That's all for now, folks – see here for a link to the paper itself! (fin/n) www.cell.com/cell/fulltex...

Much appreciation & acknowledgement to all involved - co-first authors @drshay.bsky.social Marc Sherman and co-senior authors Wolfram Goessling Ömer Yilmaz @shaleklab.bsky.social and many more who enabled this work @broadinstitute.org @mitkochinstitute.bsky.social @ragoninstitute.bsky.social

Much more in the paper and discussion – links to epigenetic priming and tissue memory, generalizing across insults like infection and alcohol, decoupling inter-organ crosstalk, and beyond! (12/n)

Our work supports a model where environmental stress -> cell-vs-tissue rebalancing (driven by ↑SOX4 and ↓HMGCS2) -> multicellular cell-intrinsic and cell-extrinsic circuits -> converge back on hepatocytes to reinforce stress responses and prime tumorigenesis (11/n)

Through geographic regression, we discovered spatially-structured multicellular hubs and signaling circuits, where cell-vs-tissue rebalancing and stress responses depended on the composition of neighboring cells and nearby microenvironmental signals (10/n)

Finally, we studied how communication across the liver’s multicellular communities shape stress responses and activity balancing, by constructing a human liver tissue microarray and conducting spatial transcriptomics (9/n)

Validating MATCHA predictions in vitro and in vivo, we showed that SOX4 was sufficient to drive cell-vs-tissue gene expression changes, increase cells’ proliferative capacity in metabolically stressful environments, and even predict human cancer risk years in the future (8/n)

We wanted to know what drives this rebalancing. So we created the new computational method MATCHA to uncover “central hub” transcription factors that co-regulate multiple gene programs and cellular activities, all at once. Available on Github at github.com/ctzouanas/MATCHA (7/n)

Beyond HMGCS2, we investigated the generalizability of stress-driven cell-vs-tissue rebalancing. We found robust dysregulation after tumorigenesis, recapitulation across multiple human cohorts spanning 100’s of patients, and stratification of human cancer survival (6/n)

So, changes in what metabolic pathways a cell can use are sufficient to regulate what genes a cell expresses, how the cell contributes to tissue (dys)function, how prone it is to becoming cancerous, and how aggressively tumors will grow after cancer starts (5/n)

Testing the enzyme HMGCS2 (rate-limiting for ketogenesis, ↓ with long-term metabolic stress), we found that knocking out this enzyme was sufficient to drive this cell-vs-tissue gene expression rebalancing and even prime tumorigenesis months later once oncogenes show up (4/n)

We found long-term metabolic stress drives cells to rebalance activities, prioritizing what benefits individual cells (pro-survival, developmental pathways) over service to collective tissue function (mature lineage identity, metabolic processing, protein secretion) (3/n)

Much prior work focused on what causes cells to die under stress, but we were interested in cells that survive. How do cells navigate challenges from their environments, how do cellular choices shape tissue-level function and disease outcomes, and what are core mechanisms? (2/n)