Johns Hopkins University (complete)

E-cadherin Promotes Metastasis in Invasive Ductal Breast Cancer

The CTD² scientists at Johns Hopkins University showed that loss of E-cadherin increased invasion and dissemination and reduced metastatic potential in invasive ductal carcinoma models, suggesting context-dependent roles for the protein.

Experimental Approaches

Genetic studies were conducted in the MMTV-PyMT and C3(1)-Tag invasive ductal carcinoma (IDC) mouse models, as well as a TNBC PDX model. The molecular basis of reduced metastasis in E-cad^- tumors were investigated by comparing the transcriptomes (RNA-seq data) of adeno-Cre-treated MMTV-PyMT; Cdh1+/+ and MMTV-PyMT; Cdh1fl/fl organoids. The next-generation RNA-seq data is available from NCBI Gene Expression Omnibus with accession number GSE114011. Differential gene expression analysis data and a semi-custom code for RNA-seq analysis is made available through GitHub.

Organoid Spectral Invasion and Population Genetics

CTD² scientists at the Johns Hopkins University adapted population genetics methods that compare the most extreme siblings in a family to study invasiveness of organoids generated from the same patient. This method could be used for better power to detect the molecular mechanisms of breast cancer metastasis.

Experimental Approaches

Tumor invasiveness and metastasis risk vary between individuals, and even within a single tumor vary between different regions and cells. Although variation within a tumor introduces clinical challenges, it can also enable within-tumor statistical tests that are powerful because they subtract each patient’s individual background. Organoids are used as a model system to probe variation in invasiveness between tumors and within tumors. Invasiveness itself is quantified through spectral analysis of the organoid boundary.

A variance components model estimates that between-tumor variation contributes 28% of organoid-to-organoid variation in invasiveness, and within-tumor variation contributes the remaining 72%. Thus, individual organoids are close to independent observations, and, for example, a study of 100 tumors each generating 1000 organoids may have similar power to a bulk study of 10,000 to 100,000 individuals. All data and methods are available at GitHub.