(1302-C) High content image analysis of human cortical adherent organoids: 3D MICro-brains
Tuesday, February 6, 2024
12:00 PM – 1:00 PM EST
Location: Exhibit Halls AB
Abstract: Scalable high-throughput screening approaches for human brain organoids remain a challenge for disease modeling and drug discovery. Using induced pluripotent stem cells (iPSC), we have established a robust 3D miniaturized format of the cerebral cortex neurodevelopmental architecture. These 3D MICro-brains demonstrate layered radial organization in functional neuronal networks containing all major brain cell types: neurons, astrocytes, myelin-producing oligodendrocytes and microglia. One of the major objectives of our consortium is to further establish 3D MICro-brains as a standardized model that is amenable to high content (HC) screening for the assessment of disease phenotypes and therapeutic agents in human neurodevelopment.
Adherent cortical organoids were fixed and stained with various markers and HC image acquisition was performed in z-stacks. Volumetric segmentation was performed by using the IBA1 (Ionized calcium binding adaptor molecule 1) channel to segment microglia and the DAPI channel to segment whole organoids and cell nuclei. Using automated image analysis we extracted 100+ features from each of the three objects detected: organoids, microglia and nuclei. These features included intensity, morphology, volume and topography scores. Image quantification data was subsequently analyzed using a defined and reproducible data analytics workflow. The numeric dataset was automatically scanned for feature redundancy based on: binary vs. continuous data type, variation, amount of missing values and Spearman's rho correlation coefficient. Heavily skewed variables were detected and transformed. All data were scaled using a robust Z-score to account for variation in feature-to-feature scales. Due to the large number of features, principal component analysis was performed to reduce complexity. This revealed that 3D-related features had a high contribution in variance within our dataset, which indicates the advantage of performing volumetric image analysis in HC screening.
Our analysis pipeline was able to robustly detect differences in organoid morphology when comparing organoids with and without microglia. Similarly, preliminary results indicate a change in microglia morphology when organoids were treated with CSF1Ri (Colony stimulating factor 1 receptor inhibitor) and LPS (Lipopolysaccharide). We aim to continue to quantify the effects of different treatments on microglia phenotype, cell proliferation and differentiation.
Our results demonstrate the feasibility and reproducibility of using the 3D MICro-brain platform for high throughput screening to support the discovery of therapeutics for a wide array of neurological diseases. They also suggest that our overall goal of a scalable animal free in vitro model is practical and achievable to deploy in a modern drug discovery environment.