(650c) Derivation of Transcription Factor Distribution Profiles From Green Fluorescent Protein Reporter Data

Mathematical models of signal transduction pathways are characterized by a large degree of uncertainty in the values of the model parameters as well as in the pathway structure. Availability of quantitative data can reduce these uncertainties, however, it is non-trivial to derive a significant amount of appropriate data. Transcription factor data are especially important as it is becoming evident that the dynamic behavior of transcription factors has a direct effect on the response of cells to stimuli (Kholodenko, 2006). While it is possible to monitor the dynamics of transcription factors using available techniques such as western blotting, electrophoretic mobility shift assay or chromatin immunoprecipitation (Elnitski et al., 2006), only qualitative or semi-quantitative information is obtained from these techniques. One alternative to these approaches is the one based upon green fluorescent protein (GFP) reporter systems presented by Huang et al. (2008) which derives transcription factor profiles by analyzing fluorescence microscopy images and solving an inverse problem. This inverse problem involves a model that describes changes in the fluorescence intensity due to transcription, translation, and post-translational activation of green fluorescent proteins in response to the presence of transcription factors in the nucleus. One drawback of this technique is that the approach is based upon the average fluorescence intensity over all cells and does not take into account the distribution of fluorescence intensity among a population of cells. However, inspection of fluorescence microscopy images shows that there is a clear heterogeneity in the fluorescence intensity exhibited by the cells. Information about phenotype heterogeneity among individual cells, i.e. the fluorescence intensity distribution, plays an important role for the dynamics of the underlying signal transduction pathways (Smits et al., 2005; Efroni et al., 2007). This phenotypic heterogeneity is due to the stochasticity of the gene expression but also because of stochastic variations in the concentrations of components of the signaling network (Raser & O'Shea, 2004).

These observations form the motivation behind this work as the goal is to derived transcription factor distribution profiles from experimental data. The presented work first develops an algorithm for determining boundaries of individual fluorescent cells from fluorescence microscopy images. The focus of this task is on dealing with cells that have non-regular shapes and images that have low contrast and a significant amount of noise. Once the boundaries of the cells have been determined, a fluorescence intensity distribution of the cells can be computed. This fluorescence intensity distribution is then used in another algorithm, also presented in this work, to compute the distribution of the transcription factor concentration.

The presented techniques are applied to experimental data for the TNF-α ~ NF-κB signaling pathway. The results show that the presented image analysis algorithm can correctly identify individual cells, the derived population balance model can appropriately describe the fluorescence intensity heterogeneity, and the determined distribution of the NF-κB concentration can accurately predict the fluorescence intensity distribution.

References

Efroni, S., Schaefer, C. F., & Buetow, K. H. (2007). Identification of key processes underlying cancer phenotypes using biologic pathway analysis. PLoS ONE, 2(5), e425.

Elnitski, L., Jin, V.X., Farnham, P.J., & Jones, S.J. (2006). Locating mammalian transcription factor binding sites: A survey of computational and experimental techniques. Genome Research, 16, 1455?1464.

Huang, Z., Senocak, F., Jayaraman, A., & Hahn, J. (2008). Integrated modeling and experimental approach for determining transcription factor profiles from fluorescent reporter data. BMC Systems Biology, 2:64.

Kholodenko, B.N (2006). Cell-signalling dynamics in time and space. Nature Reviews Molecular Cell Biology, 7, 165?176.

Raser, J. M., & O'Shea, E. K. (2004). Control of stochasticity in eukaryotic gene expression. Science, 304, 1181?1184.

Smits, K.W., Caroline, E.C., Kim, S.A., Sierd, B., Oscar, K.P., Leendert, H.W. (2005). Stripping bacillus: comK auto-stimulation is responsible for the bistable response in competence development. Molecular microbiology, 56(3), 604?614.