1 × 10⁶ cells/mL were suspended in high-glucose DMEM (Gibco) supplemented with 5% FBS. The cells were then incubated at 37 °C for 90 min with 10 µg/mL Hoechst 33342 dye (B2261; Sigma-Aldrich) either alone or in the presence of 100 µM verapamil (V4629; Sigma-Aldrich). Cell suspensions were washed with cold PBS, centrifuged and resuspended in 1 mL 1x PBS supplemented with 10 µL propidium iodide (PI; 556547; BD Biosciences) to label dead cells. Side population cells were determined using a CytoFLEX LX Flow Cytometer (Beckman-Coulter). The SP gate was defined as the region where cells disappeared when cells were exposed to verapamil that blocks Hoechst 33342 efflux.

Reverse transcription and real-time PCR (qPCR)

TRIzol reagent (Invitrogen) was used to isolated total RNA from cells according to the manufacturer’s protocol. GoScript Reverse Transcription System (Promega) was used to reversely transcribed 2 µg RNA into cDNA. qPCR was performed using Platinum SYBR Green qPCR SuperMix-UDG (Invitrogen) with a LightCycler 480 PCR platform (Roche). Specific primers used were as follows: COLEC10: forward, 5′-GGCTTTGCATCCTTGCTTCG-3′ and reverse, 5′-CCTTTAATTCCTTTCGGCCCC-3′; AFP: forward, 5′-CTTTGGGCTGCTCGCTATGA-3′ and reverse, 5′- GCATGTTGATTTAACAAGCTGCT-3′; WIF-1: forward, 5′-CTGATGGGTTCCACGGACC-3′ and reverse, 5′-AGAAACCAGGAGTCACACAAAG-3′; Axin2: forward, 5′-TATCCAGTGATGCGCTGACG-3′ and reverse, 5′-TTACTGCCCACACGATAAGG-3′; EpCAM: forward, 5′- AATCGTCAATGCCAGTGTACTT-3′ and reverse, 5′-TCTCATCGCAGTCAGGATCATAA-3′.

Flow cytometry

To measure the level of cell surface EpCAM, cells were cultured in 6-well plates and grown to 70% confluence. Cells were washed and trypsinized, and 5 × 10⁵ cells were incubated with 100 µL of diluted EpCAM or IgG antibody on ice. After 45 min of incubation, cells were washed with 1x PBS for three times and incubated with diluted secondary anti-mouse-PE antibody (12-4010-82; eBioscience™) for 30 min on ice. Then cells were washed, resuspended in 1x PBS, and analyzed using a CytoFLEX LX Flow Cytometer (Beckman-Coulter). Data were analyzed using the FlowJo software.

Limiting dilution assay

Male NCG mice (NOD/ShiLtJGpt-Prkdc^em26Il2rg^em26/ Gpt, 4–5 weeks old, healthy and naïve to any laboratory procedure) were obtained from the GemPharmatech company (Guangdong, China). All animal procedures were performed according to the Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of South China Agricultural University (ethics code: 2023D054, 2023). Due to ethical and practical constraints, we decided to use the minimum number of animals possible. Therefore, 6 mice per subgroup were used in our experiment. A total number of 48 mice were randomly divided into two groups: vector and COLEC10-Flag (n = 24); and each group contained 4 subgroups (n = 6) with gradient concentration cells: 1*10^+ 5, 1*10^+ 4, 1*10^+ 3 and 1*10^+ 2. The vector/ COLEC10-Flag group received a gradient number of vector/COLEC10-Flag overexpression PLC/PRF/5 HCC cells that resuspended in 1:1 DMEM (Invitrogen) /matrigel (354234, Corning) by subcutaneous injection into the armpit. The experiment operators and principal investigator were aware of the allocations. The order of treatments and measurements, or animal/cage location, was not systematically biased by any known or unknown factors. Tumor growth was regularly assessed through size and volume measurements. Mice developing any sign of infection or other complications during the experiment were excluded. One mouse in the vector group died for unknown reasons and was excluded from this study. The mice were sacrificed by cervical dislocation under anesthesia. Subcutaneous tumor volume and tumor incidence were recorded. Tumor-initiating frequency was calculated using Extreme Limiting Dilution Analysis (http://bioinf.wehi.edu.au/software/elda/).

RNA sequencing and data analysis

Total RNA was extracted using a Trizol reagent kit (Invitrogen, USA) according to the manufacturer’s protocol. RNA quality was assessed on an Agilent 2100 Bioanalyzer (Agilent Technologies, USA) and checked using RNase-free agarose gel electrophoresis. Total mRNA was enriched by oligo (dT) beads (Epicentre, USA) and reverse-transcribed into cDNA using random primers. The mRNA was ligated with proprietary 5′ and 3′ adapters. The ligation products were reverse-transcribed by PCR amplification to generate a cDNA library, which was sequenced using an Illumina HiSeq 2500 by Gene Denovo Biotechnology Co. (Guangzhou, China). Raw reads were further filtered by fastp (version 0.18.0) to get high-quality clean reads. The parameters were as follows: (1) removing reads containing adapters; (2) removing reads containing more than 10% of unknown nucleotides; (3) removing low-quality reads containing more than 50% of low-quality. RNAs differential expression analysis was performed by edgeR between two groups and drawn by ggplot2 R package, and genes with the parameter of false discovery rate (FDR) below 0.05 and absolute fold change ≥ 1 were considered differentially expressed.

Reagents

Commercially available antibodies used were as follows: KLHL22 (16214-1-AP; Proteintech), β-catenin (51067-2-AP; Proteintech), AFP (14550-1-AP; Proteintech), EpCAM (2929; Cell Signaling Technology), GSK3β (22104-1-AP; Proteintech), WIF-1 (sc-373780; Santa Cruz), CK1α (sc-74582; Santa Cruz), Axin (16541-1-AP; Proteintech), Tubulin (66031-1-Ig; Proteintech), anti-Flag tag (F3165; Sigma-Aldrich), and anti-HA tag (3724; Cell Signaling Technology) antibodies. Rabbit polyclonal COLEC10 antibody was prepared by GenScript Biotech. Drugs used were as follows: recombinant Human Wnt-3a Protein (5036-WN; R&D), lithium chloride (L9650; Sigma-Aldrich), MG132 (S2619; Selleck), chloroquine (C6628; Sigma-Aldrich). All other chemical reagents were obtained from Sigma-Aldrich unless otherwise indicated.

Statistical analysis

Statistical analyses were performed by GraphPad Prism 6 software (Dotmatics). t-test was used to determine the significance of variances between two groups. Each experiment was performed 3 times. Unless otherwise indicated, all error bars indicate the SD. All statistical tests were 2-sided, and P < 0.05 was considered statistically significant. P < 0.05, P < 0.01, and P < 0.001 are indicated by asterisks (*, **, and ***, respectively) in the figures. The methods of Western blotting, immunofluorescence, immunoprecipitation, immunohistochemistry and colony formation assay have been described previously [15].

Results

COLEC10 expression is negatively correlated to HCC stemness

To investigate the protein expression of COLEC10 in HCC patients, we searched the CPTAC database and found that the protein level of COLEC10 was significantly downregulated in HCC tumors (Fig. 1A). The Kaplan-Meier survival analysis indicated that higher COLEC10 expression was significantly associated with longer recurrence-free survival (RFS) in TCGA-HCC cohort (Fig. 1B), implying that COLEC10 expression was negatively related to HCC recurrence. The COLEC10 expression of EpCAM⁺ HCC patients was significantly lower than that of EpCAM^- patients in the GSE5975 dataset (Fig. 1C). The expression of COLEC10 and AFP was negatively correlated in both mRNA and protein levels (Fig. 1D). Furthermore, in the TCGA-HCC cohort, we found that COLEC10 expression was negatively correlated with mRNA and DNA stem index (Fig. 1E). mRNA stem index (mRNAsi) reflects the expression feature of stem cells, while the DNA stem index (DNAsi) shows epigenetics based on methylation patterns [16]. Gene Set Enrichment Analysis (GSEA) of the TCGA-HCC cohort showed the enrichment of undifferentiated cancer and embryonic stem cell core gene sets in HCC patients with low COLEC10 expression (Fig. 1F). In sum, our bioinformatic analysis results suggested that COLEC10 expression is negatively correlated to HCC stemness.

Fig. 1.

COLEC10 is negatively correlated with HCC stemness. A COLEC10 protein level was analyzed using the CPTAC database (https://proteomic.datacommons.cancer.gov/pdc/). NT, non-tumor; T, tumor. B High COLEC10 expression was significantly associated with better recurrence-free survival in the TCGA HCC cohort. C COLEC10 mRNA level was analyzed between tumor tissues of EpCAM⁺ and EpCAM⁻ patients in the GSE5975 HCC cohort. T/N denotes the ratio of COLEC10 expression level of tumors versus para-tumor tissue. D The correlations of mRNA (left) and protein (right) expression levels between COLEC10 and AFP were analyzed respectively in the GSE14520 and CPTAC HCC cohorts. The Pearson correlation coefficient and the P value are shown. E The correlation between COLEC10 level and mRNAsi (left) or DNAsi (right) was analyzed in the TCGA HCC cohort. F GSEA was performed between the COLEC10 low- and high-expression groups in the TCGA HCC cohort. The enrichment score (ES) and its P value were shown. CL10, COLEC10; TCGA, The Cancer Genome Atlas; GSEA, gene set enrichment analysis

COLEC10 overexpression inhibits the stemness of HCC cells

To confirm the regulatory role of COLEC10 in HCC stemness, we established COLEC10-overexpressing cells using PLC/PRF/5 and HepG2 cell lines, and the efficiency was detected by western blotting assay (Supplementary Fig. S1). COLEC10 overexpression reduced the colony and sphere formation abilities of HCC cells (Fig. 2A, B). The percentage of stem-like cells was reduced after COLEC10 overexpression in side population assay (Fig. 2C). We also found that the levels of EpCAM and AFP were significantly downregulated in both COLEC10-overexpressing cells (Fig. 3A–E, Supplementary Fig. S2). In vivo limited dilution assay showed that COLEC10 overexpression reduced tumor-initiating cell (TIC) frequency (Table 1). Our HCC cohort also demonstrated higher EpCAM and AFP expression levels in low COLEC10 expression patients by immunohistochemistry (Fig. 3F, G), in echo with our previous finding that patients with low COLEC10 expression had high serum levels of AFP [15]. Besides, we decreased the COLEC10 level using siRNA in the COLEC10-overexpressing PLC/PRF/5 cell line, and the efficiency was confirmed by western blotting assay (Fig. 3H). The percentage of stem-like cells and the level of EpCAM as well as AFP were recovered after silencing COLEC10 (Fig. 3I, J). Therefore, our experiments confirmed that COLEC10 overexpression can inhibit HCC stemness both in vitro and in vivo.

Fig. 2.

COLEC10 overexpression attenuates HCC stemness in vitro and in vivo. A–C The stemness of PLC/PRF/5 and HepG2 cells after COLEC10 overexpression was assessed using colony formation (A), sphere formation (B), and side population assays (C). The representative images of these assays and their quantification are shown. V Vector

Fig. 3.

COLEC10 overexpression reduces the levels of EpCAM and AFP in HCC. (A–D) The EpCAM level of COLEC10 overexpression was examined using qPCR (A), immunofluorescence (B), western blotting (C), and flow cytometry (D) in PLC/PRF/5 and HepG2 cells. E The AFP level after COLEC10 overexpression was examined using qPCR (left and middle) and western blotting (right) in PLC/PRF/5 and HepG2 cells. F, G Representative images and quantification of immunohistochemistry analysis of EpCAM (F) and AFP (G) protein levels in COLEC10-high or -low human HCC tissues. (H) The efficiency of COLEC10 siRNA was confirmed by western blotting. (I) Representative images of side population assays (left) and the quantification (right) after silencing COLEC10 in PLC/PRF/5 cells. (J) The mRNA level of EpCAM and AFP after silencing COLEC10 in PLC/PRF/5 cells was detected using qPCR. Full-length blots are presented in Supplementary Figure S7. Scale bar, 50 μm

Table 1.

The result of limiting dilution tumor initiation assay in PLC/PRF/5 cells

Number of cells injected	Tumor incidence
	Vector	CL10-Flag
1*10 + 5	6/6	6/6
1*10 + 4	5/5	6/6
1*10 + 3	6/6	5/6
1*10 + 2	5/6	1/6
TIC Frequency	1 in 55.8	1 in 556.1
TIC 95%CI	1/151-1/20.6	1/1356-1/228.1
P Value		0.00137

TIC tumor initiation cell, CI confidence interval

COLEC10 suppresses the activity of Wnt/β-catenin signaling

EpCAM is a direct transcription target of the Wnt/β-catenin pathway, and it regulates HCC stemness and drug resistance via a β-catenin/TCF1 positive feedback loop [17]. Our above results showed that COLEC10 overexpression could downregulate EpCAM expression, so we examined whether COLEC10 plays a role in regulating β-catenin signaling.

The β-catenin activity reporter model was established to explore the effect of COLEC10 on Wnt/β-catenin signaling. The median fluorescence intensity (MFI) of GFP was obviously reduced after silencing β-catenin (Fig. 4A, B), which suggested that our cell model was able to reflect the activity of β-catenin signaling specifically. Overexpressing COLEC10 significantly inactivated the Wnt/β-catenin signaling in HEK293T and PLC/PRF/5 cells (Fig. 4C, D). Overexpressing COLEC10 weakened the upregulation of Axin2 after adding Wnt/β-catenin pathway activator Wnt3a (Fig. 4E). The Wnt/β-catenin pathway activator Wnt3a could not increase the stemness of COLEC10-overexpressing PLC/PRF/5 cell in colony formation, sphere formation and side population assays (Fig. 4F, G). The above results suggested that COLEC10 could inhibit the Wnt/β-catenin pathway and may work downstream of or against to Wnt3a.

Fig. 4.

COLEC10 inhibits the activity of the Wnt/β-catenin pathway in HCC cells. A The efficiency of β-catenin siRNA was confirmed by western blotting. B Successful establishment of Wnt/β-catenin pathway reporter cells was confirmed using flow cytometry. C, D The activity of the Wnt/β-catenin pathway was measured in COLEC10 overexpression cells using real-time cell imaging (C) and flow cytometry (D). E The activity of the Wnt/β-catenin pathway was determined by measuring Axin2 mRNA level in COLEC10 overexpression cells. F–H) The stemness of COLEC10-overexpressing PLC/PRF/5 cells with or without Wnt3a was assessed using colony formation (F), sphere formation (G), and side population assays (H). Full-length blots are presented in Supplementary Figure S7. NC negative control, MFI median fluorescence intensity

COLEC10 inhibits Wnt/β-catenin signaling by upregulating WIF1 and decreasing β-catenin

In order to find out how COLEC10 suppresses the activity of Wnt/β-catenin signaling, we performed transcriptome sequencing analysis in COLEC10-overexpressing PLC/PRF/5 cells. We found that the mRNA level of WIF1 was significantly upregulated after COLEC10 overexpression (Fig. 5A), which was confirmed by qPCR in PLC/PRF/5 and HepG2 cell lines (Fig. 5B). The upregulated level of WIF1 was also detected in the supernatant of overexpressing-COLEC10 PLC/PRF/5 cells (Supplementary Fig. S3). Consistently, in our HCC tissue cohort, patients with high COLEC10 level had increased WIF1 expression (Fig. 5C). In addition, COLEC10 overexpression reduced the total level of β-catenin in PLC/PRF/5 cells (Fig. 5D). As is well known, β-catenin level is mainly under the control of the destruction complex and degraded by the ubiquitin-proteasome mechanism in cells [11]. Therefore, immunoprecipitation assays were performed and found that exogenous β-catenin combined more CK1α, a component of the destruction complex when overexpressing COLEC10 in HEK293T cells (Fig. 5E). This result was further confirmed in a HCC cell line. In COLEC10-overexpressing PLC/PRF/5 cells, endogenous β-catenin interacted with more CK1α in the whole cell and cytoplasm, where the destruction complex exactly works (Fig. 5F, G). Overexpression of β-catenin could increase the activation of Wnt/β-catenin signaling to some extent in COLEC10-overexpressing cells (Fig. 5H). The level of COLEC10 was negatively correlated with β-catenin, especially nucleus β-catenin level in HCC tissues (Fig. 5I). Therefore, the above results suggested that COLEC10 might suppress the activity of Wnt/β-catenin signaling by upregulating WIF1 expression and reducing β-catenin level.

Fig. 5.

COLEC10 inhibits Wnt/β-catenin signaling by upregulating WIF1 and decreasing β-catenin. A The volcano plot depicts differentially expressed genes in COLEC10-overexpressing and control cells. Red dots represent genes expressed at higher levels in COLEC10-overexpressing cells while blue dots represent downregulated genes compared to control cells. B The WIF1 mRNA level was examined using qPCR in HCC cells. C Representative images (left) and quantification (right) of immunohistochemistry analysis of WIF1 level in COLEC10-high or -low HCC tissues. D The β-catenin protein level was examined using western blotting in COLEC10-overexpressing PLC/PRF/5 cells. E Immunoprecipitation and western blotting analysis were performed on HEK293T cell lysates exogenously expressing COLEC10-His and β-catenin-Flag using anti-Flag agarose. F, G Immunoprecipitation and western blotting analysis were performed on PLC/PRF/5 total cell lysates (F) and only the cytoplasmic lysates (G) exogenously expressing COLEC10-Flag using β-catenin antibody. H The activity of the Wnt/β-catenin pathway was measured in COLEC10 overexpression cells with or without β-catenin overexpression using flow cytometry. I Representative images (left) and quantification (right) of immunohistochemistry analysis of β-catenin level in COLEC10-high or -low HCC tissues. Different localization of β-catenin in cells (total, nucleus and membrane) were measured. Full-length blots are presented in Supplementary Figure S8. IP immunoprecipitation, Sup. supernatant. Scale bar, 50 μm

KLHL22 promotes the degradation of COLEC10

Our previous work has preliminarily explored the interacting proteins of COLEC10 in HEK293T cells by a 2-step tandem affinity purification technique and MS detection [15]. To investigate the underlying mechanism for COLEC10 reduction in HCC, we further analyzed the candidates of the MS result. We uploaded the top 20 combination candidates to the CRAPome database and found that KLHL22 was a rather specific combination protein (Supplementary Table 3). The KLHL22 protein coverage and best-matched unique peptide were shown (Fig. 6A, B). Immunoprecipitation assays confirmed the combination of exogenous COLEC10 and KLHL22, and the deletion of CLR, Collagen, or Neck domain of COLEC10 abolished their interaction to some extent (Fig. 6C). KLHL22 is a substrate-specific adapter of a BCR (BTB-CUL3-RBX1) E3 ubiquitin ligase complex [18]. We assumed that KLHL22 may promote COLEC10 degradation. Overexpressing KLHL22 reduced and silencing KLHL22 increased the level of COLEC10 in HEK293T and HCC cell lines (Fig. 6D, E and Supplementary Fig. S4). Interestingly, the ubiquitination assay showed that KLHL22 could monoubiquitinate, but not polyubiquitinate, exogenous Flag-tagged COLEC10 (Fig. 6F). The lysosome-dependent protein degradation inhibitor chloroquine could reverse the reduction of COLEC10 level, while the proteasome inhibitor MG132 could not (Fig. 6G). In general, the above data suggested that KLHL22 interacts with COLEC10 and promotes its degradation.

Fig. 6.

KLHL22 promotes the degradation of COLEC10. A Protein coverage of KLHL22 in the LC/MS results. The blue lines under the amino acid sequences indicate the supporting peptides matching KLHL22. B The best unique peptide (indicated by a red asterisk in (A)) of KLHL22 protein in the LC/MS result. C Immunoprecipitation and western blotting were performed using either anti-Flag (left) or anti-HA (right) agarose on HEK293T cell lysates exogenously co-expressing HA-KLHL22 and the COLEC10 truncation mutants. D The expression level of Flag-tagged COLEC10 was reduced with increasing amounts of KLHL22 in HEK293T (left) and PLC/PRF/5 (right) cells. E The expression level of Flag-tagged COLEC10 was increased after silencing KLHL22 in PLC/PRF/5 cells. F KLHL22 stimulated the monoubiquitination of COLEC10 in vivo. HEK293T cells overexpressing COLEC10-Flag were transfected with the indicated plasmids and incubated for 24 h. Cell lysates were subjected to immunoprecipitation with anti-Flag agarose. Immunoblotting analysis was conducted for the indicated proteins. G KLHL22 reduced the accumulation of exogenously expressed COLEC10, but this did not occur in the presence of the lysosome inhibitor chloroquine. COLEC10 overexpression PLC/PRF/5 cells were transfected with the indicated plasmids and incubated for 24 h. The cells were then incubated in fresh medium with none, 10 µM MG-132, or 50 µM chloroquine for additional 12 h. The cell lysates were analyzed by immunoblotting using the indicated antibody. Full-length blots are presented in Supplementary Figure S9. LC/MS liquid chromatography-mass spectrometry, CLR cross-linking region, CRD carbohydrate recognition domain, CQ chloroquine

Furthermore, we found that KLHL22 was upregulated in HCC compared to non-tumor tissues in TCGA and GSE25097 HCC cohorts (Supplementary Fig. S5A). HCC patients with high KLHL22 expression suffered poor disease-free survival and KLHL22 positively correlated with mRNAsi in the TCGA-HCC cohort (Supplementary Fig. S5B-C). These results suggest that KLHL22 may function upstream of COLEC10 and counteract its tumor-suppressive effect in HCC.

Discussion

Previous work from our team has confirmed that COLEC10 was downregulated and associated with poor prognosis in HCC patients. Here, we provide another mechanism for how COLEC10 inhibits HCC recurrence by regulating cell stemness for the first time. Mechanically, COLEC10 overexpression inhibits β-catenin signaling by promoting the degradation of β-catenin and upregulating Wnt inhibitory factor WIF1 (Fig. 7).

Fig. 7.

Schematic diagram for how COLEC10 inhibits the stemness of HCC cells

The recurrence of HCC after hepatic resection is a weighty obstacle, with a 70% recurrence rate at 5 years, even in HCC patients with a single tumor ≤ 2 cm [19]. The molecular features of HCC vary a lot. Although the past decades have witnessed our improvement in understanding the pathophysiology and ways of treatment of the disease, the prognosis of HCC remains unsatisfactory. Currently reported markers for LCSCs include CD13, CD24, CD44, CD47, CD90, CD133, ICAM1, EpCAM, LGR5, OV6 and et al. [2]. We revealed that COLEC10 expression was negatively correlated with AFP and EpCAM both in clinical specimens and cell experiments. EpCAM⁺ HCC exhibited a distinct molecular signature similar to hepatic progenitor cells, including the presence of stem markers (e.g., c-Kit, CK19, and AFP) and the activation of Wnt/β-catenin signaling [17]. Notably, patients with the EpCAM⁺AFP⁺ HCC subtype exhibit a young age of onset, aggressive tumor growth, and poor prognosis [20]. However, since normal liver stem cells like hepatic progenitor cells also highly express EpCAM and AFP, targeting these markers directly might inhibit hepatic regeneration and lead to liver failure. COLEC10 is highly expressed in normal liver cells but low in HCC, which may serve as an alternative factor for targeting the EpCAM⁺AFP⁺ HCC subtype appropriately.

Recently, a genome-wide association study (GWAS) identified WNT3A-WNT9A as a susceptibility locus, which suggests that the Wnt/β-catenin pathway plays an early role in hepatocellular carcinogenesis [21]. It is reported that β-catenin was strongly expressed in human HCC tissue samples and LCSCs, resulting in poor prognosis [22]. Consistent evidence supports that aberrant activation of β-catenin signaling is oncogenic and functions as an essential driver in HCC stemness [23]. Several therapeutic approaches target the Wnt/β-catenin pathway in HCC, such as Wnt ligand inhibitors, stabilizing the destruction complex, and Wnt/TCF inhibitors. However, due to the safety and depressor effect defects, no clinically approved drugs have targeted this pathway efficiently so far [24]. Immunohistochemistry of our HCC tissues showed that β-catenin was upregulated in the low COLEC10 expression group, which suggests that COLEC10 negatively correlates with the β-catenin pathway. We found that COLEC10 inhibited the Wnt/β-catenin pathway by possibly promoting the combination of β-catenin and the destruction complex. LiCl activates the β-catenin pathway by inhibiting the action of GSK3β which phosphorylates β-catenin after CK1α [25]. Our results show that COLEC10 can inhibit the activation of β-catenin signaling induced by LiCl and overexpressing β-catenin can only partially reverse the effect caused by COLEC10, suggesting there are other ways for COLEC10 to inhibit β-catenin besides CK1α. According to current research findings, Wnt inhibitory protein-1 (WIF1) plays a crucial role by binding to and sequestering soluble Wnt ligands. This action effectively prevents their interaction and binding to FZD receptors, consequently suppressing Wnt/β-catenin signaling [26, 27]. RNA microarray analysis has revealed downregulation of WIF-1 in various cancers, including prostate, breast, lung, and bladder tumors [28]. Interestingly, our observations indicate that WIF1 expression increases with elevated level of COLEC10 in HCC cell lines and tissues. This suggests that COLEC10 may downregulate β-catenin signaling to inhibit cancer stemness through multiple mechanisms, with the upregulation of WIF1 being one of them. In sum, our study implies a promising role of COLEC10 in the treatment of HCC patients with aberrant activation of β-catenin pathway.

The expression of COLEC10 in HCC is pretty low both in mRNA and protein levels. We found that KLHL22 combined with COLEC10 and promoted its mono-ubiquitination and degradation for the first time, which might partly explain the low protein level of COLEC10 in HCC. Interestingly, KLHL22, as a substrate-specific adapter of a BCR E3 ubiquitin ligase complex, promoted COLEC10 degradation via lysosome rather than proteasome. Physiologically, KLHL22 interacts with Polo-like kinase 1 (PLK1) to regulate chromosome alignment and PLK1 kinetochore localization but not PLK1 stability [29]. The role of KLHL22 in tumors remains elusive. On one hand, KLHL22 activated amino-acid-dependent mTORC1 signaling by degrading DEPDC5 to promote breast carcinogenesis [12]. On the other hand, KLHL22 degraded PD-1 in T cells and thus prevented excessive T cell suppression in colorectal cancer (CRC), which inhibits CRC development [30]. In our research, the mRNA expression level of KLHL22 was upregulated and positively correlated with the stem index, suggesting its oncogenic role in HCC. Besides, in accordance with Beck J et al [13], we confirmed that KLHL22 interacted with CK1α in the HEK293T cell line (Supplementary Fig. S6), but whether COLEC10 competitively binds KLHL22 to release CK1α and then promotes β-catenin degradation requires further investigation.

The low expression level of COLEC10 has been associated with a poor prognosis in HCC, although its precise role in diagnosis, treatment, and prognosis prediction remains unclear. Notably, COLEC10 protein has been detected in both cytosolic fractions and extracellular space [31, 32]. We measured the COLEC10 level in the plasma of HCC patients (data not shown), which revealed a decrease in COLEC10 level in the late stage of HCC patients. This indicates that peripheral blood COLEC10 levels may serve as an assistant biomarker for diagnosing HCC and monitoring disease progression. Moving forward, further research will focus on investigating whether secreted COLEC10 protein exerts therapeutic effects on HCC and aids in suppressing recurrence. If proven effective, administering purified COLEC10 protein to HCC patients could represent a promising and innovative therapeutic strategy.

Conclusion

In summary, our study reveals COLEC10 as a tumor suppressor that inhibits CSC properties in HCC. These findings underscore COLEC10 as a promising target for HCC treatment strategies.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Y.-F.L., Y.-H.H. and M.-N.C. conceived the whole project. M.-N.C., D.-M.C., and X.-R.C. completed most of the experimental work. C.-H.L. and L.-X.X. helped conduct the immunohistochemistry assays and the animal experiments. J.-L.W. helped collect patient samples and information. Y.-F.L., B.-L.L. and M.-N.C. analyzed the data. Y.-F.L. and M.-N.C. wrote the manuscript. Y.-F.L., Y.-H.H., and B.-L.L. supervised the project and provided funding. All authors have read and approved the final manuscript.

Funding

This work was supported by the Natural Science Foundation of Guangdong Province (grant number: 2024A1515013208 and 2021A1515010306), National Natural Science Foundation of China (grant numbers: 81902397 and 82070612), Major Talent Training Project of the Third Affiliated Hospital of Sun Yat-sen University (granted to Y.-F.L.), Research and Development Planned Project in Key Areas of Guangdong Province (grant number: 2019B110233002), General Planned Project of Guangzhou Science and Technology (grant numbers: 202201010950 and 202201020422), and Guangzhou Science and Technology Program Key Projects (grant numbers: 2023B03J0154 and 2023B01J1007).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The study was approved by the Institute Research Ethics Committee at the Third Affiliated Hospital of Sun Yat-sen University. Written informed consent was obtained from each patient. (1) Title of the approved project: COLEC10 inhibiting hepatocellular carcinoma by suppressing β-catenin pathway; (2) Approval number: A2023-423-01; (3) Date of approval: 19/06/2023.

Competing interests

The authors declare no competing interests.