Abstract The progression and outcome of bladder cancer (BLCA) are critically affected by the propensity of tumor metastasis. Our previous study revealed that activation of the mevalonate (MVA) pathway promoted migration of BLCA cells; however, the exact mechanism is unclear. Here we show that elevated expression of MVA pathway enzymes in BLCA cells, correlating with poorer patient prognosis by analyzing single-cell and bulk-transcriptomic datasets. Inhibition of the MVA pathway, either through knockdown of farnesyl diphosphate synthase (FDPS) or using inhibitors such as zoledronic acid or simvastatin, led to a marked reduction in BLCA cell migration. Notably, this effect was reversed by administering geranylgeranyl pyrophosphate (GGPP), not farnesyl pyrophosphate (FPP) or cholesterol, indicating the specificity of geranylgeranylation for cell motility. Moreover, we found that RhoB, a Rho GTPase family member, was identified as a key effector of the impact of the MVA pathway on BLCA metastasis. The post-translational modification of RhoB by GGPP-mediated geranylgeranylation influenced its protein stability through the ubiquitin-proteasome pathway. Additionally, overexpression of RhoB was found to block the membrane translocation of integrin β1 in BLCA cells. In summary, our findings underscore the role of the MVA pathway in BLCA metastasis, providing insights into potential therapeutic targets of this malignancy. Subject terms: Bladder cancer, Prognostic markers, Cancer metabolism __________________________________________________________________ RhoB plays an important role in the effect of MVA pathway inhibition on the migratory capacity of bladder cancer cells and its protein stability is closely related to GGPP-mediated geranylgeranyl modification. Introduction Bladder cancer (BLCA) is the 10th most commonly diagnosed cancer worldwide, with more than 200,000 deaths annually^[40]1. Due to its recurring nature, BLCA requires long-term monitoring, increasing its economic burden^[41]2. Patients with metastatic BLCA have a poor prognosis, and lymph node-positive disease is an independent predictor of worse survival^[42]3. In addition, ~5% of metastatic cancer patients survive for at least 5 years postdiagnosis^[43]4. Thus, metastasis is an important factor contributing to the worsening prognosis of patients with BLCA and is one of the most important causes of death^[44]5. Therefore, studies on the molecular mechanisms underlying metastasis and the identification of novel targets or drugs are needed for the clinical management of BLCA. Lipids are essential sources of energy during tumor development and metastasis and promote intercellular communication in the tumor microenvironment^[45]6. In a prospective cohort study, metabolic factors such as BMI, cholesterol, and triglycerides were found to be positively associated with BLCA risk^[46]7. Cholesterol, an important class of lipids, is involved in the regulation of the fluidity and permeability of lipid bilayers and is also an important component of lipid rafts^[47]8. In BLCA, compared to that in RT4 cells (noninvasive BLCA cells), the cholesterol content in T24 cells (high-grade invasive BLCA cells) is greater^[48]9. The mevalonate (MVA) pathway plays an important role in cholesterol synthesis and has been reported to be involved in the regulation of tumor metastasis^[49]10,[50]11. Liu et al. discovered that activation of the MVA pathway promotes cholesterol biosynthesis and contributes to BLCA growth^[51]12. Based on our previous finding that simvastatin, an inhibitor of key enzymes of the MVA pathway, inhibits the proliferation and metastasis of BLCA cells, we confirmed that it affects the cell cycle distribution of BLCA cells through the PPARγ signaling pathway^[52]13. However, the mechanism through which the MVA pathway regulates BLCA cell metastasis is not yet clear. Intermediate metabolites of the MVA pathway, such as farnesyl pyrophosphate (FPP) and geranylgeranyl pyrophosphate (GGPP), act as substrates for protein isoprenylation and are involved in the post-translational modification of small GTPases, such as Ras and Rho family GTPases, which are essential for the invasion and metastasis of a variety of cancers^[53]14–[54]16. Farnesyl diphosphate synthase (FDPS) is an enzyme of the MVA pathway that catalyzes the synthesis of FPP and GGPP from isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP)^[55]17. It is also the target of nitrogen-containing bisphosphonates (N-BPs), a class of bone antiresorptive drugs used to treat osteoporosis and metastatic bone disease^[56]18. A study in patients with bone metastases from BLCA showed that zoledronic acid (ZOL), an N-BPs, improved overall survival compared to the effect of the placebo^[57]19. In addition, several studies have reported that ZOL inhibits the proliferation of BLCA cells^[58]20,[59]21. However, whether the effect of ZOL on BLCA is associated with mevalonate pathway inhibition and the mechanism by which ZOL improves the prognosis of patients with metastatic BLCA need to be further investigated. In this study, we investigated the impact of inhibiting the MVA pathway on the metastatic potential of BLCA and the underlying mechanisms involved. We discovered that FDPS, an enzyme whose expression is elevated in BLCA, is governed by the PSME3-mediated, ubiquitin-independent proteasome system. Intriguingly, while the migratory inhibition caused by targeting FDPS or using inhibitors such as ZOL or simvastatin was counteracted by GGPP, FPP and cholesterol did not produce the same effect. Further investigations revealed that the protein RhoB, a member of the Rho GTPases, is a critical effector of MVA pathway inhibition on BLCA cell migration. The stabilization of the RhoB protein appears to be modulated by GGPP through the ubiquitin-proteasome pathway. These findings highlight the complex role of the MVA pathway and RhoB in BLCA metastasis and suggest potential targets for therapeutic intervention. Results Genetic alterations of MVA pathway-related enzymes across urologic tumors Based on the findings of previous studies^[60]22, the following enzymatic components of the MVA pathway were identified as MVA pathway-related enzymes (MREs): ACAT1 (acetyl-CoA acetyltransferase 1), ACAT2 (acetyl-CoA acetyltransferase 2), FDFT1 (farnesyl-diphosphate farnesyltransferase 1), FDPS (farnesyl diphosphate synthase), GGPS1 (geranylgeranyl diphosphate synthase 1), HMGCL (3-hydroxy-3-methylglutaryl-CoA lyase), HMGCR (hydroxy-3-methylglutaryl-CoA reductase), HMGCS1 (3-hydroxy-3-methylglutaryl-CoA synthase 1), IDI1 (isopentenyl-diphosphate delta isomerase 1), IDI2 (isopentenyl-diphosphate delta isomerase 2), MVD (MVA diphosphate decarboxylase), MVK (MVA kinase) and PMVK (phosphomevalonate kinase). In our study, the genomic data of five urologic tumor types comprising BLCA, kidney chromophobe (KICH), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP) and prostate adenocarcinoma (PRAD) from the TCGA dataset, including genetic variation, somatic copy number alteration (SCNA), and mRNA expression data, were analyzed to discern MREs dysregulation patterns. The frequency of nonsynonymous mutations was greater in BLCA than in other urologic tumors (Fig. [61]1a). Additionally, compared to other MREs, HMGCR and HMGCS1 had relatively high mutation frequencies in numerous urologic tumors, including BLCA, KIRC, KIRP and PRAD (Fig. [62]1b). The percentages of SCNA were analyzed and found that SCNA occurred at high rates (over 5% of all samples) in the BLCA, however, in KICH and KIRP, most MREs had a lower frequency of SCNA (Fig. [63]1c). Fig. 1. Genetic alterations of MVA pathway-related enzymes across urologic tumors. [64]Fig. 1 [65]Open in a new tab a Mutation frequency of MREs across cancers in TCGA database. The X-axis represents each cancer, and the Y-axis represents the frequency of mutations in MREs. The data were downloaded from the Xena Browser ([66]https://xenabrowser.net/). b Mutation types, mutation frequencies and tumor mutation burden (TMB) of MREs across cancers. The X-axis represents each cancer, the Y-axis represents the different MREs and the number of samples with mutations, and the top is the TMB. c Histogram shows the frequency of SCNAs for each MRE in each cancer type. d The gene expression patterns of MREs across cancers. The X-axis represents each cancer, the Y-axis represents the different MREs, and the top represents the statistics of the number of upregulated and downregulated MREs in different types of tumors. e The Spearman’s correlation between somatic copy number alterations and the expression of MREs. MPIs were validated in multiple datasets: [67]GSE2450 (f), [68]GSE252007 (g), [69]GSE40355 (h), [70]GSE3167 (i), and TCGA-BLCA (j). Survival analysis of BLCA patients with high MPI or low MPI from the TCGA dataset (k) and [71]GSE13507 dataset (l). Statistical significance was ascertained by two-tailed unpaired Student’s t-tests (g–j) and the log-rank test of Kaplan–Meier analysis (k and l). The data are shown as the means ± SD. BLCA bladder urothelial carcinoma, KICH kidney chromophobe, KIRC kidney renal clear cell carcinoma, KIRP kidney renal papillary cell carcinoma, PRAD prostate adenocarcinoma. Beyond genetic alterations, we examined the gene expression patterns of MREs in tumors and normal tissues for every cancer type. There was differential expression of each MRE in at least one type of cancer. There were differences in MREs expression patterns among different urinary tumors. In BLCA, most MREs were upregulated, while in KICH, KIRC and KIRP, most were downregulated (Fig. [72]1d). SCNA in tumors is closely associated with the regulation of gene expression, so we further assessed the effect of SCNA on gene expression in MREs. The results showed that the expression of most MREs was associated with SCNA (Fig. [73]1e). The above results based on the analysis of the TCGA dataset showed that genetic or expression alterations of MREs are specific to BLCA compared to other urologic tumors and deserve further investigation. To further reveal the potential role of MREs in regulating tumorigenesis, we analyzed BLCA single-cell sequencing results from public databases ([74]GSE190888). A total of 36,424 filtered cells in the dataset were subjected to bioinformatics analysis, and clustering analysis identified 11 cell clusters (Supplementary Fig. [75]1a). Further analysis of the canonical cell type-specific markers revealed seven classical cell types, namely, basal tumor cells, urothelial cells, endothelial cells, T cells, macrophages, muscle cells and fibroblasts (Supplementary Fig. [76]1b). The MREs ACAT1, ACAT2, FDFT1, FDPS, HMGCL and PMVK were highly expressed in basal tumor cells and urothelial cells (Supplementary Fig. [77]1c–h). MVA pathway activation in BLCA and indicates a worse prognosis To explore the role of the MVA pathway in tumorigenesis and identify factors or biological processes associated with this pathway, we calculated the MVA potential index (MPI) through ssGSEA using the enrichment score of core machine components. In this study, we calculated the MPI using an independent GEO dataset ([78]GSE2450) of HUVECs treated with atorvastatin, a drug reported to be an inhibitor of MVA synthesis^[79]23. The MPI was significantly decreased by atorvastatin, but the addition of MVA reversed the MPI (Fig. [80]1f). In addition, we calculated the MPI of BLCA T24 cells transfected with siRNA targeting FDPS, a key enzyme of the MVA pathway^[81]17. As shown in Fig. [82]1g, FDPS knockdown clearly decreased the MPI compared to that of the controls. Since the inhibition of the MVA pathway by atorvastatin treatment or FDPS knockdown was unequivocal, analyses based on the validation of the above two independent datasets, revealed that the MPI could be used to represent the potential level of MVA pathway activity based on transcriptomic data. Furthermore, we found that the MPI differed between tumor tissues and normal tissues when using independent BLCA gene expression datasets. The MPI was significantly greater in BLCA samples than in normal samples, as shown in Fig. [83]1h–j. Therefore, we concluded that the MVA pathway was activated in BLCA tissues. To further understand the clinical relevance of MVA pathway activation in cancer, we investigated the role of MPI in the survival of patients with BLCA. Our results showed that BLCA patients with high MPIs had worse overall survival in both the TCGA-BLAC cohort (Fig. [84]1k) and the [85]GSE13507 cohort (Fig. [86]1l), suggesting that MPI is a risk factor for BLCA. As a result, further research into the functional roles of the MVA pathway in cancer progression is warranted. FDPS is highly expressed in BLCA and regulated by PSME3 for protein stability Our previous study revealed that targeted inhibition of HMGCR, a key enzyme of the MVA pathway, attenuated the proliferation and metastasis of BLCA cells^[87]13. FDPS is a downstream enzyme of HMGCR in the MVA pathway; however, its expression and function in BLCA have not been described. Therefore, we analyzed the expression of FDPS in several publicly available BLCA datasets and in collections of BLCA samples. The results showed significantly elevated FDPS mRNA expression in BLCA tissues compared to paracancerous or normal tissues, as evident in the TCGA-BLCA dataset (Fig. [88]2a), the Zhongnan Hospital cohort (Fig. [89]2b), and the [90]GSE13507 cohort (Fig. [91]2c). Additionally, FDPS was significantly positively correlated with lymph node metastasis in the TCGA-BLCA cohort (Supplementary Fig. [92]2a) and with tumor T stage in the [93]GSE32548 cohort (Supplementary Fig. [94]2b). Meanwhile, we determined that BLCA patients with higher FDPS expression in both the [95]GSE13507 cohort (Fig. [96]2d) and the [97]GSE32548 cohort (Supplementary Fig. [98]2c) had poorer overall survival. Furthermore, by analyzing the tissue microarray (containing 68 BLCA specimens and 40 paracancerous tissues) (Fig. [99]2e–g), we found that FDPS protein levels were upregulated in BLCA tissues compared to paracancerous tissues (Fig. [100]2f), and patients with higher FDPS protein levels had poorer overall survival (Fig. [101]2g). The above results suggest that FDPS may play a role in BLCA tumorigenesis and progression. Fig. 2. FDPS is upregulated in BLCA and promotes BLCA cell proliferation and metastasis. [102]Fig. 2 [103]Open in a new tab a The mRNA level of FDPS in BLCA (n = 408) and normal tissues (n = 19) in TCGA-BLCA (RNA-seq data). b The mRNA levels of FDPS in BLCA (n = 15) and paracancerous tissues (n = 15) in the Zhongnan Hospital cohort were measured by qRT-PCR. c The mRNA expression level of FDPS in BLCA (n = 188) and normal tissues (n = 68) in the [104]GSE13507 cohort (RNA-seq data). d OS analysis of patients with BLCA who had different FDPS mRNA levels in the [105]GSE13507 dataset. e Representative images of IHC staining analysis of FDPS protein in BLCA and paracancerous tissues from tissue microarray. The scale bars are 200 and 50 μm. f A statistical graph of staining scores of FDPS expression in BLCA (n = 68) and paracancerous tissues (n = 40). g The overall survival of patients with different FDPS protein levels in tissue microarray. The patients were divided into a high FDPS protein level group (n = 28) and a low FDPS protein level group (n = 28) according to the median staining scores of FDPS expression. The patients with missing survival data were not included. h Schematic representation of MVA pathway inhibition by targeting FDPS with siRNA or shRNA. Representative images (i) and statistical analysis of transwell migration assays results for the indicated groups of UM-UC-3 cells (j) and T24 cells (k) with FDPS knockdown (n = 3). The scale bar is 200 μm. l Schematic representation of the mouse pulmonary metastasis model constructed using LV-T24-shNC or LV-T24-shFDPS cells. Images of lung fluorescence after T24-shNC or T24-shFDPS cells were injected into the tail veins of BALB/C-nude mice for 6 weeks (m), and the fluorescence intensity of the lung metastases was quantified (n = 3) (n). o Statistical analysis of the number of metastatic nodules in H&E-stained mouse lung tissue sections (n = 3). The n number represents n biologically independent experiments in each group. Statistical significance was ascertained by two-tailed paired Student’s t-tests (b), two-tailed unpaired Student’s t-tests (a, c, f, n, and o) and one-way ANOVA with Dunnett’s multiple comparisons test (j and k) and the log-rank test of Kaplan–Meier analysis (d and g). The data are shown as the means ± SD. Previous studies have primarily explored the reasons for the high expression of FDPS in tumors at the transcriptional level^[106]24, but rarely at the post-translational modification level. Therefore, in this study, we explored potential FDPS-associated proteins by IP–MS, and the top-ranked protein, PSME3, which is an important activator of the 20S proteasome and regulates the degradation of proteins, was of interest (Supplementary Fig. [107]3a). Co-IP analysis further confirmed the interaction between FDPS and PSME3 (Supplementary Fig. [108]3b). Moreover, we found colocalization of Flag-FDPS and HA-PSME3 in T24 and UM-UC-3 cells by immunofluorescence staining (Supplementary Fig. [109]3c). After T24 and UM-UC-3 cells were transfected with the HA-PSME3 plasmid, FDPS protein was reduced (Supplementary Fig. [110]3d). FDPS protein levels were decreased by PSME3 in a dose-dependent manner, which was blocked by the addition of the proteasome inhibitor MG132 (Supplementary Fig. [111]3e). To further determine whether PSME3 affects the protein stability of FDPS, a cycloheximide (CHX) assay was performed, which revealed that PSME3 overexpression accelerated the degradation of FDPS (Supplementary Fig. [112]3f-g). We also observed that PSME3 did not affect FDPS polyubiquitination (Supplementary Fig. [113]3f–h). Accordingly, we suggest that FDPS protein stability is regulated by the PSME3-mediated ubiquitin-independent proteasome system, which is consistent with its previously reported function^[114]25. In addition, overexpression of PSME3 led to reduced migration in BLCA cells, whereas co-overexpression of FDPS and PSEM3 reversed the migratory inhibition induced by PSME3 overexpression (Supplementary Fig. [115]3i–k). MVA pathway inhibition by FDPS knockdown affects the proliferation and metastasis of BLCA cells We explored the effect of MVA pathway inhibition on the BLCA phenotype by transfecting two FDPS siRNAs to knock down FDPS in BLCA cells (Fig. [116]2h). The knockdown efficiency of two FDPS siRNAs in BLCA cells was confirmed by qRT-PCR (Supplementary Fig. [117]4a) and Western blotting (Supplementary Fig. [118]4b). The cell viability determined by MTT (Supplementary Fig. [119]5a-c) and clonogenic survival (Supplementary Fig. [120]5d–f) assays suggested that the proliferation of BLCA cells was significantly inhibited after FDPS knockdown. A transwell migration assay was used to measure the migratory ability of BLCA cells, and the results showed that FDPS knockdown significantly inhibited the migration of BLCA cells (Fig. [121]2i–k and Supplementary Fig. [122]5g, h). To further explore the effect of FDPS on BLCA metastasis in vivo, we established a mouse pulmonary metastasis model using T24 cells with stable FDPS knockdown (Fig. [123]2i). The knockdown efficiency of LV-shNC and LV-shFDPS in BLCA cells was confirmed by qRT-PCR (Supplementary Fig. [124]5i) and western blotting (Supplementary Fig. [125]5j). After 6 weeks of tail vein injection, the fluorescence intensity in the lungs of the LV-shFDPS group mice (n = 3) was relatively lower than that in the LV-shNC group (n = 3) (Fig. [126]2m, n). Moreover, H&E staining of lung tissues from the two groups of mice showed that the size and number of lung metastatic nodules were reduced in the LV-shFDPS group (Fig. [127]2o and Supplementary Fig. [128]5k). ZOL-mediated inhibition of the MVA pathway affects the proliferation and metastasis of BLCA cells Subsequently, ZOL, an FDPS inhibitor, was used to further explore the effects of MVA pathway inhibition on BLCA cells (Fig. [129]3a). BLCA cells UM-UC-3, 5637 and T24 were treated with ZOL at different concentrations (0, 1, 2, 5, 10, 20, 40, 60 and 80 μM) for 24 h (Supplementary Fig. [130]6a), 48 h (Fig. [131]3b) or 72 h (Supplementary Fig. [132]6b). The cell viability determined by the MTT assay suggested that the proliferation of UM-UC-3 (the IC50 values at 48 and 72 h were 10.86 and 0.9406 μM, respectively), 5637 (the IC50 values at 48 and 72 h were 14.13 and 4.380 μM, respectively) and T24 cells (the IC50 values at 48 and 72 h were 16.71 and 1.866 μM, respectively) were significantly inhibited by ZOL treatment in a time- and dose-dependent manner. By analyzing the IC50 values of the three BLCA cell lines (UM-UC-3, 5637 and T24) treated with different concentrations of ZOL in the above results, and combining the ZOL concentrations reported in previous studies^[133]26,[134]27, we selected 10 and 20 μM ZOL-treated BLCA cells for our subsequent experiments. Fig. 3. MVA pathway inhibition by ZOL affects the proliferation and metastasis of BLCA cells in vivo and in vitro. [135]Fig. 3 [136]Open in a new tab a Schematic representation of MVA pathway inhibition by ZOL. b An MTT assay was performed to detect changes in the proliferation of BLCA cells (UM-UC-3, T24 and 5637) after treatment with different concentrations (0, 1, 2, 5, 10, 20, 40, 60 and 80 μM) of ZOL for 48 h (n = 6). Representative images (c) and statistical analysis (d and e) of colony formation assays from the indicated groups after treatment with ZOL at different concentrations (0, 5 and 10 μM) in UM-UC-3 cells (d) and T24 cells (e) for 48 h (n = 3). The scale bar is 1 cm. Representative images (f) and statistical graph (g and h) of transwell assays from the indicated groups after treatment with ZOL at different concentrations (0, 10 and 20 μM) in UM-UC-3 cells (g) and T24 cells (h) for 48 h (n = 3). The scale bar is 200 μm. i Western blot analyses of EMT-related proteins in ZOL-treated UM-UC-3, T24 and 5637 cells. GAPDH was used as the loading control. j Schematic representation of the effects of ZOL-mediated inhibition of the MVA pathway on BLCA proliferation and metastasis in a mouse xenograft model and pulmonary metastasis model. k Gross view of a subcutaneous tumor; the upper side represents the control group (n = 3), while the lower side represents the ZOL treatment group (n = 3). l Tumor volume was calculated during the experiment. Images of lung fluorescence in the control group and ZOL treatment group (m) and quantification of the fluorescence intensity of lung metastases (n = 3) (n). Images of dissected whole lungs (o) and representative images of H&E-stained mouse lung tissue sections (n = 3) (p); the scale bars are 4 mm and 100 μm. Statistical analysis of the number of metastatic nodules in H&E-stained mouse lung tissue sections (q). The n number represents n biologically independent experiments in each group. Statistical significance was ascertained by two-tailed unpaired Student’s t-tests (l, n and q) and one-way ANOVA with Dunnett’s multiple comparisons test (d, e, g, and h). The data are shown as the mean ± SD. The clonogenic survival assay showed that, compared to that in the control group, the colony formation efficiency of ZOL-treated BLCA cells was significantly inhibited (Fig. [137]3c–e and Supplementary Fig. [138]6c, d). When we performed the clonogenic survival assay after treating UM-UC-3 cells with 10 μM ZOL, it was already difficult for the cells to form colonies, so 5 and 10 μM ZOL were chosen for the treatment of BLCA cells in this study. To assess the impact of ZOL on BLCA cell metastasis, transwell migration and wound healing assays were employed. After 24 h of treatment with different ZOL concentrations (0, 10 and 20 μM), the migration was significantly reduced in all ZOL-treated BLCA cells (UM-UC-3, 5637 and T24) (Fig. [139]3f–h and Supplementary Fig. [140]6e, f). Similar results were replicated by wound healing assay (Supplementary Fig. [141]7), reconfirming the significant decrease in migration caused by ZOL treatment. Western blotting was used to analyze the changes in proteins involved in the epithelial–mesenchymal transition (EMT) process, revealing upregulation of E-cadherin and downregulation of N-cadherin, Vimentin and Slug in BLCA cells after ZOL treatment (Fig. [142]3i). To explore the in vivo effect of ZOL on BLCA cell proliferation and metastasis, xenograft and pulmonary metastasis models were established (Fig. [143]3j). In the xenograft model, compared with that in the control group (n = 3), the tumor growth in the ZOL group (n = 3) was significantly inhibited (Fig. [144]3k, l). Pulmonary metastasis models were established by tail vein injection of T24 cells, with mice divided into a control group (PBS injection) (n = 3) and a ZOL injection group (n = 3). As shown in Fig. [145]3m, n, ZOL treatment significantly suppressed in vivo migration compared to the control group. H&E staining of lung tissues showed that the size and number of lung metastatic nodules were reduced in the ZOL treatment group (Fig. [146]3o–q). GGPP restores migration inhibition of MVA pathway in BLCA To investigate the mechanism of MVA pathway inhibition in BLCA, two MVA pathway inhibitors, simvastatin (an HMGCR inhibitor) and ZOL (an FDPS inhibitor), were applied to BLCA cells. The treated cells, along with control cells, underwent an MVA pathway intermediate metabolite assay and proteomic analysis (Fig. [147]4a). GGPP and FPP are important intermediate metabolites of the MVA pathway. Our results showed that the levels of FPP and GGPP were reduced in BLCA cells treated with simvastatin or ZOL (Fig. [148]4b, c). Fig. 4. GGPP-mediated geranylgeranylation of RhoB protein is associated with the migration ability of BLCA cells. [149]Fig. 4 [150]Open in a new tab a Flowchart for the detection of MVA pathway intermediate metabolite (FPP and GGPP) levels and proteomic changes in BLCA cells after MVA pathway inhibition by ZOL or simvastatin. The contents of FPP (b) and GGPP (c) in T24 cells treated with ZOL (20 μM) (n = 3) or simvastatin (5 μM) (n = 3) were measured by LC–MS/MS analysis. d Heatmaps of proteins whose expression significantly changed in T24 cells after MVA pathway inhibition by ZOL or simvastatin, respectively. e Western blotting was performed to detect RhoB protein in BLCA cells after treatment with ZOL. GAPDH was used as the loading control. Representative images (f) and statistical graph (g and h) of transwell migration assays from the indicated groups after treatment of BLCA cells (UM-UC-3 and T24) with RhoB siRNA and ZOL (20 μM), respectively, or in combination (n = 3). The scale bar is 200 μm. i Schematic representation of BLCA cells treated with ZOL (20 μM), GGPP (5 μM) and GGTI298 (10 μM), alone or in combination. Representative images (j) and statistical analysis (k and l) of transwell migration assays from the indicated groups after treatment of BLCA cells (UM-UC-3 and T24) with ZOL (20 μM), GGPP (5 μM) and GGTI298 (10 μM), separately or in combination (n = 3). The scale bar is 200 μm. Changes in RhoB protein expression in different groups of UM-UC-3 (k) and T24 (l) cells were also detected by western blotting. GAPDH was used as the loading control. The n number represents n biologically independent experiments in each group. Statistical significance was ascertained by two-tailed unpaired Student’s t-test (b, c, k and l) and one-way ANOVA with Dunnett’s multiple comparisons test (g and h). The data are shown as the mean ± SD. To investigate whether FPP and GGPP are involved in the effect of BLCA on migration capacity through MVA pathway inhibition, FPP (Supplementary Fig. [151]8a–c) and GGPP (Supplementary Fig. [152]8d–f) were added to ZOL-treated BLCA cells. Transwell migration assays revealed that after incubating with GGPP, the ZOL-induced reduction in the migration rate was significantly attenuated, whereas the effect of FPP was not (Supplementary Fig. [153]8a–f). Consistent with the above findings, GGPP also reversed the suppressive effect of simvastatin on migration (Supplementary Fig. [154]8g–i) and FDPS knockdown (Supplementary Fig. [155]8j–l) in BLCA cells. The MVA pathway is the central metabolic pathway for cholesterol biosynthesis, therefore, we also explored whether cholesterol is involved in the effect of BLCA on migration capacity through MVA pathway inhibition. Cholesterol was added to ZOL-, simvastatin-treated or FDPS knockdown BLCA cells, respectively. Transwell migration assays showed a slight increase in the migration of BLCA cells after the administration of cholesterol alone, however, cholesterol did not rescue the reduction in migration rate caused by ZOL-, simvastatin-treated or FDPS knockdown (Supplementary Fig. [156]9a–f). The total cholesterol content in simvastatin- or ZOL-treated BLCA cells was analyzed, and compared to that in control cells, the cholesterol content in the drug-treated BLCA cells was not significantly reduced (Supplementary Fig. [157]9g, h). Further qRT-PCR analysis revealed that the mRNA expression of ABCA1 and ABCG1, which are responsible for the transport of cholesterol to the extracellular compartment^[158]28, were decreased, whereas the mRNA expression of LDLR, NPC1L1 and SCARB1, which are responsible for the uptake of cholesterol^[159]29, were increased in simvastatin- or ZOL-treated cells (Supplementary Fig. [160]9i, j). Role of RhoB in attenuating BLCA metastasis induced by MVA pathway inhibition In our investigation, BLCA cells were treated with simvastatin or ZOL, and subsequent proteomic analysis was performed on both treated and control cells. The heatmaps show proteins that were significantly upregulated (top 10) and proteins that were significantly downregulated (top 5) in T24 cells after simvastatin (Fig. [161]4d, left panel) and ZOL treatment (Fig. [162]4d, right panel), respectively. We examined the intersection of the significantly upregulated proteins (top 10) in the two groups and identified five proteins (Fig. [163]4d, middle panel), among which RhoB has been reported to be isoprenylated by GGPP or FPP and closely associated with tumor migration^[164]30,[165]31. Further validation confirmed that the significant upregulation of the RhoB protein following ZOL-mediated inhibition of the MVA pathway (Fig. [166]4e), simvastatin treatment (Supplementary Fig. [167]10a) or FDPS knockdown (Supplementary Fig. [168]10b) in BLCA cells. Importantly, we also found that MVA pathway inhibition did not increase RhoB mRNA expression (Supplementary Fig. [169]10c). Moreover, our results showed that the addition of GGPP to the culture media could reduce the upregulation of the RhoB protein induced by ZOL (Supplementary Fig. [170]10d), simvastatin (Supplementary Fig. [171]10e), or FDPS knockdown (Supplementary Fig. [172]10f). To further investigate whether RhoB is involved in the MVA pathway-mediated regulation of the migratory ability of BLCA cells, we transfected RhoB siRNA into ZOL-treated BLCA cells or FDPS knockdown BLCA cells (Fig. [173]4f and Supplementary Fig. [174]10g–j). The transwell migration assay results revealed a significant attenuation of the inhibition of cell migration caused by ZOL (Fig. [175]4f–h) or FDPS knockdown (Supplementary Fig. [176]10h–j) upon RhoB siRNA transfection in T24 and UM-UC-3 cells. GGPP-mediated geranylgeranylation of RhoB protein associates with BLCA cell migration GGTI298, a geranylgeranyl transferase inhibitor (GGTI), can inhibit the geranylgeranyl protein transferase (GGTase)-mediated geranylgeranyl acylation reaction. Moreover, RhoB proteins can also undergo geranylgeranylation. After the administration of ZOL, GGPP, or GGTI298 separately or in combination, changes in the migratory capacity of BLCA cells were examined via a transwell migration assay (Fig. [177]4i). The results showed that the effect of ZOL on the migration of BLCA cell lines (T24 and UM-UC-3) was restored by the addition of GGPP. However, this effect was subsequently reversed by the addition of GGTI298 (Fig. [178]4j–l). Western blot analysis revealed the RhoB protein levels were significantly higher in the ZOL, GGPP and GGTI298 combination groups compared to the ZOL and GGPP combination groups (Fig. [179]4j–l). The above results further confirmed that GGPP affects the protein levels of RhoB by mediating its geranylgeranylation, thus implicating it in the regulation of BLCA cell migration capacity via the MVA pathway. Geranylgeranylated RhoB protein susceptibility to ubiquitination-mediated degradation Since the RhoB protein in BLCA cells was significantly increased after MVA pathway inhibition, a CHX assay was performed to determine the half-life of RhoB protein degradation after ZOL administration at different concentrations (10 and 20 μM). The results showed that the degradation rate of the RhoB protein decreased significantly with increasing ZOL concentration in BLCA cells (Fig. [180]5a, b). To further determine the degradation pathway of the RhoB protein, we treated BLCA cells with ZOL, MG-132 (proteasome inhibitor) or chloroquine (CQ, lysosomal pathway inhibitor), alone or in combination. Western blot analysis revealed that the RhoB protein primarily underwent degradation through the ubiquitin-proteasome pathway (Fig. [181]5c). In vitro ubiquitination assays were then performed to assess the impact of ZOL on RhoB protein ubiquitination. Immunoprecipitation revealed that ZOL significantly decreased the polyubiquitination of RhoB protein, while MG-132 significantly increased the ubiquitination of the RhoB protein, but no effect was observed with CQ (Fig. [182]5d, e). Fig. 5. Geranylgeranylated RhoB protein is more susceptible to degradation through ubiquitination pathway. [183]Fig. 5 [184]Open in a new tab Western blot analysis of the effect of ZOL on RhoB degradation in BLCA cells (T24 and 5637) incubated with CHX (50 μg/mL) for the indicated durations (a) and statistical analysis (n = 3) (b). c BLCA cells (T24 and 5637) with or without ZOL were treated with CQ (50 μM) or MG132 (10 μM) for 8 h, after which RhoB protein expression was detected via western blotting. T24 (d) and 5637 (e) cells with or without ZOL were treated with CQ (50 μM) or MG132 (10 μM) for 8 h, and then ubiquitination experiments were performed to analyze the polyubiquitination of RhoB. T24 (f) and 5637 (g) cells with or without ZOL (20 μM) were treated with GGPP (5 μM) for 24 h, after which ubiquitination experiments were performed to analyze the polyubiquitination of RhoB. h 293 T cells were transfected with the described plasmids for 48 h and then treated with MG132 (10 μM) for 8 h, after which ubiquitination experiments were performed. Considering our previous results indicating that GGPP could reverse the ZOL-induced increase in RhoB protein, we further examined changes in RhoB protein ubiquitination levels after treating BLCA cells separately or in combination with ZOL and GGPP. The results showed that GGPP could reverse the inhibitory effect of ZOL on RhoB protein ubiquitination (Fig. [185]5f, g). In addition, we found that RhoB-GG (geranylgeranylated-only RhoB) exhibited greater susceptibility to ubiquitination compared to RhoB (both geranylgeranylated and farnesylated RhoB) and RhoB-F (farnesylated-only RhoB) (Fig. [186]5h). Taken together, these results suggested that MVA pathway inhibition could inhibit the ubiquitination and degradation of the RhoB protein, which was associated with the depletion of GGPP in BLCA cells. RhoB inhibits BLCA cell proliferation and metastasis in vivo and in vitro RhoB protein expression was closely related to the migration ability of BLCA cells. To explore its clinical relevance in BLCA, we analyzed the expression level of RhoB in several publicly available BLCA datasets and found that RhoB expression was significantly reduced in BLCA (Fig. [187]6a, b and Supplementary Fig. [188]11a, b). To further study the functions of RhoB in BLCA, we constructed a RhoB overexpression plasmid to upregulate RhoB in BLCA cells. qRT-PCR was used to confirm the overexpression efficiency of the RhoB overexpression plasmid in BLCA cells (Supplementary Fig. [189]11c). Subsequently, MTT and transwell migration assays showed that the proliferation and migration of BLCA cells were significantly inhibited in cells transfected with the RhoB overexpression plasmid (Fig. [190]6c–g and Supplementary Fig. [191]11d–f). Western blotting of EMT-related proteins (E-cadherin, N-cadherin, Vimentin and Slug) revealed an upregulation of E-cadherin and a downregulation of N-cadherin, Vimentin and Slug in BLCA cells after RhoB overexpression (Fig. [192]6h), consistent with observations after ZOL treatment (Fig. [193]3i). Fig. 6. RhoB inhibits BLCA cell proliferation and metastasis. [194]Fig. 6 [195]Open in a new tab a The mRNA level of RhoB in BLCA (n = 408) and normal tissues (n = 19) in TCGA-BLCA (RNA-seq data). b The mRNA expression level of RhoB in BLCA (n = 188) and normal tissues (n = 68) in the [196]GSE13507 cohort (RNA-seq data). An MTT assay was performed to detect changes in the proliferation of BLCA UM-UC-3 (c) and T24 (d) cells with or without RhoB overexpression (n = 5). Representative images (e) and statistical analysis (f and g) of colony formation assays of UM-UC-3 and T24 cells with or without RhoB overexpression (n = 3). The scale bar is 200 μm. h Western blot analyses of EMT-related proteins in UM-UC-3, T24 and 5637 cells with or without RhoB overexpression. GAPDH was used as the loading control. i Detection of RhoB expression at the mRNA and protein levels in LV-NC cells and LV-RhoB-OE cells by qRT-PCR and western blotting. j Schematic representation of the mouse pulmonary metastasis model constructed using T24 LV-NC cells and T24 LV-RhoB-OE cells. Images of lung fluorescence after T24 LV-NC cells and T24 LV-RhoB-OE cells were injected into the tail veins of BALB/C-nude mice for 6 weeks (k), and the fluorescence intensity of the lung metastases was quantified (n = 3) (l). Images of dissected whole lungs (m) and representative images of H&E-stained mouse lung tissue sections (n); the scale bars are 4 mm and 100 μm. o Statistical analysis of the number of metastatic nodules in H&E-stained mouse lung tissue sections (n = 3). The n number represents n biologically independent experiments in each group. Statistical significance was ascertained by two-tailed unpaired Student’s t-test (a–d, f, g, i, l, and o). The data are shown as the means ± SD. Pulmonary metastasis models were established using T24 cells stably overexpressing RhoB (LV-RhoB-OE) and control cells (LV-NC) to investigate the effect of RhoB on BLCA cell metastasis in vivo (Fig. [197]6i, j). After 6 weeks of injection, the fluorescence intensity of pulmonary metastatic tumors was measured to evaluate the migration capacity. The LV-RhoB-OE group (n = 3) exhibited relatively lower fluorescence intensity in lungs tissues compared to the LV-NC group (n = 3) (Fig. [198]6k, l). H&E staining of lung tissues showed that the size and number of lung metastatic nodules were reduced in the LV-RhoB-OE group (Fig. [199]6m–o). Impact of RhoB on integrin β1 translocation in BLCA cells We performed KEGG pathway enrichment analysis of DEGs between FDPS knockdown and control T24 cells (Fig. [200]7a), as well as differentially expressed proteins between simvastatin-treated and control T24 cells (Fig. [201]7b), respectively, and found that both were enriched in pathways involving ECM-receptor interaction and focal adhesion (Fig. [202]7a, b). Furthermore, we performed RNA-seq analysis on RhoB overexpressed and control T24 cells. KEGG pathway enrichment analysis of DEGs revealed enrichment in pathways such as adhesion junctions, cell adhesion molecules, and steroid synthesis (Supplementary Fig. [203]12a), further supporting our results. Integrins are important molecules that mediate cell adhesion to the extracellular matrix and can be involved in the regulation of tumor metastatic capacity. Therefore, cytoplasmic and membrane proteins were extracted from BLCA cells transfected with RhoB overexpression plasmids or control plasmids. Western blot analysis was used to detect the integrin β1 and β3 proteins, with GAPDH and ATP1A1 serving as internal references for cytoplasmic and membrane proteins, respectively. The