Wozelka-Oltjan, L. | bioRxiv Science

bioRxiv2024

Aggressive KRAS mutations direct TGF-β response towards partial EMT in patient-derived colorectal cancer tumoroids

Transforming growth factor beta (TGF-{beta}) exhibits complex and context-dependent cellular responses. While it mostly induces tumor-suppressive effects in early stages of tumorigenesis, its tumor promoting properties are evident in advanced disease. This TGF-{beta} duality is still not fully understood, and whether TGF-{beta} supports invasion and metastasis by influencing cancer cells directly, or rather through the stromal tumor compartment remains a matter of debate. Here, we utilized a library of colorectal cancer (CRC) patient-derived tumoroids (PDTs), representing a spectrum of tumor stages, to study cancer cell-specific responses to TGF-{beta}. Using medium conditions allowing for the differentiation of PDTs, we observed TGF-{beta} induced tumor-suppressive effects in early-stage tumoroids. PDTs with TGF-{beta} pathway mutations or PDTs derived from metastatic tumors were insensitive to the treatment. Notably, one tumoroid line harboring an atypical KRASQ22K mutation underwent partial epithelial-to-mesenchymal transition (EMT), associated with morphological changes and increased invasiveness. On a molecular level, this was accompanied by elevated expression of mesenchymal genes, as well as deregulation of pathways associated with matrix remodeling and cell adhesion. Our results suggest that tumor cell intrinsic responses to TGF-{beta} are critical in determining its tumor-suppressive or -promoting effects.

cancer biology↗

bioRxiv2022

A deep learning approach for improved detection of homologous recombination deficiency from shallow genomic profiles

Homologous Recombination Deficiency (HRD) is a predictive biomarker of poly-ADP ribose polymerase 1 inhibitors (PARPi) response. Most HRD detection methods are based on genome wide enumeration of scarring events and require deep genome sequence profiles (> 30x). The cost and workflow-specific biases introduced by these genome profiling methods currently limits clinical adoption of HRD testing. We introduce the Genomic Integrity Index (GII), a Convolutional Neuronal Network, that leverages features from low pass (1x) Whole Genome Sequencing data to distinguish HRD positive and negative samples. In a cohort of 230 ovarian and breast cancer, we found GII supports accurate stratification of samples yielding results that are highly concordant with state-of-the-art HRD detection methods (0.865<AUC<0.996) which require 50x deeper coverage. We conclude that the deep learning framework supporting GII allows accurate detection of HRD from shallow genome profiles, reducing biases and data generation costs making it uniquely suited for clinical applications.

cancer biology↗