Integration of Forces and Chemical Signals 

Morphogenesis is the basis of innumerable cell functions and thus among the best studied processes in cell biology. We know probably all component processes involved in morphogenesis and the majority of their molecular parts. Missing is a quantitative understanding of how these processes are coordinated. This is the challenge our lab tackles every day with an interdisciplinary approach that welds together molecular cell biology, live cell imaging, signal processing, computer vision, and mathematics. Accordingly, we work as a team with ~20 members, big enough to establish expertise in all these disciplines, yet small enough to synergize by spontaneous self-organization and without a hierarchical overhead.  

    The processes driving cell morphogenesis are organized in a cytomechanochemical system with three key properties: i) a high level of non-linearity, i.e. feed forward and feed back interaction, between processes; ii) a high level of redundancy between processes; and iii) a separation in space and time between causative and effected processes. These three properties complicate the analysis of the system in that perturbation of any component can lead to wide-ranging adaptation. Hence, the difference between phenotype and wildtype does not necessarily inform on the actual function of the perturbed target. Moreover, the tendency for adaptation leads to heterogeneous system outputs as small genetic, epigenetic, or environmental variations can significantly rebalance the relative importance of pathways. Accordingly, cytomechanochemical system are often investigated by strong stimulation to homogenize the response, with the caveat of activating certain component processes at the expense of others and thus of obscuring their interdependences under unperturbed conditions – not in our lab.

    Cytomechanochemical systems integration in cell morphogenesis. Cytomechanochemical systems integration in cell morphogenesis.

    Even though the limitations of molecular perturbations in dissecting nonlinear and redundant systems are acknowledged in principle, the community seems surprisingly resistant to moving away from the ‘break-and-watch’ paradigm. Many of the key controversies in our field are likely the product of variable compensation responses elicited by perturbations applied under slightly different experimental conditions. To break through this impasse we follow cytomechanochemical systems as they self-organize taking advantage of experimental heterogeneity as a source of information rather than a source of uncertainty and exploiting spontaneous fluctuations over time to track information flows between processes.

    There is precedence for this approach in other disciplines of science: the accuracy of weather forecasts is well above 90%. None of these predictions relies on experimental perturbation; and econometricians determine in split seconds the causal relations between financial markets. The last experiment a Wall Street investor would run to find out how the stock markets in Shanghai influences his/her fellow investors is blow up the index in Shanghai. Our work is inspired by these incredibly sophisticated approaches these sister disciplines pursue to build predictive models of systems with the same key features.

    A fundamental difference remains, however, between cell biological data and financial data. While financial market indicators are noise-free, cell biological measurements are noisy, quire often very noisy. Therefore, we are engaged in redesigning some of the mathematical methods we borrow from econometrics for the purpose of analyzing cell biological data. 

    This is placeholder text. If you are reading this, it is here by mistake and we would appreciate it if you could email us with a link to the page you found it on. This is placeholder text. If you are reading this, it is here by mistake and we would appreciate it if you could email us with a link to the page you found it on. 

    Reconstruction of intracellular forces. Reconstruction of intracellular forces. Boundary forces ‐ i.e. the tension "felt" by the growing actin network ‐ and contraction forces (top) and adhesion forces (bottom) in neighboring protruding and retracting sectors of the cell edge. These forces are mathematically derived from the strain rate fields in the observed actin network flow.

    Cancer is a genetically pluralistic disease. While the current rush to sequencing every possible cancer genome will undoubtedly reveal additional patterns of risk genes, it has become questionable whether this is the right path to potent therapy. On the other hand, it is striking how the genomic diversity of cancers converges on a few fairly stereotypic cell behaviors that are altered from normal behaviors. Our lab asks the question, is there a stereotypic program in cell morphogenesis that confers the progression of cancer, and especially metastasis, i.e. the spreading of primary tumor cells throughout other remote tissues? Obviously, cell morphogenesis is implicated in cell migration, which is a requirement for metastatic spreading. But is migration the stereotypic program that makes a metastatic cell a metastatic cell? Other cell functions essential for metastatic progression may also be linked to a stereotypic shift in cell morphogenesis. Currently, we look for links between morphogenesis and cell survival, metabolism, and even drug resistance; and we have begun to test whether we could renew the pathologist’s perspective of morphology as a prognostic marker with radically advanced measures of cell morphogenesis to complement the heterogeneous single cell genomic and molecular profiles of cancers with a stereotypical, functionally inspired marker. In some sense our search for stereotypical functional endpoints follows from Waddington’s almost a century-old canalization paradigm: ‘developmental reactions, as they occur in organisms submitted to natural selection...are adjusted so as to bring about one definite end-result regardless of minor variations in conditions during the course of the reaction’ (Waddington, C. H. Canalization of development and the inheritance of acquired characters. Nature 150, 563-565.1942). 

    Imaging and quantifying cell morphology in 3D

    Imaging and quantifying cell morphology in 3D. Left column: Maximum intensity projection of oncogenically transformed human bronchial epithelial cells imaged at fully isotropic resolution of 0.3 x 0.3 x 0.3 um by a home-built 2-photon Bessel beam light sheet illumination microscope. Cells express mEmerald-tractin labeling the actin cortex and crawl in a collagen IV gel. Middle column: Rendering of the computationally reconstructed cell surface. Note that especially in cell A the segmentation works robustly for high and low contrast (dense and scarce actin cortex) regions – a feature of advanced computational image segmentation. Right column: Color coded surface curvature as an example of numerous geometric parameters describing cell morphology. Cell A stems from a clonal population with low expression of the oncogenic mutation KrasV12; Cell B from a clonal population with high expression of KrasV12. These studies identified expression levels the constitutively active KrasV12 signal as a cell morphogenetic switch in lung cancer.