The paper in Nature Biomedical Engineering is here: https://go.nature.com/2H037FJ
We are living in a world connected through social networks such as Twitter, Facebook and Instagram. People instantly influence others around the world by sharing their thoughts (in other words, signals). The recent global ‘MeToo’ movement initiated by Alyssa Milano’s tweeting shows the impact on the entire network from changes triggered by a single node. Interestingly, each cell, the unit of living organisms, is also regulated by an interaction network of proteins. Through the physical interactions of two or more proteins, individual cells propagate signals from the plasma membrane into the nucleus, which controls and determines the fate of the cells. An unexpected change in a single protein can influence the entire protein-protein interaction (PPI) network, resulting in deregulation in cellular decisions and triggering diseases such as cancer.
Genetic mutations are perhaps the most notable and irreversible changes that occur in individual nodes of a PPI network. Yet, it remains a daunting task to predict the way a given genetic or epigenetic modification affects the complex PPI network. There are many possible questions: do genetic mutations create novel links or strengthen certain links in the PPI network? If so, how do these aberrant PPIs relate to the fate decision of cells? The most straightforward way to answer these questions would be to examine the rewiring of the PPI networks for as many cancers as possible. Our paper just published in Nature Biomedical Engineering reports the development of a technical platform that may finally make PPI profiling possible for individual patients with cancer.
Several research groups, including ours, are developing a collection of techniques called single-molecule pull-down and co-immunoprecipitation (co-IP), which allows for the study of cellular protein complexes without prior purification steps1-4. We noted that this new single-molecule toolbox would be an ideal platform for studying PPI networks for individual cancer patients. Unfortunately, the throughput of single-molecule co-IP was still too low to examine the huge and complex PPI networks. The reaction chamber, typically used in the single-molecule imaging field, permits only four or five reactions per assembled microscope slide (Fig. 1a, top)1,2. To increase the number of reactions accommodated in a single microscope slide, I designed a small-well system with the help of Sangwoo Park and Hyunwoo Kim, who are co-authors of our new paper. We also tested many materials for the well walls that include house-made polydimethylsiloxane and silicon gaskets. We finally came up with a single-molecule imaging chip with 40 reaction chambers and minimal crosstalks and non-specific protein adsorptions (Fig. 1a, bottom).
Fig. 1 High-throughput single-molecule imaging system. (a) Photos showing the original 4-channel imaging slide (top) and the high-throughput imaging chambers containing 40 reaction chambers (bottom). (b) Schematic of automated single-molecule imaging system. Figure is reproduced from Fig. 1b of our paper. Macmillan Publishers.
Previously, each step of the single-molecule co-IP imaging was manually performed and thus required years of training for a successful experiment. Ji Young Ryu and Minju Shon, also co-authors on the paper, developed a highly automated single-molecule imaging system by integrating robotic control of the microscope stage, auto-focusing capability and image analysis for detection of single PPI complexes. When combined with our developed imaging chip, even untrained researchers can measure 100 different types of PPIs within one hour after minimal education.
Armed with technical developments, we sought to examine the changes in the PPI networks of epidermal growth factor receptor (EGFR) arising from genetic mutations in the EGFR gene. It is generally believed that the activating mutations in the EGFR kinase domain leads to hyperactivated EGFR signalling pathway. However, we found that individual tumours with the same EGFR mutation showed quite different levels of EGFR PPIs for important downstream interactors. In addition, these different levels of PPIs highly correlated with the response of individual tumours to an EGFR-targeted therapy (Fig. 2a). We found that the converse case was also true. Tumours without any activating EGFR mutations (i.e., carrying the wild-type EGFR gene) showed different levels of EGFR PPIs, which again showed a high level of correlation with the tumour responses to an EGFR-directed therapy (Fig. 2b). Our results suggest the possibility that single-molecule co-IP can be used to predict which cancer patients will most benefit from a given targeted cancer therapy by analysing their PPI networks.
Fig. 2 Correlation of EGFR PPI to the response of EGFR-targeted therapy. Correlation of EGFR-PPI level with the responses to an EGFR-targeted therapy from lung adenocarcinoma cell lines having activating EGFR mutations (a) and patient-derived tumor xenografts without any EGFR mutations (b). Figure is reproduced from Fig. 4 and 5 of our paper. Macmillan Publishers.
There is also a serendipitous finding reported in our paper. I routinely use a cocktail of tyrosine phosphatase inhibitors when preparing cell lysates to preserve any phosphotyrosine residues on the EGFRs. In the textbook model, the members of the EGFR family use tyrosine phosphorylation to elicit their signalling cascade, which is one of the most important discoveries in the study of growth factor receptors. One day, I left out the tyrosine phosphate inhibitors in the cell lysis buffer by mistake. Surprisingly, I found that the mutant EGFRs still showed strong PPIs with several interaction partners, whereas the wild-type EGFRs completely lost their PPIs. The reviewers were intrigued by our observation and gave many helpful comments, including studying association of the mutant EGFRs with heat shock protein 90 (HSP90). We found out that the mutant EGFRs build unique protein complexes around themselves, which in turn attract downstream signalling proteins in a largely phosphotyrosine-independent manner. These observations suggest a novel mechanism for how EGFR mutations lead to oncogene addiction to EGFR signalling.
Tae-Young and I recently started a start-up company named Proteina where I am continuing my efforts to validate the power of the single-molecule co-IP technique for different cohorts of cancer patients treated with various targeted therapies. We believe that the single-molecule co-IP can be a new arsenal for providing functional proteomics data of individual patients in this exciting era of precision medicine.
Our paper: Lee, H. W. et al. Profiling of protein–protein interactions via single-molecule techniques predicts the dependence of cancers on growth-factor receptors. Nat. Biomed. Eng. doi:10.1038/s41551-018-0212-3 (2018).
1. Lee, H. W. et al. Real-time single-molecule co-immunoprecipitation analyses reveal cancer-specific Ras signalling dynamics. Nat. Commun. 4, 1505, doi:10.1038/ncomms2507 (2013).
2. Lee, H. W. et al. Real-time single-molecule coimmunoprecipitation of weak protein-protein interactions. Nat. Protoc. 8, 2045-2060, doi:10.1038/nprot.2013.116 (2013).
3. Yeom, K. H. et al. Single-molecule approach to immunoprecipitated protein complexes: insights into miRNA uridylation. EMBO Rep. 12, 690-696, doi:10.1038/embor.2011.100 (2011).
4. Jain, A. et al. Probing cellular protein complexes using single-molecule pull-down. Nature 473, 484-488, doi:10.1038/nature10016 (2011).