FAM83D facilitates EMT and metastasis of cervical cancer via interaction with GSK3β and inactivation of GSK3β/stabilization of Snail signaling

Abstract

FAM83D plays a crucial role in cervical cancer (CC) metastasis, though its exact mechanism remains unclear. This study found that FAM83D was significantly upregulated in CC tissues, particularly in metastatic cases, and correlated with advanced tumor stage and poor prognosis. While it did not affect CC cell proliferation, chemosensitivity, or tumorigenesis, FAM83D overexpression enhanced migration and invasion, whereas its knockdown suppressed these effects. In vivo, silencing FAM83D markedly reduced lung and liver metastases in mice. Mechanistically, FAM83D upregulated EMT-like changes and Snail stability by directly interacting with GSK3β and promoting its Ser9 phosphorylation, thereby activating GSK3β/Snail signaling. Furthermore, AKT and PKA served as key upstream kinases in FAM83D-mediated GSK3β inactivation. Inhibition of GSK3β reversed the anti-metastatic effects of FAM83D knockdown. These results identify FAM83D as a key metastasis driver in CC, functioning through GSK3β/Snail/EMT axis, and suggest its potential as a therapeutic target to inhibit CC progression.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13062-026-00761-z.

Introduction

Cervical cancer (CC) is a common gynecological malignancy that ranks fourth in both morbidity and mortality among women worldwide. According to 2022 statistics, 661,021 women were diagnosed with CC, and 348,189 women died from CC worldwide [1]. Several studies have shown that the poor prognoses of patients with CC, such as low survival and high recurrence rates, are related to disease stage, tumor grade, lymph node metastasis, and the depth of invasion into cervical stroma [2–4]. Even though effective therapies have been established, the prognosis of patients with locally advanced CC remains poor, due to distant metastases and recurrent disease [5, 6]. The mechanisms of metastasis are highly complex biological processes, which have not been fully elucidated [7–9]. In this context, ongoing research efforts have leveraged bioinformatic and multi-omics approaches to identify key molecular drivers and potential biomarkers. For instance, recent studies have successfully identified and validated key epigenetic modification-related genes and prognostic biomarkers such as ZNF695, which correlate with tumor progression and immune infiltration in cervical squamous cell carcinoma and endocervical adenocarcinoma [10, 11]. Therefore, it is urgent to elucidate the underlying mechanism and take appropriate interventions to provide a new insight into improving CC treatment.

Previous studies have shown that aberrant activation of EMT plays an important role in cancer progression. EMT is a complex process in which epithelial cells undergo a transformation to acquire a mesenchymal phenotype, thereby conferring metastatic properties to tumor cells [12]. This transition facilitates the invasion of tumor cells into adjacent tissues and subsequent colonization of distant organs [13]. Therefore, EMT may be an attractive target to halt tumor progression and metastasis. The occurrence of EMT in cancer cells is often accompanied by changes in a range of cellular markers, including the loss of expression or function of epithelial genes (e.g., E-cadherin and occlusion zone 1) and elevated expression of genes defining mesenchymal phenotype (e.g., N-cadherin, Vimentin, fibronectin) [14, 15]. Of these, the down-regulation of E-cadherin expression is an intrinsic hallmark of EMT. EMT is regulated by a wide range of transcription factors, including Snail, ZEB, and Twist families. It has been reported that EMT transcription factors are involved in the regulation of tumor invasion and metastasis [16]. At the same time, EMT can be activated by a variety of signaling, including Wnt, TGF-β/Smads, and PI3K/AKT [17–19]. Ectopic activation of EMT-related signaling endows cancer cell metastatic potential [20, 21]. It has been demonstrated that primary CCs with an EMT phenotype are positively correlated with tumor progression, invasion, and metastasis [22]. Therefore, it is critical to investigate the underlying mechanism of EMT regulation during tumor metastasis for clinical diagnosis and treatment.

Recently, multi-omics analyses suggest that FAM83D (family with sequence similarity 83, member D) functions as a novel oncogene in a variety of human tumors [23–25]. Currently, there have been studies showing that FAM83D may mediate the involvement of EMT in hepatocarcinogenesis and metastasis [26]. In addition, FAM83D promotes pancreatic cancer tumorigenesis and gemcitabine resistance through Wnt/β-catenin signaling, as well as increases cell proliferation and invasion in glioma through activating the AKT/mTOR pathway [27, 28]. Further study has also found that upregulation of FAM83D by m⁶A methyltransferase METTL3 activates Wnt/β-catenin to accelerate cell proliferation and motility [29].

However, the biological function of FAM83D remains unknown in human CC. In this study, we found that FAM83D was overexpressed in CC tissues and is associated with metastasis and poor disease-free survival (DFS) in CC patients. Moreover, we elucidated the underlying mechanisms by which FAM83D directly interacted with GSK3β and augmented GSK3β/Snail signaling to promote EMT and metastatic potential in CC.

Materials and methods

Bioinformatics analysis

The IntACT (https://www.ebi.ac.uk/intact/home) database [30] was employed to analyze networks of proteins involved in interactions with FAM83D. The expression of FAM83D in CC tissues was analyzed based on the data from UALCAN (https://ualcan.path.uab.edu/index.html) [31], TNMplot (https://tnmplot.com/analysis/) [32], and TCGA (https://www.cancer.gov/tcga) databases.

Cell culture

Human CC cell lines SiHa and HeLa were purchased from the American Type Culture Collection (ATCC). Cells were authenticated by STR profiling and cultured in MEM (Procell, CHINA) medium supplemented with 10% (v/v) Fetal Bovine Serum (FBS) (Cyto Factory, CHINA) and 1% penicillin/streptomycin (Solarbio, CHINA). The cells were maintained in a humidified atmosphere of 5% CO₂ at 37 °C.

Lentiviral vector

To construct stable cell lines for overexpression and knockdown, we used lentiviral (Genechem, CHINA) LV-FAM83D, LV-shFAM83D, and their corresponding control lentiviral vectors to transfect the cells. We added the appropriate amount of virus based on the cell MOI and viral titer, then cultured the cells at 37 °C. After approximately 72 h, we assessed the infection efficiency. Subsequently, we continued to culture the cells in a medium containing a suitable concentration of puromycin (Beyotime, CHINA), screened them for expansion, and then collected cells for Western blotting identification. Cells with successful identification results were frozen for seed preservation and employed for examination of biological function.

Cell counting kit-8(CCK-8) assay

The effectiveness of chemotherapy drugs in killing tumor cells was evaluated using the Cycle Counting Kit-8 (CCK-8) assay (Biosharp, CHINA). When the cells reached about 80% confluence, they were placed in 96-well plates at a density of approximately 1 × 10⁴ cells per well and allowed to grow for 24 h. Subsequently, the cells were exposed to different concentrations of cisplatin, oxaliplatin, 5-fluorouracil (MedChem Express, Monmouth Junction, NJ, USA), and paclitaxel (Selleck Chemicals, Houston, TX, USA) for 48 h. Following the drug treatment, 100µL of medium containing CCK-8 reagent was added to each well, and the cells were incubated at 37 °C for about 30–60 min. The optical density (OD) of each well was measured at 450 nm using a microplate reader (Bio-Rad, Hercules, CA, USA).

Colony formation assay

In brief, cells were digested when confluence reached about 80% and inoculated into 6-well plates at a density of about 1 × 10³ cells per well. The cells were then incubated for 7–10 days. Subsequently, the cell clones were fixed using 4% paraformaldehyde (Beyotime, CHINA), stained with crystal violet (Beyotime, CHINA), and left to dry at room temperature. Finally, the cell clones were imaged and counted.

Western blotting analysis

Total protein was extracted from the cells using RIPA Strong Lysate (Beyotime, CHINA). The protein concentration was determined using the BCA Protein Assay Kit (Beyotime, CHINA). Protein (30 µg) was separated on a 10% SDS-PAGE and then transferred to PVDF membrane (Bio-Rad, Hercules, CA, USA). The membranes were then blocked using 5% nonfat dry milk (Beyotime, CHINA) in Tris-buffered saline with 0.1% Tween 20 (TBST) for 60 min at room temperature. After blocking, the membranes were cut horizontally to examine multiple proteins of different sizes on each gel. The membrane was then incubated with the primary antibody overnight at 4 °C. Following this, the membrane was washed with TBST (minimum of 3 times for 10 min each) and then incubated with an HRP-coupled secondary antibody (1:5000, ZSGB-BIO, CHINA) for 2 h at room temperature. The membrane was again washed with TBST (a minimum of 3 times for 10 min each). The signals were detected using an enhanced chemiluminescence detection kit (Invitrogen, Carlsbad, CA, USA) and Bio-Rad Molecular Imager (Bio-Rad, Hercules, CA, USA). The primary antibodies were anti-FAM83D (1:1000, ab236882, Abcam, Cambridge, MA, USA), anti-β-catenin (1:1000, sc-7963, Santa Cruz, CA, USA), anti-GSK3β and anti-Phospho-GSK3β (Ser9) (1:1000, ET1607-71, ET1607-60, HUABIO, CHINA), anti-GAPDH (1:20000, ET1601-4, HUABIO, CHINA), and anti-AKT1 (1:1000, ET1609-47, HUABIO, CHINA). Primary antibodies against E-cadherin, N-cadherin, Vimentin, Snail, PKA, Phospho-PKA (Thr 197), and Phospho-AKT(Ser 473) were gained from Cell Signaling Technology (1:1000, 3195 S, 13116T, 5741T, 3879T, 4782, 4781, 4060, Cell Signaling Technology, Danvers, MA, USA). The relative expression of the proteins was normalized against GAPDH using Image J analysis.

Wound healing assay

After inoculating cells in a 6-well plate and incubating for 24 h, allowing them to grow until they formed a single layer, a straight scratch was made across the cell layer using a sterile 10-µL pipette tip. The cell debris was removed using serum-free medium, and the wounds were imaged under a microscope (0 h). The wounds were then imaged under a microscope after replacing the medium with a medium containing 0.5% FBS for 24, 48, and 72 h. The cell wound area was calculated using Image J.

Transwell migration and invasion assay

Cell migration capacity was evaluated using the transwell assay. Invasion assays required the upper chamber to be coated with Matrigel (BD Biosciences, San Jose, CA, USA) for at least 4 hours before injecting cells, and migration assays were not performed with Matrigel. HeLa cells (1 × 10⁴ cells) and SiHa cells (2 × 10⁴ cells) were resuspended separately in 200µL of serum-free medium in the upper chamber (Corning Incorporated, Corning, NY, USA) and 500µL of complete medium containing 20% FBS in the lower chamber. Cells were incubated at 37 °C in a 5% CO2 incubator for 48 h. After washing with PBS, cells were fixed with 4% paraformaldehyde and stained with crystal violet. An orthogonal microscope was used to take pictures and count the number of cells.

Tumor xenograft and metastasis in experimental animals

Approval was granted by the Institutional Animal Care and Use Professional Committee of Chengdu Medical College. Female BALB/c nude mice, aged 5 weeks, were purchased from Sichuan Charles River Laboratories in China and maintained in specific pathogen-free environments. The sample size was estimated based on previous studies [33, 34]. For the tumor metastasis model, cells in the logarithmic growth phase (3 × 10⁶ cells per mice) were harvested, and 100 µL of cell resuspension was injected into the lateral tail vein of nude mice. After 5 weeks, the mice were euthanized, and lung tissues were collected, photographed, and assessed for lung metastatic nodules following staining with hematoxylin and eosin (H&E). In the subcutaneous transplantation model, cells in the logarithmic growth phase (5 × 10⁶ cells per mouse) were utilized, and 200µL of cell resuspension was injected subcutaneously into the right shoulder of nude mice. Once the tumors became discernible, the mice’s body weight and tumor size were recorded every 2 days using calipers, with tumor volume (mm³) calculated using the formula: v = 0.52 × L (length) × W (width)². All mice were euthanized before reaching the ethical endpoint. Subsequently, both the subcutaneous tumors and major organs of the mice were harvested, with tumors weighed and volumetrically assessed. Liver metastases were identified through staining with hematoxylin and eosin (H&E).

Co-immunoprecipitation (Co-IP)

Co-IP was performed to confirm protein-protein interaction. Cells were lysed in RIPA lysis buffer (Beyotime, CHINA), and the supernatants were centrifuged at 12,000 g for 20 min. Following this, the proteins were incubated overnight at 4 °C with anti-FAM83D (ab236882, Abcam, Cambridge, MA, USA), anti-GSK3β (ET1607-71, HUABIO, CHINA), anti-Phospho-GSK3β (Ser9) (ET1607-60, HUABIO, CHINA), anti-Snail (3879T, Cell Signaling Technology, Danvers, MA, USA) or IgG antibodies of the same species (at a ratio of 1 µg antibody to 500 µg cellular protein). Subsequently, 20–40µL of properly resuspended Protein A + G Agarose was added, and the mixture was incubated at 4 °C for 4 h. The proteins were then thoroughly washed with RIPA lysis buffer. Finally, 20–40µL of 1×SDS loading buffer was added to resuspend the sediment, which was then boiled for 10 min. The samples were then subjected to Western blotting analysis.

Ubiquitination assay

HA-Ubiquitin plasmid was transfected into CC cells with FAM83D overexpression or knockdown, respectively. The cells were pretreated with LiCl (15 µM) for 24 h, followed by treatment with MG132 (10 µM) for 6 h prior to protein extraction, as indicated. After treatment, cells were lysed using RIPA buffer. Cell lysates were subjected to immunoprecipitation with an anti‑Snail antibody, followed by Western blotting analysis with an anti‑Ubiquitin antibody (10201-2-AP, Proteintech, CHINA) to evaluate changes of Snail protein ubiquitination.

In vitro binding assay

In vitro cell-free proteins binding assay was performed using TnT^® Quick Coupled in vitro transcription/translation system (L11700, Promega, Madison, USA) according to the manufacturer’s instructions. FAM83D and GSK3β proteins were expressed in vitro using the cell-free TnT^® system in two separate reactions. Each reaction (50 µL) contained the TnT^® mix supplemented with 0.5 µg of plasmid DNA and 2 µL of methionine. Following incubation at 30 °C for 90 min, 4 µL aliquots of each reaction were analyzed by Western blotting using anti-FAM83D and anti-GSK3β antibodies to confirm protein expression. Subsequently, 40 µL aliquots of the FAM83D and GSK3β reactions were mixed and incubated on a rotating platform at 4 °C for 4 h. The interaction between FAM83D and GSK3β was then assessed by Co-IP followed by Western blotting with the corresponding antibodies.

Immunohistochemistry (IHC)

Tissue microarray HUteS120Su01 was obtained from Shanghai Outdo Biotechnology Co., LTD. The procedure began with dewaxing and rehydrating the paraffin sections. The sections were then immersed in sodium citrate buffer and heated at high temperature in a microwave for 10 min. After natural cooling, the sections were treated with a 3% hydrogen peroxide solution for 20 min to block endogenous peroxidase activity. Afterward, goat serum (ZSGB-BIO, CHINA) was incubated at 37 °C for 20 min to reduce background staining further. Next, the sections were incubated overnight at 4 °C with a primary antibody against FAM83D (1:50, bs-14678R, BIOSS, CHINA). The following day, a secondary antibody (ZSGB-BIO, CHINA) was added and incubated at 37 °C for 20 min. Finally, DAB (ZSGB-BIO, CHINA) color development was performed, with the staining duration monitored through microscopic observation. Tissues were photographed for each case under the microscope, and the staining intensity was evaluated and scored by a pathologist.

Statistical analysis

All experimental data were repeated three times, averaged, and expressed as mean ± standard deviation (SD). GraphPad Prism 8.0 software (GraphPad Software Inc., San Diego, CA, USA) was used for statistical analysis. The T-test was used to compare the means of two samples, and One-way ANOVA was used to compare the means of multiple groups. Kaplan-Meier was used to analyze the survival curves. P < 0.05 was deemed to indicate a statistically significant difference.

Result

FAM83D overexpression was associated with poor prognosis in CC

To investigate the role of FAM83D in CC, the expression of FAM83D was evaluated using the public clinical specimen database (UALCAN, TNMplot, and TCGA). As shown in Fig. 1A-B, the mRNA level of FAM83D was significantly increased in primary tumor tissues compared to normal tissues. In addition, the expressions of FAM83D in various subtypes of CC, including adenocarcinoma, squamous cell carcinoma, endocervical carcinoma, endometrioid carcinoma, and mucinous carcinoma were higher than that in normal tissues (Fig. 1C). Moreover, we found that the expression of FAM83D was significantly elevated in T3-T4 tissues in comparison to T1-T2 (Fig. 1D), and up-regulated in pathologic stage I-IV compared to normal tissues, while exhibiting a progressive increase from pathologic stage I to II-III (Fig. 1E-F). Then we also found that FAM83D was much higher expressed in samples with lymph node invasion than that without lymph node invasion (Fig. 1G-H). Furthermore, higher expression of FAM83D was associated with metastasis status in CC (Fig. 1I). To further validate the above results, we assessed the protein expression of FAM83D between CC tissues and adjacent normal tissues from clinical specimens using a tissue array HUteS20Su01 (Supplementary Table S1). The data demonstrated that the protein levels of FAM83D significantly increased in CC tissues (Fig. 1J-K). Notably, the overexpression of FAM83D protein was associated with poor disease-free survival prognosis (Fig. 1L).

Fig. 1.

FAM83D was highly expressed in CC tissues and was associated with poor patient prognosis. (A-C) Comparison of mRNA expression levels of FAM83D in CC tissues and normal tissues based on UALCAN and TCGA database. (D-F) Analysis of mRNA expression levels of FAM83D in relation to T-staging and Stage staging using TCGA and UALCAN databases (G-I). Evaluation of mRNA expression levels of FAM83D in metastatic CCs based on UALCAN, TNMplot, and TCGA databases. (J-K) Detection of protein expression levels of FAM83D in CC tissues and adjacent normal tissues by immunohistochemistry using HUteS20Su01, and the score of staining intensity. (L) Disease-free survival of CC patients with high or low FAM83D expression. *p < 0.05, **p < 0.01, ***p < 0.001,****p<0.0001

FAM83D was involved in cell metastasis in CC

Based on the above findings, we reasoned that FAM83D potentially played a role in the regulation of malignancy progression in CC. We next established stable cell lines with overexpression or knockdown of FAM83D in HeLa and SiHa cells, which were confirmed by Western blotting for protein expression of FAM83D (Fig. 2A). To assess its role in cell proliferation, we performed colony formation assays using these cell lines. The results showed that neither overexpression nor knockdown of FAM83D significantly affected the colony-forming ability of CC cells (Fig. 2B–C). Moreover, altered FAM83D expression did not influence cellular chemosensitivity to 5-fluorouracil, cisplatin, oxaliplatin, or paclitaxel (Fig. 2D). In contrast, FAM83D significantly promoted CC metastasis in vivo. Knockdown of FAM83D markedly reduced the incidence of liver metastasis without substantially affecting the volume or weight of subcutaneous xenografts (Fig. 2E–J). Histological analysis further revealed extensive tumor foci in the livers of control mice, whereas liver tissues from FAM83D-knockdown mice remained largely normal (Fig. 2K–L). Taken together, these data demonstrated that FAM83D specifically enhanced the metastatic potential of CC cells without affecting their proliferation or chemosensitivity.

Fig. 2.

FAM83D was associated with metastasis of CC. (A) Western blotting was employed to examine the expression of FAM83D in HeLa and SiHa cells with stable overexpression or knockdown of FAM83D. (B-C) Colony formation assay was conducted to evaluate the clonogenic ability of HeLa and SiHa cells following the overexpression or knockdown of FAM83D. (D) CCK-8 assay was used to detect cell viability in FAM83D-silenced HeLa and SiHa cells treated with different concentrations of 5-fluorouracil, cisplatin, oxaliplatin, and paclitaxel for 48 h. The inhibition rate was subsequently calculated. (E-I) Subcutaneous injection of FAM83D-silenced and control HeLa cells into BALB/c Nude mice (n = 6) was used to determine the tumorigenicity. The tumor growth curves, tumor volumes, and tumor weights were measured. (J) And then, the nude mice that developed liver metastases were counted. (K) Representative images of liver metastases developed from FAM83D-silenced and control HeLa cells were shown, while (L) H&E staining of liver metastases in nude mice was presented. *p < 0.05, **p < 0.01, ***p < 0.001

FAM83D enhanced the metastasis potential of CC cells in vitro and in vivo

To further verify the role of FAM83D in CC metastasis, we examined the cell migration and invasion in vitro and in vivo. As shown in Fig. 3A-D, the wound healing assay showed that overexpression of FAM83D significantly promoted cell migration, whereas knockdown of FAM83D suppressed the migration ability in HeLa and SiHa cells. Moreover, the results from the transwell assay showed that FAM83D positively regulated the cell migration and invasion of CC cells (Fig. 3E-H). To further investigate whether FAM83D promoted cell metastasis in vivo, CC cells with or without knockdown of FAM83D were injected into nude mice via the tail vein. The results showed that all nude mice in the control group developed lung metastasis, while four mice with knockdown of FAM83D developed lung metastasis (Fig. 3I-J). Furthermore, the enumeration of lung nodules from these mice revealed a significant reduction in nodule count in the FAM83D knockdown group compared with control mice (Fig. 3K). Histology study by H&E staining also demonstrated that the area of tumor tissues was decreased in the FAM83D knockdown group (Fig. 3L). The above data demonstrated that FAM83D accelerated CC metastasis in vitro and vivo.

Fig. 3.

FAM83D promoted the metastasis of CC cells in vitro and vivo. (A-D) Wound healing assay was conducted to determine the migratory ability, and (E-H) transwell assay was utilized to determine the migratory and invasive ability, respectively, in HeLa and SiHa cells exhibiting stable overexpression or silencing of FAM83D. (I) Lung metastasis in nude mice (n = 6) by tail vein injection of HeLa cells with or without silencing of FAM83D. (J-K) The statistical data of lung metastasis occurring in the mouse model. (L) Representative images and H&E staining of lung metastasis in nude mice. *p < 0.05, **p < 0.01, ***p < 0.001

FAM83D enhanced GSK3β/Snail signaling and EMT in CC

It has been reported that EMT may be a driver of tumor aggressiveness and metastatic phenotype [35]. Therefore, we investigated the effect of FAM83D on EMT markers. Our results showed that epithelial marker (E-cadherin) was down-regulated, while mesenchymal markers (N-cadherin, Vimentin) were up-regulated by overexpression of FAM83D in HeLa and SiHa cells compared to control cells. As expected, knockdown of FAM83D exerted an opposite effect (Fig. 4A-D). Moreover, Snail, a key transcription factor of EMT [36, 37], was significantly overexpressed in CC cells with FAM83D ectopic expression, while decreased by FAM83D knockdown (Fig. 4E-H). Studies have shown that Wnt/β-catenin signaling is critical for EMT during tumor metastasis [18]. However, our data showed that FAM83D had no significant effect on the expression of β-catenin. Interestingly, FAM83D positively regulated GSK3β and p-GSK3β (Ser9) (Fig. 4E-H). Considering that p-GSK3β (Ser9) can directly stabilize Snail protein expression [38], we further detected whether FAM83D modulated the ubiquitination and degradation of Snail. In CC cells treated with cycloheximide (CHX), FAM83D overexpression delayed Snail degradation (Fig. 4I-J). Consistently, FAM83D overexpression reduced Snail ubiquitination in CC cells, while FAM83D knockdown enhanced it. Moreover, treatment with LiCl, known as an inhibitor of GSK3β [39], reversed ubiquitination and downregulation of Snail mediated by FAM83D knockdown (Fig. 4K-L), indicating that FAM83D inhibited ubiquitination and degradation of Snail, and that GSK3β inactivation was critical for FAM83D-improved Snail stability. Collectively, the above results suggested that FAM83D enhanced GSK3β/Snail signaling axis and EMT in CC.

Fig. 4.

FAM83D promoted GSK3β/Snail signaling and EMT. (A-D) Western blotting for E-cadherin, N-cadherin, Vimentin, and FAM83D in HeLa and SiHa cells following stable overexpression or silencing of FAM83D. (E-H) Western blotting for Snail, β-catenin, GSK3β, and phosphorylated GSK3β (Ser9), and FAM83D protein changes in CC cells with stable overexpression or silencing of FAM83D. (I-J) Cycloheximide (CHX) chase assays for Snail stability in CC cells with stable overexpression of FAM83D. (K-L) Ubiquitination assay for Snail protein ubiquitination level in CC cells with stable overexpression or silencing of FAM83D in the presence or absence of LiCl. *p < 0.05, **p < 0.01, ***p < 0.001

FAM83D directly interacted with GSK3β

To elucidate the molecular mechanism underlying FAM83D-mediated regulation of the GSKβ/Snail signaling axis, we performed a comprehensive screening for potential FAM83D-binding proteins using the Protein Interaction Database (IntAct). The results revealed a potential interaction between FAM83D and GSK3β, as annotated in the IntAct resource. (Fig. 5A), which was further validated by Co-IP assay showing that endogenous FAM83D and GSK3β could form a complex in HeLa and SiHa cells (Fig. 5B). At the same time, the interaction between the two proteins was attenuated by FAM83D knockdown (Fig. 5C-D). Moreover, in both cell lines, Co-IP results showed that FAM83D also interacted with phosphorylated GSK3β (Ser9) (Fig. 5E-F). To validate the interaction specificity between FAM83D and GSK3β, in vitro cell-free proteins binding assay was performed using TnT^® Quick Coupled in vitro transcription/translation system according to the manufacturer’s instructions. Our results revealed that single FAM83D and GSK3β proteins co-immunoprecipitated in the in vitro cell-free expression system. This suggested that FAM83D directly bound to GSK3β in vitro (Fig. 5G). Therefore, the interaction between FAM83D and GSK3β is specific. Taken together, we speculated that FAM83D directly bound to GSK3β, inducing its inactivation, which in turn resulted in promotion of Snail expression and EMT process.

Fig. 5.

FAM83D interacted with GSK3β. (A) The possible interaction between FAM83D and GSK3β was screened from the IntACT database. (B-D) Co-IP assay for interaction between FAM83D and GSK3β in HeLa and SiHa cells with or without silencing of FAM83D. (E-F) Co-IP assay for interaction between FAM83D and phosphorylated GSK3β (Ser9) in HeLa and SiHa cells. (G) In vitro cell-free proteins binding assay for validation of interaction specificity between FAM83D and GSK3β. *p < 0.05, **p < 0.01, ***p < 0.001

AKT and PKA served as key upstream kinases in FAM83D-mediated GSK3β inactivation

It has been reported that kinases such as AKT and PKA can phosphorylate GSK3β at Ser 9 and inhibit the activity of GSK3β [40, 41]. We detected the impact of FAM83D on activities of AKT and PKA, and further identified their contributions to FAM83D-induced GSK3β Ser9 phosphorylation. Acquired data showed that FAM83D overexpression significantly enhanced phosphorylation of AKT (S473) and PKA (Thr197) in HeLa and SiHa cells. Moreover, the AKT inhibitor MK2206 or PKA inhibitor H-89 also counteracted FAM83D-induced phosphorylation of GSK3β (Ser9), respectively (Fig. 6A-D). These results indicated that AKT and PKA were key upstream kinases in FAM83D-mediated GSK3β inactivation.

Fig. 6.

FAM83D promoted GSK3β phosphorylation by activating AKT and PKA. (A-B) Western blotting for expressions of AKT, p-AKT (Ser473), GSK3β, p-GSK3β (Ser9), and FAM83D in CC cells with stable FAM83D overexpression, as well as treated with MK2206 (0.5 µM) for 48 h. (C-D) Western blotting for expressions of PKA, p-PKA (Thr 197), GSK3β, p-GSK3β (Ser9), and FAM83D in CC cells with stable FAM83D overexpression, as well as treated with H-89 (10 µM) for 24 h. *p < 0.05, **p < 0.01, ***p < 0.001

GSK3β inactivation was crucial for FAM83D-induced EMT and metastasis

To further elucidate whether FAM83D regulated EMT and metastasis via inactivation of GSK3β, we treated HeLa and SiHa cells with a GSK3β inhibitor, LiCl. As shown in Fig. 7A-D, LiCl reversed shFAM83D-increased protein expression of E-cadherin, while restoring shFAM83D-reduced protein expressions of N-cadherin, Vimentin, Snail, GSK3β, and p-GSK3β. In the meantime, LiCl reversed the downregulation of migration and invasion abilities induced by FAM83D knockdown (Fig. 7E-H). These results suggested that FAM83D promoted EMT and metastasis via inactivation of GSK3β.

Fig. 7.

The role of GSK3β/Snail axis in FAM83D-regulated EMT and metastasis of CC. (A-D) Western blotting for expressions of E-cadherin, N-cadherin, Vimentin, Snail, GSK3β, p-GSK3β (Ser9), and FAM83D in CC cells following stable FAM83D knockdown, as well as treated with LiCl (10 mM and 15 mM in HeLa and SiHa cells, respectively) for 24 h. (E-H) Following the treatment of HeLa cells with LiCl (10 mM) for 24 h and SiHa cells with LiCl (15 mM) for 24 h, a transwell assay was conducted to assess the migratory and invasive capabilities of HeLa and SiHa cells with stable FAM83D knockdown. The counts of migratory and invasive cells were recorded and subjected to statistical analysis. *p < 0.05, **p < 0.01, ***p < 0.001

Discussion

Cervical cancer (CC) remains one of the leading causes of cancer-related deaths among women worldwide and poses a significant threat to women’s health on a global scale [1]. Although early-stage CC can be effectively addressed with surgical resection and adjuvant chemoradiotherapy, CC typically presents no symptoms in its early stage [42–45]. In the present study, the expression of FAM83D was significantly elevated in early-stage CC tumor tissues, suggesting that FAM83D is a potential biomarker for identifying the initial state of CC. Metastasis represents the primary cause of mortality in patients with advanced-stage CC, accounting for the majority of cancer-related deaths in this population [5]. Unfortunately, 37% CC patients are diagnosed as advanced stage [46]. Importantly, the underlying mechanism regulating CC metastasis remains poorly understood. In our study, FAM83D was identified as a key driving force to induce cell migration, invasion, and metastasis of CC cells in vitro and vivo. Clinically, elevated FAM83D expression was significantly involved in the poor prognosis of CC patients. Knockdown of FAM83D dramatically inhibited the metastatic potential of CC cells, as evidenced by markedly reduced tumor colonization in both liver and lung tissues of the nude mice model. These findings suggested that expression of FAM83D was crucial for CC metastasis.

As documented in the literature, FAM83D, a member of the FAM83 family with established links to tumorigenesis, was originally identified as a gene encoding a spindle-associated protein that plays an essential role in mitotic regulation and cell cycle progression [25, 27]. Emerging evidence demonstrates that FAM83D not only drives tumor cell proliferation in various cancer types, including breast, glioblastoma, and pancreatic adenocarcinoma [23, 27, 47], but also increases metastatic potential in non-small cell lung cancer, colorectal cancer, and ovarian cancer [48–50]. However, our current investigation revealed a non-canonical biological function of FAM83D in CC, where it specifically promoted cell migration, invasion, and metastatic dissemination without affecting cellular proliferation or chemotherapeutic resistance in primary CC cells. This tissue-specific functional profile suggests that FAM83D may exert context-dependent oncogenic functions that vary according to the tumor origin.

EMT is a fundamental biological process that plays critical roles in normal embryonic development and tissue regeneration [51]. However, the aberrant reactivation of EMT is orchestrated with several malignant properties during tumor progression and metastasis, particularly in enhancing cellular migration, invasiveness, and cancer stemness [13]. Notably, EMT has arisen as a crucial regulator of metastasis and recurrence across various cancer types by promoting tumor cell motility and invasion [19, 52]. Therefore, to elucidate the regulatory mechanism of FAM83D during cervical cancer metastasis, EMT will serve as an important breakthrough point. EMT is tightly regulated by multiple transcription factors, including Snail, Slug, Twist1/2, and Zeb1/2 [53]. Importantly, Snail is almost involved in the entire process of driving EMT. Experimental evidence has demonstrated that silencing Snail effectively inhibits EMT transformation and EMT-induced tumor metastasis characteristics [54]. Here, our study verified that FAM83D increased EMT key transcription factor Snail expression and promoted EMT-like molecular changes, including downregulation of E-cadherin as well as upregulation of N-cadherin and Vimentin. This suggested that FAM83D enhanced EMT transformation and metastasis of CC cells.

It is well known that ectopic activation of Wnt/β-catenin signaling is vital for EMT progression [55]. GSK3β plays a significant role in Wnt/β-catenin activity. The phosphorylation of N terminal serine-9 residue of GSK-3β severs to auto-inhibit its activity [43]. The inhibition of GSK3β can suppress the destruction complex formation of β-catenin to increase β-catenin accumulation, resulting in activation of Wnt/β-catenin and expression of EMT transcription factor including Snail [56–58]. It has been reported that Wnt/β-catenin signaling is aberrantly activated in CC [59]. Therapeutic strategies targeting Wnt pathway are being explored in pre-clinical or clinical trials. Surprisingly, in this study, FAM83D did not significantly regulate β-catenin accumulation, although it obviously suppressed GSK3β activity by increasing p-GSK3β (Ser9). These data indicated that FAM83D-induced Snail expression was independent of Wnt/β-catenin signaling. Previous studies have confirmed that GSK3β activity directly affects the posttranslational regulation of Snail stability. The inhibition of GSK3β via Ser9 phosphorylation contributes to the non-phosphorylation state and protein stability of Snail [60–63]. Extending these findings, our study demonstrated that FAM83D decreased the ubiquitination and degradation of Snail protein, while pharmacological inhibition of GSK3β impaired shFAM83D-induced Snail ubiquitination. Considering that FAM83D reinforced the protein levels of p-GSK3β (Ser9) and Snail, the above data suggested that FAM83D inhibited GSK3β to stabilize Snail independent of Wnt/β-catenin signaling.

As inhibition of GSK3β is a crucial factor for tumor proliferation, metastasis, and therapy resistance, reactivation of GSK3β presents promising therapeutic potential in clinical applications [64]. Here, we further explored how FAM83D inhibited GSK3β activity. It is well known that protein-protein interaction is an important way to exert its biological function [65]. Previous studies have confirmed that GSK3β activity is controlled by its interacting proteins [66, 67]. We checked whether FAM83D bound to GSK3β. Interestingly, the physical association between FAM83D and GSK3β was found in the IntAct database. Co-IP assay also showed that FAM83D interacted with GSK3β and p-GSK3β (Ser9). Silencing FAM83D impaired co-immunoprecipitation between FAM83D and GSK3β. Notably, in vitro cell-free proteins binding assay confirmed the interaction specificity between FAM83D and GSK3β. Therefore, we supposed that FAM83D directly interacted with GSK3β and promoted the Ser9 phosphorylation to inhibit GSK3β activity.

Previous studies have shown that GSK3β activity is negatively regulated by upstream kinases such as AKT and PKA through phosphorylation [40, 68]. In line with this, we found that FAM83D promoted the phosphorylation of AKT at Ser473 and PKA at Thr197, thereby activating these two kinases. Once activated, they can phosphorylate GSK3β at Ser9, leading to its functional inactivation [69]. As expected, the inhibitor of AKT or PKA decreased FAM83D-induced GSK3β (Ser9) phosphorylation, indicating that AKT and PKA served as key upstream kinases in FAM83D-mediated GSK3β inactivation. Additionally, given that FAM83D itself directly binds to GSK3β, this suggests that FAM83D might act as an allosteric regulator or a scaffold to enhance the phosphorylation efficiency of AKT and PKA at GSK3β Ser9, or directly interfere with GSK3β activity. Functionally, treatment with LiCl, an established GSK3β inhibitor [70], effectively reversed shFAM83D-inhibited Snail expression, EMT, and cellular metastasis. This finding reinforced that GSK3β inhibition was central to the oncogenic function of FAM83D. Collectively, the above data suggested that FAM83D enhanced EMT and metastasis via interaction with GSK3β and activation of GSK3β/Snail signaling, which provided a new strategy for successful CC treatment.

Nevertheless, several questions remain to be addressed. For instance, it will be important to determine whether FAM83D exerts its pro-metastatic effects via crosstalk with other signaling pathways. Moreover, the present study is limited by its reliance on two HPV-positive cases, which may not fully represent the heterogeneity of CC, particularly HPV-negative cases or other histological subtypes. Therefore, future studies are recommended to incorporate a broader panel of experimental models—including HPV-negative CC cell lines, additional HPV-positive cell lines representing diverse histological subtypes, and ideally patient-derived xenografts or organoid models—to validate and extend the conclusions drawn here, thereby enhancing the translational relevance of the research.

In conclusion, the present study unveiled a novel function and mechanistic insight into the role of FAM83D in CC. We demonstrated that FAM83D was overexpressed in primary CC tissues and positively related to poor patient prognosis. Functionally, FAM83D promoted the metastatic potential of CC cells in vitro and in vivo. Mechanistically, FAM83D directly interacted with and inhibited GSK3β, thereby enhancing GSK3β/Snail signaling and EMT in CC (Fig. 8). Our research provided that FAM83D/GSK3β/Snail axis-induced EMT should be a potential therapy target to overcome metastasis in CC cells.

Fig. 8.

Schematic of the mechanism by which FAM83D promoted EMT and metastasis of cervical cancer via GSK3β

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Data collecting: J.F.L, F.T.L, X.M. and Z.L.Y. Writing: J.F.L, F.T.L, X.M, Z.L.Y and K. Z. Data analysis: Z.G.L, J.F.L, H.L.T, W.L.Z, H.M.L, X.Q.H, X.Y.L, M.C, Z.Y.H, K. Z, J.Y.W and Y.H.Y. Design: K. Z, J.Y.W and Y.H.Y. All authors have read and approved the final manuscript.

Funding

This work was supported by National Natural Science Foundation of China (Nos. 82272658 and 81872451), Sichuan Science and Technology Program (Grant Nos. 2022JDRC0042 and 2022NSFSC0777), CMC Excellent-talent Program (Grant No. 2024kjTzn03), Organized Research Projects of Chengdu Medical College (Grant No. CYYZZ24-01), Clinical Science Research Foundations of Collaborative Innovation Center of Sichuan for Elderly Care and Health (Grant No. 2022LHTD-02), Sichuan Higher Education Institute Key Laboratory of Major Disease Target Discovery and Protein Drug Development (Grant Nos. 23LHNBZZD07, 24LHBBYY1-08, 24LHBBYY1-09), and Development and Regeneration Key Lab of Sichuan Province (Grant No. 24LHFYSZ1-27), the above funds were awarded to Professor Kun Zhang.

Data availability

The data supporting the conclusions of this article are available upon request.

Declarations

Ethical approval

Animal study was approved by the Institutional Animal Care and Use Committee at Chengdu Medical College. All authors followed the applicable ethical standards to maintain research integrity. Clinical trial number: not applicable.

Consent for publication

All authors have read the final version of the manuscript and agree with its publication.

Competing interests

The authors declare no competing interests.