iTRAQ LC–MS/MS identified a unique membrane-associated protein profile in the LC cells derived from the long-term slow growing metastatic lung nodules of TNBC

In order to identify proteins that are essential for TNBC metastatic colonization, we applied a quantitative proteomic approach iTRAQ LC–MS/MS technique to systematically compare the protein expression profiles among MDA-MB-231-PT and MDA-MB-231-derived IV2 and LC cells as illustrated in Fig. 1A. As described previously, 231-PT, IV2 and LC cells were isolated from the 231-derived primary tumors, early lung metastases, and long-term slow growing lung metastases, respectively, of the orthotopic breast cancer mouse model [18]. The cellular proteins of 231-PT, IV2 and LC cells were fractionated into three compartments, membrane, cytosol and nucleus followed by iTRAQ-based LC–MS/MS proteomic analysis of peptides from the three compartments (Fig. 1B). A total of 7538 proteins were successfully quantified from three compartments and Venn diagram analysis showed that, among the identified proteins, 988 were membrane-associated proteins, 701 were cytosol-associated proteins, and 858 were nucleus-associated proteins (Fig. 1C). The averaging MS intensity ratio (log2) of the identified proteins were shown in Fig. 1D. We identified 71 and 83 membrane-associated proteins that were upregulated at least 1.5-fold in the LC cells IV2 cells, respectively, as compared to the 231-PT cells (Fig. 1E; Additional file 1: Tables S1, S2); By applying a supervised cluster analysis to the LC/IV2-enriched proteome in TCGA breast cancer dataset using the UCSC Xena platform, we identified a group of candidates that were exclusively upregulated in TNBC (Additional file 1: Fig. S1). Moreover, Venn diagram analysis revealed that 64 membrane-associated genes were uniquely expressed in LC cells, while 76 membrane-associated genes were exclusively expressed in IV2 cells, and 7 common genes were identified in both IV2 and LC cells (Fig. 1F; Additional file 1: Tables S1, S2).

Fig. 1

iTRAQ-based LC MS/MS approach identifies a unique membrane-related protein profiles in the metastatic TNBC cells derived from the metastatic lung nodules. A A schematic of identification of metastasis-related membrane proteins using iTRAQ-coupled LC MS/MS. B Subcellular protein fractionation of 231-PT, IV2 and LC cells. C MS-generated proteomic data were quantified and identified by proteome discoverer software and Mascot search engine, respectively. A total of 7538 proteins from three cellular compartments of 231-PT, IV2, LC cells was successfully identified and quantified. D Average of MS intensity ratio (log2) of IV2 versus 231-PT and LC versus 231-PT. E Venn diagram analysis of the proteins with 1.5-fold increase from three compartments. The metastasis-related proteins that were upregulated at least 1.5-fold in the membrane fraction of IV2 cells (shown as the blue circle) and LC cells (shown as the orange circle) as compared to that of the 231-PT cells were identified. F Venn diagram analysis of IV2-specific and LC-specific membrane-associated protein

Subsequently, we selected 64 genes that were uniquely expressed in LC cells, which could be relevant to the lung colonization potential of TNBC cells, for further clinical association studies.

High expression of Sciellin (SCEL) is significantly associated with TNBC and the poor survival of BC patients

Next, we analyzed the clinical association of the LC cell-specific membrane-associated proteins. First, the correlation between the candidate proteins and the overall survival of BC patients was analyzed by UCSC Xena. The P-values of log-rank test of the Kaplan–Meier survival analysis were retrieved from TCGA database using UCSC Xena. As shown in Fig. 2A, high expression of NT5E, RAP1GAP2, SDPR, SCEL, LPCAT2, SMURF2, and CTGF were significantly associated with the poor survival of BC patients among the candidate proteins (Fig. 2A). Among these 7 candidates, the expression of SCEL was highly associated with the TNBC subtype (Fig. 2B). In addition, the Kaplan–Meier survival curve analysis showed that high mRNA expression of SCEL was significantly correlated with poor overall survival (OS) (P = 0.035) and shorter progression-free survival (PFS) (P = 0.026) of BC patients in TCGA BRCA dataset (Fig. 2C). Similarly, high expression of SCEL was also associated with poor OS (P < 0.001, hazard ratio (HR): 1.54, confidence interval (CI):1.27–1.86) and shorter distant metastasis-free survival (DMFS) (P < 0.001, HR:1.48, CI:1.26–1.76) of BC in GSE datasets (Fig. 2C). Further analysis of the correlation between SCEL expression and the outcomes of TNBC patients using GSE datasets revealed that high expression of SCEL was also significantly associated with poor OS (P = 0.0015, HR:2.13, CI 1.32–3.44), shorter DMFS (P < 0.001, HR:1.93, CI 1.31–2.82), and shorter relapse-free survival (RFS) of patients with TNBC (P < 0.0034, HR:1.48, CI 1.14–1.94) (Fig. 2C). Next, we examined the protein expression of SCEL in TNBC tumor specimens. Immunohistochemistry staining (IHC) analysis of TNBC tissue microarray showed that high expression of SCEL was significantly associated with the late-stage TNBC (Fig. 2D–F). The comparison of SCEL protein expression in TNBC and non-TNBC tumor tissues revealed a significant upregulation of SCEL expression in the TNBC tissues as compared to the non-TNBC tissues (Fig. 2G). In addition, the expression of SCEL was not found to be correlated with higher-grade or advanced stage of non-TNBC patients (Additional file 1: Fig. S2).

Fig. 2

SCEL expression is significantly upregulated in TNBC tumors and correlates with the poor outcome of patients with TNBC. A Analysis of the survival correlation of the candidate proteins in TCGA breast cancer dataset using UCSC Xena platform. B Analysis of mRNA expression levels of top seven candidates, NT5E, RAP1GAP2, SDPR, SCEL, LPCAT2, SMURF2, and CTGF, in TNBC tumors versus non-TNBC tumors using UCSC Xena. *P < 0.01. C The Kaplan–Meier survival analysis of BC patients and TNBC patients based on SCEL expression. The Kaplan–Meier survival curves of overall survival (OS) and progression-free survival (PFS) of breast cancer patients stratified by SCEL expression were generated by UCSC Xena using TCGA breast cancer dataset. The Kaplan–Meier survival curves of overall survival (OS) and distant metastasis-free survival (DMFS) of breast cancer patients stratified by SCEL expression were generated by KM plotter using 48 GSE datasets. The Kaplan–Meier survival curves of overall survival (OS), distant metastasis-free survival (DMFS), and relapse-free survival (RFS) of TNBC patients stratified by SCEL expression were similarly generated by KM plotter using 48 GSE datasets. D Immunohistochemistry (IHC) staining of SCEL protein expression in TNBC tissue microarrays (n = 191). Pearson Chi’s square test of the observed values and the expected values of the early/late-stage tumors with low and high SCEL expression (Χ2 = 22.32, P < 0.00001). Pearson Chi’s square test was similarly performed in the lower and higher-grade tumors with low and high SCEL expression (Χ2 = 0.38, P < 0.536). E H score of SCEL IHC staining of the early-stage and late-stage TNBC tumors. *P < 0.001. F H score of SCEL IHC staining of the lower and higher grade TNBC tumors. G H scores of SCEL IHC staining of TNBC (n = 191) and non-TNBC tumors (n = 82). *P < 0.001

Altogether, we identified a novel metastasis-related protein SCEL with clinical implication and significance from the long-term slow growing metastatic lung nodules and it could play an important role during TNBC metastatic lung colonization.

SCEL is a membrane-bound protein that is specifically expressed in the LC sublines derived from the long-term metastatic lung nodules

Next, we examined the expression of SCEL in the LC sublines, 231-PT cells, and IV2 sublines (derived from short term lung metastatic nodules) [21]. Western blotting analysis demonstrated robust SCEL expression in LC sublines, whereas it was undetectable in both 231-PT and IV2 sub-lines under the short exposure condition (5 s). However, upon extended exposure (60 s), a minor amount of protein expression was observed in the 231-PT sublines (Fig. 3A). Moreover, investigation of SCEL expression in a series of breast cancer cell lines including luminal subtype (MCF-7 and T-47D), HER2 subtype (SK-BR-3 and HCC1419), and TNBC subtype (MDA-MB-468, BT-549, Hs578T), revealed that SCEL is specifically expressed in the long-term metastatic lung nodules-derived LC subline (Fig. 3A). In addition, we further analyzed the localization of SCEL protein expression in the mouse metastatic lung nodules. We found that SCEL expression was predominantly observed in both the cell membrane and cytoplasm of the majority of SCEL-positive metastatic cells (> 95%), with notably strong membrane staining (Fig. 3B, C); membrane staining indicated by the black arrows). Protein fractionation analysis further showed that SCEL was primarily enriched in the membrane fraction and less abundant in the cytoplasmic fraction (Fig. 3D). To examine the membrane localization of SCEL, an HA-SCEL fusion protein expression construct was introduced into the Hs578T cells for 24 h followed by immunofluorescent (IF) staining, and confocal imaging. As shown in Fig. 3E, three conditions were used to investigate the membrane localization of SCEL: live cell staining, cell fixation without permeabilization, and cell fixation with permeabilization. The results showed that no green fluorescence signals were detected from the HA-SCEL-transfected Hs578T cells under live cell staining condition (top, Fig. 3E). Under the fixation and non-permeabilization condition, we observed a green fluorescence staining at the peripheral membrane and within the cytoplasm of the transfected cells (middle, Fig. 3E). Under the permeabilization condition, we observed a more pronounced membrane and intracellular staining of SCEL in the transfected cells (bottom, Fig. 3E). These results suggested that SCEL protein was mainly localized to the inner part of the cell membrane and the cytoplasm. Taken together, our results confirmed that high expression of SCEL was specifically expressed in the LC cells derived from the long-term slow growing metastatic lung nodules. Additionally, the localization of the SCEL protein to the inner part of the cell membrane and cytosolic region may suggest the potential SCEL function associated with membrane receptors of LC cells.

Fig. 3

SCEL is a membrane associated protein that are exclusively expressed in the LC sublines derived from the metastatic lung nodules. A The protein expression of SCEL was compared in the multiple clones of MDA-MB-231-PT cells, and its sublines IV2, and LC cells isolated from the TNBC xenograft mouse model. Lower panel. The protein expression of SCEL was compared in a series of BC cell lines including luminal, HER2, and TNBC subtypes. B Immunohistochemistry staining of SCEL expression in LC cells-derived lung metastatic nodules in the tail vein injection mouse model. C Analysis of SCEL localization in LC cells-derived lung nodules using Image J software. Three nodules per mouse and a total of five mice were included in the analysis. Membrane staining of SCEL indicated by black arrows. D Subcellular localization analysis of SCEL in LC cells by protein fractionation. E The confocal immunofluorescence images of the HA-SCEL-transfected Hs578T cells. Membrane localization of SCEL indicated by white arrows. Scale bar, 25 μm. All cell-based experiments were performed in triplicates and repeated at least three times. The representative results are shown. *P < 0.01

Downregulation of SCEL expression significantly suppresses anchorage-independent growth and lung colonization potential of lung-tropic LC cells, as well as decreased metastasis-caused death resulting from LC cell-derived metastatic tumors

In order to address the possible role of SCEL in lung metastatic colonization of TNBC cells, we first investigated the effect of SCEL knockdown on the in vitro metastasis-related traits of cells including motility, invasiveness and 3D colony-forming ability. Western blotting confirmed the shRNA knockdown of SCEL in two stable SCEL knockdown LC lines (Fig. 4A). The Transwell migration and invasion assay showed that downregulation of SCEL expression did not significantly affect cell motility and invasiveness as compared to the control cells (Fig. 4B). On the other hand, the SCEL-downregulated cells showed an impaired ability for anchorage-independent growth when compared with the control cells (Fig. 4C). Based on these observations, we next examined whether SCEL downregulation could negatively affect the lung colonization potential of LC cells in a mouse model. The experimental lung metastasis mouse model was then performed by injecting the control or the SCEL-downregulated LC cells into the mouse tail vein and examined for lung metastasis one month later. As shown in the upper panel of Fig. 4D, the lungs resected from the mice injected with the control LC cells displayed many visible metastatic nodules, while the lungs from those injected with the SCEL-downregulated LC cells displayed relatively normal appearance. HE staining of the mouse lung sections revealed that LC cells effectively formed several sizable nodules in the mouse lung parenchyma (indicated by white arrows). In contrast, the SCEL-downregulated LC cells only formed a few minuscule metastatic lesions in the lungs (indicated by black arrows, Fig. 4D). Analysis of the area of lung parenchyma that was taken up by metastatic tumors showed that the SCEL-downregulated cells formed 90% fewer metastatic lesions than the control cells (Fig. 4E). Moreover, all mice injected with the SCEL-downregulated LC cells were able to survive through the experiment, while those injected with the control LC cells all died when reaching the endpoint (Fig. 4F). Taken together, our data indicated that SCEL played an essential role in lung metastatic colonization of TNBC cells but did not affect migration and invasion ability assayed in vitro. These observations suggest the effect of SCEL on lung metastasis may be subsequent to the arrival of cancer cells at lungs.

Fig. 4

Knockdown of SCEL expression significantly reduces lung colonization potential of lung-tropic LC cells and prevents metastasis-caused death in a mouse model. A Western blot analysis of SCEL stable knockdown LC cells. B Transwell migration and invasion analysis of the control LC cells and two SCEL stable knockdown cells. C Analysis of anchorage-independent growth of SCEL-downregulated LC cell and the control LC cells using 3D-colony-forming assay. *P < 0.05. D Upper images showed the mouse lungs dissected from SCID mice intravenously injected with the control LC cells (n = 4) or SCEL-downregulated LC cells (n = 4). Bottom images showed HE staining of the mouse lung sections from the control group and the SCEL knockdown group. E Quantitative analysis of lung metastasis of mice intravenously injected with the control LC cells or SCEL-downregulated LC cells. *P < 0.001. F The Kaplan-Meir survival analysis of mice intravenously injected with the control LC cells or SCEL-downregulated LC cells. For the cell-based experiments, each experiment was performed in triplicates and was repeated at least 3 times

Knockdown of SCEL expression significantly impairs TNF-α-induced NF-κB activation/nuclear translocation leading to the inhibition of downstream anti-apoptotic and proliferation genes

Upon reaching distant organ such as the lungs, disseminated cancer cells immediately face the survival challenges exerted by the new environment including lack of cell–cell adherence and nutrient providing niche. Here, we focused on growth factors and cytokines known to be present in the lung parenchyma where fibroblasts and macrophages are the main source of growth factors such as IGF-1, EGF and HGF, as well as inflammatory cytokines such as TNF-α and IL-6. In particular, TNF-α secreted from proinflammatory macrophages can act as a double-edged sword in the context of metastasis [22, 23]. Therefore, we compared the status of the signaling pathways induced by those factors in SCEL-downregulated LC cells and control LC cells, and examined if SCEL depletion could cause the cells to respond differently to those environmental cues. As shown in Additional file 1: Fig. S3, depletion of SCEL did not affect IGF-1, EGF and HGF-mediated growth promoting signaling in the LC cells. Notably, we observed a dramatic reduction of p65 phosphorylation 30 min after TNF-α stimulation in the SCEL-downregulated LC cells as compared with the control, while depletion of SCEL did not affect IL-6-induced Stat3 oncogenic signaling (Additional file 1: Fig. S3). We further confirmed this finding with multiple stable SCEL knockdown cells to show that depletion of SCEL indeed attenuated TNF-α-induced p65 phosphorylation (Fig. 5A). Given the known dual role of TNF-α in regulating pro-survival and pro-apoptotic signals in cancer cells [24, 25], we further explored the effect of SCEL on the molecular signaling regulated by TNF-α in lung-tropic LC cells. To do so, we transfected the LC cells with the NF-κB luciferase reporter plasmids followed by TNF-α stimulation. The NF-κB reporter assay showed similar results that TNF-α treatment significantly increased the luciferase activity in the NF-κB reporter plasmid-transfected LC cells as compared to the SCEL-downregulated LC cells (Fig. 5B). Kinetic analysis of p65 phosphorylation revealed that the control LC cells showed strong p65 activation upon TNF-α stimulation at the early time point of 5 min, and the p65 activation persisted through 48 h (Fig. 5C). However, SCEL-downregulated LC cells showed impaired p65 activation upon TNF-α stimulation at the early time point, and the activation dwindled down prematurely one hour later (Fig. 5C). Given that p65 phosphorylation triggers its nuclear translocation and transcription activation function, we next examined if SCEL depletion suppressed the nuclear translocation of p65 upon TNF-α stimulation. The results of immunofluorescence staining showed that abundant nuclear translocation of the phosphorylated p65 was observed in the control LC cells treated with TNF-α. On the contrary, the fluorescent intensity of the phosphorylated p65 was significantly reduced in SCEL-downregulated LC cells upon TNF-α stimulation, and so was the nuclear translocation of the phosphorylated p65 (Fig. 5D, E). Additionally, subcellular fractionation experiment confirmed that depletion of SCEL significantly inhibited the nuclear translocation of phosphorylated p65 upon TNF stimulation (Fig. 5F). TNF-α-induced NF-κB nuclear translocation has been known to activate the downstream anti-apoptotic genes/proliferation genes to promote cancer cell survival and growth in various human cancers [26]. Therefore, we examined if depletion of SCEL could impair the expression of the TNF-α/NF-κB-driven anti-apoptotic and cell proliferation genes in response to TNF-α. Quantitative real-time PCR (qRT-PCR) analysis showed that the expression of the NF-κB-driven anti-apoptotic genes including IER3 (1EX-1L), TGM2 (TGM2), and CFLAR (c-FLIP) as well as the NF-κB-driven proliferation genes such as CCND1 (Cyclin D1) and MYC (Myc) were found to be increased in TNF-α-treated LC cells. However, these genes did not show significant elevation in SCEL-downregulated LC cells treated with TNF-α (Fig. 5G, H), Taken together, our data showed that knockdown of SCEL could impair TNF-α/NF-κB-driven anti-apoptotic and proliferation signals in the lung-tropic TNBC cells.

Fig. 5

Knockdown of SCEL expression significantly decreases TNF-α-induced NF-κB activation and the expression of NF-κB-driven anti-apoptotic genes. A Western blotting analysis of p65 activation upon TNF-α stimulation in multiple SCEL stable knockdown LC clones and the control LC cells. B Analysis of NF-κB reporter assay using the SCEL stable knockdown LC cells and the control LC cells. *P < 0.001. C Analysis of TNF-α-dependent p65 activation kinetics in SCEL-downregulated LC cell and the control LC cells. *P < 0.01. D The representative IF images of TNF-α-induced p65 phosphorylation and nuclear translocation in the SCEL-downregulated LC cells and the control LC cells. Scale bar, 20 μm. E Quantitative result of phosphorylated p65 nuclear translocation. *P < 0.01. F Analysis of nuclear translocation of phosphorylated p65 in the control cells and the SCEL-downregulated cells upon TNF-α stimulation using subcellular fractionation. G qRT-PCR analysis of anti-apoptotic gene expression in response to TNF-α-stimulation in the control LC and SCEL-downregulated LC cells. H qRT-PCR analysis of proliferation-related gene expression in response to TNF-α-stimulation in the control LC and SCEL-downregulated LC cells. *P < 0.01. Each experiment was performed in triplicates and was repeated at least 3 times

SCEL is required for regulating the switches between TNF-α-mediated cell survival and apoptosis in lung-tropic TNBC cells

TNF-α binding to its receptor TNFR1 has pleiotropic effects that could trigger two distinct cellular signals: the pro-survival pathway, which is governed by the complex I (TNFR1/TRAF2) that activates NF-κB-driven anti-apoptotic pathway, and the pro-apoptotic pathway, which is regulated by the complex II (FADD/Caspase) that induces caspase signaling cascade [24, 26, 27]. Given that our data showed that SCEL depletion significantly impaired 3D colony-forming ability, suppressed in vivo lung colonization capability and blocked TNF-α-induced NF-κB pro-survival signals of LC cells, we inquired about whether SCEL depletion could switch TNF-α-induced cell survival to apoptosis in the TNBC cells. Thus, we examined the effect of SCEL depletion on the expression of the major effector of TNF-α/NF-κB-activated anti-apoptotic pathway, which is the anti-apoptotic protein, c-FLIP. Two isoforms, the long form, c-FLIPL, and the short form, c-FLIPS, are predominantly expressed in human, and both can function to inhibit caspase 8-mediated apoptosis cascade. The results of western blotting showed that the expression of c-FLIPL/c-FLIPS was significantly elevated in LC cells after 24-h of TNF-α treatment as compared to SCEL knockdown LC cells (Fig. 6A), which confirmed the qRT-PCR result (Fig. 5F). Moreover, we found that TNF-α treatment significantly induced Akt phosphorylation at the 30-min time point, and downstream Erk1/2 activation was persisted up to 48 h in the control LC cells. Conversely, SCEL-downregulated cells exhibited impaired Akt phosphorylation upon TNF-α treatment as compared to the control LC cells (Fig. 6A). However, downstream Erk1/2 activation was unaffected during the early time points (30 min and 1 h), but its activation was reduced 24 h after TNF-α stimulation (Fig. 6A). Furthermore, we found that SCEL-downregulated cells exhibited diminished Bc1-2 phosphorylation and a decrease in the total Bcl-2 protein at the steady state level compared to the control cells (Fig. 6A). SCEL-downregulated cells showed a reduction in Bcl-2 at both phosphorylation and total protein levels during the 48-h TNF-α treatment, and displayed a significant increase in cleaved caspase 3 as compared to the control cells (Fig. 6A). In addition, the immunofluorescence (IF) staining confirmed a significant increase in cleaved caspase 3 in TNF-α-treated SCEL-downregulated LC cells as compared to TNF-α-treated control cells (Fig. 6B, C). Next, we examined if depletion of SCEL could affect cell growth in the soft agar culture and monolayer culture in response to TNF-α stimulus. The 3D colony-forming assay showed that TNF-α treatment promoted the colony-forming ability of the control LC cells. Conversely, TNF-α impaired the 3D colony-forming ability of SCEL-downregulated LC cells (Fig. 6D). The MTS cell proliferation assay showed that LC cells displayed an increased cell proliferation in response to TNF-α treatment (Fig. 6E), while TNF-α treatment significantly impaired the cell proliferation of the SCEL-downregulated LC cells (Fig. 6E).

Fig. 6

Loss of SCEL switches TNF-α-mediated cell survival to apoptosis in the lung-tropic metastatic TNBC cells. A Western blotting analysis of TNF-α-induced NF-κB-mediated pro-survival/apoptotic and proliferation signals in SCEL stable knockdown LC and the control LC cells. B The representative immunofluorescence (IF) images of cleaved caspase 3 in TNF-α-treated control LC and SCEL stable knockdown LC cells. Scale bar, 20 μm. C Quantitative result of the IF staining of TNF-α-induced cleaved caspase 3 in TNF-α-treated control LC and SCEL stable knockdown LC cells. *P < 0.01. D Evaluation of the capacity for anchorage-independent growth in control LC and SCEL-downregulated LC cells following TNF-α stimulation. *P < 0.05. E MTS cell proliferation analysis of the control LC and SCEL-downregulated LC cells in the presence or absence of TNF-α. *P < 0.05. F Western blotting analysis of TNF-α-induced NF-κB-mediated pro-survival/apoptotic and proliferation signals in SCEL-overexpressing 231 and the control 231 cells. G The representative immunofluorescence (IF) images of cleaved caspase 3 in TNF-α-treated SCEL-overexpressing 231 and the control 231 cells. Scale bar, 20 μm. H Quantitative result of the IF staining of TNF-α-induced cleaved caspase 3 in TNF-α-treated SCEL-overexpressing 231 and the control 231 cells. *P < 0.01. I Evaluation of the capacity for anchorage-independent growth in SCEL-overexpressing 231 cells and the control 231 following TNF-α stimulation. *P < 0.05. J MTS cell proliferation analysis of SCEL-overexpressing 231 and the control 231 cells in the presence or absence of TNF-α. *P < 0.05. Each experiment was performed at least three times

Subsequently, we investigated whether overexpression of SCEL could enhance the TNF-α-induced anti-apoptotic and proliferation signals in the parental 231 cells. First, 231 cells were transduced with an SCEL-expressing lentiviral vector, followed by puromycin selection to establish stable SCEL-expressing 231 cells. We then examined the effect of SCEL overexpression on TNF-α-induced anti-apoptotic and proliferation signals. Overexpression expression of SCEL could significantly promote TNF-α-induced p65 and Akt/Erk1/2 phosphorylation and also maintained the level of Bcl-2 phosphorylation at the early timepoints of TNF-α treatment (Fig. 6F). Furthermore, overexpression of SCEL could significantly increase the protein levels of endogenous Bcl-2 (Fig. 6F). The SCEL-expressing 231 cells exhibited elevated cFLIP (c-FLIPL/c-FLIPS) protein expression after 24 h of TNF-α treatment as compared to the control 231 cells (Fig. 6F). Additionally, overexpression of SCEL effectively suppressed TNF-α-induced caspase 3 activation (Fig. 6G). The soft agar assay showed that the ectopic expression of SCEL significantly promoted the ability of anchorage-independent growth (AIG) of 231 cells (Fig. 6I). Notably SCEL-expressing cells exhibited the highest AIG capacity in response to TNF-α treatment. Conversely, the control cells displayed impaired AIG in the presence of TNF-α. In addition, TNF-α treatment elicited a modest promotion of cell proliferation in SCEL-expressing cells while inhibiting the growth of cells in a monolayer setting.

Collectively, our findings suggest that SCEL is required for regulating the switches between TNF-α-induced pro-survival and apoptosis, and it may play a crucial role in mediating the TNF-α-induced NF-κB/c-FLIP survival axis in TNBC cells.

SCEL interacts with TNFR1 to promote its protein stability

Previous studies have shown that the membrane localization of TNFR1 is important for sustaining complex I/NF-κB pro-survival machinery, and TNFR1 protein instability triggers complex II-mediated apoptotic signals [24]. Thus, we examined whether depletion of SCEL in lung-tropic LC cells had an effect on the stability of TNFR1 protein and complex II protein such as FADD and caspase 8. First, we assessed if TNFR1 protein expression was altered in SCEL-downregulated LC cells. The result of western blotting revealed a reduction in TNFR1 protein levels, concomitant with elevated levels of complex II protein, FADD and caspase 8, in the SCEL-downregulated cells as compared with the control LC cells (Fig. 7A). Next, we investigated if SCEL depletion influenced TNFR1 protein stability. The control LC cells and SCEL-downregulated LC cells were treated with medium containing 20 μg/mL cycloheximide to block protein synthesis. Cells were harvested at the indicated time points and subjected to western blotting analysis. The results showed that TNFR1 protein stability significantly decreased in the SCEL-downregulated as compared to the control LC cells (Fig. 7B). TNFR1 protein degradation kinetics analysis revealed that TNFR1 in SCEL-downregulated cells degraded much faster than that in the control cells (Fig. 7C). In SCEL-downregulated cells, about 40% and nearly 100% of TNFR1 was degraded, respectively, after 1-h and 6-h of CHX treatment. On the other hand, in the control cells, less than 5% of TNFR1 was degraded after 1-h of CHX treatment, and 40% of TNFR1 was still present at the 8-h time point. Next, we examined whether overexpression of SCEL had effect on the stability of TNFR1 protein and the levels of TNFR1 complex II in the parental 231 cells, which have no SCEL expression. The result of western blotting showed that overexpression of SCEL increased the stability of TNFR1 and reduced the levels of TNFR1 complex II, FADD and caspase 8 (Fig. 7D). Meanwhile, SCEL-overexpressing 231 cells showed increased TNFR1 protein stability following CHX treatment as compared to the control 231 cells (Fig. 7E). Analysis of protein degradation kinetics of TNFR1 demonstrated that overexpression of SCEL significantly reduced the degradation rate of TNFR1 following CHX treatment (Fig. 7F).

Fig. 7

SCEL interacts with TNFR1 and plays a crucial role in maintaining its protein stability. A Western blotting analysis of protein expression of TNFR1 complex II in the SCEL stable knockdown LC cells and the control LC cells. B Protein degradation analysis of TNFR1 in the control LC cells and the SCEL-downregulated cells following 20 μg/mL cycloheximide (CHX) treatment. C Quantification of protein degradation kinetics of TNFR1 in the control LC and the SCEL-downregulated cells following 20 μg/mL cycloheximide (CHX) treatment. *P < 0.05. D Western blotting analysis of protein expression of TNFR1 complex II in SCEL-overexpressing 231 and the control 231 cells. E Protein degradation analysis of TNFR1 in in SCEL-overexpressing 231 and the control 231 cells in response to 20 μg/ml cycloheximide (CHX) treatment (top). Quantification of protein degradation of TNFR1 (bottom). *P < 0.01. F Analysis of protein interaction between SCEL and TNFR1 using reciprocal co-immunoprecipitation (co-IP) G Analysis of protein interaction between SCEL and TNFR1 using reciprocal co-immunoprecipitation (co-IP). H Examination of interaction between SCEL and TNFR1 through GST pull-down assay. Each experiment was repeated at least 3 times

These results led us to hypothesize the possibility that the presence of SCEL could stabilize TNFR1 through protein interaction. To test this idea, we then performed co-immunoprecipitation to test interaction between SCEL and TNFR1. To do so, we transfected HA-tagged SCEL into LC cells and 48 h later harvested the cells for protein extraction and analysis of co-immunoprecipitation (Co-IP) using control IgG, anti-HA antibody or anti-TNFR1, respectively. The Co-IP result showed that TNFR1 could be pulled down with HA-SCEL using anti-HA antibody, suggesting the potential interaction between TNFR1 and SCEL. The TNFR1-SCEL interaction was confirmed by the reciprocal co-IP using anti-TNFR1 antibody (Fig. 7G). Furthermore, GST pull-down assay was carried out to investigate the protein interaction between SCEL and TNFR1. To do so, the recombinant GST-SCEL fusion protein was prepared using E.-coli BL-21 strain (please refer to “Method” section) followed by purification. Next, His-tagged TNR1 expression construct was introduced into 293 T cells, and cell lysate was harvested for the in vitro protein binding by incubation with GST-SCEL fusion protein. Subsequently, protein complex was pulled down using Glutathione Sepharose beads. Western blotting analysis of GST pull-down assay indicated that His-tagged TNFR1 could be successfully detected in the GST-SCEL pull-down sample as compared to the control GST pull-down sample (Fig. 7H). Coomassie blue staining indicated equivalent loading amounts of the control GST and GST-SCEL fusion proteins (Fig. 7H).

Together, our results suggest that SCEL may interact with TNFR1 and the loss of SCEL could lead to instability of the TNFR1 protein, which could potentially explain why LC cells that lack SCEL are prone to undergo apoptosis in the presence of TNF-α.

Depletion of SCEL significantly inhibits cell proliferation and induces apoptosis in the lung metastatic loci

Next, we examined whether the effect of SCEL depletion on the TNF-α mediated survival could be also observed in the metastatic loci of the mouse lung. To do so, we first performed IHC staining to observe the presence of tumor-associated macrophages (TAMs) including CD68+ M1 and CD206+ M2 macrophages in the lung metastatic loci. CD68+ M1 macrophages, known for its tumor-inhibiting ability, are the major TNF-α-producing immune cells in the tumor microenvironment [28]. The result of IHC staining showed that an increased number of CD68+ M1 macrophages were observed in the lung metastatic loci derived from the control LC cells as compared to that of in the lung metastatic loci derived from SCEL-downregulated LC cells (Fig. 8A, B). In addition, we also observe an increased number of CD206+ M2 macrophages, the tumor-promoting macrophages, in the lung metastatic loci of the control LC cells as compared to that of in SCEL-downregulated LC cells (Fig. 8A, B). These results indicated that depletion of SCEL in LC cells significantly reduced tumor-associated macrophage (TAMs) at the lung metastatic loci. Notably, elevated expression of cleaved caspase 3 and decreased expression of Ki67 were observed in the SCEL-downregulated LC cells-derived lung nodules as compared to those derived from the control LC cells (Fig. 8A, B). Furthermore, we demonstrated that shSCEL cells-derived lung nodules showed a decrease in TNFR1 expression as compared to the control cells-derived lung nodules whereas the levels of TNF-α were similar in both shSCEL and control nodules (Fig. 8A, B). Our in vivo evidence suggested SCEL expression is essential for the progression of metastatic lung colonization of TNBC through inhibiting apoptosis and promoting proliferation.

Fig. 8

Depletion of SCEL suppresses cell proliferation and induces apoptosis in the lung metastatic loci. A Representative images of immunohistochemistry staining (IHC) for SCEL, CD206, CD68, cleaved caspase 3, Ki67 expression, TNFR1, and TNF-α of the metastatic lung nodules derived mice injected with the control LC and SCEL-downregulated LC cells, respectively. B Quantification of IHC staining of SCEL, CD206, CD68, cleaved caspase 3 and Ki67, TNFR1, and TNF-α expression in the metastatic lung nodules using Image J. *P < 0.05

The protein expression of SCEL significantly correlates with TNFR1 protein levels in TNBC specimens

Next, the clinical correlation between the protein expression levels of SCEL, TNF-α, and TNFR1 in TNBC tissues was investigated. IHC analysis demonstrated a significant correlation between the expression of SCEL and TNRF1. Pearson’s correlation coefficient analysis showed that elevated SCEL expression was significantly associated with increased TNFR1 levels, whereas there was no correlation between SCEL expression and TNF-α (Additional file 1: Fig. S4B). or between TNFR1 expression and TNF-α expression (Additional file 1: Fig. S4C). Together, IHC analysis of clinical TNBC tissues supports our cell-based and mechanistic study that SCEL expressio