Adenoviral gene therapy for bladder cancer.

Enhanced by polyamide surfactant Syn3, intravesical administration of rAd-IFNα2b results in transduction of the virus into the bladder epithelium, resulting in the synthesis and expression of local IFNα2b cytokine. Upon secretion, IFNα2b binds to the IFNα receptor on bladder cancer and other cells, resulting in signaling via the JAK-STAT pathway. A plethora of induced IFN-stimulated genes containing IFN-sensitive response elements that contribute to activation of pathways restrict cancer growth.

Can Personalized Neoantigens Raise the T Cell Bar?

Ott et al. report the results of a phase 1B study in which a personalized neoantigen vaccine was combined with programmed death receptor-1 blockade in patients with advanced cancers. The study provides a framework for combinatorial vaccine therapies that could mount robust T cell responses, enhance tumor killing, and provide clinical benefit.

Intratumoral CD4+ T Cells Mediate Anti-tumor Cytotoxicity in Human Bladder Cancer.

Responses to anti-PD-1 immunotherapy occur but are infrequent in bladder cancer. The specific T cells that mediate tumor rejection are unknown. T cells from human bladder tumors and non-malignant tissue were assessed with single-cell RNA and paired T cell receptor (TCR) sequencing of 30,604 T cells from 7 patients. We find that the states and repertoires of CD8+ T cells are not distinct in tumors compared with non-malignant tissues. In contrast, single-cell analysis of CD4+ T cells demonstrates several tumor-specific states, including multiple distinct states of regulatory T cells. Surprisingly, we also find multiple cytotoxic CD4+ T cell states that are clonally expanded. These CD4+ T cells can kill autologous tumors in an MHC class II-dependent fashion and are suppressed by regulatory T cells. Further, a gene signature of cytotoxic CD4+ T cells in tumors predicts a clinical response in 244 metastatic bladder cancer patients treated with anti-PD-L1.

A Phase Ib Trial of Personalized Neoantigen Therapy Plus Anti-PD-1 in Patients with Advanced Melanoma, Non-small Cell Lung Cancer, or Bladder Cancer.

Neoantigens arise from mutations in cancer cells and are important targets of T cell-mediated anti-tumor immunity. Here, we report the first open-label, phase Ib clinical trial of a personalized neoantigen-based vaccine, NEO-PV-01, in combination with PD-1 blockade in patients with advanced melanoma, non-small cell lung cancer, or bladder cancer. This analysis of 82 patients demonstrated that the regimen was safe, with no treatment-related serious adverse events observed. De novo neoantigen-specific CD4+ and CD8+ T cell responses were observed post-vaccination in all of the patients. The vaccine-induced T cells had a cytotoxic phenotype and were capable of trafficking to the tumor and mediating cell killing. In addition, epitope spread to neoantigens not included in the vaccine was detected post-vaccination. These data support the safety and immunogenicity of this regimen in patients with advanced solid tumors (Clinicaltrials.gov: NCT02897765).

Revealing the Activity of Trimeric G-proteins in Live Cells with a Versatile Biosensor Design.

Heterotrimeric G-proteins (Gαßγ) are the main transducers of signals from GPCRs, mediating the action of countless natural stimuli and therapeutic agents. However, there are currently no robust approaches to directly measure the activity of endogenous G-proteins in cells. Here, we describe a suite of optical biosensors that detect endogenous active G-proteins with sub-second resolution in live cells. Using a modular design principle, we developed genetically encoded, unimolecular biosensors for endogenous Gα-GTP and free Gßγ: the two active species of heterotrimeric G-proteins. This design was leveraged to generate biosensors with specificity for different heterotrimeric G-proteins or for other G-proteins, such as Rho GTPases. Versatility was further validated by implementing the biosensors in multiple contexts, from characterizing cancer-associated G-protein mutants to neurotransmitter signaling in primary neurons. Overall, the versatile biosensor design introduced here enables studying the activity of endogenous G-proteins in live cells with high fidelity, temporal resolution, and convenience.

Tumor Evolution and Drug Response in Patient-Derived Organoid Models of Bladder Cancer.

Bladder cancer is the fifth most prevalent cancer in the U.S., yet is understudied, and few laboratory models exist that reflect the biology of the human disease. Here, we describe a biobank of patient-derived organoid lines that recapitulates the histopathological and molecular diversity of human bladder cancer. Organoid lines can be established efficiently from patient biopsies acquired before and after disease recurrence and are interconvertible with orthotopic xenografts. Notably, organoid lines often retain parental tumor heterogeneity and exhibit a spectrum of genomic changes that are consistent with tumor evolution in culture. Analyses of drug response using bladder tumor organoids show partial correlations with mutational profiles, as well as changes associated with treatment resistance, and specific responses can be validated using xenografts in vivo. Our studies indicate that patient-derived bladder tumor organoids represent a faithful model system for studying tumor evolution and treatment response in the context of precision cancer medicine.

Comprehensive Molecular Characterization of Muscle-Invasive Bladder Cancer.

We report a comprehensive analysis of 412 muscle-invasive bladder cancers characterized by multiple TCGA analytical platforms. Fifty-eight genes were significantly mutated, and the overall mutational load was associated with APOBEC-signature mutagenesis. Clustering by mutation signature identified a high-mutation subset with 75% 5-year survival. mRNA expression clustering refined prior clustering analyses and identified a poor-survival "neuronal" subtype in which the majority of tumors lacked small cell or neuroendocrine histology. Clustering by mRNA, long non-coding RNA (lncRNA), and miRNA expression converged to identify subsets with differential epithelial-mesenchymal transition status, carcinoma in situ scores, histologic features, and survival. Our analyses identified 5 expression subtypes that may stratify response to different treatments.

Gut microbiota carcinogen metabolism causes distal tissue tumours.

Exposure to environmental pollutants and human microbiome composition are important predisposition factors for tumour development1,2. Similar to drug molecules, pollutants are typically metabolized in the body, which can change their carcinogenic potential and affect tissue distribution through altered toxicokinetics3. Although recent studies demonstrated that human-associated microorganisms can chemically convert a wide range of xenobiotics and influence the profile and tissue exposure of resulting metabolites4,5, the effect of microbial biotransformation on chemical-induced tumour development remains unclear. Here we show that the depletion of the gut microbiota affects the toxicokinetics of nitrosamines, which markedly reduces the development and severity of nitrosamine-induced urinary bladder cancer in mice6,7. We causally linked this carcinogen biotransformation to specific gut bacterial isolates in vitro and in vivo using individualized bacterial culture collections and gnotobiotic mouse models, respectively. We tested gut communities from different human donors to demonstrate that microbial carcinogen metabolism varies between individuals and we showed that this metabolic activity applies to structurally related nitrosamine carcinogens. Altogether, these results indicate that gut microbiota carcinogen metabolism may be a contributing factor for chemical-induced carcinogenesis, which could open avenues to target the microbiome for improved predisposition risk assessment and prevention of cancer.

Y chromosome loss in cancer drives growth by evasion of adaptive immunity.

Loss of the Y chromosome (LOY) is observed in multiple cancer types, including 10-40% of bladder cancers1-6, but its clinical and biological significance is unknown. Here, using genomic and transcriptomic studies, we report that LOY correlates with poor prognoses in patients with bladder cancer. We performed in-depth studies of naturally occurring LOY mutant bladder cancer cells as well as those with targeted deletion of Y chromosome by CRISPR-Cas9. Y-positive (Y+) and Y-negative (Y-) tumours grew similarly in vitro, whereas Y- tumours were more aggressive than Y+ tumours in immune-competent hosts in a T cell-dependent manner. High-dimensional flow cytometric analyses demonstrated that Y- tumours promote striking dysfunction or exhaustion of CD8+ T cells in the tumour microenvironment. These findings were validated using single-nuclei RNA sequencing and spatial proteomic evaluation of human bladder cancers. Of note, compared with Y+ tumours, Y- tumours exhibited an increased response to anti-PD-1 immune checkpoint blockade therapy in both mice and patients with cancer. Together, these results demonstrate that cancer cells with LOY mutations alter T cell function, promoting T cell exhaustion and sensitizing them to PD-1-targeted immunotherapy. This work provides insights into the basic biology of LOY mutation and potential biomarkers for improving cancer immunotherapy.

IDH2 stabilizes HIF-1α-induced metabolic reprogramming and promotes chemoresistance in urothelial cancer.

Drug resistance contributes to poor therapeutic response in urothelial carcinoma (UC). Metabolomic analysis suggested metabolic reprogramming in gemcitabine-resistant urothelial carcinoma cells, whereby increased aerobic glycolysis and metabolic stimulation of the pentose phosphate pathway (PPP) promoted pyrimidine biosynthesis to increase the production of the gemcitabine competitor deoxycytidine triphosphate (dCTP) that diminishes its therapeutic effect. Furthermore, we observed that gain-of-function of isocitrate dehydrogenase 2 (IDH2) induced reductive glutamine metabolism to stabilize Hif-1α expression and consequently stimulate aerobic glycolysis and PPP bypass in gemcitabine-resistant UC cells. Interestingly, IDH2-mediated metabolic reprogramming also caused cross resistance to CDDP, by elevating the antioxidant defense via increased NADPH and glutathione production. Downregulation or pharmacological suppression of IDH2 restored chemosensitivity. Since the expression of key metabolic enzymes, such as TIGAR, TKT, and CTPS1, were affected by IDH2-mediated metabolic reprogramming and related to poor prognosis in patients, IDH2 might become a new therapeutic target for restoring chemosensitivity in chemo-resistant urothelial carcinoma.

Enfortumab Vedotin and Pembrolizumab in Untreated Advanced Urothelial Cancer.

BACKGROUND: No treatment has surpassed platinum-based chemotherapy in improving overall survival in patients with previously untreated locally advanced or metastatic urothelial carcinoma. METHODS: We conducted a phase 3, global, open-label, randomized trial to compare the efficacy and safety of enfortumab vedotin and pembrolizumab with the efficacy and safety of platinum-based chemotherapy in patients with previously untreated locally advanced or metastatic urothelial carcinoma. Patients were randomly assigned in a 1:1 ratio to receive 3-week cycles of enfortumab vedotin (at a dose of 1.25 mg per kilogram of body weight intravenously on days 1 and 8) and pembrolizumab (at a dose of 200 mg intravenously on day 1) (enfortumab vedotin-pembrolizumab group) or gemcitabine and either cisplatin or carboplatin (determined on the basis of eligibility to receive cisplatin) (chemotherapy group). The primary end points were progression-free survival as assessed by blinded independent central review and overall survival. RESULTS: A total of 886 patients underwent randomization: 442 to the enfortumab vedotin-pembrolizumab group and 444 to the chemotherapy group. As of August 8, 2023, the median duration of follow-up for survival was 17.2 months. Progression-free survival was longer in the enfortumab vedotin-pembrolizumab group than in the chemotherapy group (median, 12.5 months vs. 6.3 months; hazard ratio for disease progression or death, 0.45; 95% confidence interval [CI], 0.38 to 0.54; P<0.001), as was overall survival (median, 31.5 months vs. 16.1 months; hazard ratio for death, 0.47; 95% CI, 0.38 to 0.58; P<0.001). The median number of cycles was 12 (range, 1 to 46) in the enfortumab vedotin-pembrolizumab group and 6 (range, 1 to 6) in the chemotherapy group. Treatment-related adverse events of grade 3 or higher occurred in 55.9% of the patients in the enfortumab vedotin-pembrolizumab group and in 69.5% of those in the chemotherapy group. CONCLUSIONS: Treatment with enfortumab vedotin and pembrolizumab resulted in significantly better outcomes than chemotherapy in patients with untreated locally advanced or metastatic urothelial carcinoma, with a safety profile consistent with that in previous reports. (Funded by Astellas Pharma US and others; EV-302 ClinicalTrials.gov number, NCT04223856.).

Treatment of muscle-invasive and advanced bladder cancer in 2020.

Bladder cancer accounts for nearly 170,000 deaths worldwide annually. For over 4 decades, the systemic management of muscle-invasive and advanced bladder cancer has primarily consisted of platinum-based chemotherapy. Over the past 10 years, innovations in sequencing technologies have led to rapid genomic characterization of bladder cancer, deepening our understanding of bladder cancer pathogenesis and exposing potential therapeutic vulnerabilities. On the basis of its high mutational burden, immune checkpoint inhibitors were investigated in advanced bladder cancer, revealing durable responses in a subset of patients. These agents are now approved for several indications and highlight the changing treatment landscape of advanced bladder cancer. In addition, commonly expressed molecular targets were leveraged to develop targeted therapies, such as fibroblast growth factor receptor inhibitors and antibody-drug conjugates. The molecular characterization of bladder cancer and the development of novel therapies also have stimulated investigations into optimizing treatment approaches for muscle-invasive bladder cancer. Herein, the authors review the history of muscle-invasive and advanced bladder cancer management, highlight the important molecular characteristics of bladder cancer, describe the major advances in treatment, and offer future directions for therapeutic development.

ctDNA guiding adjuvant immunotherapy in urothelial carcinoma.

Minimally invasive approaches to detect residual disease after surgery are needed to identify patients with cancer who are at risk for metastatic relapse. Circulating tumour DNA (ctDNA) holds promise as a biomarker for molecular residual disease and relapse1. We evaluated outcomes in 581 patients who had undergone surgery and were evaluable for ctDNA from a randomized phase III trial of adjuvant atezolizumab versus observation in operable urothelial cancer. This trial did not reach its efficacy end point in the intention-to-treat population. Here we show that ctDNA testing at the start of therapy (cycle 1 day 1) identified 214 (37%) patients who were positive for ctDNA and who had poor prognosis (observation arm hazard ratio = 6.3 (95% confidence interval: 4.45-8.92); P < 0.0001). Notably, patients who were positive for ctDNA had improved disease-free survival and overall survival in the atezolizumab arm versus the observation arm (disease-free survival hazard ratio = 0.58 (95% confidence interval: 0.43-0.79); P = 0.0024, overall survival hazard ratio = 0.59 (95% confidence interval: 0.41-0.86)). No difference in disease-free survival or overall survival between treatment arms was noted for patients who were negative for ctDNA. The rate of ctDNA clearance at week 6 was higher in the atezolizumab arm (18%) than in the observation arm (4%) (P = 0.0204). Transcriptomic analysis of tumours from patients who were positive for ctDNA revealed higher expression levels of cell-cycle and keratin genes. For patients who were positive for ctDNA and who were treated with atezolizumab, non-relapse was associated with immune response signatures and basal-squamous gene features, whereas relapse was associated with angiogenesis and fibroblast TGFß signatures. These data suggest that adjuvant atezolizumab may be associated with improved outcomes compared with observation in patients who are positive for ctDNA and who are at a high risk of relapse. These findings, if validated in other settings, would shift approaches to postoperative cancer care.

Gold nanorod-assisted theranostic solution for nonvisible residual disease in bladder cancer.

Residual nonvisible bladder cancer after proper treatment caused by technological and therapeutic limitations is responsible for tumor relapse and progression. This study aimed to demonstrate the feasibility of a solution for simultaneous detection and treatment of bladder cancer lesions smaller than one millimeter. The α5ß1 integrin was identified as a specific marker in 81% of human high-grade nonmuscle invasive bladder cancers and used as a target for the delivery of targeted gold nanorods (GNRs). In a preclinical model of orthotopic bladder cancer expressing the α5ß1 integrin, the photoacoustic imaging of targeted GNRs visualized lesions smaller than one millimeter, and their irradiation with continuous laser was used to induce GNR-assisted hyperthermia. Necrosis of the tumor mass, improved survival, and computational modeling were applied to demonstrate the efficacy and safety of this solution. Our study highlights the potential of the GNR-assisted theranostic strategy as a complementary solution in clinical practice to reduce the risk of nonvisible residual bladder cancer after current treatment. Further validation through clinical studies will support the findings of the present study.

Tumor-neutrophil cross talk orchestrates the tumor microenvironment to determine the bladder cancer progression.

The immune landscape of bladder cancer progression is not fully understood, and effective therapies are lacking in advanced bladder cancer. Here, we visualized that bladder cancer cells recruited neutrophils by secreting interleukin-8 (IL-8); in turn, neutrophils played dual functions in bladder cancer, including hepatocyte growth factor (HGF) release and CCL3highPD-L1high super-immunosuppressive subset formation. Mechanistically, c-Fos was identified as the mediator of HGF up-regulating IL-8 transcription in bladder cancer cells, which was central to the positive feedback of neutrophil recruitment. Clinically, compared with serum IL-8, urine IL-8 was a better biomarker for bladder cancer prognosis and clinical benefit of immune checkpoint blockade (ICB). Additionally, targeting neutrophils or hepatocyte growth factor receptor (MET) signaling combined with ICB inhibited bladder cancer progression and boosted the antitumor effect of CD8+ T cells in mice. These findings reveal the mechanism by which tumor-neutrophil cross talk orchestrates the bladder cancer microenvironment and provide combination strategies, which may have broad impacts on patients suffering from malignancies enriched with neutrophils.

Nivolumab plus Gemcitabine-Cisplatin in Advanced Urothelial Carcinoma.

BACKGROUND: No new agent has improved overall survival in patients with unresectable or metastatic urothelial carcinoma when added to first-line cisplatin-based chemotherapy. METHODS: In this phase 3, multinational, open-label trial, we randomly assigned patients with previously untreated unresectable or metastatic urothelial carcinoma either to receive intravenous nivolumab (at a dose of 360 mg) plus gemcitabine-cisplatin (nivolumab combination) every 3 weeks for up to six cycles, followed by nivolumab (at a dose of 480 mg) every 4 weeks for a maximum of 2 years, or to receive gemcitabine-cisplatin alone every 3 weeks for up to six cycles. The primary outcomes were overall and progression-free survival. The objective response and safety were exploratory outcomes. RESULTS: A total of 608 patients underwent randomization (304 to each group). At a median follow-up of 33.6 months, overall survival was longer with nivolumab-combination therapy than with gemcitabine-cisplatin alone (hazard ratio for death, 0.78; 95% confidence interval [CI], 0.63 to 0.96; P = 0.02); the median survival was 21.7 months (95% CI, 18.6 to 26.4) as compared with 18.9 months (95% CI, 14.7 to 22.4), respectively. Progression-free survival was also longer with nivolumab-combination therapy than with gemcitabine-cisplatin alone (hazard ratio for progression or death, 0.72; 95% CI, 0.59 to 0.88; P = 0.001). The median progression-free survival was 7.9 months and 7.6 months, respectively. At 12 months, progression-free survival was 34.2% and 21.8%, respectively. The overall objective response was 57.6% (complete response, 21.7%) with nivolumab-combination therapy and 43.1% (complete response, 11.8%) with gemcitabine-cisplatin alone. The median duration of complete response was 37.1 months with nivolumab-combination therapy and 13.2 months with gemcitabine-cisplatin alone. Grade 3 or higher adverse events occurred in 61.8% and 51.7% of the patients, respectively. CONCLUSIONS: Combination therapy with nivolumab plus gemcitabine-cisplatin resulted in significantly better outcomes in patients with previously untreated advanced urothelial carcinoma than gemcitabine-cisplatin alone. (Funded by Bristol Myers Squibb and Ono Pharmaceutical; CheckMate 901 ClinicalTrials.gov number, NCT03036098.).

Erdafitinib or Chemotherapy in Advanced or Metastatic Urothelial Carcinoma.

BACKGROUND: Erdafitinib is a pan-fibroblast growth factor receptor (FGFR) inhibitor approved for the treatment of locally advanced or metastatic urothelial carcinoma in adults with susceptible FGFR3/2 alterations who have progression after platinum-containing chemotherapy. The effects of erdafitinib in patients with FGFR-altered metastatic urothelial carcinoma who have progression during or after treatment with checkpoint inhibitors (anti-programmed cell death protein 1 [PD-1] or anti-programmed death ligand 1 [PD-L1] agents) are unclear. METHODS: We conducted a global phase 3 trial of erdafitinib as compared with chemotherapy in patients with metastatic urothelial carcinoma with susceptible FGFR3/2 alterations who had progression after one or two previous treatments that included an anti-PD-1 or anti-PD-L1. Patients were randomly assigned in a 1:1 ratio to receive erdafitinib or the investigator's choice of chemotherapy (docetaxel or vinflunine). The primary end point was overall survival. RESULTS: A total of 266 patients underwent randomization: 136 to the erdafitinib group and 130 to the chemotherapy group. The median follow-up was 15.9 months. The median overall survival was significantly longer with erdafitinib than with chemotherapy (12.1 months vs. 7.8 months; hazard ratio for death, 0.64; 95% confidence interval [CI], 0.47 to 0.88; P = 0.005). The median progression-free survival was also longer with erdafitinib than with chemotherapy (5.6 months vs. 2.7 months; hazard ratio for progression or death, 0.58; 95% CI, 0.44 to 0.78; P<0.001). The incidence of grade 3 or 4 treatment-related adverse events was similar in the two groups (45.9% in the erdafitinib group and 46.4% in the chemotherapy group). Treatment-related adverse events that led to death were less common with erdafitinib than with chemotherapy (in 0.7% vs. 5.4% of patients). CONCLUSIONS: Erdafitinib therapy resulted in significantly longer overall survival than chemotherapy among patients with metastatic urothelial carcinoma and FGFR alterations after previous anti-PD-1 or anti-PD-L1 treatment. (Funded by Janssen Research and Development; THOR ClinicalTrials.gov number, NCT03390504.).

Inference of gene regulatory networks based on directed graph convolutional networks.

Inferring gene regulatory network (GRN) is one of the important challenges in systems biology, and many outstanding computational methods have been proposed; however there remains some challenges especially in real datasets. In this study, we propose Directed Graph Convolutional neural network-based method for GRN inference (DGCGRN). To better understand and process the directed graph structure data of GRN, a directed graph convolutional neural network is conducted which retains the structural information of the directed graph while also making full use of neighbor node features. The local augmentation strategy is adopted in graph neural network to solve the problem of poor prediction accuracy caused by a large number of low-degree nodes in GRN. In addition, for real data such as E.coli, sequence features are obtained by extracting hidden features using Bi-GRU and calculating the statistical physicochemical characteristics of gene sequence. At the training stage, a dynamic update strategy is used to convert the obtained edge prediction scores into edge weights to guide the subsequent training process of the model. The results on synthetic benchmark datasets and real datasets show that the prediction performance of DGCGRN is significantly better than existing models. Furthermore, the case studies on bladder uroepithelial carcinoma and lung cancer cells also illustrate the performance of the proposed model.

IDMIR: identification of dysregulated miRNAs associated with disease based on a miRNA-miRNA interaction network constructed through gene expression data.

Micro ribonucleic acids (miRNAs) play a pivotal role in governing the human transcriptome in various biological phenomena. Hence, the accumulation of miRNA expression dysregulation frequently assumes a noteworthy role in the initiation and progression of complex diseases. However, accurate identification of dysregulated miRNAs still faces challenges at the current stage. Several bioinformatics tools have recently emerged for forecasting the associations between miRNAs and diseases. Nonetheless, the existing reference tools mainly identify the miRNA-disease associations in a general state and fall short of pinpointing dysregulated miRNAs within a specific disease state. Additionally, no studies adequately consider miRNA-miRNA interactions (MMIs) when analyzing the miRNA-disease associations. Here, we introduced a systematic approach, called IDMIR, which enabled the identification of expression dysregulated miRNAs through an MMI network under the gene expression context, where the network's architecture was designed to implicitly connect miRNAs based on their shared biological functions within a particular disease context. The advantage of IDMIR is that it uses gene expression data for the identification of dysregulated miRNAs by analyzing variations in MMIs. We illustrated the excellent predictive power for dysregulated miRNAs of the IDMIR approach through data analysis on breast cancer and bladder urothelial cancer. IDMIR could surpass several existing miRNA-disease association prediction approaches through comparison. We believe the approach complements the deficiencies in predicting miRNA-disease association and may provide new insights and possibilities for diagnosing and treating diseases. The IDMIR approach is now available as a free R package on CRAN (https://CRAN.R-project.org/package=IDMIR).

Creation of bladder assembloids mimicking tissue regeneration and cancer.

Current organoid models are limited by their inability to mimic mature organ architecture and associated tissue microenvironments1,2. Here we create multilayer bladder 'assembloids' by reconstituting tissue stem cells with stromal components to represent an organized architecture with an epithelium surrounding stroma and an outer muscle layer. These assembloids exhibit characteristics of mature adult bladders in cell composition and gene expression at the single-cell transcriptome level, and recapitulate in vivo tissue dynamics of regenerative responses to injury. We also develop malignant counterpart tumour assembloids to recapitulate the in vivo pathophysiological features of urothelial carcinoma. Using the genetically manipulated tumour-assembloid platform, we identify tumoural FOXA1, induced by stromal bone morphogenetic protein (BMP), as a master pioneer factor that drives enhancer reprogramming for the determination of tumour phenotype, suggesting the importance of the FOXA1-BMP-hedgehog signalling feedback axis between tumour and stroma in the control of tumour plasticity.