We collected a unique set of paired, metastatic tumor specimens from 11 high-grade serous patients treated in @TyksVsshp (by @mijohy & colleagues) before and after neo-adjuvant chemotherapy, and analysed dissociated tumors with scRNA-seq. (2/9)
Unlike stroma or immune cells, cancer cells had distinctly patient-specific profiles. To find the hidden, shared states from these genetically heterogeneous tumor specimens, @KaiyangZhang@HautaniemiLab developed a new clustering method, PRIMUS (3/9)
PRIMUS simultaneously clusters and denoises data, and benchmarks well against batch removal tools, especially on unbalanced data. I know you want to test it on your data, so go ahead!! Available with tutorials: github.com/KaiyangZ/PRIMUS (4/9)
Combination of paired, clinical tumor specimens and new analysis approach revealed that a stress-related cancer cell state consistently increases during chemotherapy and is associated with inflammatory signaling with the surrounding stroma. (5/9)
We further show that the stress-associated state is subclonally enriched during chemotherapy as low-stress subclones fail to re-initiate proliferation after NACT. Increased levels of stress in treatment-naive tumors are associated with short chemoresponse.(6/9)
Our discovery can help to identify tumors that will likely have poor response to upcoming chemotherapy. Preventing chemo-induced inflammatory signaling between cancer cells and stroma could offer venues for effective combinatorial treatments. (7/9)
And if you have HGSS organoid & CAF co-cultures, we’d love to collaborate to test preventing chemo-induced inflammatory signaling between cancer cells and stroma!
• • •
Missing some Tweet in this thread? You can try to
force a refresh