Stochastic and Deterministic Elements of a cell fate decision in the C. elegans gonad
Together the laboratory of Prof. Iva Greenwald at Columbia University, New York, we investigated a Notch-mediated cell-fate decision in C. elegans gonad in which two initially equivalent so-called alpha cells interact via LIN-12/Notch such that one becomes the anchor cell (AC) and one becomes the ventral uterine precursor (VU). The AC/VU decision is random, as 50% of animals have an AC derived from either one of the two cells, but despite intense research the origin and nature of the stochasticity remained unknown.
Time-lapse microfluidics imaging has allowed us, for the first time, to obtain large numbers of gonadal cell lineages and probe the relationship between cell-lineage timing and cell-fate specification (Attner & Keil, Curr Biol 2019). We found that the birth-order difference of precursors is predictive of the cell-fate, once divisions are longer than a certain threshold time apart. To gain insight into the molecular events impacting, and impacted by, birth order, we analyzed the temporal patterns of HLH-2 (the master transcription factor for the decision), as well as LIN-12/NOTCH and Lag-2/Delta protein expression using endogenous fluorescent tags obtained with CRISPR/Cas9. Long-Term Imaging revealed that precursor cells are truly uncommitted at birth and need to interact via LIN-12/Notch to resolve their fates. Notably however, the timing of events in their ancestor cells has essentially already determined their fates in an almost deterministic manner. Besides resolving a long-standing puzzle in the field, our work suggests that perhaps for other paradigms with a reproducible outcome, what has been considered ‘‘noise’’ may just appear so because a deterministic element has not been identified.
Cell divisions during C. elegans early gonadogenesis, imaged using microfluidics animal confinement and immobilization.
Long-term live imaging of developing C. elegans larvae using microfluidics
Live-imaging is a vital tool for dissecting developmental processes. Imaging at high spatiotemporal resolution is routinely carried out in embryos of many model organisms. But for subsequent larval development, animal locomotion is critical and, as a consequence, long-term imaging is notoriously challenging. We recently overcame this challenge for the first time in C. elegans through a novel microfluidics methodology.
This required a two-layer PDMS microfluidics chip with channels as narrow as 4.5um, integrated control of microscope and microfluidics and tailored automated image processing.
As a result, it is now possible to perform multi-channel 4D imaging of the C. elegans over its entire post-embryonic development (~3 days) in up to 10 animals simultaneously (Keil et al., Dev Cell 2017). This technology paves the way for many quantitative studies of development that we are currently undertaking.
Long-term high-resolution imaging of C. elegans larvae with microfluidics.
A C. elegans animal immobilized at various post-embryonic stages from L1 (upper left)
to L4 (lower right), showing dynamic expression of the miRNA mir-47. Scale bar, 100 um.
Cell-corpse clearance after non-apoptotic cell death
(with Shaham Lab at The Rockefeller University)
One of the first applications for which the imaging technology proved vital was the study of non-apoptotic cell death of the linker cell (LC) in the C. elegans male. The LC dies in the last larval stage of development and subsequently gets phagocytosed. We wanted to test whether the conserved genes that mediate apoptotic cell removal during development, also control LC removal. Using long-term live imaging, we were for the first time able to capture all morphogenetic events of LC death at high spatiotemporal resolution and follow protein localization dynamics along. This, together with genetic and biochemical analyses led us to identify a novel network of conserved small GTPases as key coordinator of LC phagocytosis and subsequent cell degradation. (Kutscher, Keil & Shaham, Dev Cell 2018).
Following linker cell death and degradation using live-imaging with microfluidics.
The dying linker cell is pseudo-colored in green (mig-24p::Venus), the engulfing U.l/rp cells are pseudo-colored in magenta (lin-48p::mKate2). Left shows merged Nomarksi and fluorescence micrographs, right shows a 3D rendering of the fluorescence channels.