UBE2L6 promotes invasion and metastasis of triple-negative breast cancer by enhancing autophagy through STK38 ISGylation

Abstract

Although aberrant activation of autophagy is well known in triple-negative breast cancer (TNBC), its functional roles and underlying mechanisms remain largely unknown. In the present study, we found that high UBE2L6 expression was strongly associated with aggressive clinical features in TNBC. We demonstrated that UBE2L6 promoted migration, invasion, and lung metastasis of TNBC by enhancing STK38-mediated autophagy. Mechanistically, UBE2L6 stabilized STK38 by promoting its ISGylation and inhibiting its ubiquitin-proteasomal degradation. Therefore, targeting UBE2L6 and modulating the STK38 ISGylation-autophagy axis represent potential intervention points in TNBC.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s00018-026-06327-6.

Introduction

Triple-negative breast cancer (TNBC) is a distinct subtype lacking expression of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2), accounting for approximately 15% of all breast cancer cases [1]. TNBC exhibits distinct clinicopathological features, including higher rates of early recurrence and distant metastasis, and has the poorest prognosis among all breast cancer subtypes [2]. TNBC remains a major therapeutic challenge due to the lack of specific molecular targets.

Dysregulation of post-translational modifications (PTMs), particularly ubiquitin and ubiquitin-like modifications, is a hallmark of TNBC pathogenesis and a source of potential therapeutic targets [3–5]. Among these, the interferon-stimulated gene 15 (ISG15) modification system (ISGylation) has garnered attention for its marked upregulation in tumors and its involvement in critical cancer-related processes, including autophagy and immune regulation [6–10]. However, the mechanisms of ISGylation and its role in TNBC remain poorly defined. A critical yet understudied component of this machinery is the E2 conjugating enzyme UBE2L6 (UBCH8). UBE2L6 is a unique dual-function enzyme, capable of mediating both ubiquitination and ISGylation, thereby serving as a potential regulatory node that integrates these two modification pathways [11–15]. Intriguingly, analysis of TNBC single-cell sequencing data highlights UBE2L6 as a central hub within networks of ubiquitination and ISG15-mediated signaling, strongly implicating it in TNBC biology. Despite this compelling bioinformatic link, the precise role of UBE2L6 in driving TNBC progression, and whether it does so predominantly through its ISGylation activity, is entirely unknown. The present study aimed to: (1) define the clinical and functional impact of UBE2L6 in TNBC; (2) elucidate whether its oncogenic effects are mediated through the ISGylation pathway; and (3) identify the key downstream molecular mechanisms.

Autophagy exhibits dual regulatory functions in the pathogenesis and progression of TNBC. Autophagy degrades damaged cellular components to generate energy, enhancing tumor cell survival under stress and driving malignant progression. TNBC cells exhibit higher levels of autophagic activity compared to non-TNBC cells, attributable to their adaptation to hypoxic and nutrient-deprived conditions in the tumor microenvironment [16, 17]. It has been demonstrated that autophagy inhibition through genetic interventions or pharmacological inhibitors can suppress stem cell characteristics, proliferation capacity, and metastatic potential of triple-negative breast cancer cells, while also inducing cellular apoptosis and inhibiting their oncogenic signaling pathways [18–20]. Therefore, autophagy has been identified as a promising therapeutic target of TNBC.

STK38 is a member of the AGC kinase family and its activation requires phosphorylation at Ser281/Thr444 [21–24]. Moreover, STK38 kinase undergoes other post-translational modifications, such as ISGylation and acetylation [25]. In addition, STK38 promotes autophagosome formation in human cells via a kinase-dependent mechanism [26]. STK38 mRNA levels are upregulated in breast cancer, ovarian cancer, and lung cancer [27–31]. However, the specific role of STK38 in TNBC and whether it influences tumor progression through the autophagic pathway remain unclear.

This study revealed the critical regulatory role and molecular mechanisms of the UBE2L6 protein in TNBC. We identified UBE2L6 as a key risk factor in TNBC that promoted TNBC migration, invasion, and metastasis by enhancing the STK38-mediated autophagy pathway. These findings provide crucial experimental evidence for establishing a precision therapeutic framework for TNBC lung metastasis.

Results

High expression of UBE2L6 was associated with aggressive clinicopathological features in TNBC

We analyzed the single-cell transcriptomic dataset GSE161529 to characterize the molecular pathways dysregulated in TNBC. Our analysis revealed that genes differentially expressed in TNBC cells were significantly enriched in ubiquitination-related pathways relative to normal epithelial and non-TNBC cells (Fig. 1A). Ubiquitination and ubiquitin-like modifications play key roles in cancer development by regulating protein stability, degradation, and cell signaling. These findings suggest that ubiquitination and ubiquitin-like modification systems may regulate malignant progression of TNBC. We identified 92 overlapping genes through cross-analysis of ubiquitination and ubiquitin-like modification systems (Fig. 1B). Further analysis revealed that the ISGylation-associated gene UBE2L6 was predominantly upregulated in TNBC compared to non-TNBC or normal (Fig. 1C-D). UBE2L6 served as a central hub in ISG15 protein interaction network (Fig. 1E). In the overall breast cancer cohort, elevated UBE2L6 expression significantly correlated with poor pathological grade, lymph node metastasis, and low overall survival (OS) (Fig. 1F-H). Moreover, high UBE2L6 expression in TNBC was significantly associated with aggressive clinicopathological characteristics, including advanced TNM stage (III-IV), high Ki-67 proliferation index, and frequent lymph node metastasis (Fig. 1J-K and Supplementary Table 1). Furthermore, UBE2L6 was commonly upregulated in TNBC cell lines (MDA-MB-231, HCC1937, MDA-MB-468, and Hs578T) compared to hormone receptor-positive (HR+) cell lines (T47D and MCF-7) with lower malignant potential (Fig. 1L). Although dysregulation of UBE2L6 was commonly observed in breast cancer, this study aimed to specifically elucidate its function in the TNBC subtype. Based on these findings, we hypothesized that UBE2L6 promotes TNBC metastasis by regulating the ubiquitination and/or ISGylation pathways.

Fig. 1.

High expression of UBE2L6 was associated with aggressive clinicopathological features in TNBC. (A) GO enrichment analysis plot of the molecular function module. (B) Screen for common genes between ubiquitin-like protein ligase binding and ubiquitin protein ligase binding, and perform venn diagram visualization analysis. (C-D) The volcano plot visually illustrates the distribution of 92 core overlapping genes screened from figure B, and labels the genes of particular interest. (E) ISG15 core interacting genes from STRING database. (F) Pan-Cancer analysis based on the UALCAN database reveals the expression signature of the UBE2L6 Gene. (G) Expression of UBE2L6 in breast cancer based on individual cancer stages from UALCAN database. (H) Expression of UBE2L6 in breast cancer based on nodal metastasis status from UALCAN database. (I) The Kaplan-Meier analysis for overall survival (OS) performed in breast cancer patients based on different UBE2L6 mRNA expression levels. (J) UBE2L6 expression in 50 TNBC and adjacent normal breast tissues (IHC). Scale bar: 50 μm. (K) Representative western blot images of UBE2L6 expression in TNBC tissues and adjacent normal breast tissues. Comparative analysis of UBE2L6 expression between cancer tissues and paired adjacent normal tissues in 20 TNBC patients. N, adjacent normal breast tissues. T, TNBC tissues. (L) Immunoblotting analysis of UBE2L6 protein levels in different breast cell lines. Data were representative of at least three (n ≥ 3) independent biological replicates. *p < 0.05, **p < 0.01, ***p < 0.001

UBE2L6 knockdown inhibited the proliferation, migration, and invasion of TNBC cells in vitro, and suppressed lung metastasis in vivo

We established stable MDA-MB-231 and Hs578T cell lines with either knockdown or overexpression of UBE2L6. UBE2L6 overexpression increased the expression of MMP2, MMP9, and vimentin (Fig. 2A), whereas its knockdown decreased the expression of MMP2, MMP9, and vimentin (Fig. 2B). Wound healing and transwell invasion assays revealed that UBE2L6 overexpression significantly promoted migration (Fig. 2C) and invasion (Fig. 2F), whereas its knockdown inhibited migration (Fig. 2D) and invasion (Fig. 2E). To evaluate the role of UBE2L6 in lung metastasis, we injected MDA-MB-231 cells with stable UBE2L6 knockdown or control cells into nude mice via the tail vein. UBE2L6 knockdown significantly reduced the formation of lung metastatic nodules (Fig. 2G). These results demonstrated that UBE2L6 knockdown inhibited the proliferation, migration, and invasion of TNBC cells in vitro, and suppressed lung metastasis in vivo.

Fig. 2.

UBE2L6 knockdown inhibited the migration and invasion of TNBC cells in vitro, and suppressed lung metastasis in vivo. (A) WB analysis of the expression of vimentin, MMP2, and MMP9 in MDA-MB-231 and Hs578T cells with stable 3×Flag-UBE2L6 overexpression (n = 3). (B) WB analysis of the expression of vimentin, MMP2, and MMP9 in MDA-MB-231 and Hs578T cells with stable UBE2L6 knockdown (n = 3). (C-D) Wound healing assay was performed in MDA-MB-231 and Hs578T cells following UBE2L6 overexpression (C) or knockdown (D) (n = 5). Cells were pretreated with mitomycin C (7.5 µg/mL) for 6 h prior to the wound healing assay. (E-F) Transwell invasion assay was performed in MDA-MB-231 and Hs578T cells following UBE2L6 knockdown (E) or overexpression (F) (n = 5). (G) BALB/c nude mice underwent tail vein injection of MDA-MB-231 cells (Stable UBE2L6 knockdown vs. control groups). Lung specimens and metastatic nodule quantification at 10 weeks post-injection, with representative H&E staining images (n = 5). Data were representative of at least three (n ≥ 3) independent biological replicates. Statistical analysis was performed using a two-tailed Student’s t-test or one-way ANOVA. *p < 0.05, **p < 0.01, ***p < 0.001

Furthermore, EdU and colony formation assays showed that UBE2L6 knockdown inhibited cell proliferation in MDA-MB-231 and Hs578T cells (Fig. S1A-B), whereas its overexpression promoted cell proliferation (Fig. S2A-B). Although UBE2L6 exhibited a pro-proliferative effect, the poor prognosis of TNBC is primarily attributed to its high invasiveness, metastatic potential, and early recurrence. Given the lack of effective targeted therapies, we focused our investigation on how UBE2L6 regulates TNBC cell migration, invasion, and metastasis.

UBE2L6 promoted TNBC migration, invasion, and lung metastasis by activating autophagy

Compared to other breast cancer subtypes, TNBC exhibits the highest autophagic activity [16, 17, 32, 33]. Clinical data indicate that LC3B is a reliable biomarker for poor prognosis in TNBC patients [34]. Therapeutically, hydroxychloroquine (HCQ) has shown potent autophagy-inhibiting effects in preclinical studies [20]. We further analyzed LC3B protein levels in different breast cancer subtypes from clinical specimens using immunohistochemistry (IHC). LC3B protein levels were significantly higher in the epithelia of TNBC than in luminal A, luminal B, and HER2 + subtypes (Fig. S3A-B). Studies have indicated that autophagy, which was upregulated in TNBC, may promote tumor progression via mechanisms such as enhanced tumor cell survival.

Notably, KEGG analysis revealed that UBE2L6 overexpression significantly activated the autophagy pathway in MDA-MB-231 cells (Fig. 3A). Western blot analysis revealed that UBE2L6 overexpression increased the LC3-II/LC3-I ratio and decreased p62 levels (Fig. 3B), consistent with enhanced autophagic flux (Fig. 4H-K). In contrast, Baf-A1 treatment increased the LC3-II/LC3-I ratio but led to p62 accumulation, indicating autophagic flux blockade. Interestingly, the upregulation of MMP2/9 and vimentin induced by UBE2L6 overexpression was significantly rescued by treatment with the autophagy inhibitor Baf-A1 (Fig. 3B), suggesting that UBE2L6 may promote TNBC migration and invasion by enhancing autophagy. To validate that autophagy is required for UBE2L6-mediated migration and invasion, we inhibited the pathway using Baf-A1 or HCQ and assessed the consequent phenotypic changes (Figs. 3 and 4). Baf-A1 treatment significantly inhibited the migration and invasion promoted by UBE2L6 overexpression in MDA-MB-231 and Hs578T cells (Fig. 3C-D), indicating that UBE2L6 enhanced TNBC cell migration and invasion via an autophagy-mediated mechanism. In vivo studies showed that UBE2L6 overexpression significantly potentiated the lung colonization potential of MDA-MB-231 cells (Fig. 3E). The effect was effectively suppressed by HCQ treatment (Fig. 3E), indicating that UBE2L6 promoted autophagy-dependent metastasis.

Fig. 3.

UBE2L6 promoted TNBC migration, invasion, and lung metastasis by activating autophagy. (A) LC-MS/MS-based proteomic profiling of UBE2L6-overexpressing MDA-MB-231 cells versus controls. KEGG pathway enrichment analysis of screened differentially expressed proteins. (B) WB analysis of autophagy-related proteins (LC3 and p62), MMP2, MMP9, and vimentin in MDA-MB-231 and Hs578T cells after UBE2L6 overexpression for 48 h with or without Baf-A1 (0.5 µmol/L, 24 h) treatment (n = 3). (C) Wound healing assay was used to evaluate the migration ability of MDA-MB-231 and Hs578T in different groups (n = 5). Cells were pretreated with mitomycin C (7.5 µg/mL) for 6 h prior to the wound healing assay. (D) Transwell invasion assay was used to evaluate the invasive ability of MDA-MB-231 and Hs578T in different groups (n = 5). (E) BALB/c nude mice were divided into four groups and intravenously injected with MDA-MB-231 cells as follows: Vector + NS, UBE2L6 + NS, Vector + HCQ, and UBE2L6 + HCQ. Gross lung specimens and metastatic nodule quantification at 10 weeks post-injection, with representative H&E staining images. Nude mice received intraperitoneal injections of either hydroxychloroquine (HCQ, 60 µg/g/d) or normal saline (NS, 60 µg/g/d) from week 2 to week 7. Data were representative of at least three (n ≥ 3) independent biological replicates. Statistical analysis was performed using one-way ANOVA followed by Dunnett tests for multicomparisons. ANOVA, analysis of variance. *p < 0.05, **p < 0.01, ***p < 0.001

Fig. 4.

UBE2L6 promoted autophagy in TNBC cells. (A-B) WB analysis of LC3 and p62 expression in MDA-MB-231 and Hs578T cells after UBE2L6 overexpression or knockdown for 48 h (n = 3). (C-F) Confocal microscopy images and quantitative assessments of fluorescent puncta in MDA-MB-231 and Hs578T cells are presented to evaluate autophagic flux (n = 5). Following UBE2L6 knockdown, cells were transduced with mRFP-GFP-LC3 adenovirus for 48 h. Yellow puncta represent co-localization of GFP and mRFP signals, while red puncta reflect autophagosome-lysosome fusion and subsequent GFP signal quenching. (G) Autophagic structures (indicated with red arrows) in UBE2L6 knockdown and vector control groups of MDA-MB-231 and Hs578T cells were assessed using transmission electron microscopy (TEM) (n = 5). (H-K) Following UBE2L6 overexpression with or without Baf-A1 treatment, cells were transduced with mRFP-GFP-LC3 adenovirus for 48 h (n = 5). Data were representative of at least three (n ≥ 3) independent biological replicates. Statistical analysis was performed using two-tailed Student’s t-test or one-way ANOVA followed by Dunnett tests for multiple comparisons. ANOVA, analysis of variance. *p < 0.05, **p < 0.01, ***p < 0.001

UBE2L6 overexpression potentiated autophagic flux, as indicated by an increased LC3-II/I ratio and concomitant reduction in p62 levels relative to vector controls (Fig. 4A). Conversely, UBE2L6 knockdown attenuated autophagy, evidenced by a decreased LC3-II/I ratio and accumulated p62 (Fig. 4B). To directly assess autophagic flux using the mRFP-GFP-LC3 reporter, we observed that UBE2L6 knockdown inhibited the formation of both red (autolysosomes) and yellow (autophagosomes) puncta in merged images of MDA-MB-231 and Hs578T cells (Fig. 4C-F). Conversely, UBE2L6 overexpression enhanced the accumulation of these puncta (Fig. 4H-K). Electron microscopy revealed that UBE2L6 knockdown cells exhibited significantly fewer autophagosomes or autolysosomes compared to control cells (Fig. 4G). Baf-A1 treatment diminished red puncta (autolysosomes) while elevating yellow puncta (autophagosomes) in merged images of UBE2L6-overexpressing Hs578T and MDA-MB-231 cells (Fig. 4H-K). In summary, these results demonstrated that Baf-A1 treatment abolished the pro-migratory and invasive effects induced by UBE2L6, suggesting that UBE2L6 promoted TNBC cell migration and invasion in an autophagy-dependent manner.

UBE2L6 enhanced STK38 ISGylation and inhibited its ubiquitination in TNBC cells

GO enrichment analysis revealed enrichment of terms related to the ISGylation pathway, ubiquitination, and ubiquitin-like modification pathways in MDA-MB-231 cells (Fig. 5A). ISGylation is a post-translational modification process where the ISG15 protein covalently attaches to target proteins. Its key roles include regulating protein stability, modulating signaling pathways, interacting with ubiquitination, and enhancing antiviral defense [35, 36]. ISGylation regulates key cellular pathways to significantly influence diverse biological processes. As the principal E2 enzyme responsible for ISGylation, UBE2L6 warrants investigation into its specific target proteins and core regulatory pathways, which is critical for deciphering its functional mechanisms.

Fig. 5.

UBE2L6 enhanced STK38 ISGylation and inhibited its ubiquitination in TNBC cells. (A) Bioinformatics analysis of pathways related to GO. (B) Heatmap analysis of differentially expressed autophagy-related proteins upon UBE2L6 knockdown. (C-D) Exogenous (C) and endogenous (D) Co-IP assays were conducted to determine the interaction between UBE2L6 and STK38 (n = 3). (E) The colocalization analysis of UBE2L6 and STK38 in MDA-MB-231 and Hs578T cells using confocal microscopy and ImageJ software (Pearson correlation coefficient, PCC) (n = 3). (F) WB analysis of STK38 and p-STK38 expression following UBE2L6 knockdown (n = 3). (G-H) CHX chase experiments (20 µg/mL) were performed in MDA-MB-231 cells, with the 0 h value normalized to 100% (n = 3). CHX, cycloheximide. (I) WB analysis of STK38 level changes in MDA-MB-231 cells with UBE2L6 knockdown, with or without MG132 treatment (8 h) (n = 3). (J) Effect of USP18 overexpression on STK38 protein in UBE2L6 knockdown group and control group (n = 3). (K) Validation of STK38 ISGylation in HEK293T cells (n = 3). (L-M) Ubiquitination and ISGylation modifications of STK38 protein in MDA-MB-231 cells with UBE2L6 overexpression (L) or knockdown (M) were determined via IP and WB assays (n = 3). Data were representative of at least three (n ≥ 3) independent biological replicates. Statistical analysis was performed using two-tailed Student’s t-test. ns, no significance.*p < 0.05, **p < 0.01, ***p < 0.001

Our study demonstrated that UBE2L6 promoted the metastasis of TNBC by enhancing autophagy. Aberrant regulation of autophagy-related proteins was observed following UBE2L6 knockdown in our proteomic dataset, with a notable downregulation of STK38 (Fig. 5B). STK38 was identified as a central regulator in the autophagy network [26]. Subsequent co-immunoprecipitation (Co-IP) assays confirmed that UBE2L6 interacts specifically with STK38 among the autophagy-related proteins. We confirmed the endogenous and exogenous interactions between UBE2L6 and STK38 (Fig. 5C-D). Confocal microscopy analysis showed that UBE2L6 and STK38 were significantly colocalized in the cytoplasm of MDA-MB-231 and Hs578T cells (Fig. 5E). The molecular docking results revealed a binding free energy of −7.5 kcal/mol (lower than the − 4 kcal/mol threshold) between UBE2L6 and STK38 proteins, demonstrating binding stability (Fig. S4A). The docking result was consistent with the Co-IP/IF data and supported a direct or close interaction between UBE2L6 and STK38. UBE2L6 knockdown decreased STK38 and p-STK38 expression (Fig. 5F). Subsequently, we treated MDA-MB-231 cells with the protein synthesis inhibitor CHX. STK38 protein levels and stability were decreased following UBE2L6 knockdown (Fig. 5G-H), whereas mRNA levels remained unchanged (Fig. S4B). UBE2L6 knockdown decreased STK38 protein levels, whereas this effect was rescued by MG132 (Fig. 5I). These findings suggested that UBE2L6 overexpression reduced ubiquitination-mediated degradation of STK38. USP18 is a specific deISGylase that removes ISG15 from bound proteins [37]. We found that USP18 overexpression significantly reduced STK38 protein expression with or without UBE2L6 overexpression, suggesting STK38 stability may also be regulated by ISGylation (Fig. 5J). Based on these findings, we hypothesized that UBE2L6 modulated autophagic activity in TNBC cells by catalyzing ISG15-mediated modification of STK38. Subsequent experiments confirmed that UBE2L6 overexpression promoted STK38 ISGylation in HEK293T cells (Fig. 5K). These results suggested that UBE2L6 enhanced STK38 protein stability by increasing its ISGylation level. As expected, UBE2L6 knockdown reduced STK38 ISGylation, while UBE2L6 overexpression promoted STK38 ISGylation in MDA-MB-231 cells (Fig. 5L-M). Additionally, UBE2L6 knockdown significantly enhanced STK38 ubiquitination, whereas its overexpression had the opposite effect in MDA-MB-231 cells (Fig. 5L-M).

The ISGylation level of STK38 was significantly increased by UBE2L6 overexpression, whereas its ubiquitination level exhibited an opposite trend. Collectively, these results demonstrated that UBE2L6 promoted STK38 ISGylation, leading to reduced degradation of STK38 protein mediated by the ubiquitin-proteasome system in TNBC cells. This negative correlation suggested that ISGylation and ubiquitination may antagonistically regulate STK38 stability, potentially through a competitive mechanism.

UBE2L6 knockdown inhibited STK38-mediated autophagy

STK38 kinase is an evolutionarily conserved autophagy regulator [26]. However, it remains unclear whether STK38 affects the malignant progression of TNBC through the autophagy pathway. Co-IP assays demonstrated that UBE2L6 knockdown significantly reduced the interaction of STK38 with the autophagy core complex (Beclin1/ATG14L/VPS34) (Fig. S5A). Further colocalization experiments revealed that UBE2L6 knockdown significantly reduced the colocalization levels between STK38 and the components of the VPS34 complex (Beclin1/ATG14L/VPS34) (Fig. S5B-C). We observed that STK38 promoted autophagosome formation through initiation-phase regulation (Fig. S6D-F), consistent with previous research findings [26]. 3-MA is an early-stage autophagy inhibitor that blocks autophagosome formation by inhibiting VPS34. Notably, 3-MA (VPS34 inhibitor) blocked STK38-induced autophagy activation (Fig. S6D-F). These results revealed that STK38 directly targeted the VPS34 complex to regulate autophagosome initiation and promote autophagy activation. Moreover, 3-MA effectively reversed the increased invasiveness caused by STK38 overexpression in MDA-MB-231 and Hs578T cells (Fig. S6G). These results indicated that UBE2L6 knockdown inhibited STK38-mediated autophagy by destabilizing STK38 protein.

High STK38 expression is associated with a higher risk of recurrence in TNBC patients

Bioinformatics analysis using the UALCAN database revealed that STK38 mRNA expression was significantly upregulated in TNBC tissues compared to normal breast tissues (Fig. S7A). STK38 mRNA expression was significantly upregulated in TNBC subtypes (BL1, BL2, IM, M) compared to normal breast tissues (Fig. S7B). Kaplan-Meier analysis showed that TNBC patients with high STK38 expression had significantly shorter relapse-free survival (RFS) than those with low expression (Fig. S7C). The Human Protein Atlas database shows that STK38 mRNA expression is significantly higher in highly invasive TNBC cell lines (MDA-MB-231, Hs578T, MDA-MB-468, and HCC1937) compared to low-invasive breast cancer cells (MCF-7 and T47D) (Fig. S7D). WB results revealed that STK38 protein expression was significantly higher in highly invasive TNBC cell lines (MDA-MB-231, Hs578T, MDA-MB-468, and MDA-MB-453) compared to low-invasive breast cancer cell lines (MCF-7 and T47D) (Fig. S7E). These results suggested that STK38 high expression was associated with poor prognosis in TNBC patients.

UBE2L6 promoted TNBC lung metastasis via enhancing the STK38-mediated autophagy

To determine whether STK38 is involved in UBE2L6-induced autophagy and metastasis, we performed rescue experiments. Western blot analysis revealed that STK38 overexpression rescued the decrease in the LC3-II/I ratio and the increase in p62 protein levels induced by UBE2L6 knockdown (Fig. 6A). At the same time, STK38 overexpression reversed the decrease of MMP2, MMP9, and vimentin caused by UBE2L6 knockdown (Fig. 6A). Furthermore, as evidenced by mRFP-GFP-LC3 fluorescence labeling (Fig. 6B-C) and TEM (Fig. 6D), STK38 overexpression rescued the UBE2L6 knockdown-induced reduction in autophagosome formation and autolysosome generation. These results indicated that UBE2L6 activated autophagy via the STK38-dependent pathway.

Fig. 6.

UBE2L6 promoted TNBC lung metastasis via enhancing STK38-mediated autophagy. (A) Western blot analysis was performed to examine the expression of autophagy markers (LC3 and p62) and EMT markers (MMP2, MMP9, and vimentin) in different groups (n = 3). (B-C) Representative confocal microscopy images of mRFP-GFP-LC3 assay in MDA-MB-231 and Hs578T cells from Vector and UBE2L6 knockdown groups with or without STK38 overexpression (n = 5). (D) Autophagic structures (indicated by red arrows) in different groups of MDA-MB-231 cells were evaluated by TEM (n = 5). (E-F) The invasion ability of different groups was evaluated using transwell invasion assay in MDA-MB-231 and Hs578T cells (n = 5). (G-H) The migration ability of different groups was evaluated using wound healing assay in MDA-MB-231 and Hs578T cells (n = 5). Cells were pretreated with mitomycin C (7.5 µg/mL) for 6 h prior to the wound healing assay. (I) BALB/c nude mice underwent tail vein injection of MDA-MB-231 cells (shNC+Vector group, shUBE2L6 + Vector group, and shUBE2L6 + STK38 group). Lung specimens and metastatic nodule quantification at 10 weeks post-injection, with representative H&E staining images. (J) Immunohistochemical (IHC) staining of UBE2L6, STK38, LC3B, p62, MMP9, and vimentin in metastatic lung nodules (n = 5). Scale bar: 50 μm. Data were representative of at least three (n ≥ 3) independent biological replicates. Statistical analysis was performed using one-way ANOVA followed by Dunnett tests for multiple parisons. ANOVA, analysis of variance. *p < 0.05, **p < 0.01, ***p < 0.001

Wound healing and invasion assays revealed that STK38 overexpression reversed the inhibition of migration and invasion caused by UBE2L6 knockdown in MDA-MB-231 and Hs578T cells (Fig. 6E-H). In vivo studies demonstrated that STK38 overexpression significantly reversed the inhibition of lung metastasis caused by UBE2L6 knockdown in MDA-MB-231 cells (Fig. 6I). IHC assessment of metastatic lung nodules demonstrated that UBE2L6 knockdown significantly decreased LC3B, MMP9, and vimentin expression levels, while causing p62 accumulation (Fig. 6J). Notably, STK38 overexpression significantly reversed these molecular changes induced by UBE2L6 knockdown. In conclusion, these results indicated that UBE2L6 promoted TNBC lung metastasis via enhancing the STK38-mediated autophagy.

Discussion

Currently, early-stage triple-negative breast cancer (TNBC) is still primarily treated with surgery, supplemented by radiotherapy and chemotherapy [38, 39]. In addition, immunotherapy has emerged as a novel therapeutic approach in recent years and may represent an effective option for a subset of TNBC patients. Accumulating evidence indicates that the progression of TNBC is closely associated with immune regulation, particularly through the programmed cell death protein 1/programmed death-ligand 1 (PD-1/PD-L1) axis [40–42]. However, despite certain advances in the treatment of TNBC, this subtype still lacks well-defined therapeutic targets, highlighting an urgent need to further screen for and investigate the mechanisms underlying its progression in order to facilitate more precise treatment strategies.

Beyond ubiquitination, ubiquitin-like modifications, including ISGylation, critically regulate oncogenic processes [43, 44]. While ISG15 expression is elevated in breast cancer [45, 46], the functional contributions and regulatory mechanisms of the specific ISGylation enzymatic machinery, particularly in TNBC, remain poorly defined. Notably, the E2 conjugating enzyme UBE2L6 is a central component of the ISGylation cascade [47, 48], yet its role in TNBC pathogenesis has not been investigated. Thus, we sought to determine how UBE2L6 promotes TNBC progression.

Although previous studies have established the role of UBE2L6 in other disease contexts, our study was primarily motivated by its marked upregulation in TNBC and its strong association with aggressive clinical features (Fig. 1). This prompted us to hypothesize that UBE2L6 plays a TNBC-specific role beyond its known functions in general protein modification. Our functional experiments substantiated this finding, revealing a critical role for UBE2L6 in promoting migration, invasion, and metastasis (Fig. 2).

Taken together, these findings indicate that UBE2L6 promotes metastasis through autophagic pathways (Figs. 3 and 4). The elevated LC3B expression observed in aggressive TNBC [16, 34, 49] may be partially attributable to upstream regulators such as UBE2L6. More importantly, although previous studies had identified autophagy as a general vulnerability in TNBC [32, 50–52], our study revealed that the UBE2L6-STK38 axis activated this metastasis-promoting autophagic program.

Our study found that UBE2L6 stabilized STK38 by promoting its ISGylation and inhibiting its ubiquitin-proteasomal degradation (Fig. 5). Previous studies have shown that STK38 serves as a critical modulator of autophagy through multiple molecular mechanisms [26, 53, 54]. For example, STK38 regulated the nuclear export of beclin1 by phosphorylating XPO1, thereby modulating autophagy [53]. Another study showed that STK38 facilitated the association between the exocyst complex protein Exo84 and the Beclin1/RalB complex, a critical step in triggering autophagosome biogenesis [26]. The role of autophagy in cancer metastasis is complex, as it can exhibit both pro-metastatic and anti-metastatic effects [55–58]. During late-stage metastasis, autophagy promotes survival and colonization of extracellular matrix (ECM)-detached metastatic cells at distant sites. STK38 may regulate the balance between autophagy and apoptosis through distinct molecular mechanisms [26, 54]. Apoptosis resistance, particularly anoikis resistance, is critical for all stages of metastatic progression [59]. STK38-mediated autophagy may promote tumor metastasis by enhancing anoikis resistance. Further studies using appropriate cellular and animal models are needed to verify whether and how STK38-mediated autophagy supports metastasis. Our study demonstrated that STK38 promoted migration and invasion in TNBC cells by enhancing autophagy initiation (Figs. S5 and S6). The suppression of UBE2L6 significantly attenuated TNBC invasion and lung metastasis by inhibiting the STK38-mediated autophagy (Fig. 6). Thus, the UBE2L6-STK38 axis represents a pro-metastatic pathway that becomes advantageous in advanced, detached cancer cells, aligning with our in vivo lung colonization model.

This study delineates a novel UBE2L6/ISGylation/STK38/autophagy axis promoting TNBC metastasis. In conclusion, this study demonstrates that UBE2L6 promotes invasion and lung metastasis in TNBC by enhancing STK38-mediated autophagy. Mechanistically, UBE2L6 stabilized STK38 by promoting its ISGylation and inhibiting its ubiquitin-proteasomal degradation (Fig. 7). Future studies should focus on: (1) Validating this axis in patient-derived organoids and additional in vivo models; (2) Screening for small-molecule inhibitors targeting UBE2L6; (3) Investigating whether the UBE2L6-STK38 axis also contributes to the autophagy-mediated drug resistance observed in TNBC, which would significantly broaden its clinical relevance.

Fig. 7.

Model depicting the role of UBE2L6 as an ISGylation enzyme that maintains STK38 stability and regulates downstream oncogenic behaviors in TNBC

Materials and methods

Datasets and sample selection

This study analyzed the GSE161529 dataset from the GEO database (http://www.ncbi.nlm.nih.gov/geo), comprising 421,761 cells from 52 patients. The GSE161529 dataset includes the following samples: 4 cases of TNBC tumor tissues, 4 cases of BRCA1-mutated TNBC tumor tissues, 6 cases of HER2-positive tumor tissues, 19 cases of ER-positive tumor tissues, and 6 cases of ER-positive breast tumor tissues with lymph node metastasis. In addition, this dataset also includes 13 samples of whole breast cells from healthy controls, 11 samples of normal breast epithelial cells, and 4 samples of whole breast cells from precancerous lesions carrying BRCA1 mutations.

Identification of differentially expressed genes (DEGs) and functional enrichment analysis

Epithelial cells from the Tumor group were isolated and subjected to malignant (cancer cells) versus non-malignant cell identification using the copykat package (version 1.1.0). Differential analysis between sample groups was performed with thresholds set as avg_log2FC > 0.25, p_val < 0.05, and min.pct > 0.1. Volcano plots of differentially expressed genes (DEGs) were generated using the R package ggplot2 (version 3.5.1), while the R package ComplexHeatmap (version 2.20.0) was used to visualize expression heatmaps of DEGs across sample groups. Differentially expressed genes were subjected to GO and KEGG enrichment analyses using the R packages clusterProfiler (version 4.14.3) and org.Is.eg.db (version 3.20.0) to identify their shared functional characteristics and associated pathways. The Gene Ontology (GO) system comprises three categories: Cellular Components (CC), Molecular Functions (MF), and Biological Processes (BP). Enriched GO terms were filtered with an adjusted p-value cutoff < 0.05. Similarly, KEGG pathway enrichment analysis was performed using the same threshold (adjusted p-value cutoff < 0.05).

Bioinformatics analysis

The overall survival (OS) plots were downloaded from the Kaplan-Meier online database (https://kmplot.com/analysis/). The expression of UBE2L6 in different BC subtypes, pathological grades, or Pan-cancer was predicted using the UALCAN online database (https://ualcan.path.uab.edu/cgi-bin/ualcan-res.pl).

Ethics statement

All clinical samples were obtained from breast cancer patients at The First Affiliated Hospital of Harbin Medical University, none of whom had undergone prior radiotherapy or neoadjuvant chemotherapy. This study was approved by the Ethics Committee of the First Affiliated Hospital of Harbin Medical University (Approval No. 2025288) with written informed consent obtained from all participants. Experimental protocols followed the Declaration of Helsinki, and animal studies complied with NIH Guidelines for care and use of laboratory animals. All animal experimental procedures and methods were approved by the Institutional Animal Care and Use Committee of the First Affiliated Hospital of Harbin Medical University (Approval No. 2024072).

Immunohistochemistry (IHC)

The immunohistochemical (IHC) staining procedure was conducted using target-specific antibodies. Representative microscopic images were acquired at varying magnification levels (100×, 200×, and 400×) employing an Olympus digital microscopy imaging platform. Histological scores (H-scores) were provided as a function of the percentage of positive cells multiplied by the staining intensity. H-Score = ∑(pi × i) = (percentage of weak intensity cells × 1) + (percentage of moderate intensity cells × 2) + (percentage of strong intensity cells × 3). Two different investigators blindly evaluated three sections from distinct areas of each tumor. Inclusion criteria: (1) Pathologically confirmed TNBC (ER/PR negative, HER2 negative); (2) Complete clinical follow-up or pathological data; (3) No neoadjuvant therapy before surgery; (4) No other concurrent tumors. This study included tumor tissues (Ca) and paired adjacent normal tissues (Adjacent) from 50 patients with TNBC. Immunohistochemical staining was performed, and UBE2L6 expression levels were evaluated using the H-score system. The median H-score in adjacent normal tissues was 89. In tumor tissues, patients were divided into low-expression (n = 25) and high-expression (n = 25) groups using the median H-score (179.5) as the cutoff for subsequent clinicopathological association analysis. To compare UBE2L6 expression levels between adjacent normal tissues (n = 50) and tumor tissues (n = 50), the Mann-Whitney U test (a nonparametric test) was employed to assess whether the difference in medians between the two groups was statistically significant.

Cell cultures and cell transfection

MDA-MB-231 (CL-0150 A, RRID: CVCL_0062), Hs578T (CL-0114, RRID: CVCL_0332), HEK293T (RRID: CVCL_0063), and T47D (CL-0228, RRID: CVCL_0553), were kindly provided by Procell Life Science & Technology Co., Ltd. MCF-7 (RRID: CVCL_0031) and MDA-MB-453 (RRID: CVCL_0418) were obtained from the Cell Bank, Type Culture Collection, Chinese Academy of Sciences (CBTCCCAS, Shanghai, China). BT-549 (RRID: CVCL_1092) cell line was kindly provided by Haixing Bioscience (Suzhou, China). Cell lines were authenticated by short tandem repeat profiling. Hs578T, MCF-7 and HEK293T cell lines were cultivated in the Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, CA, USA) supplemented with 10% fetal bovine serum (FBS, ExCell Biol, FSP500). MDA-MB-231 and MDA-MB-453 cells were cultured in L-15 medium (Gibco, CA, USA) supplemented with 10% FBS (ExCell Biol, FSP500). T47D and BT-549 cells were cultured in RPMI-1640 medium supplemented with 10% FBS (ExCell Biol, FSP500). To avoid microbial contamination, culture medium was also supplemented with 1% penicillin-streptomycin (Gibco, CA, USA). All cells were cultured in a humidified environment with or without (MDA-MB-231 and MDA-MB-453 cells) 5% CO₂ at 37 °C. UBE2L6/STK38 knockdown or overexpression lentiviruses were purchased from Shanghai Genechem Co., Ltd (China). shRNA sequences are provided in Supplementary Table 2.

Cell migration and Matrigel invasion assays

Breast cancer (BC) cells were cultured in 6-well plates until 90% confluence. To exclude the potential influence of cell proliferation on the experimental outcome, cells were pretreated with mitomycin C (7.5 µg/mL) for 6 h prior to the wound healing assay. Linear scratches were created using sterile 200 µL pipette tips, and wound closure was assessed by quantitatively analyzing cell-free areas through digital imaging. In transwell migration assays, cells were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet after 12–24 h incubation. For transwell invasion assays, Matrigel-coated membranes were used. Cells that migrated or invaded through the membrane were quantified using ImageJ software.

Western blot (WB) and immunoprecipitation (IP) analysis

Protein concentration in cell lysates was determined using the BCA assay kit (Beyotime, China). Cells were lysed in RIPA buffer containing protease and phosphatase inhibitors. Proteins were separated by SDS-PAGE (10–12% gels), transferred to nitrocellulose membranes, blocked with 5% non-fat milk, and incubated overnight at 4 °C with primary antibodies. Following a washing step, the membranes underwent incubation with the secondary antibody (LI-COR, IRDye^® 800CW Goat anti-Mouse/Rabbit IgG, USA) for 60 min at room temperature. The bands were visualized using a molecular imaging system (LI-COR Odyssey, USA).

Cells were lysed in RIPA buffer supplemented with protease inhibitors. Lysates were centrifuged at 12,000 × g for 15 min at 4 °C. Supernatants were incubated with 2 µg specific antibody or isotype control overnight at 4 °C, followed by 2 h incubation with Protein A/G magnetic beads. After washing with lysis buffer three times, bound proteins were eluted with 2× Laemmli buffer at 95 °C for 10 min and analyzed by western blotting. The antibodies used in this study are listed in Supplementary Table 3.

Molecular docking

This study utilized GRAMM for rigid protein-protein docking, where ligand and receptor conformations remain fixed while identifying optimal binding sites. Protein sequences were retrieved from UniProtKB based on gene names, then modeled using SWISS-MODEL to generate structural PDB files. Among 10 docking results obtained, the top-ranked complex demonstrated exceptional binding stability with a free energy of −7.5 kcal/mol. The evaluation criteria for binding free energy stipulate that values below − 4 kcal/mol are generally accepted, with smaller values being preferable. This result indicates robust surface interaction and optimal binding conformation. The obtained energy confirms a highly stable protein-protein interaction system.

Immunofluorescence (IF)

The cells were fixed with 4% paraformaldehyde for 15 min, permeabilized with 0.1% Triton X-100 for 30 min, and blocked with 1% BSA for 30 min at room temperature. Next, the cells were incubated overnight at 4 °C with an appropriate concentration of antibody. Subsequently, the cells were incubated in the dark for 1 h with secondary antibodies conjugated to Alexa Fluor 488 (Thermo Fisher, A32723, RRID: AB_2633275; A21206, RRID: AB_2535792) or Alexa Fluor 555 (Thermo Fisher, A-21434, RRID: AB_2535855; A32727, RRID: AB_2633276). Then, the nuclei were stained with DAPI for 30 min. The images were captured using a confocal fluorescence microscope.

Autophagy assay

MDA-MB-231 and Hs578T cells transfected with mRFP-GFP-LC3 lentivirus were cultured in DMEM/RPMI 1640 medium containing puromycin (1 µg/mL) for one week. After seeding in glass-bottom dishes and overnight incubation, autophagic flux was analyzed using a confocal microscope [60].

Transmission electron microscopy (TEM)

MDA-MB-231 and Hs578T cells were cultured in 10 cm dishes and treated under varying conditions for 48 h. After centrifugation, cell pellets were resuspended in immunoelectron microscopy fixative and preserved at 4 °C for structural stabilization. Ultrastructural analysis was subsequently performed using a Hitachi HT7700 electron microscope (Japan).

Animal experiment

Female BALB/c nude mice (6 weeks old) were obtained from Beijing Vital River (Beijing, China). For the metastasis model, 1 × 10⁶ MDA-MB-231 cells were injected intravenously into female BALB/c nude mice via the tail vein. The mice were euthanized 10 weeks after tail vein injection of the cells. Each lung tissue sample was fixed, sectioned, and stained with H&E to quantify metastatic nodules. All experimental procedures and methods were approved by the Institutional Animal Care and Use Committee of the First Affiliated Hospital of Harbin Medical University (Approval No. 2024072).

RNA extraction and real-time PCR analysis

Total RNA was isolated from cellular and tissue specimens with AxyPrep Total RNA Medium Kit (AXYGEN, USA) and reverse-transcribed into cDNA using PrimeScript reverse transcriptase (TOYOBO, Japan) following the manufacturer’s protocol. qRT-PCR was conducted in a 20 µL reaction system on an Applied Biosystems 7500 Fast Real-Time PCR System, with cDNA as the template. Relative gene expression levels were calculated via the 2^−ΔΔCT method and normalized to the housekeeping gene GAPDH. Triplicate experiments were performed, and primer sequences are provided in Supplementary Table 4.

4D Label-free proteomics analysis

The proteomic analysis was supported by Shanghai Applied Protein Technology Co., Ltd. The experimental workflow consisted of protein extraction, tryptic digestion, and LC-MS/MS analysis. The resulting data were processed using MaxQuant (version 1.6.14) for protein identification and quantification. Subsequent bioinformatic analysis was performed as follows: differentially expressed proteins (DEPs) were annotated with Gene Ontology (GO) terms using Blast2GO and were also searched against the KEGG database to obtain KEGG orthology identifiers. The false discovery rate (FDR) was controlled at ≤ 1% for both protein and peptide identification. For label-free quantification (LFQ), razor and unique peptides were used with a minimum ratio count of 1, and a match-between-runs window of 2 min was applied. Proteins were considered differentially expressed if they met the following criteria: a Q-value < 0.05, a fold change ≥ 1.2, and support by at least two unique peptides. The experiment included three biological replicates per condition.

Statistics analysis

Statistical analyses were performed using GraphPad Prism (version 8.0) and ImageJ (version 1.8.0). Differences between the two groups and among multiple groups were assessed using Student’s t test (or nonparametric test) and one-way ANOVA, respectively. The significance of differences in clinicopathological data was determined using the chi-squared test. Pearson’s correlation coefficient (PCC) was applied to quantitatively assess the degree of colocalization of fluorescent signals. All data are presented as mean ± standard deviation (SD) or median (interquartile range, IQR) based on at least three independent biological replicates. Survival curves were analyzed by the Kaplan-Meier method and log-rank test. p value < 0.05 indicated statistical significance (ns indicates no significant difference, *p < 0.05, **p < 0.01, ***p < 0.001).

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

L.J., K.L., and J.Y. performed Western Blot and qRT-PCR. K.L., Y.H., and W.Z. performed cell culturing and transfection. K.L. and Y.Q. collected the clinical samples and parameters. J.F., R.C., and M.W. performed the IF, IHC, H&E staining and biochemical analysis. K.L., C.L. and X.H. performed statistical analysis and animal experiment. L.J. and K.L. interpreted the results and wrote the manuscript. Y.Z., X.H., and C.L. provided conceptual advice and revised the manuscript. Y.Z. conceived the project and supervised the study. All authors read and approved the final paper.

Funding

This paper was supported by grants from National Natural Science Foundation of China (Nos. 82270668 (L.J.), 81900586 (L.J.) and 82274028 (X.H.)), Natural Science Foundation of Heilongjiang Province of China (No. PL2024H072 (Y.Z.)), Youth Fund Project of Harbin Medical University (No. 2023-KYYWF-0164 (C.L.)), the Foundation of the First Affiliated Hospital of Harbin Medical University (No.2023 B11(C.L.)), Postgraduate Research & Practice Innovation Program of Harbin Medical University (No. YJSCX2023-64HYD (K.L.)), the Science Fund for Distinguished Young Scholars of First Affiliated Hospital of Harbin Medical University (No. HYD2020JQ0010 (L.J.)), and Chen Xiao-ping Foundation for the Development of Science and Technology of Hubei Province (No. CXPJJH11900018-2003 (Y.Z.)).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

AI Use declaration

The authors did not utilize AI for any generative editorial tasks or autonomous content generation.

Financial interests

All authors hereby declare that there are no financial interests, personal relationships, or other potential conflicts of interest that could influence the objectivity of this study’s design, implementation, data analysis, or manuscript preparation.