Final Report Summary - BCMOLMED (Molecular Medicine for Bladder Cancer)
Introduction: Molecular Medicine for Bladder Cancer (BC), BCMolMed, is a European funded project combing interdisciplinary expertise from leading academic and industrial research groups in the field of translational science. The main objective of the project is to provide a comprehensive training for Early Stage Researchers in cutting-edge proteomics technologies, systems biology approaches, biomarker development and oncology. More information about BCMolMed project can be found on the project’s Website http://www.bcmolmed.org/. Research activities in BCMolMed are driven by the unmet clinical needs in bladder cancer. Currently, bladder cancer places a heavy burden on the society and health-care system, accounting for 165,000 deaths worldwide. According to the histopathological characteristics, malignant tumors are grouped into: a) non-muscle invasive (NMIBC), corresponding to around 75% of cases at initial presentation, and b) muscle-invasive disease (MIBC), representing 25% of cases at initial presentation. The high recurrence and progression rates impose a need for a constant patient monitoring. Furthermore, limitations of the currently applied disease management means (invasive nature of cystoscopy, sub-optimal performance of non-invasive markers, and limited treatment options for aggressive disease) underscore the necessity for the development of new approaches. To address these clinical needs, three early stage researchers (ESRs) were recruited and the following research activities were undertaken, of which the results are outlined below:
1) Verification and assessment of currently known bladder cancer biomarkers.
2) Identification of novel molecular determinants of bladder cancer.
3) Integration of available data and development of an "in silico" disease model.
Summary overview of BCMolMed results:
The research activities of ESR1 targeted the exploration of existing potential bladder cancer biomarkers. In detail, ESR1 evaluated the urinary peptide biomarkers using capillary electrophoresis coupled to mass spectrometry (CE-MS), resulting in a publication in Clinical Cancer Research (Frantzi et al, Clinical Cancer Research, 2016). Through the interaction of ESR1 with another closely related European Project (TransBioBC), a large cohort of urine samples from bladder cancer patients could be analysed, providing very high statistical power. In the course of this study, two biomarker panels were established and tested in a total of 1,357 individual samples, aiming at diagnosis of primary (n=721) and recurrent BC (n=636). A biomarker panel for diagnosis of primary BC comprised of 116 urine peptides, differentially abundant between primary BC and urological controls. The panel exhibited excellent performance in the independent set of samples (area under the curve, AUC, of 0.87; sensitivity of 91% and specificity of 68%). A biomarker panel for diagnosis of recurrent cancer was comprised of 106 urinary peptides differentially excreted between patients with recurrent BC and patients without a sign of recurrence for at least 1 year. An AUC of 0.75 was achieved upon validation in the independent cohort (sensitivity of 87% and specificity of 51%). Accuracy of detection of cancer relapse was improved through the combination of urinary biomarkers with cytological examination (AUC=0.87). The resulting manuscript reporting on the evaluation of the urinary proteome profiling using the CE-MS approach received a high level of attention and was also featured in Nature Reviews Urology (13, 240–241, 2016). Further, ESR1 investigated the urinary and tissue levels of histone 2B (H2B) and NRC-interacting factor 1 (NIF1) using enzyme-linked immunosorbent Assay (ELISA) and immunohistochemistry (IHC), respectively. The urinary levels of NIF-1 and H2B showed a significant difference in the urine from patients with benign urological diseases in comparison to patients with bladder cancer. NIF-1 showed decreased tissue expression levels as BC cancer progresses from NMIBC to MIBC, while H2B presented the opposite trend (Frantzi et al, Journal of Proteome Research, 2013). The results were awarded with a travel award for the ESUR meeting (2013). Moreover, ESR1 investigated the biological role of Profilin-1 in bladder cancer invasion. Comprehensive characterization of the effect of Profilin-1 silencing in T24M, a BC cell line, was evaluated though the combination of in vitro (migration, proliferation, adhesion assays, flow cytometry) and in vivo assays, along with mRNA sequencing analysis. Down-regulation of Profilin-1 decreased cell adhesion as well as tumor growth. Additionally, silencing of Profilin-1 resulted in the decreased expression of proteins associated with the non-canonical Wnt/Ca2+ signaling pathway (Frantzi et al, Oncotarget, 2016).
The research conducted by ESR2 in the course of BCMolMed focused on the identification of novel molecular determinants of bladder cancer invasion, using high-resolution tissue and cellular proteomics profiling. Comprehensive characterization of intracellular and secreted proteins was performed using bladder cancer cell line models that differ in metastatic potential (T24 and T24M). Secreted proteins were extracted from conditioned medium and from fractions enriched in endoplasmic reticulum and Golgi apparatus (ER/Golgi). Employing liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) analysis, a total of 253 (for the analysis of total cell proteome), 217 (for the analysis of ER/Golgi fraction) and 256 (for the analysis of conditioned medium, CM) differentially expressed proteins were identified comparing T24 and T24M cells. Cross-correlation of proteomics and RNA sequencing data from the same cell lines allowed for shortlisting of 253 proteins, exhibiting agreement in the regulation trend using at least 2 strategies (any of: transcriptomics, proteomics cell extract, proteomics secretome, proteomics ER/Golgi). This cross-correlation of -omics datasets increased the validity of the findings, also reflected by the higher level of agreement of the 253 proteins with the literature, in comparison to individual datasets. The work was published in Scientific Reports (Latosinska et al., Scientific Reports, 2016) and was awarded with a travel award for the ESUR meeting (2014). Moreover, ESR2 performed an extensive investigation of the bladder cancer tissue proteome. Initially, a comparative analysis of two quantification strategies applied in proteomics (label-free and label-based quantification) was conducted, to allow selection of the most suitable method for identification of differentially expressed proteins in tissue samples from patients with non-muscle invasive (pTa, n=4) and muscle-invasive bladder cancer (pT2+, n=4). Application of the label-free method resulted in the identification of 910 proteins, whereas the label-based approach resulted in the identification of 1092 (when combined with additional pre-fractionation) or 332 proteins (alone). A significantly higher protein sequence coverage was observed for the label-free strategy. Although the application of pre-fractionation resulted in an increase in the protein identification rate when comparing with the label-free method, this did not result in the identification of more differentially expressed proteins (Latosinska et al, Plos One, 2015). Considering also the increased validity obtained by using the label-free approach (as reflected by the higher protein sequence coverage), this method was selected to subsequently evaluate the changes associated with bladder cancer invasion. Following the comparative analysis of 11 tissue samples (n=5 NMIBC and n=6 MIBC) a total of 144 differentially expressed proteins were shortlisted, based on the cross-correlation of two evaluation methods. In silico pathway and interactome analysis indicated that many of the observed changes are hallmarks of cancer (i.e. protein synthesis, metabolism, angiogenesis, endocytosis, cell-ECM interactions, cytoskeletal remodeling, cell adhesion, oxidative response/ xenobiotic metabolism), further confirming the validity of the approach. Based on the biological relevance of the associated molecules and the novelty of the findings, two proteins were further investigated using in vitro and in vivo models. Knockdown of these proteins in T24M cells resulted in the inhibition of cell migration and proliferation. Additionally, significant decrease in the tumor growth was noted upon silencing of one of the molecule (Latosinska et al, manuscript in preparation).
The research project of ESR3 focused on the development of in silico models of bladder cancer invasiveness and a BC-related database. To enable a comprehensive analysis of molecular alterations, molecular information on BC invasion was extracted from publicly available resources (such as PubMed, Web of Science, Google Scholar) as well as dedicated -omics repositories (Gene Expression Omnibus, GEO, or ArrayExpress). To avoid misinterpretation, only datasets of high validity were included, applying the following criteria: a) sample size >50 samples/study, b) magnitude of differential abundance (>2-fold change), c) specific phenotypic conditions (T1, T2a/b, T3a/b, T4a/b). Following these criteria, a total of 1,279 protein-coding genes were retrieved and used for the development of the interactome of muscle invasive bladder cancer. Through this approach novel processes associated with bladder cancer invasion were identified, such as regulation of actin cytoskeleton, Neurotrophin signalling pathway and Endocytosis (Bhat et al, Plos One, 2015). Additionally, mined data were used as a basis to develop BCCluster, a knowledgebase on the molecular determinants of bladder cancer invasion. BCCluster is available online ( http://bccluster.org/ ), providing information on DNA-mutations, DNA-methylations, mRNAs, miRNAs, proteins and metabolites associated with muscle invasive bladder cancer (Bhat et al. Bladder Cancer, 2016). In total, BCCluster integrates data from 112 published manuscripts, and 1,559 statistically significant features associated with BC invasion. Additionally, a total of 435 protein-protein interaction data and 92 molecular pathways significant in BC invasion were deposited in this database.
Conclusions: Overall, the BCMolMed program achieved all its aims and delivered meaningful and significant outputs, providing an insight into molecular processes underlying bladder cancer invasion, contributing to the identification of putative therapeutic targets, as well as novel urinary biomarkers for the diagnosis of primary and recurrent bladder cancer. The obtained results are expected to have a significant impact on bladder cancer, with a high potential to improve disease management. Know-how on the use of high-resolution proteomics profiling of tissue and cell lines, subcellular fractionation methods, consolidation of existing and literature data, development of in silico disease models and biomarker assay development was gained. The results were extensively disseminated to different target groups including researches, clinicians, general public and industry, to increase awareness about the ongoing research activities in the field of bladder cancer as well as about the significant impact of the European funding on the training of scientists and health research. A total of 19 articles were published in leading scientific journals, including original research articles, reviews and book chapters. ESRs have received an extensive training in the field of translational research as well as a series of complementary skills enhancing their competitiveness in the field. The overall success of the BCMolMed project is also depicted in a short article that was recently published online on the website of the German National Contact point responsible for Marie Curie project as a paradigm of Marie Curie success story (http://www.nks-kmu.de/teilnahme-erfolgsgeschichte-mosaiques-diagnostics.php). All three ESRs have obtained their PhD degree, and, also as a result of the high number of publications achieved in leading journals, all ESRs have secured postdoc positions and are now advancing in their careers.
Socio-economic impacts of the project: Considering the heavy burden of bladder cancer, the ultimate clinical need is to decrease the numbers of patients dying from the disease. Specifically, patients at advance disease stage have a significantly less favorable outcome. Thus, an improvement in the patient outcome could be achieved by timely diagnosis of primary and recurrent disease as well as by introduction of novel targets for therapeutic intervention. Therefore, the results obtained in the course of BCMolMed (i.e. urinary-based biomarker panels for diagnosis of primary and recurrent bladder cancer, along with the development of the biology driven therapeutic targets) contribute towards this improvement. Impact of the BCMolMed project may be exemplified by: (a) timely and more accurate diagnosis of bladder cancer, (b) reduction of the number of invasive cystoscopies, thus increase of compliance, (c) improvement on the bladder cancer therapeutic options through the proposed novel targets for therapeutic intervention.
Contact details: Prof. Dr. Dr. Harald Mischak, Mosaiques Diagnostics GmbH, Rotenburger Str. 20,
30659 Hannover, Germany, Tel: +49 (0)511 55 47 44 13, Fax: +49 (0)511 55 47 44 31, E-mail: mischak@mosaiques-diagnostics.com
1) Verification and assessment of currently known bladder cancer biomarkers.
2) Identification of novel molecular determinants of bladder cancer.
3) Integration of available data and development of an "in silico" disease model.
Summary overview of BCMolMed results:
The research activities of ESR1 targeted the exploration of existing potential bladder cancer biomarkers. In detail, ESR1 evaluated the urinary peptide biomarkers using capillary electrophoresis coupled to mass spectrometry (CE-MS), resulting in a publication in Clinical Cancer Research (Frantzi et al, Clinical Cancer Research, 2016). Through the interaction of ESR1 with another closely related European Project (TransBioBC), a large cohort of urine samples from bladder cancer patients could be analysed, providing very high statistical power. In the course of this study, two biomarker panels were established and tested in a total of 1,357 individual samples, aiming at diagnosis of primary (n=721) and recurrent BC (n=636). A biomarker panel for diagnosis of primary BC comprised of 116 urine peptides, differentially abundant between primary BC and urological controls. The panel exhibited excellent performance in the independent set of samples (area under the curve, AUC, of 0.87; sensitivity of 91% and specificity of 68%). A biomarker panel for diagnosis of recurrent cancer was comprised of 106 urinary peptides differentially excreted between patients with recurrent BC and patients without a sign of recurrence for at least 1 year. An AUC of 0.75 was achieved upon validation in the independent cohort (sensitivity of 87% and specificity of 51%). Accuracy of detection of cancer relapse was improved through the combination of urinary biomarkers with cytological examination (AUC=0.87). The resulting manuscript reporting on the evaluation of the urinary proteome profiling using the CE-MS approach received a high level of attention and was also featured in Nature Reviews Urology (13, 240–241, 2016). Further, ESR1 investigated the urinary and tissue levels of histone 2B (H2B) and NRC-interacting factor 1 (NIF1) using enzyme-linked immunosorbent Assay (ELISA) and immunohistochemistry (IHC), respectively. The urinary levels of NIF-1 and H2B showed a significant difference in the urine from patients with benign urological diseases in comparison to patients with bladder cancer. NIF-1 showed decreased tissue expression levels as BC cancer progresses from NMIBC to MIBC, while H2B presented the opposite trend (Frantzi et al, Journal of Proteome Research, 2013). The results were awarded with a travel award for the ESUR meeting (2013). Moreover, ESR1 investigated the biological role of Profilin-1 in bladder cancer invasion. Comprehensive characterization of the effect of Profilin-1 silencing in T24M, a BC cell line, was evaluated though the combination of in vitro (migration, proliferation, adhesion assays, flow cytometry) and in vivo assays, along with mRNA sequencing analysis. Down-regulation of Profilin-1 decreased cell adhesion as well as tumor growth. Additionally, silencing of Profilin-1 resulted in the decreased expression of proteins associated with the non-canonical Wnt/Ca2+ signaling pathway (Frantzi et al, Oncotarget, 2016).
The research conducted by ESR2 in the course of BCMolMed focused on the identification of novel molecular determinants of bladder cancer invasion, using high-resolution tissue and cellular proteomics profiling. Comprehensive characterization of intracellular and secreted proteins was performed using bladder cancer cell line models that differ in metastatic potential (T24 and T24M). Secreted proteins were extracted from conditioned medium and from fractions enriched in endoplasmic reticulum and Golgi apparatus (ER/Golgi). Employing liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) analysis, a total of 253 (for the analysis of total cell proteome), 217 (for the analysis of ER/Golgi fraction) and 256 (for the analysis of conditioned medium, CM) differentially expressed proteins were identified comparing T24 and T24M cells. Cross-correlation of proteomics and RNA sequencing data from the same cell lines allowed for shortlisting of 253 proteins, exhibiting agreement in the regulation trend using at least 2 strategies (any of: transcriptomics, proteomics cell extract, proteomics secretome, proteomics ER/Golgi). This cross-correlation of -omics datasets increased the validity of the findings, also reflected by the higher level of agreement of the 253 proteins with the literature, in comparison to individual datasets. The work was published in Scientific Reports (Latosinska et al., Scientific Reports, 2016) and was awarded with a travel award for the ESUR meeting (2014). Moreover, ESR2 performed an extensive investigation of the bladder cancer tissue proteome. Initially, a comparative analysis of two quantification strategies applied in proteomics (label-free and label-based quantification) was conducted, to allow selection of the most suitable method for identification of differentially expressed proteins in tissue samples from patients with non-muscle invasive (pTa, n=4) and muscle-invasive bladder cancer (pT2+, n=4). Application of the label-free method resulted in the identification of 910 proteins, whereas the label-based approach resulted in the identification of 1092 (when combined with additional pre-fractionation) or 332 proteins (alone). A significantly higher protein sequence coverage was observed for the label-free strategy. Although the application of pre-fractionation resulted in an increase in the protein identification rate when comparing with the label-free method, this did not result in the identification of more differentially expressed proteins (Latosinska et al, Plos One, 2015). Considering also the increased validity obtained by using the label-free approach (as reflected by the higher protein sequence coverage), this method was selected to subsequently evaluate the changes associated with bladder cancer invasion. Following the comparative analysis of 11 tissue samples (n=5 NMIBC and n=6 MIBC) a total of 144 differentially expressed proteins were shortlisted, based on the cross-correlation of two evaluation methods. In silico pathway and interactome analysis indicated that many of the observed changes are hallmarks of cancer (i.e. protein synthesis, metabolism, angiogenesis, endocytosis, cell-ECM interactions, cytoskeletal remodeling, cell adhesion, oxidative response/ xenobiotic metabolism), further confirming the validity of the approach. Based on the biological relevance of the associated molecules and the novelty of the findings, two proteins were further investigated using in vitro and in vivo models. Knockdown of these proteins in T24M cells resulted in the inhibition of cell migration and proliferation. Additionally, significant decrease in the tumor growth was noted upon silencing of one of the molecule (Latosinska et al, manuscript in preparation).
The research project of ESR3 focused on the development of in silico models of bladder cancer invasiveness and a BC-related database. To enable a comprehensive analysis of molecular alterations, molecular information on BC invasion was extracted from publicly available resources (such as PubMed, Web of Science, Google Scholar) as well as dedicated -omics repositories (Gene Expression Omnibus, GEO, or ArrayExpress). To avoid misinterpretation, only datasets of high validity were included, applying the following criteria: a) sample size >50 samples/study, b) magnitude of differential abundance (>2-fold change), c) specific phenotypic conditions (T1, T2a/b, T3a/b, T4a/b). Following these criteria, a total of 1,279 protein-coding genes were retrieved and used for the development of the interactome of muscle invasive bladder cancer. Through this approach novel processes associated with bladder cancer invasion were identified, such as regulation of actin cytoskeleton, Neurotrophin signalling pathway and Endocytosis (Bhat et al, Plos One, 2015). Additionally, mined data were used as a basis to develop BCCluster, a knowledgebase on the molecular determinants of bladder cancer invasion. BCCluster is available online ( http://bccluster.org/ ), providing information on DNA-mutations, DNA-methylations, mRNAs, miRNAs, proteins and metabolites associated with muscle invasive bladder cancer (Bhat et al. Bladder Cancer, 2016). In total, BCCluster integrates data from 112 published manuscripts, and 1,559 statistically significant features associated with BC invasion. Additionally, a total of 435 protein-protein interaction data and 92 molecular pathways significant in BC invasion were deposited in this database.
Conclusions: Overall, the BCMolMed program achieved all its aims and delivered meaningful and significant outputs, providing an insight into molecular processes underlying bladder cancer invasion, contributing to the identification of putative therapeutic targets, as well as novel urinary biomarkers for the diagnosis of primary and recurrent bladder cancer. The obtained results are expected to have a significant impact on bladder cancer, with a high potential to improve disease management. Know-how on the use of high-resolution proteomics profiling of tissue and cell lines, subcellular fractionation methods, consolidation of existing and literature data, development of in silico disease models and biomarker assay development was gained. The results were extensively disseminated to different target groups including researches, clinicians, general public and industry, to increase awareness about the ongoing research activities in the field of bladder cancer as well as about the significant impact of the European funding on the training of scientists and health research. A total of 19 articles were published in leading scientific journals, including original research articles, reviews and book chapters. ESRs have received an extensive training in the field of translational research as well as a series of complementary skills enhancing their competitiveness in the field. The overall success of the BCMolMed project is also depicted in a short article that was recently published online on the website of the German National Contact point responsible for Marie Curie project as a paradigm of Marie Curie success story (http://www.nks-kmu.de/teilnahme-erfolgsgeschichte-mosaiques-diagnostics.php). All three ESRs have obtained their PhD degree, and, also as a result of the high number of publications achieved in leading journals, all ESRs have secured postdoc positions and are now advancing in their careers.
Socio-economic impacts of the project: Considering the heavy burden of bladder cancer, the ultimate clinical need is to decrease the numbers of patients dying from the disease. Specifically, patients at advance disease stage have a significantly less favorable outcome. Thus, an improvement in the patient outcome could be achieved by timely diagnosis of primary and recurrent disease as well as by introduction of novel targets for therapeutic intervention. Therefore, the results obtained in the course of BCMolMed (i.e. urinary-based biomarker panels for diagnosis of primary and recurrent bladder cancer, along with the development of the biology driven therapeutic targets) contribute towards this improvement. Impact of the BCMolMed project may be exemplified by: (a) timely and more accurate diagnosis of bladder cancer, (b) reduction of the number of invasive cystoscopies, thus increase of compliance, (c) improvement on the bladder cancer therapeutic options through the proposed novel targets for therapeutic intervention.
Contact details: Prof. Dr. Dr. Harald Mischak, Mosaiques Diagnostics GmbH, Rotenburger Str. 20,
30659 Hannover, Germany, Tel: +49 (0)511 55 47 44 13, Fax: +49 (0)511 55 47 44 31, E-mail: mischak@mosaiques-diagnostics.com