The Nectin‑4 expression and clinical significance in OS
Nectin 1~3 were commonly enriched in normal adult tissues, while Nectin-4 was specifically up-regulated in various tumors and played significantly role in tumor occurrence and development (Deng et al. 2019, Takano et al. 2009, Derycke et al. 2010, Lin et al. 2019). Nevertheless, we had little idea about the role of Nectin-4 protein in OS. Firstly, we were eager to monitor the expression levels of Nectin-1~4 in OS tissues and adjacent normal tissues. By using RT-qPCR, we noticed that the expression of Nectin-1~3 had no significant differences between OS and normal tissues (Fig. 1A-1C). As expected, we found that Nectin-4 expression was markedly increased in OS tissues compared to normal tissues (Fig. 1D). Similarly, the significant difference in expression of Nectin-4 was also observed in 8 paired samples of OS and adjacent normal tissues both at mRNA and protein level (Fig. 1E and 1F).
In order to further investigate the expression difference of Nectin-4 between OS tissues and normal tissues, and the association between Nectin-4 and clinical features, we downloaded the related gene expression profiles and clinical data from TARGET, GTEx, and GEO database. Interestingly, the mRNA level of Nectin-4 displayed obvious overexpression in OS tissues using TARGET database compared with that in the normal muscle tissue obtained from GTEx database (Supplementary Fig. 1). Furthermore, to evaluate the association between the gene expression of Nectin-4 and clinical features in OS, we re-assayed the data acquired from the GSE21257. The mean value of Nectin-4 mRNA expression was regarded as the threshold and then the patients were divided into two cohorts (high expression cohort and low expression cohort). As shown in Table 1, we found that the Nectin-4 high expression was markedly associated with tumor metastasis (P<0.05). Strikingly, we found that more patients in the high expression group (16/19, 84.2%) exhibited tumor metastasis compared with those in the low expression group (18/34, 52.9%), which implies that Nectin-4 might regulate the tumor progression in OS.
To determine the differential expression of Nectin-4 in diverse histologic stages of OS, Nectin-4 IHC was performed on an OS TMA. Overall, 48/70 (68.6%) tumors expressed Nectin-4, with 30/48 (62.5%) tumors showing high level expression and 18/48 (37.5%) low level. High expression of Nectin-4 occurred most commonly in stage IVB OS tissues (66.7%, 4/6), followed by stage IIB (51.6%, 16/31), stage IIA (32.3%, 10/31), and stage IA OS tissues (0%, 0/2), which indicate that high expression of Nectin-4 is closely correlated with higher OS stages (Fig. 1G).
Next, three diverse OS cell lines (MG63, U2OS and 143B) and one osteoblastic cell line (hFOB1.19) were assessed for basal expression levels of Nectin-4 by RT-qPCR and Western blotting. The results demonstrated that Nectin-4 expression in MG63, U2OS, and 143B cell lines was significantly higher than that in hFOB1.19 cell lines (Fig. 1H and 1I). The highest up-regulation of Nectin-4 was found in 143B cells, followed by MG63 and U2OS cell lines. Therefore, the 143B cell line was selected for Nectin-4 knockdown, and MG63 and U2OS cell lines were selected for Nectin-4 overexpression. Thus, we concluded that Nectin-4 is highly expressed in clinical OS tissues and OS cells.
Nectin-4 promotes OS cells proliferation
To discuss whether Nectin-4 had an influence on OS cell viability, we firstly constructed and sorted out the OS cell lines with stable Nectin-4 overexpression (Nectin-4-OE) or knockdown by means of lentivirus infection. As shown in the Fig. 2A, 2B and Supplementary Fig. 2A, we validated that Nectin-4 was successfully over-expressed in Nectin-4-OE group of U2OS and MG63 cells compared with vector negative control (Vector-NC) group cells at both mRNA and protein levels. Besides, to further validate the functional roles of Nectin-4 in OS cells, we also constructed the stable Nectin-4 knockdown OS cell line. As shown in the Fig. 2C, using RT-qPCR, we found that Nectin-4 was significantly down-regulated at mRNA level in 143B cells infected with Lenti-shRNAs, where shNectin-4#2 showed the most potently down-regulated effect compared with shNectin-4#1 and shNectin-4#3. Moreover, we validated the result at protein levels via Western blotting assays (Fig. 2D and Supplementary Fig. 2B). Thus, we chose the shNectin-4#2 cells for next investigation.
Next, we investigated the function of Nectin-4 on the cell proliferation via CCK-8 and colony formation assays. As presented in Fig. 2E and 2F, in comparison with the Vector-NC groups, up-regulation of Nectin-4 remarkably enhanced cell proliferation both in U2OS and MG63 cell lines at different time points (24, 36, 48 and 72h) (each, P<0.001). Moreover, we detected the effect of down-regulated Nectin-4 on OS cell viability. As presented in Fig. 2G, the proliferation rates of shNectin-4#2 group cells were remarkably lower than that of shCtrl group cells at different time points (24, 36, 48 and 72h) through CCK-8 trial in 143B cell line (P<0.01, respectively). In addition, we also found that up-regulation of Nectin-4 could noticeably facilitate colony formation both in MG63 and U2OS cell lines compared with those in the Vector-NC groups, suggesting that overexpression of Nectin-4 could markedly promote the OS cell proliferation (Fig. 2H and Supplementary Fig. 2C, Both in MG63 and U2OS, P<0.01). Inversely, down-regulation of Nectin-4 noticeably restrained colony formation in 143B cells compared with those in the shCtrl group (Fig. 2H and Supplementary Fig. 2C, P<0.001). In conclusion, these above results indicated that Nectin-4 could significantly modulate the human OS cells proliferation.
Nectin-4 enhances human OS cells migration and invasion in vitro
As mentioned above, a close connection between Nectin-4 up-regulation and tumor metastases was confirmed in the GSE21257. To further investigate the effect of Nectin-4 on cell metastasis, we firstly adopted the transwell assays to observe whether the overexpression of Nectin-4 could enhance the migratory and invasive capacity of MG63 and U2OS cells. Transwell assay without Matrigel demonstrated that much more migrated cells tinted with dark blue were observed in the Nectin-4-OE group cells than those in the Vector-NC groups (Fig. 3A, in MG63, P<0.01 and in U2OS, P<0.001). Similarly, using transwell assay with Matrigel, we confirmed that the number of invasive cells in Nectin-4-OE groups were over twice as those in the control groups (Fig. 3B, in MG63, P<0.001 and in U2OS, P<0.01), implying that Nectin-4 could enhance the OS cell movement ability. Furthermore, we also detected the migration and invasion ability of shNectin-4#2 143B cell lines via transwell assays. As shown in Fig. 3C, knockdown of Nectin-4 remarkably restrained the movement activity of 143B cells where the migrated cells and invaded cells were markedly cut down compared with shCtrl cells (P<0.01 and P<0.001, respectively).
In addition, the wound-healing assay was also carried out to validate the function of the Nectin-4 up-regulation on the migration ability in OS cell lines. As demonstrated in Fig. 3D and Supplementary Fig. 3, the wound width of the Nectin-4-OE group was remarkably shortened than that of the Vector-NC group (in MG63, P<0.01 and in U2OS, P<0.001), confirming that up-regulation of Nectin-4 could remarkably heighten the migratory capacity of the MG63 and U2OS cells. Taken together, up-regulation of Nectin-4 promoted the cell movement activity in human OS cell lines. Moreover, we further observed the influence of Nectin-4 down-regulation on the migration activity in 143B cells. Compared with shCtrl group, shNectin-4#2 group cells presented with a much broader wound width (Fig. 3D and Supplementary Fig. 3, P<0.001). In a word, these results implied that Nectin-4 played important roles in modulating the human OS cell proliferation and movement activities in vitro.
Nectin-4 modulates epithelial-mesenchymal transition (EMT) via activating PI3K/AKT/NF-κB signal pathway
It had been reported that EMT is a classical progress in tumor cell metastasis, and we discovered that Nectin-4 played important roles in promoting OS cell movement activities. To further investigate if and how Nectin-4 could promote OS metastasis, we evaluated the expression changes of typical EMT markers in OS cell lines both at mRNA and protein levels. As shown in Fig. 4A and 4B, we discovered that up-regulation of Nectin-4 led to a reduction of ZO-1 expression, an epithelial marker, and also notably elevated the expression of mesenchymal markers including Vimentin and N-Cadherin in U2OS and MG63 cells using RT-qPCR and Western blotting analysis, respectively. Furthermore, we also detected the upstream transcriptional regulators including Zeb1 and Slug of EMT signaling pathway. As displayed in Fig. 4A and 4B, the expression of Zeb1 and Slug was much higher in the group of Nectin-4-OE cells than those in the Vector-NC group cells. Additionally, the changes of EMT-related molecule markers were also observed in 143B cells infected with lenti-shNectin-4#2. In contrast with Nectin-4 up-regulation, knockdown of Nectin-4 obviously down-regulated the expression of Vimentin and N-Cadherin, and enhanced the ZO-1 expression (Fig. 4A and 4B). Similarly, EMT-related transcription factors (Zeb1 and Slug) also had the same trend with mesenchymal markers which were down-regulated when Nectin-4 was depleted (Fig. 4A and 4B). With all of these results together, we confirmed that Necin-4 exerted a crucial influence on regulating EMT.
Numerous studies confirmed that Nectin-4 could facilitate tumor growth, angiogenesis, and metastasis via phosphoinositide 3-kinase/protein kinase B (PI3K/AKT) signaling cascade in breast cancer, gallbladder carcinoma, and colorectal cancer (Zhang et al. 2016, Zhang et al. 2018). To evaluate whether Nectin-4 could activate PI3K/AKT pathway in OS, we performed both GSVA analysis and GSEA analysis. As demonstrated in Supplementary Fig. 4A, GSVA analysis showed that PI3K pathway was markedly enriched in group of highly-expressed Nectin-4. In addition, using GSEA analysis, we also found that PI3K pathway displayed positive correlation with Nectin-4 expression in OS Target dataset (Supplementary Fig. 4B). To investigate whether Nectin-4 was involved in regulating OS-related EMT via PI3K/AKT pathway, we observed the expression levels of AKT and p-AKT both in MG63 and U2OS cells. We found that the expression of p-AKT in Nectin-4-OE group was much higher than that in the Vector-NC group. On the contrary, knockdown of Nectin-4 obviously decreased the expression of p-AKT in OS cell lines (Fig. 4C). Besides, we also detected that up-regulation of p-NF-κB P65 (p-P65) in Nectin-4-OE group compared with Vector-NC group (Fig. 4C). Based on these findings, we speculate that Nectin-4 might modulate the EMT pathway through regulating PI3K/AKT/NF-κB pathway. Then, we managed to observe if we could adopt a classical AKT inhibitor named as LY294002 to reverse the effect of Nectin-4 overexpression. In our present research, we discovered that LY294002 successfully reversed the Nectin-4 overexpression effect on the expression of markers of EMT and PI3K/AKT pathway (p-AKT) (Fig. 4D). Moreover, to further validate that Nectin-4 could regulate EMT markers by PI3K/AKT/NF-κB signal pathway in OS cell lines. We picked out a sort of NF-κB inhibitor (PDTC) to observe its effects on Nectin-4-OE MG63 and U2OS cells. The results demonstrated that PDTC could block the efficiency of Nectin-4 overexpression on EMT-related markers and p-P65 (Fig. 4E). In addition, we detected that LY294002 could reverse the Nectin-4 overexpression effect on the expression levels of p-P65 (Fig. 4D). Functionally, we also revealed that LY294002 reversed the function of Nectin-4 up-regulation on migratory and invasive activities in both MG63 and U2OS cell lines (Fig. 4F, both P<0.01). Interestingly, we observed that the high levels of AKT and P65 expression when Nectin-4 was up-regulated, while down-regulation of Nectin-4 resulted in the opposite effect. Above all, we speculated that Nectin-4 could modulate the levels of AKT and p-65 to stimulate the PI3K/AKT/NF-κB signaling.
MiR-520c-3p is a downstream regulator that targets AKT1 and p65
Next, we were eager to clarify the deeper mechanism under which Nectin-4 modulated PI3K/AKT/NF-κB signal pathway. Given that miRNAs could play significant regulatory roles by targeting mRNAs for cleavage or translational repression, we focused on the effects of Nectin-4-regulated miRNAs that both targeted AKT and NF-κB. As AKT1 played crucial roles in PI3K/AKT signaling in OS, p65 served as the most important subunit of NF-κB complex (Zhu et al. 2014, Liu et al. 2020, Zhu et al. 2017, Zou et al. 2018). Potential miRNA targeting AKT1 and P65 were predicted by six online bioinformatics analysis software databases including TargetScan, miRanda, microT, miRmap, RNA22 and PITA (https://starbase.sysu.edu.cn/agoClipRNA.php?source=mRNA). After initial evaluation, as shown in Fig. 5A and 5B, a total of 82 and 59 miRNAs targeting AKT1 and P65 were selected, respectively. Subsequently, we screened out 28 miRNAs simultaneously targeting AKT1 and p65 which were generated by taking intersection of the two miRNAs clusters using Venn diagrams (Fig. 5C). Among them, we selected 12 miRNAs simultaneously predicted by at least three databases (Supplementary Table 1). Finally, we picked out 8 miRNAs (miR-302d-3p, miR-520d-3p, miR-302c-3p, miR-302b-3p, miR-520c-3p, miR-520a-3p, miR-520b, miR-302e) with relatively high reliability (context++ score＜-0.15 and weighted context++ sco ＜-0.15 ） in TargetScan software.
To assess which miRNA is the downstream regulator of Nectin-4, we evaluated these miRNAs expression levels by using RT-qPCR in stable Nectin-4 knockdown 143B cells. The results showed that miR-520c-3p was the microRNA with the most significantly up-regulated expression (≥2-fold) in Nectin-4 knockdown 143B cells (Fig. 5D). Moreover, the results of spearman’s correlation analysis revealed that Nectin-4 expression was significantly negatively correlated with the expression of miR-520c-3p (Fig. 5E), implying that miR-520c-3p is a downstream regulator of Nectin-4. Besides, the negative correlation between Nectin-4 and miR-520c-3p expression was also confirmed in TARGET database (Supplementary Fig. 5A). In addition, we also detected the miR-502c-3p expression in 29 OS specimens and 8 adjacent normal specimens by using RT-qPCR. Both in paired and unpaired specimens, the results revealed that the miR-520c-3p was apparently lower in OS specimens compared with adjacent normal specimens (Supplementary Fig. 5B and S5C). Furthermore, we attempted to analyze the difference in miR-520c-3p expression between primary and metastatic OS samples in TARGET database. We found that the expression of miR-520c-3p was significantly lower in OS metastatic samples than primary samples (Fig. 5F). Therefore, we proposed that miR-520c-3p was the downstream regulator of Nectin-4, which thus was selected for follow-up studies.
Nectin-4 activates PI3K/AKT/NF-κB signaling through modulating miR-520c-3p
According to the prediction analysis by TargetScan, we confirmed the binding sites of miR-520c-3p in the 3′-UTR of AKT1 and P65 mRNA (Fig. 6A). The luciferase reporters were co-transfected with miR-520c-sp mimic or inhibitor into HEK-293T cells. We found that miR-520c-3p up-regulation significantly suppressed the luciferase activity of AKT1 and P65 with Wt 3′-UTRs, while down-regulation of miR-520c-3p resulted in the opposite effect. However, the miR-520c-3p mimics or inhibitor had no effect on those with Mt 3′-UTRs (Fig. 6B and Supplementary Fig. 6A). To clarify miR-520c-3p function in OS, we firstly assessed the basal expression of miR-520c-3p in different OS cell lines and osteoblastic cell line by RT-qPCR (Supplementary Fig. 6B). Subsequently, the 143B cells (lowest levels of miR-520c-3p) were transfected with miR-520c-3p mimics (miR mimics). And the MG63 and U2OS cells (higher levels of miR-520c-3p) were transfected with miR-520c-3p inhibitor (miR inhibitor). We identified the transduction efficiencies in the OS cell lines by using RT-qPCR (Supplementary Fig. 6C). Moreover, we observed miR-520c-3p overexpression could significantly reduce the expression of AKT1, p-AKT1, P65, and p-P65 in 143B cells by using Western blotting and RT-qPCR. While miR-520c-3p knockdown resulted in the opposite results in the MG63 and U2OS cells (Fig. 6C and 6D). In summary, these results demonstrated that AKT1 and P65 are direct targets of miR-520c-3p in OS cells. As mentioned above, down-regulation of miR-520c-3p was closely related to OS metastases. To investigate the effect of miR-520c-3p on OS cell metastasis, we firstly detected the functional role of miR-520c-3p on the migratory capacity of OS cells by transwell assays. The results revealed that much fewer migrated cells were observed in the miR mimics group cells than those in the NC mimics groups. Conversely, the migratory capacity in miR-520c-3p inhibitor group cells was remarkably improved compared with the NC inhibitor group (Fig. 6E and Supplementary Fig. 6D). Then, to further clarify whether Nectin-4 could activate PI3K/AKT/NF-κB signaling mediated by miR-520c-3p in OS cells, miR mimics or miR inhibitor were respectively transfected into Nectin-4-OE U2OS cells and shNectin-4 143B cells, respectively. The results of Western blotting trials demonstrated that the up-regulation of miR-520c-3p successfully reversed the Nectin-4 overexpression effect on the expression of markers of EMT (Vimentin and Zo-1) and PI3K/AKT/NF-κB pathway (AKT1, p-AKT1, P65, and p-P65) (Fig. 6F). On the contrary, the miR-520c-3p silencing reversed the Nectin-4 knockdown effect on the expression of markers of EMT and PI3K/AKT/NF-κB pathway (Fig. 6F). Consistently, the transwell trials also showed that miR-520c-3p overexpression reversed the function of Nectin-4 up-regulation on migratory activities in U2OS cells. And the knockdown of miR-520c-3p could also reverse the effect of Nectin-4 silencing on migratory ability in 143B cells (Fig. 6G and Supplementary Fig. 6E). Given these findings, we validated that Nectin-4 could promote OS cells EMT and migration by activating PI3K/AKT/NF-κB signaling mediated by miR-520c-3p.
Nectin4 enhances OS cells tumorigenesis and lung metastasis in vivo
To directly evaluate the role of Nectin-4 in OS cells tumorigenesis and growth in vivo, the subcutaneous transplantation model of human OS cells in BALB/c-nude mouse was adopted. Briefly, 143B cells infected with shNectin-4#2 or shCtrl lenti-virus were injected subcutaneously into each flank of BALB/c-nude mice. Finally, all of the mice were killed to harvest the xenograft. The results indicated that the capacity of tumorigenesis in shNectin-4#2 mice group was remarkably lower than that in the shCtrl group (P<0.01, Fig. 7A). Tumor growth of the shNectin-4#2 group was slower than that in the group (Fig. 7B). Moreover, it was obvious that the mean weight of the subcutaneous tumors generated from the Nectin-4 down-regulation group was significantly lower compared with the control group (P<0.01, Fig. 7C). The total protein was extracted from the tumors, and the expression of Vimentin, Ki67 and Nectin-4 were remarkedly reduced in the shNectin-4#2 group than the shCtrl group, while the expression of ZO-1 was opposite (Fig. 7D). Additionally, HE staining for the tumor sections was performed to evaluate the tissue morphology, which obviously revealed that much more cells nodules and masses in the shCtrl group than those in shNectin-4#2 group (Fig. 7E). Meanwhile, the IHC staining also provided the same powerful evidence that the expression of Ki-67, Nectin-4 and Vimentin in Nectin-4 knockdown group were significantly decreased compared with that in control group, also an opposite trend to the expression of ZO-1 (Fig. 7E). Given these findings, we identified that the knockdown of Nectin-4 could markedly attenuate the OS cells tumorigenesis and growth in vivo.
Then, to further verify the crucial effect of Nectin-4 on the process of OS metastasis, we also conducted a lung metastatic mouse model. Briefly, shNectin-4#2 or shCtrl 143B cells were injected into the five BALB/c-nude mice by lateral tail vein, respectively. After 3 weeks, the mice were anaesthetized, and their lungs were dissected. As shown in Fig. 7F and 7G, the quantities and incidence of pulmonary metastasis in shNectin-4#2 group mice were fewer than those in shCtrl group through visual observation. Importantly, the number of OS metastasis cells with dark stained of nucleus in shNectin-4#2 group were less than that of in the shCtrl group using HE staining (Fig. 7H). Collectively, these data clearly indicated that knockdown of Nectin-4 could prominently restrain the tumorigenesis and metastatic of OS in vivo.
To investigate whether Nectin-4 promotes tumorigenesis mediated by miR-520c-3p in vivo, we detected the function role of miR-520c-3p knockdown in tumorigenesis on the basis of Nectin-4 silencing. Firstly, the shNectin-4#2 143B cells were transfected with miR-520c-3p inhibitor or NC inhibitor and then subcutaneously injected into the flank of six nude mice. Finally, the tumors were harvested from mice at 24 days after injection. The tumors derived from the group of "shNectin-4#2+miR inhibitor" were much larger in weight and volume than those from the group of "shNectin-4#2+NC inhibitor" (Fig. 7I and Supplementary Fig. 7). The results demonstrated that miR-520c-3p silencing significantly reversed the effect in tumorigenesis and growth caused by Nectin-4 silencing. Given these findings, we confirmed that Nectin-4 and miR-520c-3p played a vital role in promoting the proliferation and metastasis of OS in vivo.