PIN1 prolyl isomerase promotes initiation and progression of bladder cancer through the SREBP2-mediated cholesterol biosynthesis pathway

Abstract

Identities of functional pSer/Thr.Pro protein substrates of the PIN1 prolyl isomerase and its effects on downstream signaling in bladder carcinogenesis remain largely unknown. Phenotypically, we found that PIN1 positively regulated bladder cancer cell proliferation, cell motility and urothelium clearance capacity in vitro and controlled tumor growth and potential metastasis in vivo. Mechanistically, we observed a negative enrichment of SREBP2-driven cholesterol metabolism pathways and a decrease in free/total cholesterol levels in PIN1-knockout bladder cancer cells. Moreover, we showed that PIN1 interacted with SREBP2 following its phosphorylation by the JNK MAP kinase at Ser455, which lies near the Site-2 cleavage site that generates the active, nuclear-form of SREBP2. Therapeutically, a combination of the sulfopin PIN1 covalent inhibitor and the simvastatin HMGCoA reductase inhibitor suppressed cell proliferation in vitro and tumor growth in vivo synergistically. Together, these findings emphasize that PIN1 can act as a driver and potential therapeutic target in bladder cancer.

Introduction

Bladder cancer (BLCA) poses a serious public health challenge and is the most common urinary system cancer, with ~83K new cases (75.8% men), and ~12K male/~4K female deaths in the US in 2024¹. Most BLCA cases are the urothelial carcinoma type and approximately 75% of these are non-muscle-invasive bladder cancers (NMIBC)². NMIBC can reduce the life quality of patients and is one of the most expensive cancers due to life-long routine surveillance and clinical treatments³. The other 25% are muscle invasive bladder cancers (MIBC), which progress rapidly to become metastatic causing high patient mortality. However, despite its prevalence and adverse impact on human health, BLCA has been relatively understudied compared to other cancers.

Phosphorylation of proteins on serine or threonine residues preceding proline (pSer/Thr-Pro) has been reported to be a principal signaling mechanism in controlling cell proliferation and transformation, and its dysregulation causes human cancers^{4, 5}. Ser/Thr-Pro bonds exist in two distinct cis and trans conformations, the cis-trans interconversion can be catalyzed specifically by peptidyl prolyl cis/trans isomerases (PPIases) including a unique and conserved phospho-dependent PPIase, Pin1^{6, 7}. The Pin1 (protein interacting with never in mitosis A-1) protein sequence consists of an N-terminal binding WW domain, a flexible linker and a C-terminal catalytic PPIase domain⁷. Subsequent studies have demonstrated that Pin1 plays a pivotal role in both initiation and progression of human cancers^7–13. However, the functional roles of PIN1 and its potential downstream targets in human BLCA remains to be investigated.

Sterol regulatory element-binding proteins (SREBPs), a family of helix-loop-helix leucine zipper transcriptional factors have been implicated in controlling cellular lipid metabolism and maintaining cholesterol homeostasis^{14, 15}. SREBPs consist of three isoforms in humans: SREBP1a and SREBP1c encoded by the SREBF1 gene, and SREBP2 encoded by the SREBF2 gene¹⁴. The SREBP2 activation process depends on two-step proteolytic cleavage by Site-1 and Site-2 proteases on the Golgi, releasing its transcriptionally active N-terminal domain from the membrane, which translocates into the nucleus and binds to sterol regulatory elements (SREs) within the promoter of genes involved in cholesterol uptake or synthesis¹⁶. Importantly, the transcriptional activity of nuclear SREBP2 is also regulated by various post-translational modifications including phosphorylation. Insulin-activated ERK-MAPK increases SREBP2 activity via phosphorylation at Ser432 and Ser455¹⁷. Glycogen synthase kinase 3 directly phosphorylates Ser443 on SREBP2 to mediate Fbw7-induced ubiquitination and degradation of nuclear SREBP2¹⁸. Whether PIN1 might regulate SREBP2-mediated cholesterol homeostasis in BLCA has not been explored.

Despite recent considerable progresses in BLCA treatment including the utilization of immune checkpoint inhibitors, targeted therapies, and antibody-drug conjugates^19–21, there is still a need for more curative treatments for aggressive BLCA. HMGCR, the rate-limiting enzyme in the cholesterol biosynthesis pathway whose inhibition suppresses the SREBP2 signaling pathway is a potential cancer target. The simvastatin HMGCR inhibitor is effective in treating hepatocellular carcinoma²². In addition, statins can synergize with other inhibitors or drugs to restrain tumor growth to a greater extent²³. Recent work indicates that the combination of simvastatin and romidepsin, a class I HDAC inhibitor, clinically approved for cutaneous T cell lymphoma treatment²⁴, kills BLCA cells synergistically²⁵. However, no studies have examined whether PIN1 inhibition combined with simvastatin might have synergistic anticancer effects in BLCA.

In the current study, using PIN1 gene knockout (KO) and re-expression in huBLCA and mouse bladder carcinoma cells, we confirmed positive regulatory roles of PIN1 in BLCA cell proliferation and cancer progression in vitro and tumor growth and potential metastasis in vivo. Importantly, the PIN1 and SREBP2 interaction links PIN1’s regulation to activation of the SREBP2-mediated cholesterol biosynthesis pathway in BLCA cells. Moreover, we find that a combination of the PIN1 inhibitor sulfopin and the HMGCR inhibitor simvastatin acts as a potent anticancer therapy suppressing tumor growth in vivo.

Results

PIN1 predominantly localizes in intermediate and umbrella urothelial cell layers and is highly expressed in human BLCA cells

To examine expression and function of PIN1 in normal urothelium or in huBLCA, we analyzed two published single-cell RNA sequencing databases. Based on the first scRNA-seq dataset²⁶, we did cell clustering using marker genes and obtained imputed expression levels of PIN1 shown in the tSNE and violin plots based on an ALRA (Adaptively-thresholded Low Rank Approximation) algorithm. PIN1 was highly expressed in luminal cells compared to that in basal cells for both normal mouse and human urothelial cells (Supplementary Fig. S1A, B). Analyses of the second scRNA-seq dataset²⁷ revealed a similar PIN1 expression pattern in the human normal urothelium (Supplementary Fig. S1C, D). Cycling luminal cells showed higher PIN1 expression levels compared to those from other urothelial cell types for normal and tumors (Supplementary Fig. S1C, D). Additionally, based on a public dataset containing 460 early-stage BLCA patients²⁸, significantly higher PIN1 expression levels were observed in both luminal CIS-like and luminal classes compared to that in the early basal-like class (Supplementary Fig. S1E). Utilizing a gene set predictor, BASE47 signature²⁹ combined with gene expression data from 460 early-stage BLCA patients, we found a similar PIN1 expression pattern in the basal and luminal urothelial cell types (Supplementary Fig. S1F). Moreover, early-stage BLCA patients with high PIN1 expression levels showed significantly shorter progression-free survival (Supplementary Fig. S1G).

To further investigate clinical correlations between PIN1 expression and BLCA, we evaluated PIN1 protein expression in huBLCA tissue microarrays. As shown in Fig. 1A and B, PIN1 expression was increased significantly in non-muscle invasive and muscle invasive huBLCA samples compared to that in adjacent/cancer adjacent bladder tissues. Moreover, muscle invasive BLCA samples possessed higher PIN1 expression than that in non-muscle invasive samples. Additionally, PIN1 was predominantly expressed in intermediate and umbrella cell layers of adjacent and cancer samples rather than in the basal cell layer (Fig. 1C and D). By immunostaining for the cell proliferation marker, Ki67, we found that PIN1⁺ urothelial cells in huBLCA samples were hyperproliferative (Fig. 1E and F).

Figure 1. PIN1 predominantly localizes in the luminal urothelial layer of the bladder and PIN1⁺ cancer cells exhibit higher proliferation.

A. IHC staining of PIN1 in human bladder cancer (huBLCA) tissue microarray. (Representative images were shown for adjacent normal bladder tissues, non-muscle invasive and muscle invasive bladder cancer samples. Scale bars, 100 μm). B. Bladder cancer samples including non-muscle and muscle-invasive cancer samples showed higher PIN1 expression levels than those of adjacent/cancer adjacent bladder tissues. Compared to the non-muscle invasive bladder cancer samples, muscle invasive cancer samples showed higher PIN1 expression levels (n= 10 adjacent normal bladder tissues, n= 22 cancer adjacent bladder tissues, n= 30 non-muscle invasive BLCA, n= 42 muscle invasive BLCA. H-scores are calculated utilizing the QuPath software). C. Triple staining of urothelial cell markers, Cytokeratin 5 (CK5), Cytokeratin 8 (CK8) and PIN1 in a huBLCA tissue microarray. (Scale bars, 50 μm). D. Quantification of the percentage of CK5⁺CK8^low/- (basal), CK5⁺CK8⁺ (intermediate) and CK5^low/-CK8⁺ (umbrella) populations in PIN1⁺ human urothelial cells. (n= 12 cancer adjacent bladder tissues, n= 12 bladder cancer tissues. PIN1⁺ urothelial cells from 10 fields in each core of the huBLCA microarray slide were counted. Of PIN1⁺ urothelial cells, cell numbers from different urothelial layers (basal, intermediate, and umbrella) were counted. The Y axis showed the percentage of cell numbers from different urothelial layers in that from PIN1⁺ urothelial cells.) E. Representative images of cell proliferation marker Ki67 immunofluorescence staining on huBLCA and cancer adjacent tissue sections. (Scale bars, 50 μm). F. Quantification of the percentage of Ki67⁺ urothelial cells in huBLCA and cancer adjacent tissues. (n= 12 cancer adjacent bladder tissues, n= 12 bladder cancer tissues. Pan-keratin⁺ urothelial cells from 10 fields in each core of the huBLCA microarray slide were counted. The Y axis represented the percentage of PIN1⁺Ki67⁺ cell numbers in that from pan-keratin⁺ urothelial cells). G. Triple immunostaining of urothelial cell markers, CK5, CK8 and Pin1 in normal mouse bladder tissues. (Scale bars, 25 μm. The arrows pointed to Pin1⁺ mouse basal, intermediate and luminal urothelial cells). H. Quantification of the percentage of CK5⁺CK8^low/-, CK5⁺CK8⁺, CK5^low/-CK8⁺ populations in Pin1⁺ mouse bladder urothelial cells. (n = 3 mice. Pin1⁺ urothelial cells located in CK5⁺ basal layer and CK5⁻ inner layers from 5 frozen sections per mouse, 5 fields per frozen section were counted. Of Pin1⁺ urothelial cells, cell numbers from different urothelial layers (basal, intermediate, and umbrella) were counted. The Y axis showed the percentage of cell numbers from different urothelial layers in that from Pin1⁺ urothelial cells.) I. Representative images of Ki67 immunostaining on normal mouse bladder tissues. (Scale bars, 25 μm. The arrow pointed to a Pin1⁺CK8⁺Ki67⁺ mouse urothelial cell.) J. Quantification of the percentage of Pin1⁺CK8⁺Ki67⁺ urothelial cells following mouse bladder tissue development (n = 3 mice. 6-week and 10-week-old, two ages were chosen. Pin1⁺CK8⁺ urothelial cells from 5 frozen sections per mouse, 5 fields per frozen section are counted. The Y axis showed the percentage of Ki67⁺ cells in Pin1⁺CK8⁺ urothelial cells. Source data are provided as a Supplementary table 3. All p values were calculated using a two-tailed unpaired Student t-test. Data are presented as mean + s.e.m.).

In addition, intermediate and umbrella cells of normal mouse urothelium showed higher Pin1 expression than basal cells (Fig. 1G and H). Next, we found that ~80% Pin1⁺ urothelial cells from 6-week-old normal mice were proliferative, with a decrease observed during development (Fig. 1I and J). Collectively, PIN1 expression patterns and proliferative state obtained from human microarray analyses and wild-type mouse bladder tissues are in line with expression data from public scRNA-seq of normal human, normal mouse bladder tissues and huBLCA samples and RNA-seq of 460 early-stage BLCA patients.

PIN1 positively affects huBLCA cell proliferation, stemness maintenance, cell invasion, migration and urothelium clearance capacity in vitro

To explore possible functional roles of PIN1 in BLCA initiation and progression, we classified 5 huBLCA cell lines into 2 groups, PIN1^low (RT4 and SW780) and PIN1^high (TCCSUP, T24 and 5637) cell lines based on PIN1 RNA and protein expression levels; SV-HUC-1 SV40-immortalized human urothelial cells were used as a normal control (Supplementary Fig. S2A, B). Then, PIN1-overexpressing SV-HUC-1 and RT4 cells were established and showed a marked promotion of cell proliferation compared to control cells (Supplementary Fig. S2C, D). Furthermore, PIN1 overexpression increased SV-HUC-1 and RT4 cell motility (Supplementary Fig. S2E, F), cell invasion and migration compared to control cells (Supplementary Fig. S2G, H). Next, we explored whether overexpressing PIN1 influenced maintenance of stemness and cancer progression. A spheroid-formation assay showed that PIN1 overexpressing SV-HUC-1 and RT4 cells formed more spheres compared to control cells (Supplementary Fig. S2I, J). Using a urothelium clearance assay, we found that PIN1 overexpressing RT4 cells exhibited increased urothelium clearance (Supplementary Fig. S2K). Taken together, our findings indicate that overexpression of PIN1 contributes to increased cell proliferation, invasion, migration, stemness maintenance in both SV-HUC-1 and RT4 cells and urothelium clearance in RT4 cells.

To test whether PIN1 has an indispensable role in controlling BLCA cell proliferation and progression, we generated PIN1 Knockout (KO) cells and re-expressed wild-type PIN1 in control and PIN1-KO T24 and 5637 cells (Supplementary Fig. S3A–D). Significantly decreased cell proliferation was detected in PIN1-KO T24 and 5637 cells and restored upon re-expressing wild-type PIN1 in PIN1-KO cells (Supplementary Fig. S3E, F). Importantly, PIN1-KO T24 and 5637 cells exhibited a remarkable decrease in spheroid-forming capacity and re-expressing WT PIN1 in PIN1-KO cells reversed this defect (Supplementary Fig. S3G–J). Further immunostaining of spheroid sections revealed reduced Ki67⁺ signals in PIN1-KO T24 and 5637 cells, that were partially restored upon re-expressing PIN1 (Supplementary Fig. S3K–N). Moreover, PIN1 ablation in T24 and 5637 cells decreased cell invasion, migration, and cell motility, which were recovered upon re-expressing PIN1 (Supplementary Fig. S4A–J). As shown in Supplementary Fig. S4K–N, PIN1 deletion in T24 and 5637 huBLCA cells suppressed urothelium clearance compared to control cells and was rescued by PIN1 re-expression.

To further determine which functional domains of PIN1 were important, we stably re-expressed full-length, WW domain binding defective mutant (PIN1 W34A) and catalytically inactive mutant (PIN1 K63A) forms of PIN1 in PIN1-KO T24 and 5637 cells (Supplementary Fig. S4O) and found that re-expressing full-length PIN1, but not the binding or catalytic mutants in PIN1-KO T24 and 5637 cells restored proliferation comparable to the control cell levels (Supplementary Fig. S4P, Q). Collectively, these findings pinpoint PIN1 as a positive regulator in controlling BLCA cell proliferation and progression in vitro.

PIN1 deletion suppresses the growth of xenograft tumors formed from huBLCA cells

To examine whether our in vitro results could be replicated in vivo, we utilized a flank xenograft implantation model. Injections of PIN1-KO T24 and 5637 cells did not affect the body weight of mice throughout tumor development (Supplementary Fig. S5A). Tumors formed from PIN1-KO 5637 and T24 cells displayed significantly lower tumor weight and volume, whereas PIN1 re-expressing cells formed tumors similar in size to control cells (Fig. 2A–C and Supplementary Fig. S5B–D). Interestingly, H&E staining of tumor sections derived from PIN1-KO T24, and 5637 cells showed a decline in urothelial cell layers, which indicates possible alterations in urothelium composition due to PIN1 knockout (Supplementary Fig. S5E, F). Indeed, triple immunostaining analyses of tumor sections revealed remarkable increases in the percentage of intermediate (CK5⁺CK8⁺) and umbrella (CK5⁻CK8⁺) cells and decreases in the percentage of basal (CK5⁺CK8⁻) cells in tumor samples prepared from PIN1-KO 5637 cells compared to those from control cells (Fig. 2D and E). Re-expression of PIN1 in PIN1-KO 5637 cells reversed these phenotypes except the percentage of CK5⁻CK8⁺ umbrella cells (Fig. 2D and E). Similar results were obtained for T24 cells (Supplementary Fig. S5G, H). These findings suggest that PIN1 deletion promotes basal-to-luminal differentiation and blunts BLCA progression. Next, we investigated whether the ablation of PIN1 influenced tumor cell proliferation and apoptosis in vivo. To this end, we stained for cell proliferation and apoptosis markers and discovered that tumor samples derived from PIN1-KO 5637 and T24 cells exhibited decreased cell proliferation and increased apoptosis (Fig. 2F–I and Supplementary Fig. S5I, J and S5K, L), which provides a possible explanation for the smaller tumor size formed from PIN1-KO cells. Notably, tumors derived from PIN1 re-expressing cells did not display these phenotypes.

Figure 2. PIN1 promotes 5637 cell tumor growth in vivo.

A-C. Flank xenografts formed from PIN1-KO 5637 cells had decreased tumor weight (B) and volume (C); re-expression of PIN1 in PIN1-KO 5637 cells restored tumor formation similar to that observed with control cells. (n= 6 mice for each group). D, E. Immunostaining of urothelial cell markers, CK5 and CK8 showing remarkable increase in the percentage of intermediate (CK5⁺CK8⁺) and umbrella (CK5⁻CK8⁺) cells and decrease in the percentage of basal (CK5⁺CK8⁻) cells in tumor masses prepared from PIN1-KO 5637 cells compared to those from control cells. Re-expression of PIN1 in PIN1-KO 5637 cells reversed these phenotypes except the percentage of umbrella cells. (n= 3 tumor masses. Urothelial cells (including cells from CK5⁺ basal layers and CK5⁻ inner layers) from 5 frozen sections per tumor mass, 5 fields per frozen section were counted. Of urothelial cells, cell numbers from different urothelial layers (basal, intermediate, and umbrella) were counted. The Y axis showed the percentage of cell numbers from different urothelial layers in that from urothelial cells.) F-I. Cell proliferation marker Ki67 immunostaining showing a significant decrease in cell proliferation in tumor masses formed from PIN1 KO 5637 cells (F and H). PIN1 re-expression in PIN1-KO 5637 cells restored cell proliferation similar to that with control cells (F and H). Immunostaining for cleaved caspase 3 (CC3) apoptosis marker showing a significant increase in apoptosis in tumor masses formed from PIN1-KO 5637 cells (G and I). PIN1 re-expression in PIN1-KO 5637 cells restored the level of apoptosis similar to that in control cells (G and I). (Scale bars, 50 μm. n= 3 tumor masses. Pan-keratin⁺ urothelial cells from 5 frozen sections per tumor mass, 5 fields per frozen section were counted. The Y axis in the H and I showed the percentage of Ki67⁺ and CC3⁺ cells in Pan-keratin⁺ urothelial cells.) J, K. Bladder orthotopic implants prepared from PIN1-KO 5637 cells displaying decreased tumor growth (lower luciferase signal) compared to those from control and PIN1 re-expressing 5637 cells. (n= 4 mice per group. All p values were calculated using a two-tailed unpaired Student t-test. Data are presented as mean + s.e.m. ns, not significant (p > 0.05)). L. H&E staining analyses confirmed the tumor mass formed within the bladder lumen. (Scale bars, 500 μm. The arrow points to the tumor mass).

As a more physiological model of BLCA, we used an orthotopic transplantation model. As shown in Fig. 2J and K, we observed smaller tumors formed in the bladder lumen with PIN1-KO 5637 cells compared to those from control and PIN1 re-expressing cells. Additionally, H&E staining analyses confirmed the smaller tumor mass derived from PIN1-KO 5637 cells formed within the bladder lumen (Fig. 2L and Supplementary Fig. S6A). Further immunostaining on orthotopic graft sections confirmed a reduction in the percentage of basal (CK5⁺CK8⁻) cells and an increase in the percentage of intermediate (CK5⁺CK8⁺) cells in grafts formed from PIN1-KO 5637 cells (Supplementary Fig. S6B, C). Moreover, orthotopic grafts prepared from PIN1-KO 5637 cells displayed decreasing proliferation, Ki67⁺ cells (Supplementary Fig. S6D, E) and increasing apoptosis, CC3⁺ cells (Supplementary Fig. S6F, G), which strongly supports the results of the subcutaneous transplantation model. Collectively, these findings underscore that PIN1 plays a positive role in BLCA growth in vivo.

The PIN1 inhibitor, sulfopin suppresses huBLCA cell proliferation, spheroid-formation, cell invasion and migration in vitro and tumor growth in vivo

Previous work has suggested that pharmacological inhibition of PIN1 has the potential to simultaneously block multiple cancer-driving pathways¹³ and suppress oncogenesis. Several PIN1 inhibitor compounds, including juglone³⁰, all-trans retinoic acid (ATRA)³¹, arsenic trioxide (ATO)³² and KPT-6566³³, have been shown to exhibit anti-cancer activity and have been utilized to investigate the role of PIN1 in tumor initiation and progression. However, these inhibitors are not very specific, and here we chose a newly developed highly selective and specific PIN1 covalent inhibitor, sulfopin. Sulfopin has been shown to negatively affect cell viability in vitro and reduce tumor initiation and progression in murine and zebrafish models of MYCN-driven neuroblastoma³⁴. We set out to assess the effects of sulfopin on wild-type 5637 and T24 cells and found that sulfopin reduced PIN1 protein expression levels in a time- and dose-dependent way (Fig. 3A and Supplementary Fig. S7A). In line with results obtained from PIN1-KO cells, sulfopin-treated wild-type 5637 and T24 cells displayed significantly lower cell proliferation compared to those with DMSO treatment (Fig. 3B and Supplementary Fig. S7B), as was also the case for sulfopin (1 μM) treatment of PIN1 re-expressing cells (Fig. 3C, D and Supplementary Fig. S7C, D). However, the proliferation of PIN1-KO 5637 and T24 cells was unaffected by sulfopin (1 μM) treatment (Fig. 3E and Supplementary Fig. S7E). Meanwhile, sulfopin treatment not only decreased spheroid-formation (1 μM) (Supplementary Fig. S7F–I) but also cell invasion and migration (2 μM) of 5637 and T24 control and PIN1 re-expressing cells (Supplementary Fig. S7J–Q). Likewise, spheroid-formation capacity, cell invasion and migration of PIN1-KO 5637 and T24 cells were not influenced by sulfopin treatment (Supplementary Fig. S7F–Q).

Figure 3. Treatment with sulfopin, a PIN1 specific inhibitor, reduces 5637 cell proliferation in vitro and tumor growth in vivo.

A. IB confirming time- and dose-dependent inhibition of PIN1 protein expression following sulfopin treatment of 5637 cells. Eight hours later, a nearly 90% decrease in PIN1 protein expression was observed in sulfopin-treated 5637 cells. (Numbers labeled above the PIN1 protein band are normalized to DMSO treatment). B. Reduced cell proliferation in wild-type 5637 cells treated with different concentrations of sulfopin. Increased cells death was observed with a high concentration (5 μM) of sulfopin after 6 days. (n= 4 technical replicates. The Y axis represented the absorbance at 450 nm, normalized to the absorbance at 0 h). C-E. Sulfopin-treated control (C) and PIN1 re-expressing 5637 cells (D) exhibit lower cell proliferation, whereas sulfopin-treated and untreated PIN1-KO cells (E) exhibit similar proliferation (n= 4 technical replicates. Con, control; KO, knockout; RE, rescue. The Y axis represented the absorbance at 450 nm, normalized to the absorbance at 0 h). F-H. Sulfopin treatment significantly reduced the weight (G) and volume (H) of tumors formed from 5637 control cells, whereas no inhibitory effect was observed with tumors from 5637 PIN1-KO cells. (n= 7 mice each group). I, J. Sulfopin treatment led to a significant increase in the percentage of intermediate (CK5⁺CK8⁺) and umbrella (CK5⁻CK8⁺) cells and a decrease in the percentage of basal (CK5⁺CK8⁻) cells in tumor masses prepared from 5637 control cells. Sulfopin untreated and treated PIN1-KO cells showed no difference in percentages of the 3 typical urothelial subtypes. (n= 3 tumor masses. Urothelial cells (including cells from CK5⁺ basal layers and CK5⁻ inner layers) from 5 frozen sections per tumor mass, 5 fields per frozen section were counted. Of urothelial cells, cell numbers from different urothelial layers (basal, intermediate, and umbrella) were counted. The Y axis showed the percentage of cell numbers from different urothelial layers in that from urothelial cells.) K-N. Cell proliferation marker Ki67 immunostaining showing a significant decrease in cell proliferation in tumor masses formed from 5637 control cells following sulfopin treatment (K and M). Immunostaining for the cleaved caspase 3 (CC3) apoptosis marker showing a strong increase in cell apoptosis in tumor masses formed from 5637 control cells following sulfopin treatment (L and N), but sulfopin did not affect cell proliferation and apoptosis in tumors formed from PIN1-KO 5637 cells. (Scale bars, 50 μm. n= 3 tumor masses. Pan-keratin⁺ urothelial cells from 5 frozen sections per tumor mass, 5 fields per frozen section were counted. The Y axis in the M and N represented the percentage of Ki67⁺ and CC3⁺ cells in pan-keratin⁺ urothelial cells. All p values were calculated using a two-tailed unpaired Student t-test. Data are presented as mean + s.e.m. ns, not significant (p > 0.05)).

Similar to phenotypes obtained with PIN1-KO cells, we observed that sulfopin administration suppressed the growth of 5637 and T24 control cell tumors but did not further suppress the growth of PIN1-KO cell tumors (Fig. 3F–H and Supplementary Fig. S8A–C). No significant changes in body weight were observed following BLCA cell injections and sulfopin treatment (Supplementary Fig. S8D). H&E staining analyses on tumor sections derived from T24 and 5637 control cells showed that sulfopin treatment led to a decline in urothelial cell layers (Supplementary Fig. S8E, F). In sulfopin-treated 5637 tumors, we detected an increase in the percentage of intermediate (CK5⁺CK8⁺) and umbrella (CK5⁻CK8⁺) cells and a decrease in the percentage of basal (CK5⁺CK8⁻) cells (Fig. 3I and J). Decreased cell proliferation and increased apoptosis were noted in tumors from mice treated with sulfopin (Fig. 3K–N), confirming anti-tumor effects of sulfopin. However, no sulfopin-dependent alterations in cell proliferation and apoptosis were observed in tumor samples derived from PIN1-KO 5637 cells, indicating lack of off-target effects of sulfopin. Similar results were obtained for T24 cell tumors (Supplementary Fig. S8G–L). In conclusion, these data demonstrate that the PIN1 inhibitor, sulfopin negatively affects huBLCA cell proliferation, spheroid-formation, invasion and migration in vitro and reduces tumor growth through inhibiting cell proliferation and promoting apoptosis in vivo.

Pin1 promotes mouse urothelial carcinoma cell proliferation and spheroid-formation in vitro and tumor growth in vivo

To further examine functional roles of Pin1 in mouse bladder cells, we utilized MB49, a mouse bladder carcinoma cell line and established stable Pin1-KO and Pin1 re-expressing MB49 cells (Fig. 4A). To test whether Pin1 is required for MB49 cell proliferation and spheroid-formation, we assessed control and Pin1-KO MB49 cells and observed decreased cell proliferation and spheroid-formation capacity in Pin1-KO MB49 cells (Fig. 4B–D) that were reversed by Pin1 re-expression (Fig. 4B–D).

Figure 4. Pin1 positively affects MB49 mouse bladder carcinoma cell proliferation, spheroid-formation in vitro and tumor growth in vivo.

A. IB confirming efficient Pin1 ablation in sgRNA-transfected MB49 cells (obtained through a limiting dilution and clonal selection assay) and Pin1 restoration in MB49 rescue cells. (Con, control; KO, knockout; RE, rescue). B. Cell proliferation assay showing lower cell proliferation in Pin1-KO MB49 cells; re-expression of Pin1 in Pin1-KO MB49 cells restored proliferation to a level comparable with control cells (n= 4 technical replicates. Y axis represented the absorbance at 450 nm, normalized to the absorbance at 0 h). C, D. In vitro spheroid-formation assay showing fewer spheroids produced from Pin1-KO MB49 cells; re-expression of Pin1 in Pin1-KO MB49 cells restored spheroid-formation comparable with control cells (n= 4 separate wells. Scale bars, 40 μm for magnified images and 200 μm for low magnification images. KO, knockout). E-G. Tumors formed from Pin1-KO MB49 cells had decreased tumor weight (F) and volume (G); re-expression of Pin1 in Pin1-KO MB49 cells restored tumor weight and volume comparable to that from control cells (n= 8 xenografts per group). H-K. Tumor masses produced from Pin1-KO MB49 cells showed decreased cell proliferation (H and J) and increased cell apoptosis (I and K). Pin1 re-expression in Pin1-KO MB49 cells restored cell proliferation and apoptosis comparable with control cells. (Scale bars, 50 μm. n= 3 tumor masses. Pan-keratin⁺ urothelial cells from 5 frozen sections per tumor mass, 5 fields per frozen section were counted. The Y axis in J and K showed the percentage of Ki67⁺ and CC3⁺ cells in pan-keratin⁺ urothelial cells.) L, M. Bladders instilled with Pin1-KO MB49 cells showed smaller tumor size (weaker luciferase signal). Bladders instilled with control and Pin1 re-expressing MB49 cells showed comparable tumor size (similar luciferase signals) (Control group, n= 10, Pin1 Knockout group, n= 9, Rescue group, n= 10 mice). N-P. Bladders with Pin1-KO MB49 cell instillation displayed reduced weight (O) and smaller size (P) compared to those with control and Pin1 re-expressing MB49 cell instillation. (Control group, n= 10, Pin1 Knockout group, n= 9, Rescue group, n= 10). Q. In vitro bioluminescence imaging confirmed the metastasis of luciferase expressing MB49 cells to kidneys and lungs. R. Kaplan-Meier curve showing prolonged survival of mice with Pin1-KO MB49 cell instillation compared to those with control and Pin1 re-expressing MB49 cell instillation. (Control group, n= 10, Pin1 Knockout group, n= 9, Rescue group, n= 10. Survival analysis was performed using the Mantel-Cox log-rank test. All p values were calculated using a two-tailed unpaired Student t-test. Data are presented as mean + s.e.m. ns, not significant (p > 0.05)).

To explore the role of Pin1 in MB49 cell tumor growth, we subcutaneously inoculated C57/B6 mice with control, Pin1-KO and Pin1 re-expressing MB49 cells. Tumors formed from Pin1-KO MB49 cells displayed a significant decrease in tumor weight and volume, compared to control cells, and this defect was largely reversed with Pin1 re-expression (Fig. 4E–G). However, there was no change in the percentage of different urothelial cell types in tumors formed by Pin1-KO MB49 cells (Supplementary Fig. S9A, B). Furthermore, Pin1-KO MB49 cell tumors had decreased cell proliferation and increased apoptosis (Fig. 4H–K).

To strengthen results obtained with the subcutaneous allograft model, we performed an orthotopic implantation model. We observed a reduction in body weight of mice injected with control and Pin1 re-expressing MB49 cells compared to those with Pin1-KO cells (Supplementary Fig. S9C, D), which may be due to the loss of appetite caused by the growing tumor mass within the bladder lumen and tumor clogs in the urethra. Additionally, larger tumors were found in bladder lumens of mice injected with control and Pin1 re-expressing MB49 cells compared to those with Pin1-KO cells (Fig. 4L and M). Compared with control and Pin1 re-expressing MB49 cell injections, bladders harvested from mice with Pin1-KO cell injection were reduced in weight and volume (Fig. 4N–P). Meanwhile, we investigated the metastatic burden in distant organs of mice with MB49 cell instillation. Lymph nodes of normal mice are small and difficult to distinguish from surrounding adipose and connective tissues, whereas enlarged lymph nodes can be easily found in mice with acute inflammation or advanced cancer³⁵. Although no metastasis was observed in lymph nodes near the bladder, lymph nodes harvested from mice injected with Pin1-KO MB49 cells were smaller than the enlarged nodes from mice with control and Pin1 re-expressing cells injection (Supplementary Fig. S9E). Of note, in vitro bioluminescence imaging and immunohistochemical staining confirmed the metastasis of luciferase-expressing control but not Pin1-KO MB49 cells to lungs and kidneys (Fig. 4Q and Supplementary Fig. S9F). Furthermore, survival analyses indicated that mice inoculated with Pin1-KO MB49 cells survived significantly longer than those inoculated with control or Pin1 re-expressing cells (Fig. 4R). We conclude that Pin1 plays a positive role in promoting mouse bladder carcinoma cell proliferation and stemness maintenance in vitro and tumor growth and metastasis in vivo, similar to the role for PIN1 in huBLCA initiation and progression.

Pin1 facilitates the growth of organoids produced from primary mouse urothelial cells

To confirm the indispensable role of Pin1 in monolayer culture, we performed CRISPR/Cas9 genome editing for Pin1 gene in organoids derived from mouse primary urothelial cells and observed that fewer organoids were generated from Pin1-KO primary cells compared to those from control cells (Supplementary Fig. S10A–C). In secondary and tertiary organoid formation assays we found that re-expressing Pin1 in Pin1-KO mouse primary urothelial cells partially reversed organoid-formation defects resulting from Pin1 knockout (Supplementary Fig. S10B–D). Moreover, organoids produced from mouse primary urothelial cells contained both basal and luminal cell types confirmed by immunostaining (Supplementary Fig. S10E). Organoids derived from Pin1-KO primary urothelial cells exhibited a significantly decreased percentage of Pin1⁺ cells and mean Pin1-positive signal (Supplementary Fig. S10E–G). Additionally, organoids formed from Pin1-KO primary urothelial cells showed declining proliferation in total urothelial cells and different urothelial cell types (Supplementary Fig. S10H–K). Collectively, these data highlight that Pin1 plays a pivotal role in controlling mouse urothelial cell proliferation and organoid formation.

Based on RNA-sequencing data, GSEA enrichment analyses indicate that cholesterol metabolism-associated pathways are positively correlated with PIN1 expression levels

To elucidate underlying molecular mechanisms by which PIN1 mediated cis-trans isomerization of phosphosites in target proteins induces functional alterations in downstream signaling in BLCA cells, we performed bulk RNA-sequencing using T24 and 5637 huBLCA cells. We quantified reads derived from raw data and re-confirmed PIN1 expression levels of each sample for sequencing (Supplementary Fig. S11A, B). In line with our previous results obtained from huBLCA cells, GSEA enrichment analyses confirmed that stemness-related and cell proliferation-associated gene signatures were negatively enriched in PIN1-KO cells compared to control cells (Supplementary Fig. S11C, D), along with negative enrichment of cell invasion, migration-related and malignant progression-related gene signatures in PIN1-KO cells (Supplementary Fig. S11E, F).

Interestingly, based on GSEA enrichment analyses, we observed that cholesterol metabolism-associated pathways (including the crucial Bloch and Kandutsch-Russell pathways) were significantly downregulated in PIN1-KO cells compared to control cells (Supplementary Fig. S11G). Previous studies have demonstrated that cholesterol is an essential lipid involved in modulating normal biological functions, inflammation, metabolic disease, and cancer³⁶. Hence, we asked whether PIN1 affected free/total cholesterol levels in huBLCA cells. As shown in Fig. 5A and Supplementary Fig. S11H, free cholesterol levels were strongly decreased in PIN1-KO 5637 and T24 cells compared to those in control cells and restored with PIN1 re-expression. Next, we assessed changes in total cholesterol levels and found that PIN1-KO 5637 and T24 cells had lower total cholesterol levels compared to control cells, that were reversed upon PIN1 re-expression (Fig. 5B and Supplementary Fig. S11I).

Figure 5. PIN1 promotes the maintenance of cell cholesterol levels and positively correlates with expression levels of SREBP2 and its downstream targets.

A. Fillipin III staining assay showing significantly lower free cholesterol level in PIN1-KO 5637 cells comparably to control cells; re-expression of PIN1 in PIN1-KO 5637 cells restored free cholesterol levels comparably to control cells (n= 6 3 fields per well, 2 wells were analyzed. Scale bars, 40 μm. KO, knockout). B. Significantly lower total cholesterol levels were detected in PIN1-KO 5637 cells compared with control cells; re-expression of PIN1 in PIN1-KO 5637 cells restored total cholesterol levels comparably to control cells (n= 4 separate wells from a 96-well plate. The luminescence values recorded using a plate reader were normalized to cell numbers first and then to control). C. Compared to control and PIN1 re-expressing cells, PIN1-KO 5637 cells cultured in medium with lipoprotein-depleted FBS showed significantly lower cell proliferation; both control and rescue 5637 cells with media containing lipoprotein-depleted FBS showed comparable cell proliferation with media containing normal FBS. PIN1-KO 5637 cells cultured in medium with lipoprotein-depleted FBS exhibited increasing cell proliferation when supplemented with a cholesterol lipid concentrate (n= 4 wells from a 96-well plate at 96 h and 120 h, two timepoints. Normal, culture media with normal FBS; Depleted, culture media with lipoprotein-depleted FBS; Cholesterol, addition of a cholesterol lipid concentrate to culture media with lipoprotein-depleted FBS. The Y axis represented the absorbance at 450 nm, normalized to the absorbance at 0 h). D. Usage of the HMGCR (a rate-limiting enzyme in the cholesterol biosynthesis process) inhibitor, simvastatin reduced the wild type 5637 cell proliferation (n= 4 technical replicates. The Y axis represented absorbance at 450 nm, normalized to the absorbance at 0 h). E-G. IB confirming reduction in protein expression levels of SREBP2 (E), a key transcription factor in cholesterol metabolism-related pathways and its downstream targets, HMGCS1 (F) and MVD (G) following PIN1 deletion and restoration with re-expression of PIN1. (Numbers labeled above the bands of SREBP2 precursor (P) and nuclear (N) forms are normalized to control. Con, control; KO, knockout; RE, rescue). H, I. IB confirming efficient knockdown of SREBP2 protein expression in 5637 cells. Knockdown of SREBP2 reduced 5637 cell proliferation. (Con, control; shSRE #1, shSREBP2 #1; shSRE #5, shSREBP2 #5. n= 4 technical replicates. The Y axis of panel I represented the absorbance at 450 nm, normalized to the absorbance at 0 h). J. In vitro cell proliferation assays showing overexpression of SREBP2 in PIN1-KO 5637 cells restored cell proliferation comparable to control cells. (n= 4 technical replicates. KO, Knockout; RE, re-expression; OE, overexpression. The Y axis represented the absorbance at 450 nm, normalized to the absorbance at 0 h). K-P. Immunostaining of tumor xenografts formed from PIN1-KO 5637 cells showing decreased SREBP2 (K and N) and its downstream targets, HMGCS1 (L and O) and MVD (M and P) expression levels. Xenografts from PIN1 rescue 5637 cells showing expression levels of these cholesterol metabolism-associated genes similar to those from 5637 control cells. (Scale bars, 50 μm. n= 3 tumor masses. Pan-keratin⁺ urothelial cells from 5 frozen sections per tumor mass, 4 fields per frozen section were counted. The Y axis in N, O and P showed the percentage of nuclear SREBP2⁺, HMGCS1⁺ and MVD⁺ cells in pan-keratin⁺ urothelial cells. All p values were calculated using a two-tailed unpaired Student t-test. Data are presented as mean + s.e.m. ns, not significant (p > 0.05)).

To further examine whether decreased cholesterol levels influenced huBLCA cell proliferation, we checked proliferation of control, PIN1-KO and PIN1 re-expressing cells in culture media with either normal FBS or lipoprotein-depleted FBS. As shown in Fig. 5C and Supplementary Fig. S11J, PIN1-KO 5637 and T24 cells cultured in delipidated media showed significantly lower cell proliferation, whereas control and PIN1 re-expressing cells largely maintained cell proliferation, presumably due to intrinsic cholesterol biosynthesis. When lipoprotein-depleted media were supplemented with a cholesterol lipid concentrate, control, PIN1-KO and PIN1 re-expressing 5637 and T24 cells all exhibited similar cell proliferation to cells cultured in media with normal FBS (Fig. 5C and Supplementary Fig. S11J). As another means of assessing a requirement for cholesterol, we used simvastatin to inhibit HMGCR, the rate-limiting enzyme in cholesterol biosynthesis, and found that 5 μM simvastatin reduced proliferation of both 5637 and T24 cells (Fig. 5D and Supplementary Fig. S11K). These results imply that PIN1-dependent maintenance of normal cholesterol level plays a principal role in controlling huBLCA cell proliferation.

PIN1 positively modulates expression levels of SREBP2 and its downstream targets in vitro and in vivo

To explore the relationship between PIN1 and cholesterol biosynthesis pathway, we analyzed alterations in expression levels for key enzymes and regulators in Bloch and Kandutsch-Russell biosynthesis pathways, following PIN1 knockout and re-expression. We discovered that RNA levels of 21 pathway genes were significantly diminished in PIN1-KO cells but restored to near control cell levels upon re-expression of PIN1 (Supplementary Fig. S11L). Besides the heatmap analyses, 18 of 21 pathway genes (except ACOT2, NR1H3 and FDPS genes) RNA expression levels in PIN1-KO T24 and 5637 cells were consistently decreased with PIN1 deletion and restored following PIN1 re-expression (Supplementary Fig. S11M–P). Moreover, we eliminated three fatty acid metabolism closely related genes, FADS1/2 and FASN from the 18-gene list.

To identify candidate PIN1 target proteins among the remaining 15 cholesterol biosynthesis pathway proteins, we used the PhosphoSitePlus online motif analysis tool and found that 7 (SREBP2, HMGCR, HMGCS1, MVD, MVK, CYP51A1 and DHCR7) of the 15-cholesterol metabolism-related protein sequences are reported to contain pSer/Thr-proline sites, potential PIN1 binding sites. Intriguingly, Pin1-KO mice were reported to be resistant to high-fat-diet-induced increases in mRNA expression levels of several genes involved in lipid metabolism pathway including SREBP2 in adipose tissues³⁷, leading us to focus on investigating the relationship between PIN1 and the SREBP2-mediated cholesterol biosynthesis pathway in BLCA (Supplementary Fig. S11Q). Initially, we assessed protein expression levels of SREBP2 and found decreased SREBP2 expression in PIN1-KO 5637 and T24 cells that was restored following PIN1 re-expression (Fig. 5E and Supplementary Fig. S12A–D). Notably, expression of the nuclear-form of SREBP2 showed a remarkable decrease, compared to that of the precursor-form of SREBP2 (Fig. 5E and Supplementary Fig. S12D), which suggests that PIN1 may regulate the cleavage of SREBP2 protein. Additionally, protein levels of SREBP2’s targets HMGCS1, HMGCR, MVD and MVK declined because of PIN1 ablation and recovered after re-expressing PIN1 in PIN1-KO 5637 (Fig. 5F, G and Supplementary Fig. S12E, F) and T24 cells (Supplementary Fig. S12G–J). To assess whether SREBP2 reduction affected cell proliferation, we first knocked down SREBP2 in 5637 and T24 cells and detected an efficient decrease in expression of SREBP2 and its targets, HMGCS1 and MVD, as well as a slight reduction in PIN1 expression (Fig. 5H and Supplementary Fig. S12K–N). Next, we observed a significant decrease in proliferation of SREBP2 knockdown 5637 and T24 cells (Fig. 5I and Supplementary Fig. S12O). Notably, overexpression of SREBP2 in PIN1-KO 5637 and T24 cells increased cell proliferation to control cell levels (Fig. 5J and Supplementary Fig. S12P).

To confirm expression alterations of SREBP2 and its targets in tumor xenografts, we performed a triple immunostaining assay and observed a decrease in the percentage of active/nuclear SREBP2⁺ cells and its targets, HMGCS1 and MVD in PIN1-KO 5637 and T24 cell tumors, which was reversed in PIN1 re-expressing cells (Fig. 5K–P and Supplementary Fig. S12Q–V). In sum, these findings confirm that PIN1 positively regulated expression levels of SREBP2 and its targets, HMGCS1, HMGCR, MVD and MVK that are involved in the cholesterol biosynthesis process in huBLCA cells in vitro and in vivo.

To further elucidate whether Pin1 influences the SREBP2-mediated cholesterol biosynthesis pathway in MB49, mouse urothelial carcinoma cells, we tested changes in total cholesterol level and found that Pin1-KO MB49 cells possessed lower total cholesterol levels compared to those from control and Pin1 re-expressing cells (Supplementary Fig. S13A). Immunoblotting confirmed decreased expression levels of SREBP2 and its targets, HMGCS1 and MVD in Pin1-KO MB49 cells and reversal of their expression levels following Pin1 re-expression (Supplementary Fig. S13B–D). Next, we assessed expression levels of SREBP2 and its targets in MB49 cell allograft tumors and observed reduced expression of SREBP2 and its targets in tumor samples from Pin1-KO cells that were restored by re-expression of Pin1 in Pin1-KO MB49 cells (Supplementary Fig. S13E–J). Collectively, Pin1 plays a crucial role in regulating the SREBP2-mediated cholesterol biosynthesis pathway in MB49 cells, which is consistent with results from 5637 and T24 huBLCA cell xenografts.

To further investigate how PIN1 regulates SREBP2 in BLCA cells, we examined alterations in SREBP2 protein stability and SREBP2 transcriptional activity following PIN1-KO and re-expression. A CHX (a protein synthesis inhibitor) chase assay showed that PIN1 deletion did not affect the rate of protein degradation of SREBP2 and MVD in T24 and 5637 cells (Supplementary Fig. S14A–D). As shown in Supplementary Fig. S14E, F, the SREBP2 DNA-binding activity declined in PIN1-KO T24 and 5637 cells and was restored following PIN1 re-expression. Taken together, these results indicate that PIN1 regulated the cholesterol biosynthesis pathway through modulating SREBP2 transcriptional activity in BLCA cells.

The PIN1 WW domain binds to the JNK-dependent phospho-Ser455 SREBP2

Previous work has shown that PIN1 regulates the activity or stability of several transcription factors, including p53, c-Myc and p65 in a phosphorylation-dependent manner^38–40. As shown by the PhosphoSitePlus online motif analysis tool, the SREBP2 protein sequence contains several potential PIN1 binding sites, which were reported to be phosphorylated by MAP kinases, such as ERK and JNK (c-Jun N-terminal kinase). JNK activation has been reported to promote SREBP2 processing in hepatocytes⁴¹. Moreover, the inflammatory cytokine, tumor necrosis factor α (TNFα) activated JNK by activation of Rho kinase in human pulmonary microvascular endothelial cells⁴².

Based on these findings, we hypothesized that PIN1 binds to JNK-phosphorylated SREBP2 upon TNFα stimulation of BLCA cells. First, we tested alterations in the JNK pathway following TNFα stimulation and observed a rapid and transient increase in expression levels of phosphorylated-JNK and its targets, p-ATF2 and p-cJun (Fig. 6A and Supplementary Fig. S15A). Second, we utilized an enzyme-catalyzed proximity labeling approach⁴³ and validated fusion protein expression by anti-Flag blotting and IF staining (Fig. 6B and Supplementary Fig. S15B–E). As shown in Fig. 6C and Supplementary Fig. S15F, PIN1 interacted with SREBP2 in 5637 and T24 huBLCA cells.

Figure 6. JNK-mediated phosphorylation of Ser455 in SREBP2 is required for interaction between PIN1 and SREBP2 in 5637 cells.

A. Immunoblot showing increased phosphorylation of SAPK/JNK and its downstream targets, p-ATF2 and p-cJun following TNFα treatment of 5637 cells. Total SAPK/JNK protein expression did not change following TNFα treatment. (Numbers labeled above the band of phospho-JNK, p-ATF2 and p-cJun are normalized to protein expression levels at 0 min). B. Scheme showing Turbo ID-biotinylated proximal labeling interactome for PIN1 in BLCA cells. C. 5637 cells were stably transfected with the pLVX-Flag, pLVX-Flag-Turbo, and pLVX-Flag-Turbo-PIN1 constructs. Streptavidin-magnetic beads were utilized to enrich Turbo expressing 5637 cell extracts. Immunoblotting assay confirmed the interaction between PIN1 and SREBP2 proteins following TNFα stimulation. (Numbers labeled above the band of SREBP2 (including both precursor (P) and nuclear (N) forms) are normalized to control). D. IB showing the inhibition of phosphorylated SAPK/JNK and its downstream targets, p-ATF2 and p-cJun expression with TNFα treatment and JNK inhibitor, JNK-IN-8 in 5637 cells at the 10 min timepoint. Total SAPK/JNK protein expression did not change following TNFα and JNK-IN-8 treatment. (Numbers labeled above the bands of phospho-SAPK/JNK, p-ATF2 and p-cJun are normalized to control). E. The interaction between PIN1 and SREBP2 was abolished when treated with JNK inhibitors, SP600125 and JNK-IN-8 even with TNFα stimulation. (Numbers labeled above the band of SREBP2 (including both precursor (P) and nuclear (N) forms) are normalized to control). F. The PIN1 WW binding domain, but not the PPIase catalytic domain interacted with SREBP2 following TNFα treatment. (Numbers labeled above the band of SREBP2 are normalized to negative controls). G. Co-IP of the single W34A PIN1 binding domain and the K63A PIN1 catalytic domain mutants, and the W34A/K63A double mutant was used to confirm that the SREBP2 interaction required the PIN1 WW binding domain. (Numbers labeled above the band of SREBP2 (including both precursor (P) and nuclear (N) forms) are normalized to control). H. PIN1 bound to N-terminal domain of SREBP2 with TNFα treatment in 5637 cells. (Numbers labeled above the band of PIN1 are normalized to eluates from cells infected with Flag-tagged full-length SREBP2). I. Phosphorylation at Ser455 in SREBP2 is required for interaction between PIN1 and SREBP2. (Numbers labeled above the band of PIN1 are normalized to eluates from cells infected with Flag-tagged normal full-length SREBP2).

To further validate whether JNK-mediated phosphorylation of SREBP2 is required for the interaction between PIN1 and SREBP2, we used two JNK inhibitors: JNK-IN-8⁴⁴, the first selective and irreversible covalent inhibitor, and SP600125⁴⁵, a frequently used reversible ATP-competitive inhibitor to abolish JNK phosphorylation and activation. We confirmed that JNK-IN-8 and SP600125 both reduced expression levels of phosphorylated-JNK and its targets, p-ATF2 and p-cJun in 5637 and T24 cells (Fig. 6D and Supplementary Fig. S16A–C). Additionally, JNK-IN-8 and SP600125 treatment abolished the interaction between PIN1 and SREBP2 even with TNFα stimulation (Fig. 6E and Supplementary Fig. S16D). Collectively, these findings demonstrated that JNK phosphorylation of SREBP2 was required for the interaction between PIN1 and SREBP2. To determine which domain of PIN1 interacts with SREBP2, we generated PIN1 truncation mutants, N-terminal WW domain and C-terminal PPIase domain and observed that the WW domain interacted with SREBP2 in 5637 and T24 cells, similar to the interaction between full-length PIN1 and SREBP2 (Fig. 6F and Supplementary Fig. S16E). Moreover, the W34A non-binding mutant, and double mutant, W34A/K63A failed to interact with SREBP2 in 5637 and T24 cells (Fig. 6G and Supplementary Fig. S16F).

Conversely, to examine whether SREBP2 binds to PIN1, we overexpressed a Flag-tagged SREBP2 construct in 5637 and T24 cells and confirmed its expression by IB (Supplementary Fig. S16G). Previous work has demonstrated that SREBP2 is composed of three segments: an N-terminal segment that is a transcription factor, a membrane localization segment, and a C-terminal segment that plays a regulatory role^{46, 47}. To test which domain of SREBP2 binds to PIN1, we expressed N-terminal domain (NTD) and C-terminal domain (CTD) fragments in huBLCA cells. As shown in Fig. 6H and Supplementary Fig. S16H, full-length SREBP2 and the NTD but not the CTD of SREBP2 interacted with PIN1. To further determine which phosphosite(s) in the SREBP2 NTD is required for interaction between SREBP2 and PIN1, we used the PhosphoSitePlus online kinase library prediction tool to define candidate JNK sites; Ser106 and Ser455 in the SREBP2 NTD are most likely to be phosphorylated by JNK, and we generated alanine point mutants for these two sites. As shown in Fig. 6I and Supplementary Fig. S16I, only the Ser455 site, and not the Ser106 site in the SREBP2 NTD was indispensable for the interaction between SREBP2 and PIN1 in BLCA cells. Together, these findings indicate that PIN1 WW domain binds to pSer455 in the NTD of SREBP2, and vice versa.

SREBP2 Ser455 phosphorylation is required for PIN1’s activity in regulating cholesterol biosynthesis, cell proliferation and migration in vitro and tumor growth in vivo

To further determine the importance of SREBP2 S455 phosphorylation for the function of PIN1 in BLCA cells, we first re-expressed either wild-type SREBP2 or SREBP2 S455A mutant in SREBP2 knockdown (with shRNA targeting the SREBP2 3’ UTR) T24 and 5637 cells (Supplementary Fig. S17A, B). Next, we either treated cells with sulfopin, the PIN1 inhibitor (Supplementary Fig. S17A, B) or knocked out the PIN1 gene in control, SREBP2 knockdown and SREBP2 S455A mutant re-expressing T24 and 5637 cells (Supplementary Fig. S17C, D). As shown in Supplementary Fig. S18A, B, unlike re-expression of wild-type SREBP2 in SREBP2 knockdown cells, we found that re-expressing SREBP2 S455A mutant only partially restored total cholesterol levels (26.0% restoration for T24, 17.5% for 5637) compared to those of SREBP2 knockdown cells. Moreover, sulfopin PIN1 inhibitor treatment of SREBP2 S455A mutant re-expressing cells slightly reduced the total cholesterol level (14.5% reduction for T24, 13.9% for 5637) compared to that of SREBP2 S455A mutant cells (Supplementary Fig. S18A, B). In addition, cell proliferation measurements showed that re-expression of the SREBP2 S455A mutant but not wild-type SREBP2 in SREBP2 knockdown T24 and 5637 cells slightly restored cell proliferation (4.4% restoration for T24; 20.3% for 5637 at 96 h) compared to that of SREBP2 knockdown cells. Further sulfopin treatment in SREBP2 S455A mutant cells slightly reduced the cell proliferation (27.35% reduction for T24, 17.76% for 5637 at 96 h) compared to that of SREBP2 S455A mutant cells (Supplementary Fig. S18C–F). Furthermore, re-expression of SREBP2 S455A mutant in SREBP2 knockdown T24 and 5637 cells partially restored cell migration (34.9% restoration for T24; 28.8% for 5637) compared to that of SREBP2 knockdown cells. Sulfopin treatment of SREBP2 S455A mutant cells slightly reduced cell migration (20.6% reduction for T24, 23.1% for 5637) compared to that of SREBP2 S455A mutant cells (Supplementary Fig. S18G–J). In addition to PIN1 inhibition by sulfopin, deleting the PIN1 gene in control, SREBP2 knockdown and SREBP2 S455A mutant re-expressing T24 and 5637 cells showed similar alterations in total cholesterol level, cell proliferation and migration in vitro (Supplementary Fig. S19A–J). In sum, these results demonstrated that PIN1 regulates the cholesterol level, BLCA cell proliferation and migration mainly through the interaction between PIN1 and phosphorylated SREBP2, and SREBP2-mediated cholesterol biosynthesis pathway.

To strengthen the results obtained in vitro, we conducted a subcutaneous xenograft transplantation assay. We found that injections of SREBP2 knockdown, SREBP2 S455A mutant re-expressing, SREBP2 S455A mutant plus PIN1 KO and PIN1 KO 5637 cells did not affect the body weight of mice throughout tumor development (Supplementary Fig. S20A, B). As shown in Supplementary Fig. S20C–E, tumors formed by SREBP2 knockdown 5637 cells displayed lower tumor weight and volume, and, unlike SREBP2 knockdown 5637 cells re-expression of wild-type SREBP2, cells re-expressing SREBP2 S455A mutant did not regain similar tumor size to control cells. PIN1 deletion in SREBP2 S455A mutant re-expressing 5637 cells slightly reduced tumor weight (24.8% reduction) and volume (23.0% reduction). Moreover, deleting PIN1 in 5637 cells with wild-type SREBP2 significantly reduced the tumor size compared to those formed with SREBP2 S455A mutant cells (Supplementary Fig. S20C–E). Taken together, these data confirmed that SREBP2 Ser455 phosphorylation was important for PIN1’s capacity to affect tumor growth in vivo.

Sulfopin-simvastatin combination suppresses huBLCA cell proliferation in vitro and tumor growth in vivo synergistically

Considering the individual inhibitory effects of sulfopin, a PIN1 inhibitor, and simvastatin, an HMGCR inhibitor on cell proliferation, we tested whether a sulfopin-simvastatin combination could kill BLCA cells synergistically. For this purpose, we generated synergy plots of the effects of combined sulfopin and simvastatin treatment of T24 and 5637 cells analyzed by the R package Synergyfinder and ascertained a synergistic effect of the sulfopin-simvastatin combination on cell proliferation (synergy score>5) (Fig. 7A and B). A significant decrease in cell proliferation and increase in necroptosis (p-MLKL⁺) were observed in T24 (Fig. 7C and D) and 5637 (Fig. 7E and F) cells treated with sulfopin plus simvastatin combination, compared to that with single inhibitor treatment. Moreover, 5637 cells treated with the sulfopin plus simvastatin combination showed lower total cholesterol levels (Supplementary Fig. S21A). To evaluate the effects of the sulfopin-simvastatin combination on tumor growth in vivo, we used a xenograft transplantation model with 5637 cells. Either single inhibitor therapy or combination therapy led to significantly smaller tumor masses, and sulfopin plus simvastatin combination therapy was more efficient in suppressing tumor growth throughout treatment cycles compared to mice treated with sulfopin or simvastatin alone (Fig.7G and H). Of note, the sulfopin plus simvastatin combination regimen was better in restraining tumor development than that of the romidepsin HDAC inhibitor plus simvastatin therapy²⁵ (Fig.7G and H). What is more, tumor masses treated with sulfopin plus simvastatin combination regimens showed significantly decreased cell proliferation and increased apoptosis in comparison with those administered the single inhibitor therapy (Fig.7I–L). As shown in Supplementary Fig. S21B–G, tumors treated with sulfopin plus simvastatin combination regimens exhibited remarkably decreased expression levels of cholesterol pathway genes, SREBP2, HMGCS1 and MVD, compared to those with the single inhibitor treatment.

Figure 7. Sulfopin-simvastatin combination suppresses huBLCA cell proliferation in vitro and tumor growth in vivo synergistically.

A, B. Synergy plots of sulfopin and simvastatin treatment of T24 (A) and 5637 (B) huBLCA cells showing a synergistic suppressive effect of the sulfopin and simvastatin combination on cell proliferation. C-F. Cell proliferation assay showing significantly lower cell proliferation in T24 (C) and 5637 (E) cells treated with 5 μM sulfopin plus 12.5 μM simvastatin (the red line) compared to that with single inhibitor treatment. (The Y axis in C and E represented the absorbance at 450 nm, normalized to the absorbance at 0 h). A necroptosis marker phosphorylated mixed lineage kinase domain-like protein (p-MLKL) immunostaining exhibiting significantly higher necroptosis in T24 (D) and 5637 (F) cells treated with sulfopin plus simvastatin in comparison to that with single inhibitor therapy. (n= 3 wells. Hoechst⁺ T24 and 5637 cells from 5 fields per well were counted. The Y axis in D and F represented the percentage of p-MLKL⁺ cells in Hoechst⁺ T24 and 5637 cells.) G, H. Sulfopin plus simvastatin combination therapy was more effective in suppressing 5637 cells-formed tumor growth (tumor weight (G) and volume (H)) during treatment cycles, compared to the single inhibitor therapy. (n= 5 mice each group). I-L. Cell proliferation marker Ki67 immunostaining showing a significant decrease in cell proliferation in tumor masses treated with sulfopin plus simvastatin compared to that with no inhibitor and single inhibitor treatment (I and K). Immunostaining for cleaved caspase 3 (CC3) apoptosis marker showing a significant increase in apoptosis in tumor masses treated with sulfopin plus simvastatin compared to that with no inhibitor and single inhibitor treatment (J and L). (Scale bars, 50 μm. n= 3 tumor masses. Pan-keratin⁺ urothelial cells from 5 frozen sections per tumor mass, 5 fields per frozen section were counted. The Y axis in K and L represented the percentage of Ki67⁺ and CC3⁺ cells in Pan-keratin⁺ urothelial cells. Sul, sulfopin; Sim, simvastatin; Sul + Sim, sulfopin plus simvastatin; Romi + Sim, romidepsin plus simvastatin. All p values were calculated using a two-tailed unpaired Student t-test. Data are presented as mean + s.e.m. ns, not significant (p > 0.05)). M. A working model depicting how PIN1 affects bladder cancer initiation and progression through disrupting cholesterol metabolism homeostasis. PIN1 binds to JNK-dependent phospho-Ser455 SREBP2, which is located close to the second cleavage site. Ser455 phosphorylation and PIN1-mediated isomerization of the Ser455.Pro456 bond may facilitate the cleavage and translocation of the SREBP2 N-terminal domain from the Golgi membrane to the nucleus, where it can activate gene expressions involved in cholesterol biosynthesis pathways. Moreover, PIN1 inhibitor sulfopin plus HMGCR inhibitor simvastatin combination therapy effectively suppresses cell proliferation and tumor growth. (Created in BioRender. Lee, D. (2025) https://BioRender.com/v48v783).

As shown in our working model (Fig. 7M), PIN1 bound to JNK-dependent phosphorylated-SREBP2 upon TNFα stimulation, which facilitates SREBP2 processing and activation on the Golgi and active SREBP2-driven cholesterol metabolism pathway gene expressions in the nucleus. Based on our results, we propose that elevated cholesterol levels in urothelial cells contribute to BLCA initiation and progression and a combination therapy targeting PIN1 with sulfopin and cholesterol biosynthesis with simvastatin can effectively suppress BLCA growth.

Discussion

The PIN1 phospho-dependent prolyl isomerase has been implicated in promoting cancer phenotypes through its ability to target and activate key signaling nodes that are phosphorylated by Pro-directed protein kinases, such as the MAP kinases. Here, we provide direct evidence that PIN1 promotes BLCA cell proliferation and progression in culture and drives tumor growth and lung metastasis in mouse models. Importantly, re-expression of WT PIN1, but not PIN1 mutants defective for WW domain or catalytic domain activity in PIN1-KO BLCA cells reversed these phenotypes. These results imply that PIN1’s prolyl isomerase and pSer/pThr.Pro binding activities are important for the tumorigenic properties of human and mouse BLCA cell lines, and that PIN1 exerts its tumor-promoting function in these cells by isomerizing one or more pSer/pThr.Pro sites in target proteins.

A major finding of our studies is that PIN1 isomerase function is important for cholesterol biosynthesis, a process required for maximal BLCA cell proliferation. PIN1-KO cells exhibited reduced free/total cholesterol levels, which correlated with reduced BLCA cell proliferation in delipidated medium and could be restored by exogenous cholesterol. Previous bulk RNA-seq and public datasets analyses have demonstrated that BLCA with higher rates of cholesterol metabolism pathways were more aggressive and positively correlated with higher risk of recurrence than those with a quiescent subtype, which supports our experimental findings that decreased expression of cholesterol biosynthesis pathway genes caused by PIN1 deletion in human and mouse BLCA cells were associated with lower cell proliferation and less tumor burden^{48, 49}. Our finding of the importance of cholesterol biosynthesis in BLCA parallels reports in other cancer types, where cancer cells have been shown to require excess cholesterol and intermediates of the cholesterol biosynthesis pathway to maintain cell proliferation⁵⁰. Both increased endogenous cholesterol synthesis and high cholesterol exposure contribute to cancer progression⁵¹. In addition, glioblastoma stem cells (GSCs) have been demonstrated to depend on de novo cholesterol biosynthesis⁵². In the case of BLCA, external cholesterol levels in the tumor microenvironment may be suboptimal meaning that endogenous cholesterol synthesis in tumor cells would be required for tumor growth.

In terms of underlying molecular mechanisms through which PIN1-mediated cis-trans isomerization promotes cholesterol synthesis, we focused on the SREBP2 transcription factor because it is a key step in the pathway driving transcription and expression of pathway genes including that of the rate-limiting enzyme, HMGCR. Moreover, active and nuclear-form of SREBP2 expression levels were reduced in PIN1-KO BLCA. We used TNFα to stimulate JNK in BLCA cells and showed that TNFα treatment induced close association of SREBP2 with WT PIN1 but not with the WW domain PIN1 mutant, implying that the PIN1/SREBP2 interaction is mediated by the WW domain. Conversely, the N-terminal region of SREBP2 (residues 1 to 481) was required for JNK-dependent association with PIN1. This led us to focus on possible JNK sites in this region of SREBP2, and through mutagenesis of Ser103 and Ser455 predicted JNK sites we showed that the Ser455Ala mutation largely abolished PIN1/SREBP2 association. On this basis, we propose that PIN1 binds to pSer455.Pro in SREBP2 following activation of the JNK MAP kinase. Human SREBP2 activation depends on two-step proteolytic cleavage by Site-1 and Site-2 proteases of full length SREBP2 protein, releasing the N-terminal domain from the membrane. A tetrapeptide sequence, R(Arg519)XXL(Leu522) in the middle of the 31-residue hydrophilic loop is the recognition signal for the Site-1 protease that mediates the sterol-regulated first cleavage^{47, 53}, and the Leu484-Cys485 bond at the junction between the cytoplasmic N-terminal fragment of SREBP2 and the first transmembrane segment is the second cleavage site for the Site-2 protease⁵⁴. Because the Ser455 site is located close to the second cleavage site, Ser455 phosphorylation and PIN1-mediated isomerization of the Ser455.Pro456 bond may facilitate the cleavage and activation of SREBP2 by altering local conformation.

Our elucidation of a key role for PIN1 in BLCA has potential therapeutic implications. Current BLCA therapies include surgery such as transurethral resection, radiation therapy, platinum-based chemotherapy, and immunotherapy using drugs like BCG (Bacillus Calmette-Guerin). Despite recent progress in targeted therapy and immunotherapy, BLCA finally develops resistance to these therapies and relapses. Our work suggests that a combination of a PIN1 inhibitor with a statin to inhibit cholesterol biosynthesis could be useful in BLCA treatment. Simvastatin is a clinically approved HMGCR inhibitor and has been reported to antagonize tumor growth synergistically with other anticancer agents. Here, we tested a newly developed highly selective and PIN1 covalent peptide inhibitor, sulfopin³⁴ in combination with simvastatin for BLCA treatment and observed a significant suppression of tumor growth compared to that treated with single inhibitors, sulfopin or simvastatin. To validate the anticancer effect of the sulfopin plus simvastatin combination therapy, the administration schedule and dosages will need to be optimized to obtain better tumor-killing.

In summary, our findings provide evidence for an essential role of PIN1 in initiation and progression of BLCA. One functional target for PIN1 in BLCA cells is JNK-phosphorylated SREBP2, a key transcription factor in the cholesterol biosynthesis pathway, which is activated in a PIN1-dependent manner to elevate endogenous cholesterol levels. However, there are certainly other functional PIN1 phosphoprotein targets in BLCA cells that promote cholesterol biosynthesis, proliferation, migration, and tumorigenesis that remain to be identified.

Materials and Methods

Cell lines

The SV40 immortalized human bladder epithelial cell line SV-HUC-1 (RRID:CVCL_3798), human bladder cancer cell lines RT-4 (RRID:CVCL_0036), SW780 (RRID:CVCL_1728), T24 (RRID:CVCL_0554), 5637 (RRID:CVCL_0126) and TCCSUP (RRID:CVCL_1738) were purchased from American Type Culture Collection (ATCC). All functional experiments using SV-HUC-1 (RRID:CVCL_3798), RT4 (RRID:CVCL_0036), SW780 (RRID:CVCL_1728) and TCCSUP (RRID:CVCL_1738) cell lines were completed within 6 months after resuscitation. They are Mycoplasma negative. SV-HUC-1 (RRID:CVCL_3798) is an epithelial cell expressing uroepithelial keratins isolated from the uroepithelium of an 11-year-old, male. RT4 (RRID:CVCL_0036) is a cell line exhibiting epithelial morphology that was isolated from urinary bladder tissue derived from a 63-year-old, white, male patient with transitional cell papilloma. SW780 (SW-780, SW 780, RRID:CVCL_1728) is a cell line exhibiting epithelial morphology that was established from an 80-year-old, white, female patient with grade I transitional cell carcinoma. TSSCUP (RRID:CVCL_1738) is a cell line exhibiting epithelial morphology that was established from a 67-year-old, female patient with grade IV transitional cell carcinoma. SV-HUC-1 (RRID:CVCL_3798) cells were maintained in F-12K complete medium. RT4 (RRID:CVCL_0036) cells were cultured in McCoy’s 5A complete medium. SW780 (RRID:CVCL_1728) cells were cultured in L-15 complete medium. TCCSUP (RRID:CVCL_1738) cells were cultured in EMEM complete medium. SV-HUC-1, RT4, SW780 and TCCSUP cell lines cultured in 10 cm dish were passaged every 3–4 days. They were used within 6 passages. T24 (RRID:CVCL_0554) is a cell line exhibiting epithelial morphology that was isolated from the urinary bladder of an 81-year-old, white female patient. 5637 (RRID:CVCL_0126) is a cell line exhibiting epithelial morphology that was isolated from the urinary bladder of a 68-year-old, white male patient with grade II carcinoma. T24 cells were cultured in McCoy’s 5A complete medium. 5637 cells were maintained in RPMI-1640 complete medium. T24 and 5637 cell lines cultured in 10 cm dish were passaged every 3 days. They were used within 6 passages. MB49⁵⁵ (RRID:CVCL_7076) mouse bladder carcinoma cell line (derived from a male C57BL/6 mouse by exposing primary bladder cells to a chemical carcinogen, 7, 12-dimethylbenz[a]anthracene (DMBA) for 24 hours and subsequent culturing) was purchased from Sigma-Aldrich (SCC148) and maintained in DMEM complete medium. HEK293T (RRID:CVCL_0063) cells were cultured in DMEM complete medium. MB49 and 293T cell lines cultured in 10 cm dish were passaged every 3 days. They were used within 6 passages. T24, 5637, MB49 and 293T cell lines were thawed after 3–4 months freezing. T24 and 5637 parental cell lines were profiled by Radil-IDEX to verify their origin, lack of contamination (mycoplasma, other cells, bacteria, fungus, virus) before performing in vivo experiments. T24, 5637, 293T and MB49 cell authentication was re-confirmed by IDEXX BioAnalytics on 9-23-2024. They are consistent with the cell line of origin. The Mycoplasma testing for T24, 5637, 293T and MB49 cells was done following instructions from the manufacturer (Mycoplasma detection kit, rep-mys-100, InvivoGen) on 9-23-2024. They are Mycoplasma negative. The detailed information about cell authentication and Mycoplasma testing are listed in Supplementary report. All complete media were supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin. All cells were propagated in an incubator with 5% CO₂ at 37 °C.

Mice

Nude mice (Foxn1^nu, RRID:IMSR_JAX:002019) and C57BL/6J (RRID:IMSR_JAX:000664) were obtained from the Jackson Laboratory. All animal experiments were conducted in accordance with IACUC protocols approved by Salk Institute and in compliance with all relevant ethical regulations.

Plasmids and lentivirus production

The single guide RNA sequences are GAAGATCACCCGGACCAAGG upstream and CCTTGGTCCGGGTGATCTTCC downstream of the 2^nd exon of the human PIN1 gene. To make lentivirus, a transfer plasmid (lentiCRISPRv2 hygro, RRID:Addgene_98291 with sgRNA sequences targeting huPIN1) was co-transfected into HEK293T cells with 2^nd packaging plasmids pVSVG (RRID:Addgene_31947) and psPAX2 (RRID:Addgene_12260). 2 h prior to transfection, the culture media were removed and replaced with fresh pre-warmed growth medium. The mixture of PEI reagents (Kyfora Bio, #23966–1): DNA (3:1 v/w ratio) were mixed thoroughly, incubated at room temperature for 15 min and added to the dish. Cells with transfection mix were put back to the incubator with 5% CO₂ at 37 °C overnight. 72 h post-transfection, the media were collected and filtered through a 0.45 μm strainer. The virus suspension was mixed with 1/3 volume Lenti-X concentrator (TaKaRa, 631231) and incubated overnight. The next day, the mixture was centrifuged at 1500 g, 4 °C for 45 min. The virus pellet was re-suspended with growth medium, used directly on cultured cells or aliquoted, and stored in a −80 °C freezer. A Pin1 overexpression plasmid was obtained by cloning the Pin1 full-length cDNA sequence into the pLVX-Puro backbone from RRID:Addgene_180635. The pLVX-EF1α-Turbo-N-IRES-Puro backbone was kindly provided Steve Fuhs and Ray Whitson from Genomics Institute of the Novartis Research Foundation, La Jolla, CA. The huPIN1 full-length cDNA sequence was fused to C-terminal of Turbo with a linker between two sequences. The PIN1 N-terminal WW domain (amino acids 1 to 54) and C-terminal PPIase domain (amino acids 47 to 163) were cloned into the pLVX-EF1α-Turbo-N-IRES-Puro backbone, replacing the huPIN1 cDNA sequence. Point and double mutants, W34A and K63A were constructed utilizing the QuikChange II Site-Directed Mutagenesis Kit (Agilent Technologies, 200523–5).

The human SREBF2 vector encoding full length SREBP2 was obtained from Vector Builder (vector ID, VB220714–1469). Vectors to express N-terminal domain (NTD, 1–481) and C-terminal domain (CTD, 587–1141) fragments of full-length huSREBP2 were constructed based on the huSREBP2 backbone. The huSREBP2 (Ser106A, Ser455A) point mutants were constructed using the QuikChange II Site-Directed Mutagenesis Kit. The primers used here are listed in Supplementary Table 1.

The 5 human SREBP2 shRNA sequences were purchased from Millipore-Sigma (TRC clone ID: TRCN0000020664, TRCN0000020665, TRCN0000020666, TRCN0000020667, and TRCN0000020668). pLKO.1-puro non-target shRNA control plasmid DNA was purchased from Millipore-Sigma (SHC016). The #1 (TRCN0000020664) and #5 (TRCN0000020668) were validated to know down SREBP2 efficiently in 5637 huBLCA cells via qRT-PCR and immunoblotting assays. The #1 (TRCN0000020664) and #4 (TRCN0000020667) were validated to knock down SREBP2 efficiently in T24 huBLCA cells via qRT-PCR and immunoblotting assays.

RNA extraction and quantitative-PCR analysis

Total RNA was extracted from control, PIN1 KO and PIN1 re-expressing cells using the NucleoSpin RNA purification kit (Macherey-Nagel, 740955.50) following instructions from the manufacturer. cDNA was synthesized from the extracted total RNA using the SuperScript III First-Strand Synthesis kit (Invitrogen, 18080400). qPCR was performed utilizing the SYBR green PCR master mix (4309155, Thermo Fisher). Relative transcript abundance was determined by the comparative CT method using GAPDH as a reference gene. Primers used in RT-PCR are listed in Supplementary Table 1.

Cell proliferation

To explore the effect of PIN1 on BLCA cell proliferation, a colorimetric assay (following instructions given in the Cell Counting Kit-8 (CCK-8) kit, CK04–11, Dojindo) was conducted. Briefly, CCK-8 (WST-8), water-soluble tetrazolium salt, can be reduced by dehydrogenase in cells to give a yellow-color formazan dye. The amount of the formazan dye, generated by the activities of dehydrogenases in cells, is directly proportional to the number of living cells. 5,000–6,000 BLCA cells were seeded into a well of a 96-well plate with 100 μL growth medium with different concentrations of the PIN1 inhibitor, sulfopin (Focus Biomolecules, 10–4450), HMGCR inhibitor, simvastatin (Santa Cruz, sc-200829), or with added lipoprotein-depleted FBS (KALEN Biomedical, 880100–1) or cholesterol lipid concentrate (250X, ThermoFisher, 12531018) per well. Cells were incubated with 10 μL CCK8 reagent for 3 h at 37 °C, and then the absorbance increase at 450 nm was measured using a microplate reader (Tecan Infinity M1000).

Cycloheximide (CHX) chase assay

50 mg CHX powder was dissolved with 1 mL DMSO to make 50 mg/mL in stock, freshly made to ensure maximal activity. Cell media were removed and replaced with complete media with 50 μg/mL CHX working concentration. Cells were treated with either DMSO (for 0 h timepoint) or CHX (50 μg/mL, for 4 h, 8 h, 15 h, 24 h, 48 h and 72 h timepoints). Then, cells were harvested and cell proteins extracted for immunoblotting.

Immunoprecipitation and Immunoblotting

T24 and 5637 cells were starved overnight and treated with 15 ng/mL recombinant human TNFα (PeproTech, 300–01A) for 15 min. Here, we utilized an enzyme-catalyzed proximity labeling approach and enriched the Turbo ID-biotinylated proximal endogenous proteins using streptavidin magnetic beads. Compared to negative controls (cells infected with both empty vector and vector with Turbo only), the highly streptavidin-enriched (at least 2-fold higher) proteins in eluates from cells infected with Turbo-PIN1 fusion vector represent the interactome for PIN1 in BLCA cells. For immunoprecipitation, T24 and 5637 cells were washed with ice-cold PBS, harvested using a cell scraper and lysed in a lysis buffer (50 mM Tris, 150 mM NaCl, 0.1% (wt/vol) SDS, 0.5% (wt/vol) sodium deoxycholate, 1% (vol/vol) Triton X-100, PH7.5) supplemented with protease (#11836170001, Sigma) and phosphatase inhibitors (#4906845001, Sigma). Cell lysate was incubated on ice for at least 20 min and clarified by centrifugation at 12,000 rpm at 4 °C for 15 min. The clarified lysate was transferred to a fresh microcentrifuge tube. Protein concentrations were assessed by BCA protein assay. 10% clarified lysate was transferred to a new tube as input. Then 500 μg protein with an additional 500 μL lysis buffer was incubated with pre-activated 35 μL streptavidin (New England Biolabs, S1420S) or anti-flag (Millipore sigma, M8823) magnetic beads at 4 °C for overnight. The next day, beads were collected using a magnetic rack. The supernatant was the corresponding unbound (flow-through) sample and could be used for immunoblotting analyses. The beads were washed 3 times with lysis buffer and the final wash was collected for immunoblotting analyses. The beads were boiled in 35 μL of 3X protein loading buffer at 98 °C for 10 min. The input, flow-through and final wash were combined with 6X protein loading buffer and boiled at 98 °C for 10 min. For immunoblotting, the whole-cell lysates were combined with 6X protein loading buffer and boiled for 10 min at 98 °C. 10–30 μg protein were loaded, separated on a 10–12% SDS-PAGE electrophoresis gel, and then transferred to PVDF membranes. Membranes were blocked with 5% skimmed milk in 1X TBST (TBS buffer containing 0.1% Tween-20) and incubated with primary antibodies overnight at 4 °C. The next day, the membranes were rinsed with 1X TBST buffer 3 times and incubated with secondary antibodies for 1 h at room temperature. After washing 3 times with 1X TBST, the membranes were imaged using a LICOR Odyssey scanner. Antibodies used in the study are listed in Supplementary Table 2.

Nuclear proteins extraction and SREBP-2 transcription factor binding assay

Preparation of nuclear extracts is the first step in examining transcription factor activity. Nuclear proteins were extracted from control, PIN1 KO and PIN1 re-expressing cells using a Cayman’s Nuclear Extraction Kit (Item No. 10009277) following instructions from the manufacturer. Briefly, cells (10 cm dish) were collected and added with 500 μL ice-cold 1X Complete Hypotonic Buffer to allow cells to swell on ice for 15 min. 100 μL of 10% NP-40 Assay Reagent was added to the tube, mixed and centrifuged at 14,000 × g for 30 seconds at 4 °C. The supernatant containing the cytosolic fraction was transferred to a new tube and stored at −80 °C. The pellet was resuspended in 100 μL ice-cold Complete Nuclear Extraction Buffer, incubated on ice for 30 minutes (vortex vigorously every 10 min) and centrifuged at 14,000 × g for 10 min at 4 °C. The supernatant containing the nuclear fraction was aliquoted and stored at −80 °C to avoid freeze/thaw cycles. The protein concentration was quantified using the BCA protein assay. Specific transcription factor DNA-binding activity in nuclear extracts was tested using Cayman’s SREBP-2 Transcription Factor Assay Kit (Item No. 10007819) following instructions from the manufacturer. Briefly, wells of a 96-well plate were added with Complete Transcription Factor Binding Assay Buffer (CTFB), competitor dsDNA, positive control, or nuclear extracts and incubated overnight at 4 °C. Each well was washed five times with 200 μL of 1X Wash Buffer followed by addition of 100 μL of diluted SREBP2 primary antibody (except blank wells) for 1 h at room temperature and the plate agitated on an orbital shaker. Then, each well was washed five times with 200 μL of 1X Wash Buffer followed by addition of 100 μL of diluted secondary antibody (except blank wells) for another 1 h at room temperature on an orbital shaker. Subsequently, each well was washed five times with 200 μL of 1X Wash Buffer followed by addition of 100 μL of Developing Solution for 45 min at room temperature on an orbital shaker. Finally, 100 μL of Stop Solution was added to each well and the absorbance at 450 nm was measured using a plate reader.

Immunofluorescence staining

Mouse bladders were fixed in 4% paraformaldehyde for 1 h and dehydrated in 30% sucrose solution overnight. Then tissues were embedded in Optimal Cutting Temperature (O.C.T) compound and put in a −80 °C freezer overnight. The next day, frozen tissues were cut to a thickness of 6 μm. Sections were washed with PBS and subjected to the permeabilization step in PBS containing 0.5% Triton X-100 for 15 min. After washing in PBS, sections were transferred to a blocking solution (PBS with 0.3% TritonX-100 and 5% donkey serum) for 1 h at room temperature. Primary antibodies diluted in blocking solution were added to sections that were kept at 4 °C overnight. The next day, primary antibodies were washed away with PBS three times. Sections were incubated with secondary antibodies, conjugated to Alexa Fluo-488, 568 or 647 for 1 h at room temperature. Sections were thoroughly washed with PBS and mounted with Vector Shield mounting medium after incubation for 5 min in 1 μg/mL Hoechst 33342 (Thermo Fisher, 62249). The source data for immunostaining experiments are provided in Supplementary Table 3 with different sheets. Antibodies used in the study are listed in Supplementary Table 2.

Immunohistochemical staining

Paraffin-embedded human bladder cancer tissue (BL244a and BL807 from TissueArray.com) sections were deparaffinized, rehydrated and subjected to a heat-induced epitope retrieval step in 0.01M sodium citrate (pH 6.0, 10X, Sigma-Aldrich, C9999). Treatment with 3% hydrogen peroxide for 10 min was used to ablate endogenous peroxidase activity. Then sections were blocked with the blocking solution (PBS with 0.3% TritonX-100 and 5% donkey serum) and incubated with diluted primary antibodies overnight. After washing, horseradish peroxidase conjugated secondary antibodies were added to sections for 1 h at room temperature. Then DAB staining was performed according to the manufacturer’s instructions (ImmPACT DAB substrate Kit, SK-4105, from Vector Labs). Sections were washed with water for 4–6 min and then counterstained with hematoxylin for about 2 min, followed by dehydration and mounting with the neutral balsam mounting medium. The QuPath 0.4.4 software was used to quantify the DAB staining of human bladder cancer microarray. The stains were automatically deconvoluted by using the estimate vector stains function. The positive cell function was used to count the number of positively stained cells. Using nuclear hematoxylin staining, cells were automatically detected by the software. Three intensity thresholds were set for positive DAB-stained cells, threshold 1+ was set at 0.2 OD, threshold 2+ at 0.3 OD, and threshold 3+ at 0.4 OD. The cell DAB OD max parameter was used for the score compartment. The rest of the settings were kept as default, and the same settings were used for all images. To measure both the intensity and percentage of cells stained, the H-score was used and calculated automatically by the QuPath software.

Spheroid-formation assay

Several studies have utilized in vitro spheroid culture to propagate bladder cancer stem cells that were presumed to initiate tumor growth and contribute to metastasis^56–58. Cells were collected and counted. A single cell suspension (2× 10³ cells suspended in DMEM/F12 culture medium supplemented with 2% B27, 20 ng/mL recombinant human EGF and 20 ng/mL FGF, mixed 50% (vol/vol) Matrigel) was added to one well of a 96-well ultra-low attachment plate (CLS 3474, Corning). The plate was put in a 37 °C incubator and monitored every 3 days under a microscope. The medium was refreshed every 3 days. After approximately 2 weeks culture, spheres were counted. 2 mg/mL Dispase II solution was added to each well to digest the Matrigel. One hour later, spheres were collected by centrifugation at 400 g for 5 min and then washed several times with ice-cold PBS to remove the Matrigel. 0.25% Trypsin/EDTA was applied to dissociate spheres into single cells, then inactivated with complete medium and centrifuged. Finally, the single cell pellet was resuspended in culture medium. Spheres can be passaged to obtain secondary and tertiary passages. To prepare sections, spheres were collected, fixed with 4% paraformaldehyde (fresh) for 1 h, washed with pre-cold PBS twice, centrifuged and resuspended with 50 μL collagen type I and put back into a 37 °C incubator for at least 1 h to solidify the sphere/collagen mixture. The mixture was transferred to 4% paraformaldehyde (fresh) for fixation, 30% sucrose for dehydration and O.C.T for embedding and finally sectioning. 5 μm frozen sections were utilized for immunostaining assays.

Wound healing assay

The ibidi Culture-Insert 2 Well in a μ-Dish 35 mm (80209, Ibidi) was utilized to determine cell motility. The ibidi Culture-Insert 2 well provided two well culture reservoirs, each separated by a 500 μm wall. Cells were seeded in both reservoirs. The insert was removed after 48 h attachment of cells. The cell layer was washed twice with cell-free media to remove cell debris and non-attached cells. Cell patches were overlayed with basic culture medium. The dish was put back into the incubator and photos were taken at different time points to record cell migration. ImageJ software (RRID:SCR_003070) was used to calculate the area of wound at 0 h and the last time point. Wound closure percentage was calculated using the formula (area (0 h)-area (last time point))/area (0 h).

Transwell assay (with or without Matrigel)

For cell migration assay, cells (1–2× 10⁵/mL) were washed with PBS, resuspended in 200 μL serum-free culture medium, and put into the top chambers of the transwell plates (transwell polycarbonate membrane cell culture inserts with 8 μm pore, CLS3422–48EA, Corning). Then 750 μL medium supplemented with 20% FBS as attractant was added to the well of the plate (lower compartment). The plate was incubated at 37 °C for 24 h. Cells that migrated to the lower surface of the membrane were fixed with 4% paraformaldehyde (fresh) for 10–20 min at room temperature, followed by permeabilization with 100% methanol for 10 min at room temperature and staining with 0.1% crystal violet (61135–25G, Sigma-Aldrich, dissolved in methanol) for 20–30 min at room temperature. The insert then was washed 3 times in PBS to remove the excess dye. Cells in the upper chamber were removed thoroughly with a cotton swab and the insert was dried completely before counting the cells on the lower surface of the chamber under a microscope. Different views were chosen randomly. For the cell invasion assay, 100 μL diluted Matrigel (1:5 diluted in ice-cold serum-free medium) was added to the upper chamber of the insert overnight at 37 °C. The next day, an additional 100 μL serum-free medium was applied to the upper chamber. 750 μL medium supplemented with 20% FBS as attractant was added to the well of the plate (lower chamber). The plate was put back in the 37 °C incubator for 24 h. Next, the medium in the upper chamber was removed, and the insert was rinsed 3 times with PBS. Cells that invaded to the lower surface of the insert were fixed, permeabilized and stained with crystal violet and observed using the steps in cell migration protocol mentioned above.

Urothelium clearance assay

Based on previous studies, tumor spheroid attachment and spreading on a urothelial monolayer promotes clearance of urothelial cells from the area underneath the spheroid^{59, 60}. To obtain 3D cell clusters, 6-well plates were pre-coated using 10 mg/mL poly-2-hydroxyethyl methacrylate (poly-HEMA) overnight and rinsed 3 times with PBS the next day before use. Clusters from each cell line were prepared by seeding 4–5× 10⁵ cells in 2 mL of media in poly-HEMA pre-coated 6-well plates. GFP-expressing SV-HUC-1 cells were seeded on 24-well plates and maintained until confluent. In the co-culture experiment, a single sphere (via limiting dilution) was placed onto a confluent urothelial monolayer (using the microscope to confirm only one sphere each well). Images under the bright-field and fluorescent microscope were captured at different time-points. To quantify the clearance of urothelial cells by a sphere, the area devoid of the urothelial monolayer was measured and divided by the initial area of the sphere using ImageJ software (RRID:SCR_003070).

Subcutaneous injections with human/mouse bladder cancer cells

To assess tumor growth in vivo, T24 (1× 10⁷) and 5637 (5× 10⁶) cells were subcutaneously inoculated in the flanks of nude mice and monitored for 7 weeks. Six mice were randomly assigned to each group. For the allograft implantation model, C57/B6 mice were injected in the flank with MB49 (3× 10⁶) mouse urothelial carcinoma cells and monitored for 4 weeks. Tumor height and width were measured with a caliper every 2 days. The maximum single tumor cannot exceed 1.5 cm in diameter for subcutaneous tumor growth. The tumor volume was calculated using the formula width² × height × 0.52. Here, to optimize T24 cells-formed xenografts growth, we tried to propagate more viable T24 cells in flanks of mice. Briefly, primary xenografts derived from 2D cultured T24 cells were harvested and enzymatically digested using Collagenase IV/Dispase II/Dnase I solution to produce single cell suspensions, which were re-injected into flanks of mice to obtain secondary xenografts. This cycle was repeated to generate tertiary xenografts for further analyses. To examine the effect of the PIN1 inhibitor, sulfopin, on tumor growth in vivo, tumor masses (reaching 0.5 cm in diameter) formed from control and PIN1 KO cells were treated with vehicle, sulfopin (30 mg/kg, oral gavage, every other day) for 14 days. Formulation was 5% NMP, 5% solutol, 20% DMSO. Body weights were measured, and bladders collected for toxicity testing. PIN1 immunostaining on bladder tissues was performed for dosing. After 14 days’ treatment, tumor masses were collected, and their weight, height, and width measured. The in vivo efficacy of the sulfopin-simvastatin combination was assessed using nude mice bearing 5637 huBLCA cell subcutaneous xenografts. 5× 10⁶ 5637 cells were implanted subcutaneously into nude mice and treatment was initiated when the mice had 0.5 cm diameter tumors. The mice were randomly divided into the vehicle and the treatment groups (n= 5 each group). The vehicle group received intraperitoneal injections of sulfopin dilution solution, 5% NMP, 5% solutol, 20% DMSO, and the treatment groups received 30 mg/kg sulfopin or 15 mg/kg simvastatin or both once a day for 2 weeks to evaluate the effects on tumor growth in vivo. The injection of simvastatin + romidepsin (0.5 mg/kg, Cayman Chemical, #17130) was used as a positive control.

Orthotopic transplantation of human/mouse bladder cancer cells

Comparatively, the male mouse urethra is longer and more tortuous, making transurethral catheterization more difficult. Thus, only female nude mice were used for the orthotopic injection here. 7–8 weeks old female nude mice were anesthetized using inhalable isoflurane. The mice were prepared for surgery by cleaning the surgical area and applying ophthalmic ointment to their eyes. Then the mouse was catheterized using a 24-gauge catheter through the urethra. The bladder was emptied and washed twice by injecting 100 μL sterile water. Then, the bladder was emptied and instilled with 100 μL 0.1% Poly-L-Lysine for 20 min to facilitate the adhesion of cells. The bladder was emptied and washed 3 times with sterile water. 25 μL 5× 10⁶ 5637 (control, PIN1 KO and PIN1 re-expressing) luciferase expressing bladder cancer cell suspension mixed with 25 μL Matrigel basement membrane matrix (354234, Corning) was injected into the bladder through the catheter. The urethra was clamped using a suture and the mouse remained under maintenance anesthesia for 2 h. The suture was removed, and the mouse recovered from the anesthesia after 2 h. Following a similar orthotopic implantation procedure, the 7–8 weeks old female C57BL/6J mice were used for orthotopic injection with 3× 10⁵ MB49 (control, Pin1 KO and Pin1 re-expressing) luciferase expressing mouse bladder carcinoma cells.

RNA-seq and data analysis

Control, PIN1 KO and PIN1 re-expressing T24, and 5637 cells were collected and lysed using the RNA lysis buffer provided in the kit. Total RNA was extracted according to the manufacturer’s instructions from the NucleoSpin RNA purification kit (Macherey-Nagel, 740955.50). Quality control and library preparation were conducted in the NGS Core at the Salk Institute. The libraries were sequenced on an Illumina Hiseq4000 platform and single reads were obtained. Raw data were obtained from the Salk NGS core. The alignment data were obtained in the San Diego Supercomputer center and reads were quantified. To establish generality, expression data from T24 and 5637 cell lines were combined to generate the TPM abundance data. GSEA enrichment analyses were performed utilizing TPM abundance data. The expression-based heatmap was obtained using the Heatmapper online tool.

Single cell RNA-seq data processing and integration analysis

The raw UMI count matrices of each sample from GSE129845^26> and GSE135337²⁷ were processed using the Seurat (v4.3.0.1)⁶¹ script in R. Files from GSE129845 were original cellRanger output data, unfiltered. Each data was separately loaded into a Seurat object and analyzed using the Seurat pipeline (additional parameters: nfeatures=3000, dims=50, resolution=0.6). Doublet cells were identified using DoubletFinder (v2.0.3)⁶² (RRID:SCR_018771) scripts in R. Cells with nFeature_RNA ≥ 4,000, nCount_RNA ≤ 500, percent.mt ≥ 15, or doublet-labeled were rejected from further analysis. After quality control, all mice and human samples were merged, separately. As the files from GSE135337 were already quality-controlled, no additional filtering was applied. Each data was separately loaded into a Seurat object and the default Seurat pipeline was executed after merged all data (additional parameters: nfeatures=3000, dims=50, resolution=0.6). Given the existence of cross-sample variation, we used the HMY method (RunHarmony), integrated in the harmony (v1.0.1)⁶³ package, to remove differences between batches/samples for each merged data set. The top 25 dimensions of Principal Components (PCs) were then used for t-Stochastic Neighbor Embedding (tSNE) and graph-based cell clustering. The identified clusters were validated using markers from previous studies. To recover the true expression of cells from dropout noise, an imputation was performed by using build-in function from SeuratWrappers package (v0.3.1) named RunALRA⁶⁴ with default parameters. Violin plots, scatter plots and box plots were utilized to visualize Pin1 expression level in different cell clusters.

Filipin III fluorescence staining of free cholesterol in cultured cells

Cells were rinsed 3 times with PBS and fixed with 3% paraformaldehyde (fresh) for 1 h at room temperature. Cells were washed 3 times with PBS and incubated with 1 mL of 1.5 mg glycine/mL PBS for 10 min at room temperature to quench the paraformaldehyde. Cells were stained with 1 mL filipin III (Sigma F-9765, 25 mg/mL in DMSO) working solution (0.05 mg/mL in PBS/10% FBS) for 2 h at room temperature. After PBS washing, cells in PBS were viewed by fluorescence microscopy using a UV filter set.

Total cholesterol level testing

The Cholesterol/Cholesterol Ester-Glo^™ Assay (Promega, J3190) measures cholesterol using a cholesterol dehydrogenase that links the presence of cholesterol to the production of NADH and the activation of proluciferin that produces light with luciferase. The protocol was used according to the instructions from the kit. Briefly, cells were washed twice with 100 μL PBS. Cells were lysed using 50 μL cholesterol lysis solution and incubated 30 min at 37 °C. Then 50 μL cholesterol detection reagent with esterase was added to each well. The plate was incubated at room temperature for 1 h using a plate shaker at a low rpm. Finally, the luminescence of each well was recorded on a plate-reading luminometer (Tecan Infinity M1000). Meanwhile, cell proliferation of the duplicates with the same coating cell number was tested using CCK8 reagent. OD450 values represent the viable cell number. The relative total cholesterol levels were normalized to the OD value of the corresponding well first and then to the control.

CRISPR/Cas9 mediated Pin1 knockout mouse bladder organoids

Based on previous studies describing bladder⁶⁵ and prostate⁶⁶ organoid culture methods, bladders from 4-week-old wild type mice (both female and male) were harvested and cut into halves. The innermost urothelial cell layer was stripped off and digested using the collagenase/Dispase II/Dnase I solution for at least 1 h with a 37 °C shaker. Then a single cell suspension was obtained using a 40 μm filter. The LentiCRISPRv2-puro plasmid containing cas9 and puro fragments was purchased from RRID:Addgene_98290. The single guide RNA sequences are CAGATGTGAGCAGCGCACCT upstream and AGGTGCGCTGCTCACATCTGC downstream of the 2^nd exon of the mouse Pin1 gene. The knockout efficiency of moPin1-cas9 lentivirus was tested using MB49 mouse bladder carcinoma cells. Then, moPin1-cas9 lentivirus was transfected into primary mouse urothelial cells using a spinoculation method at 200 g and room temperature for 4 h. The plate was put back to 37 °C incubator for 2 h. Finally, the cell suspension was collected and centrifuged at 500 g, room temperature for 5 min. The cell pellet was resuspended with organoid culture media and reseeded in a 96-well low-attachment plate. The mouse bladder organoid culture media⁶⁵ contains advanced DMEM/F-12 (ThermoFisher, 12634010), FGF10 (100 ng/mL of Peprotech, 100–26), FGF7 (25 ng/mL of Peprotech, 100–19), A83–01 (500 nM of Cayman Chemical, 9001799), B27 (2% ThermoFisher, 17504001), 100 U/mL penicillin,100 μg/mL streptomycin, and 50% (vol/vol) Matrigel. ROCK inhibitor (Y-27632, 10 μM) was added to the media after passaging to prevent cell death. Organoids were frozen in freezing media containing 50% FBS, 10% DMSO and 40% advanced DMEM/F12. Similarly, 2 mg/mL Dispase II solution was added to each well to digest the Matrigel. 1 h later, 1^st organoids were collected through centrifugation at 400 g for 5 min and then washed several times with ice-cold PBS to remove the Matrigel. 0.25% Trypsin/EDTA was applied to dissociate organoids into single cells, then inactivated with complete medium and centrifuged. Next, single cell suspensions digested from the primary organoids were infected with a lentiviral expression construct (using pLenti CMV GFP Hygro (656–4) backbone, RRID:Addgene_17446) containing full-length Pin1 to obtain Pin1 re-expressing (rescue) mouse primary urothelial cells and re-seed in a 96-well low-attachment plate for secondary and tertiary organoid formation.

Statistical analysis and reproducibility

The QuPath 0.4.4 software was used to quantify the H-score for DAB staining of human bladder cancer microarray. The ImageJ, RRID:SCR_003070 software was used to quantify positive stained cells in immunostaining images. Mice for transplantation and treatment were assigned to experimental groups randomly but with each group containing males and females, except for the orthotropic tumor transplantation experiments, which for technical reasons were all females. All experiments were blind to group assignment and outcome assessment. Based on previous work in terms of sample size calculation^{67, 68}, we assigned each group at least 4 mice to increase the chance of getting more accurate and significant results for animal studies. To largely reduce the variation within each group, we chose at least 3 replicates to do statistical analyses for in vitro data. To accurately obtain alterations in control and different experimental groups for in vivo studies, we tried to calculate data from more sections, more fields and more cells. Negative control and PIN1 over-expressing SV-HUC-1 and RT4 cells with similar passages were used for all functional analyses at the same time, which represents results for one experiment. Control, PIN1 knockout and PIN1 re-expressing T24 and 5637 cells with similar passages were chosen for all functional analyses in vitro and in vivo at the same time, which represents results for one experiment. For the immunostaining of spheroids produced from T24 and 5637 cells and organoids from mouse primary urothelial cells, we included both compact and lumen-containing spheroids and organoids for statistical analyses. Microsoft Excel (RRID:SCR_016137) and GraphPad Prism 9.3.1 (RRID:SCR_002798) were used for data compilation and graphical representation. All the in vivo and cell culture experiments were repeated at least twice independently with similar results. Data are expressed as mean + standard error of the mean (S.E.M.). All p values were calculated using a two-tailed unpaired Student’s t-test and a p-value <0.05 was defined to be significant. ns, not significant (p > 0.05)

Significance:

This study provides deeper insights into the regulatory role of the PIN1 phospho-dependent prolyl isomerase in bladder cancer. The identification of the link between PIN1 and SREBP2-mediated transcription and cholesterol biosynthesis offers the potential for developing novel therapeutic strategies for bladder cancer.

Data availability

The RNA-sequencing data that supports the findings of this study have been deposited in the Gene Expression Omnibus (GEO) database (RRID:SCR_005012) of NCBI under accession number GSE262761. The single cell RNA-sequencing data analyzed in this study were obtained from Gene Expression Omnibus (GEO) at GSE129845 and GSE135337. The codes for analyzing the single cell RNA-sequencing data are deposited to the website https://github.com/WangX200/BLCA.scRNAdataset. The processed data for single cell RNA-sequencing are deposited to figshare https://figshare.com/s/cac9c270d8a160b55567 or DOI: 10.6084/m9.figshare.27198105. All other statistical data are available in the main text or the supplementary materials. The source data for all immunofluorescence staining experiments underlying Figure 1D, 1F, 1H, 1J, 2E, 2H, 2I, 3J, 3M, 3N, 4J, 4K, 5N, 5O, 5P, 7D, 7F, 7K, 7L and Supplementary Fig. S3L, S3N, S5H, S5J, S5L, S6C, S6E, S6G, S8J–S8L, S9B, S10F, S10I, S10K, S12A, S12C, S12T–S12V, S13H–S13J, S21C, S21E and S21G are provided in Supplementary Table 3 with different sheets.