Making Sense of the Noise: How Cells Convert Stimulus to Action
Author: Sarah Yunes
Editor: Rajiv Shrestha
To perform their biological function, cells constantly take in multiple diverse stimuli and, by activating signaling pathways, determine what the outcome will be. Sometimes, despite all of the cells being capable of responding to the stimulus, there can be a lot of heterogeneity in how the cells respond. For example, tumor necrosis factor (TNF) stimulus can result in different outcomes while acting through the pro-inflammatory transcription factor family NF-κB. To better understand how cells determine how to react to this signaling pathway on a single cell level, Dr. Suzanne Gaudet spoke at the Boston QSP’s July event. Dr. Gaudet is an Associate Professor at the Department of Cancer Biology at Dana-Farber Cancer Institute and at the Department of Genetics at the Blavatnik Institute at Harvard Medical School. Her talk was titled “Transcriptional dynamics as a driver of phenotypic diversity".
Dr Gaudet (left) and Scientific Writer Sarah Yunes (right) during the interview
NF-κB is a family of transcription regulators that contain three transcriptional activators (RelA, RelB, and c-Rel) and two repressors (p50 and p52). NF-κB proteins generally form dimers with other proteins from the family. TNF activates NF-κB signaling by binding its receptor and starting a signaling cascade, leading to the degradation of the NF-κB inhibitor IκBα. When IκBα is degraded, NF-κB dimers are able to move into the nucleus and activate transcription. To measure NF-κB activity, the movement into the nucleus of a fluorescently tagged RelA was monitored. Even when comparing cells in the same cell line, there is a fair amount of variability in how much RelA had relocated to the nucleus at a given point in time after stimulus. As a result, there is a substantial overlap in the amount of RelA in the nucleus in the untreated and the TNF-treated cells, even when measured over time. This is in contrast to the transcriptional response in untreated and TNF-treated cells, which had very little overlap, indicating very switch-like regulation.
The next question is how does this variability in RelA nuclear translocation translate into the transcriptional response of NF-κB signaling. The model developed by the Gaudet lab using single cell data showed that the maximum fold change in RelA nuclear localization best matches the transcriptional response for the three studied genes. Surprisingly, this model best matches to an incoherent feedforward loop, where NF-κB increases the expression of both its own inhibitor and a competitor for the NF-κB binding sites. The model can be tuned to better match the biological response by changing only the affinity of the competitor for NF-κB binding sites. The model was able to match expression profiles for five genes in a different cell line. Further tuning the model to account for differences in the synthesis and degradation rates of different transcripts led to an even better fit to the data. An incoherent feed forward loop allows a level of robustness in response to variable start conditions, like initial nuclear levels of RelA.
One example where this variability is directly relevant to disease response is the reactivation of the human immunodeficiency virus (HIV). HIV can lay dormant in T cells for years before reactivating and leading to acquired immunodeficiency syndrome (AIDS), a serious illness characterized by a weakened immune system, which leads to an increase in tumors and opportunistic infections. Some people develop AIDS faster than others and what causes this is still unclear. Reactivation of HIV can be caused by NF-κB-dependent pro-inflammatory activity. However, even in cell clones with the same NF-κB binding site in the HIV promoter, there can be a different rate of HIV reactivation in response to the same stimulus. Even in this case, the fold change of RelA in the nucleus largely determines the amount of HIV transcripts produced, but different clones have different fold changes in RelA. These differences are caused by the HIV genome integrating into different parts of the chromosomes, some of which is restrictive to gene expression. This detailed model of cell-to-cell heterogeneity in cell response to NF-κB signaling is invaluable to learning more about how the cellular response to TNF can lead to diverse outcomes in different cell backgrounds and disease states.
Dr. Gaudet was always drawn to the more quantitative aspects of biology, expressing an interest in biostatistics as an undergraduate student. After finishing her PhD in biochemistry, her postdoc saw her interfacing with applied mathematicians, engineers, and other more quantitatively-focused people in a large, interdisciplinary lab group. Over time, she became much more involved in the modeling aspects, eventually creating her own models, thanks to peers willing to teach and explain modeling. Dr. Gaudet and her lab uses models as an explanation tool. She says, “when you observe phenomena and they’re hard to explain, but the model allows you to test an idea about a mechanism”, it provides insights into how the processes work. As to the relative absence of modeling more broadly in biology research, particularly in the academic sphere, Dr. Gaudet brought up the common theme of students turning to life sciences to avoid the more quantitative side of science. With the rapidly expanding fields of genomics, transcriptomics, proteomics, and other omics, biostatistics and bioinformatics are becoming even more important, leading biology to be better connected to mathematics and modeling fields. This could act to expand the presence of modeling in the field. When she’s not pursuing her research, Dr. Gaudet enjoys dancing Lindy hop, a type of swing dancing.
Dr. Gaudet’s talk was followed by a mixer event where community members discussed and socialized over food and drinks. Here are some moments from the event captured by our photographer. Enjoy!
We would like to thank Applied BioMath for sponsoring the event and the CIC for sponsoring the venue.
Please subscribe to our website to get updates on job openings in QSP and related fields, blogs, and upcoming events. Our next event, part of our Modeling and Simulation in Drug Development event series, will be on September 18. Please check here for more details.