1.KDELR1 was specifically highly expressed in CS tumor cells
In order to illuminate the expression and transcriptional regulation characteristics of CS genes, we constructed the single-cell transcriptional landscape of CS by performing single-cell RNA sequencing (scRNA-seq) on CS and OS samples. And normal cancellous bone tissue was served as control bone (CB). A total of 102 380 cells were acquired after quality control exclusions and batch effect elimination. The Seurat package and Seurat alignment method canonical correlation analysis were used to normalise data, dimensionality reduction, clustering, differential expression and integrated analysis of datasets[27, 28]. Based on the expression of classic markers, mesenchymal cells (COL1A1, COL1A2), myeloid cells (CD14), osteoclasts (CTSK), T cells (CD3D), B and plasma cells (CD79A), vascular endothelial cells (PECAM1) and erythroid cells (HBB, HBA2) were identified (Fig. 1A and B, Figure S1A and B). We further reduced the dimensionality of mesenchymal cells to accurately identify and analyze CS cells, and its cell subtypes were identified including OS cells, CS cells, osteobalsts, adipocytes, chondrocytes, bone marrow derived mesenchymal stem cells (BMSCs) (Fig. 1C and D, Figure S1C). The most significant markers of the 6 clusters are shown in Figure S2A. Then, we analyzed the expression of biomarkers of osteoblasts (COL1A1, BGLAP and RUNX2), chondrocytes (COL2A1, COMP and SOX9) and sarcoma (POSTN and DCN), adipocyte (LPL and PPARG) and BMSCs (Fig. 1E, Figure S2B and C). Cell proportion analysis further clarified that CS cells mainly originate from CS samples, while OS cells originate from OS samples (Fig. 1F). GSVA analysis showed that CS was significantly enriched in pathways such as cartilage formation and differentiation, while OS was significantly enriched in osteogenic related pathways, further confirming the accuracy of the dimensionality reduction. Furthermore, we found that in tumor-related pathways such as DNA replication, the enrichment intensity of CS was weaker than that of OS, indicating a lower malignancy level of CS. Besides, ECM secretion and Golgi-ER-related protein synthesis were enriched in CS (Fig. 1G). Subsequently, we analyzed CS-specific differentially expressed genes (DEGs), a total of 455 and 145 DEGs were identified in CS versus chondrocytes and CS versus OS, respectively. Furthermore, we identified the overlapped markers between two list of DEGs, in which KDELR1 came out in front (Fig. 1H). It was consistent with the expression levels of KDELR1 in mesenchymal cells (Fig. 1I). Volcano plots showed that KDELR1 was elevated in CS cells compared with both chondrocytes and OS cells (Fig. 1J and K). These results demonstrated that KDELR1 is a CS-specific molecule and might be involved in the pathogenesis of CS.
2. Elevated expression of KDELR1 is linked to high-grade CS.
Initially, the expression of KDELR1 in CS was examined in 110 samples, comprising osteochondroma specimens (n = 29), and CS specimens with various grades including Grade I (n = 28), Grade II (n = 28), Grade III (n = 25). The results of Immunohistochemistry staining of these specimens exhibited that the expression level of KDELR1 positively correlated with the grade malignancy of CS (Fig. 2, A and B). The Western blot results aligned with the IHC findings (Fig. 2C). We also determined the expression of KDELR1 in human normal cartilage cell line CHON-001, and CS cell lines SW1353 and Hs 819.T. The results showed a higher expression of KDELR1 in SW1353 and Hs 819.T CS cells than CHON-001 cells (Fig. 2D). We further detected that KDELR1 was also correlated with other indicators that can reflect the malignancy of CS, including recurrence (Fig. 2E), survival (Fig. 2F), MCS (Fig. 2G), and metastasis (Fig. 2H). Finally, we analyzed the relationship of KDELR1 to Kaplan-Meier free survival of patients. Patients with higher expression levels of KDELR1 had a poorer prognosis; this fact strongly supports the role of KDELR1 in the disease progression (Fig. 2I). All these results indicated that the expression level of KDELR1 can reflect the malignancy extent of CS.
3. KDELR1 influences malignant behaviors of CS.
To further investigate the impact of KDELR1 on CS, we generated CS sublines that overexpressed KDELR1, or siRNA targeted KDELR1, or control. The transfection efficiency was verified by western blot (Fig. 3A). Then, CCK8 assay revealed that cell viability of CS cells was not affected by overexpressed KDELR1 but was dramatically suppressed in KDELR1-kd CS cells (Fig. 3B). Cisplatin plays a key component of CS chemotherapy drugs in the systemic therapy of CS[29, 30]. Considering the close relationship between membrane proteins and tumor drug resistance, and KDELR1’s involvement in the process of membrane proteins modification and processing, we further explored how KDELR1 regulates the chemotherapeutic effect of cisplatin in CS therapy. We found that KDELR1-kd had actions of promoting sensitivity of CS cells to cisplatin and reduced cisplatin-resistance in CS therapy (Fig. 3C). Then, transwell assay showed that KDELR1-oe has no effect on CS migration, but KDELR1-kd attenuates the migratory ability of CS (Fig. 3D). Together, these data collectively prove that KDELR1 is closely related to malignant behaviors and chemotherapy resistance of CS.
4. Knocking down KDELR1 leads the YAP1 accumulation in cytoplasm and its inactivation.
KDELR1 recognizes, binds, and subsequently retrieves the HDEL-containing protein from Golgi to endoplasmic reticulum, thus regulating the synthesis and processing of membrane proteins[31]. Therefore, we utilized LC-MS analysis to explore the downstream regulators of KDELR1. The volcano plot in Fig. 4A shows that YAP1 was one of the proteins with the most significant expression differences between KDELR1-kd and ctrl-siRNA SW1353 cells. GO enrichment analysis indicated that KDELR1 was associated with the expression of membrane protein and secretory protein involving in components such as ribonucleoprotein complex, coated vesicle membrane, and actin cytoskeleton (Fig. 4B). Dephosphorylated YAP1 translocates from cytoplasm to nucleus and targets TEA domain (TEAD) transcription factor family to regulate the expression of its targeted genes, while phosphorylation deactivates YAP1 and promotes its cytoplasmic localization. Therefore, we first detected the expression of different forms of YAP1 in KDELR1-kd and ctrl-siRNA SW1353 and Hs 819.T cells. The results showed knocking down KDELR1 upregulated the total YAP1 and phosphorylated YAP1 and suppressed the active-YAP1 (Fig. 4C). For further confirmation, the nuclear and cytoplasmic protein were extracted and its total YAP1 was determined. The results demonstrated that KDELR1-kd decreased dephosphorylated YAP1 in nucleus, meanwhile upregulated phosphorylated YAP1 in cytoplasm (Fig. 4D). Then, immunofluorescence assay was conducted and the figures illustrate that KDELR1-kd elevated p-YAP1 and led its elevation of cytoplasmic accumulation (Fig. 4E). All this showed that KDELR1-kd promoted YAP1 phosphorylation, which inhibited its nuclear translocation and subsequently caused its accumulation and deactivation.
5. Knocking down KDELR1 activates Hippo-Yap signaling pathway.
To elucidate the YAP1-phosphorylation mechanism of knocking down KDELR1, the total RNA of KDELRI-kd and siRNA-ctrl cells were extracted and differentially expressed genes were analysed by RNA-seq. These differentially expressed genes were showed in Fig. 5A and B; functional enrichment was carried out and the results indicated that—besides these processes and pathways associated to the processing of membrane proteins and secreted proteins, as a vital regulatory signal of Hippo-YAP1 signaling pathway was a potential contributor to the accumulation and deactivated of YAP1 (Fig. 5C). Next, the data of qRT-PCR and western blot exhibited that a higher mRNA and protein expression of these downstream targets of Hippo-YAP1 signaling pathway comparing to siRNA-ctrl cells (Fig. 5, D and E). These findings suggest that KDELR1 plays a significant role in modulating the Hippo-YAP signaling pathway.
6.The regulation of Hippo-YAP pathway by KDELR1 is achieved by promoting the expression of MAP4K4.
Subsequently, we delved deeper into understanding how KDELR1 influences the Hippo-YAP signaling pathway. As a downstream nuclear effector of the mammalian Hippo signaling pathway, YAP/TAZ was regulated by TAO kinases (TAOKs), large tumor suppressor 1/2 (LATS1/2), mammalian STE20-like protein kinase (MST1/2), MAP4Ks and NF2-depended MAP4K 4/6[32]. Thus, we performed a comprehensive detection of YAP-related kinases in KDELR1-oe and KDELR1-kd CS cells, which showed that the YAP1-phosphorylation was not regulated by TAOKs, MST1/2, and the scaffold proteins NF2, but by p-LATS1 and MAP4K4 among the MAP4Ks family (Fig. 6, A-D). Furthermore, we found that reduced KDELR1 promoted the expression of MAP4K4 and the phosphorylation of LATS1 and YAP1 in CS cells (Fig. 6E). The above results revealed that MAP4K4 may serve as a pivotal kinase in mediating KDELR1's impact on Hippo-YAP signaling.
7.KDELR1 regulates the expression of MAP4K4 by affecting integrins.
MAP4K4 belongs to the serine/threonine protein kinase family, not a membrane or secreted protein. Therefore, we analyzed KDELR1 mediated for regulating MAP4K4 in CS. The upstream regulators of MAP4K4 including PP2A, RAP2, PLCγ, PYK2 have been well-proved[33, 34]. We found that PLCγ and RAP2 were the key regulators of KDELR1-kd-induced MAP4K4 (Fig. 7A). The general function of RAP2 was regulating cell adhesion and spreading mediated by integrin[35]. Besides, PLCγ is a critical part of integrin-mediated transduction pathways[36]. Notably, as membrane proteins, Integrins are synthesized and processed by the ER-Golgi system and could be subject to regulation by KDELR1. Integrins, in turn, are crucial signaling molecules implicated in tumor aggressiveness and resistance to drugs. We performed a comprehensive detection of integrin family in KDELR1-kd and siRNA-ctrl CS cells; the results showed that integrin-α3 was upregulated by KDELR1-kd (Fig. 7B). Furthermore, we isolated the ER and Golgi and subsequently the expression of integrins family in ER and Golgi were measured. The results exhibited that integrin-α3, integrin-α5 and integrin-β5 were trapped in ER and Golgis (Fig. 7C). In summary, these results showed that the MAP4K4-regulation of KDELR1 was mediated by affecting the synthesis, processing, and plasma membrane translocation of integrin-α3, integrin-α5 and integrin-β5.
8.KDELR1 knockdown impedes human CS cells progression in vivo.
To further substantiates the role of KDELR1 in regulating drug resistance and malignant behavior in CS. we established subcutaneous tumor xenograft models. The tumors were excised, collected, photographed (Fig. 8A) and sectioned after euthanasia. The tumor volume analysis showed that KDELR1 knockdown inhibited CS growth with or without the synergy of Cisplatin (Fig. 8B). The protein of parts of tumors was extracted, and the expression of MAP4K4, YAP1 and p-YAP1 was determined. The results confirmed that KDELR1-kd led YAP1-phosphorylation were associated with increased MAP4K4 (Fig. 8C). The results of IF staining showed that the expression of YAP1 was significantly increased in KDELR1-kd tumor xenograft (Fig. 8D). Taken together, the results of these in vivo studies confirmed that KDELR1-kd ameliorates CS proliferation, metastasis, and drug-resistance by activating Hippo-YAP1signaling pathway via ITG-α3/5β5-PLCγ-MAP4K4 axis.