Characteristics of primary cell from the enrolled patients
The primary cells of tumor tissue from 24 patients with endometrial cancer were collected in this study, and their clinicopathological features are shown in Table 1. Among them, 91.67% (22 cases) were pathologically classified as endometrioid adenocarcinoma, 91.67% (22 cases) were positive for ER (estrogen receptor) expression, 70.83% (17 cases) were in the early stage (stage I) of endometrial cancer, and 29.17% (7 cases) were in the advanced stage (stage II-IV). The proportion of patients with pathological grading of G1, G2 and G3 was 20.83%, 45.83% and 33.33%, respectively. There were 8 patients with MELF infiltration (33.33%) and 4 patients with cervical infiltration (16.67%). In addition, 45.83% of the patients with LVSI positive, 8.33% of the patients with lymph node metastasis positive, and 16.67% of the patients with fallopian tubes and ovaries involved. The traction force difference between primary EC cells and normal endometrial cells was further measured and analyzed by single-cell traction test (Figure 1A). It was found that the traction force of primary EC cells in 24 cases was 79.25 ± 11.25Pa, and that of normal endometrial primary cells was 208.0 ± 28.15Pa, which was significantly higher than that of primary EC cells (Figure 1B). Area under ROC curve (AUC) analysis was used to evaluate accuracy of TF value in determining the nature of endometrium. The results showed that the AUC was 0.802, 95% CI was 0.670-0.933, p<0.001 (Figure 1C). The results showed that the TF value of different primary cells could be used as an predictive tool for predicting EC. What’s more, we also compare the TF value between patients with different metastatic features. The results indicated that primary metastatic cell performed lower value of TF (Figure S1). Taken together, these findings suggest that the higher the metastatic ability of EC cells, the lower the TF value, and TF value is a predictor for judging the invasion and metastasis for EC.
Associations between clinicopathological features and TF
To further investigated the associations between TF values and clinical features in EC patients, we compare the TF values in different clinicopathological characteristics (Table 2), including FIGO stage, MELF (Microcystic, elongated, and fragmented), histological type, ER (estrogen receptor), CI (cervical invasion), ovary involvement, tumor grade, and LVSI (lymph-vascular space invasion). As shown in Figure 2, there was no significant difference in primary cell traction in different histological types and ER expression. However, the traction force of primary cells in FIGO stage (stage I vs stage II-IV), MELF, CI, ovary involvement, grade, and LVSI were significantly different, and the traction force of primary cells with late stage, positive MEFL, positive CI, ovary involvement, higher grade, and positive LVSI were significantly lower than that of the control groups (all p <0.05). These results collectively indicated that poor clinical features showed lower TF values.
High TF promotes the metastatic capability for EC cells
In order to study the effect of cell traction on its invasion and metastasis ability, four representative EC cell lines, naming ishikawa, HEC-1-A, HEC-50B, and AN3CA were selected according to the tissue origin, pathological characteristics and pathological grading of EC cell lines. The results of single-cell traction test showed that HEC-1-A cells were taken as the control group. The migration ability of the four kinds of cells was first studied by scratch test, as shown in Figure 3A. It was found that the migration ability of HEC-50B cells and AN3CA cells with low traction force was significantly stronger than that of ishikawa cells and HEC-1A cells with higher traction force, and the migration ability of AN3CA cells with the lowest cell traction force was the strongest, which is statistically significant (p<0.05). Transwell invasion experiment is also used to investigate the invasive ability of EC cells. and the results are shown in Figure 3B. The invasive ability of HEC-50B cells and AN3CA cells is significantly stronger than that of ishikawa cells and HEC-1A cells, which is also statistically significant. We then compare the traction force among the four cell lines. There is no significant difference in traction force between ishikawa cells and HEC-1-A cells (57.43 ± 19.52Pa vs 103.9 ± 27.40Pa) (p>0.05). Towever, the traction force of HEC-50B cells (37.34 ± 9.933Pa) and AN3CA cells (17.97 ± 10.48Pa) was significantly lower than that of HEC-1-A cells. To further explore whether TF features were correlated with progesterone resistance, we investigated their TF values in progesterone sensitive and resistance groups. We found that the progesterone resistance could obviously increase the invasion and migration abilities of EC cells (Figure 3D-E). What’s more, traction force measured by TFM also indicated that cells in progesterone resistance groups performed lower TF values (Figure 3F). Together, these data demonstrated that cell lines with higher metastatic ability had lower TF values.
Bioinformatics analysis for SLC8A1 in EC
Our previous study suggested that SLC8A1 played an essential role in mechanical-stimulus induced progression for EC patients. Therefore, we then explored whether the TF promote metastasis through SLC8A1 or not. We divided total patients into the low- and high-SLC8A1 group by its median threshold. We discovered that the high expression of SLC8A1 group possessed worse survival than the low expression group (Figure4A, P=0.02). Then we investigated the association between SLC8A1 and different clinical variables. The results showed that EC patients, higher EC stage, and higher EC grade were prone to have a high expression of SLC8A1 (Figure 4B-D). What’s more, expression of SLC8A1 also increased in positive lymph node metastasis, positive peritoneal cytology, dead status, and histological type of SEA (Figure S2A-D). We then selected DEGs between the low- and the high-SLC8A1 group with |Log FC|≥1 and FDR < 0.05. Finally, 445 DEGs were identified and these genes were plotted in heatmap and volcano (Figure 4E-F). As for functional analysis in Figure 4G, we found that most DEGs were enriched in regulation of ion transmembrane transport, calcium ion homeostasis, cellular calcium ion homeostasis, and regulation of membrane potential by GO analysis. KEGG pathway analysis showed that DEGs were mostly enriched in neuroactive ligand−receptor interaction, calcium signaling pathway, and Cell adhesion molecules (Figure 4H). In summary, these results indicated that SLC8A1 may act as a biomarker for the progression and survival for patients with EC.
Validation of SLC8A1 expression and function in patients in PKUPH
In order to verify the effect of SLC8A1 on the survival rate of EC and distribution of SLC8A1 in different clinicopathological types, we selected 24 EC patients from our hospital for verification. Survival analysis indicated that high expression of SLC8A1 is also associated with worse prognosis for EC patients (Figure 5A). What’s more, SLC8A1 expression is higher in patients with high grade, positive LNM, positive peritoneal cytology, and positive LVSI (Figure 5B-E). Functional analysis revealed that DEGs from high and low expression of SLC8A1 are enriched in regulation of membrane potential, regulation of Wnt signaling pathway, filopodium, and ion channel activity (Figure 5G). KEGG results suggested that these genes are enriched in Wnt signaling pathway, calcium signaling pathway, MAPK signaling pathway, and Hippo signaling pathway et al (Figure 5H). These findings illustrate that SLC8A1-related genes played an important role in the progression of EC, and Wnt signaling pathway may be induced by SLC8A1.
SLC8A1 induces the metastasis in EC by mediating Wnt signaling pathway and TF
In order to further study the effect of SLC8A1 on the function of EC cells, this study further investigated the expression of SLC8A1 in EC cell lines, including HEC-1-A, ishikawa, HEC-50B, and AN3CA. The results showed that the expression of SLC8A1 was the highest expression in ishikawa cell lines, with a more significantly different than other cell lines (Figure 6A). Thus, we chose ishikawa cell line to create knockdown and overexpression cell line, respectively. Next, we knockdown and overexpress SLC8A1 in ishikawa cells by transfecting small interferon plasmid. The transfection efficiency was shown in Figure 6B by western blot, and the results showed that the plasmid could significantly decrease or increase the expression of SLC8A1 compared with the control group. We further investigated whether SLC8A1 affected metastasis of EC cells. Transwell assay showed the invasive ability of ishikawa cells in SLC8A1 low expression group were significantly weakened compared with the control group, and its ability was enhanced in high expression group (Figure 6C). The migration speed of si-SLC8A1 in ishikawa cells was remarkedly slower, and that of OE-SLC8A1cells was significantly faster control group (Figure 6D). In order to study the regulatory mechanism of SLC8A1 on the mechanical stimulus of EC cells, we further explored the effect of SLC8A1 on the traction force in ishikawa by single-cell traction test. The results indicated that the traction force of endometrial cancer cells significantly decreased after overexpression of SLC8A1, and increased in the knockdown group of SLC8A1 (Figure 6E). Our previous study suggested that mechanical force and TF are closely related to cytoskeleton in cellular surface. Therefore, associations between structure or expression of cytoskeleton, F-actin, and expression of SLC8A1 are analyzed. Immunofluorescence results suggested that si-SLC8A1 and OE- SLC8A1 groups could obviously decrease or increase expression of F-actin (Figure 6F). According to enrichment of SLC8A1-associated DEGs, we investigate whether the downstream of Wnt-β-catenin pathway have an effect on F-actin. Western blot showed that the expression of Wnt, β-catenin, and F-actin decreased compared with si-SLC8A1 group. We also confirmed that overexpression of SLC8A1 is up-regulated, while its expression was significantly downregulated after adding the inhibitor of Wnt (Figure 6G). Our data strongly suggested that SLC8A1 could promote the progression and mechanical force, especially F-actin of EC through Wnt-β-catenin pathway.