Final Report Summary - CANCER KINOME (Systems Biology of Cancer Kinome)
Project context and objectives
In this project the goal was to use systems biology methods to characterise and decipher the phosphorylation mediated informational networks (i.e.signalling networks) of the 120 protein kinases causally linked to the onset of cancer. The project was divided to three specific goals:
-1) Identify and characterise the topology of the human cancer kinome protein interaction network;
-2) Functional annotation of the cancer kinome network;
-3) Database development and integration of obtained network models with existing biochemical and clinical knowledge. These three goals were scheduled to be reached by end of the project. The first goal consists of identification of kinase complexes, identification of novel regulatory proteins and post-translational modifications (priority on phosphorylation) within kinases. To identify the kinase complexes, we generated 120 transgenic epitope tagged HEK293 cell lines expressing the corresponding kinases and identified the interacting proteins using high-throughput AP/MS -based protein complex analysis strategy. Using this strategy, we could purify and identify kinase complexes for 107 (89.2 %) of the 120 kinases. The remaining 13 unanalysable kinases are linked to cell cycle regulation and cell survival, and are expressed in extremely low levels in the generated cell lines, while the activation of the kinase expression causes these cells to undergo apoptosis. Because of this problem we could not collect enough cells for the purifications and had to exclude these 13 kinases from AP/MS analyses.
The second goal was to gain functional insight on the identified cancer signalling networks using high-content microscopic analyses, to identify kinase substrates with protein microarrays, and to use mass spectrometry to identify phosphopeptides and phosphosites on the kinases themselves, as well as on the identified kinase interacting proteins.
The third goal was to develop a database from the signalling networks and to integrate existing information into this network. Due to the massive amount of data generated with AP/MS based protein complex analysis, we had to subdivide the project into smaller publishable units. The first project manuscript describes the findings from the analysis of 57 kinase, and the second and third of 32 and 25 kinases, respectively. By the end of this year the first manuscript with the findings from the analysis of the 57 kinases is to be submitted. In this analysis we have identified 481 proteins interacting with 57 bait proteins, consisting of a total of 652 interactions. Many of the newly identified interacting proteins are linked (by GO annotation) to either cell communication or nucleotide metabolism (namely RNA processing and transcription). In addition, to our great surprise the majority (two-thirds) of CMGC kinases also form complexes with other kinases, revealing a larger kinase signalling informational 'highway'. Furthermore, ~8% (38/481) of the kinase interacting proteins are known cancer- linked proteins, further illustrating that these kinases and their regulation is extremely critical for cancer development. Our AP/MS -based protein complex analysis also allows us to detect possible post-translational modifications in the kinases and in their interacting proteins. For the first set of 57 kinases we could detect 407 unique phosphorylation sites, 104 of which are novel.
Project outcomes
As it was now obvious that we had established a state-of-the-art high-throughput AP/MS -based protein complex analysis workflow, we also wanted to make our knowledge available to the whole interactomics research community. To demonstrate the applicability and reproducibility of the workflow in different laboratories, we teamed up with a pioneer of the interactomics field, Professor Giulio Superti-Furga (Center for Molecular Medicine of the Austrian Academy of Sciences (CeMM), Vienna, Austria). The professor's research group obtained 32 kinase-expressing cell lines from us. In addition, they used our standard operating procedures to compare and evaluate the AP/MS -based protein complex analysis workflow. The comparison of the two analyses shows an extremely high overlap (>78 %) of identified interacting proteins. We have one manuscript ready to be submitted from this collaboration and another under preparation.
To study the dynamics of protein-protein interactions we decided to use a set of 25 receptor tyrosine kinases (RTK) as a model. Using an irreversible protein-tyrosine phosphatase inhibitor, pervanadate, we have been able to activate the RTKs and study the effect of the length of activation with the dynamics of the protein-protein interaction, using label-free quantitative AP/MS. The data from this part of the project is currently under analysis.
For the database development, we have now gathered the data obtained from the project and have integrated additional information from already established databases (COSMIC, HPRD, PhosphoSitePlus, Panther and PINA). This information and the corresponding database will be published simultaneously with the manuscripts. The manuscripts are entitled 'Neutrophil p38d-PKD1 signaling fine-tunes PTEN activity to control acute inflammation of the lung', 'Dissecting the CMGC kinase signaling network', 'Robustness in large-scale affinity purification-mass spectrometry interaction proteomics by an inter-laboratory cross-wise comparative analysis' and 'Modularity of the human Hippo pathway interaction proteome'.
Taken together, we have now studied more than 100 cancer-linked human protein kinases using state-of-the-art proteomics methods and have gained detailed insight on how these kinases can exert their influence on cancer development. This knowledge on the physical and functional signalling networks of the cancer kinases will be extremely valuable when trying to understand the subtle chances in cellular homeostasis that can drive the cells to become cancerogenic. Furthermore, the large amount of data gathered on the project will be highly valuable for the wider bioresearch community and will initiate, in addition to the collaborative projects started by us, its own a plethora of projects. In addition to the more global impact, the project helped the researcher to establish an independent research career. The researcher is currently acting as Group Leader and the Director of the Proteomics Facility at the Institute of Biotechnology, University of Helsinki, Finland.
In this project the goal was to use systems biology methods to characterise and decipher the phosphorylation mediated informational networks (i.e.signalling networks) of the 120 protein kinases causally linked to the onset of cancer. The project was divided to three specific goals:
-1) Identify and characterise the topology of the human cancer kinome protein interaction network;
-2) Functional annotation of the cancer kinome network;
-3) Database development and integration of obtained network models with existing biochemical and clinical knowledge. These three goals were scheduled to be reached by end of the project. The first goal consists of identification of kinase complexes, identification of novel regulatory proteins and post-translational modifications (priority on phosphorylation) within kinases. To identify the kinase complexes, we generated 120 transgenic epitope tagged HEK293 cell lines expressing the corresponding kinases and identified the interacting proteins using high-throughput AP/MS -based protein complex analysis strategy. Using this strategy, we could purify and identify kinase complexes for 107 (89.2 %) of the 120 kinases. The remaining 13 unanalysable kinases are linked to cell cycle regulation and cell survival, and are expressed in extremely low levels in the generated cell lines, while the activation of the kinase expression causes these cells to undergo apoptosis. Because of this problem we could not collect enough cells for the purifications and had to exclude these 13 kinases from AP/MS analyses.
The second goal was to gain functional insight on the identified cancer signalling networks using high-content microscopic analyses, to identify kinase substrates with protein microarrays, and to use mass spectrometry to identify phosphopeptides and phosphosites on the kinases themselves, as well as on the identified kinase interacting proteins.
The third goal was to develop a database from the signalling networks and to integrate existing information into this network. Due to the massive amount of data generated with AP/MS based protein complex analysis, we had to subdivide the project into smaller publishable units. The first project manuscript describes the findings from the analysis of 57 kinase, and the second and third of 32 and 25 kinases, respectively. By the end of this year the first manuscript with the findings from the analysis of the 57 kinases is to be submitted. In this analysis we have identified 481 proteins interacting with 57 bait proteins, consisting of a total of 652 interactions. Many of the newly identified interacting proteins are linked (by GO annotation) to either cell communication or nucleotide metabolism (namely RNA processing and transcription). In addition, to our great surprise the majority (two-thirds) of CMGC kinases also form complexes with other kinases, revealing a larger kinase signalling informational 'highway'. Furthermore, ~8% (38/481) of the kinase interacting proteins are known cancer- linked proteins, further illustrating that these kinases and their regulation is extremely critical for cancer development. Our AP/MS -based protein complex analysis also allows us to detect possible post-translational modifications in the kinases and in their interacting proteins. For the first set of 57 kinases we could detect 407 unique phosphorylation sites, 104 of which are novel.
Project outcomes
As it was now obvious that we had established a state-of-the-art high-throughput AP/MS -based protein complex analysis workflow, we also wanted to make our knowledge available to the whole interactomics research community. To demonstrate the applicability and reproducibility of the workflow in different laboratories, we teamed up with a pioneer of the interactomics field, Professor Giulio Superti-Furga (Center for Molecular Medicine of the Austrian Academy of Sciences (CeMM), Vienna, Austria). The professor's research group obtained 32 kinase-expressing cell lines from us. In addition, they used our standard operating procedures to compare and evaluate the AP/MS -based protein complex analysis workflow. The comparison of the two analyses shows an extremely high overlap (>78 %) of identified interacting proteins. We have one manuscript ready to be submitted from this collaboration and another under preparation.
To study the dynamics of protein-protein interactions we decided to use a set of 25 receptor tyrosine kinases (RTK) as a model. Using an irreversible protein-tyrosine phosphatase inhibitor, pervanadate, we have been able to activate the RTKs and study the effect of the length of activation with the dynamics of the protein-protein interaction, using label-free quantitative AP/MS. The data from this part of the project is currently under analysis.
For the database development, we have now gathered the data obtained from the project and have integrated additional information from already established databases (COSMIC, HPRD, PhosphoSitePlus, Panther and PINA). This information and the corresponding database will be published simultaneously with the manuscripts. The manuscripts are entitled 'Neutrophil p38d-PKD1 signaling fine-tunes PTEN activity to control acute inflammation of the lung', 'Dissecting the CMGC kinase signaling network', 'Robustness in large-scale affinity purification-mass spectrometry interaction proteomics by an inter-laboratory cross-wise comparative analysis' and 'Modularity of the human Hippo pathway interaction proteome'.
Taken together, we have now studied more than 100 cancer-linked human protein kinases using state-of-the-art proteomics methods and have gained detailed insight on how these kinases can exert their influence on cancer development. This knowledge on the physical and functional signalling networks of the cancer kinases will be extremely valuable when trying to understand the subtle chances in cellular homeostasis that can drive the cells to become cancerogenic. Furthermore, the large amount of data gathered on the project will be highly valuable for the wider bioresearch community and will initiate, in addition to the collaborative projects started by us, its own a plethora of projects. In addition to the more global impact, the project helped the researcher to establish an independent research career. The researcher is currently acting as Group Leader and the Director of the Proteomics Facility at the Institute of Biotechnology, University of Helsinki, Finland.