Machine Learning Enables Live Label-Free Phenotypic Screening in Three Dimensions

Imaging-Based Machine Learning Analysis of Patient-Derived Tumor Organoid Drug Response

Three-quarters of compounds that enter clinical trials fail to make it to market due to safety or efficacy concerns. This statistic strongly suggests a need for better screening methods that result in improved translatability of compounds during the preclinical testing period. Patient-derived organoids have been touted as a promising 3D preclinical model system to impact the drug discovery pipeline, particularly in oncology. However, assessing drug efficacy in such models poses its own set of challenges, and traditional cell viability readouts fail to leverage some of the advantages that the organoid systems provide. Consequently, phenotypically evaluating complex 3D cell culture models remains difficult due to intra- and inter-patient organoid size differences, cellular heterogeneities, and temporal response dynamics. Here, we present an image-based high-content assay that provides object level information on 3D patient-derived tumor organoids without the need for vital dyes. Leveraging computer vision, we segment and define organoids as independent regions of interest and obtain morphometric and textural information per organoid. By acquiring brightfield images at different timepoints in a robust, non-destructive manner, we can track the dynamic response of individual organoids to various drugs. Furthermore, to simplify the analysis of the resulting large, complex data files, we developed a web-based data visualization tool, the Organoizer, that is available for public use. Our work demonstrates the feasibility and utility of using imaging, computer vision and machine learning to determine the vital status of individual patient-derived organoids without relying upon vital dyes, thus taking advantage of the characteristics offered by this preclinical model system.

The Trifecta of Single-Cell, Systems-Biology, and Machine-Learning Approaches

Together, single-cell technologies and systems biology have been used to investigate previously unanswerable questions in biomedicine with unparalleled detail. Despite these advances, gaps in analytical capacity remain. Machine learning, which has revolutionized biomedical imaging analysis, drug discovery, and systems biology, is an ideal strategy to fill these gaps in single-cell studies. Machine learning additionally has proven to be remarkably synergistic with single-cell data because it remedies unique challenges while capitalizing on the positive aspects of single-cell data. In this review, we describe how systems-biology algorithms have layered machine learning with biological components to provide systems level analyses of single-cell omics data, thus elucidating complex biological mechanisms. Accordingly, we highlight the trifecta of single-cell, systems-biology, and machine-learning approaches and illustrate how this trifecta can significantly contribute to five key areas of scientific research: cell trajectory and identity, individualized medicine, pharmacology, spatial omics, and multi-omics. Given its success to date, the systems-biology, single-cell omics, and machine-learning trifecta has proven to be a potent combination that will further advance biomedical research.

Automated 3D Microphysiometry Facilitates High-Content and Highly Reproducible Oxygen Measurements within 3D Cell Culture Models

Microphysiometry is a powerful technique to study metabolic parameters and detect changes to external stimuli. However, applying this technique for automated label-free and real-time measurements within cell-laden three-dimensional (3D) cell culture constructs remains a challenge. Herein, we present an entirely automated microphysiometry setup that combines needle-type microsensors with motorized sample and sensor positioning systems inside a standard tissue-culture incubator. The setup records dissolved oxygen as a metabolic parameter along the z-direction within cell-laden 3D constructs in a minimally invasive manner. The microphysiometry setup was applied to characterize the spatial oxygen distribution within thick cell-laden 3D constructs, study the time-dependent changes on the oxygen tension within 3D breast cancer models following a chemotherapeutic treatment, and identify kinetics and recovery effects after drug exposure over 5 weeks. Our data suggest that the microphysiometry setup enables highly reproducible measurements without human intervention, due to the high degree of automation and positional accuracy. The results demonstrate the applicability of the setup to provide valuable long-term insights into oxygenation within 3D models using minimally invasive, label-free, and entirely automated analysis methods.

Image-based high-content screening in drug discovery

While target-based drug discovery strategies rely on the precise knowledge of the identity and function of the drug targets, phenotypic drug discovery (PDD) approaches allow the identification of novel drugs based on knowledge of a distinct phenotype. Image-based high-content screening (HCS) is a potent PDD strategy that characterizes small-molecule effects through the quantification of features that depict cellular changes among or within cell populations, thereby generating valuable data sets for subsequent data analysis. However, these data can be complex, making image analysis from large HCS campaigns challenging. Technological advances in image acquisition, processing, and analysis as well as machine-learning (ML) approaches for the analysis of multidimensional data sets have rendered HCS as a viable technology for small-molecule drug discovery. Here, we discuss HCS concepts, current workflows as well as opportunities and challenges of image-based phenotypic screening and data analysis.

A High-Throughput Automated Confocal Microscopy Platform for Quantitative Phenotyping of Nanoparticle Uptake and Transport in Spheroids

There is a high demand for advanced, image-based, automated high-content screening (HCS) approaches to facilitate phenotypic screening in 3D cell culture models. A major challenge lies in retaining the resolution of fine cellular detail but at the same time imaging multicellular structures at a large scale. In this study, a confocal microscopy-based HCS platform in optical multiwell plates that enables the quantitative morphological profiling of populations of nonuniform spheroids obtained from HT-29 human colorectal cancer cells is described. This platform is then utilized to demonstrate a quantitative dissection of the penetration of synthetic nanoparticles (NP) in multicellular 3D spheroids at multiple levels of scale. A pilot RNA interference-based screening validates this methodology and identifies a subset of RAB GTPases that regulate NP trafficking in these spheroids. This technology is suitable for high-content phenotyping in 3D cell-based screening, providing a framework for nanomedicine drug development as applied to translational oncology.

Machine learning and image-based profiling in drug discovery

The increase in imaging throughput, new analytical frameworks and high-performance computational resources open new avenues for data-rich phenotypic profiling of small molecules in drug discovery. Image-based profiling assays assessing single-cell phenotypes have been used to explore mechanisms of action, target efficacy and toxicity of small molecules. Technological advances to generate large data sets together with new machine learning approaches for the analysis of high-dimensional profiling data create opportunities to improve many steps in drug discovery. In this review, we will discuss how recent studies applied machine learning approaches in functional profiling workflows with a focus on chemical genetics. While their utility in image-based screening and profiling is predictably evident, examples of novel insights beyond the status quo based on the applications of machine learning approaches are just beginning to emerge. To enable discoveries, future studies also need to develop methodologies that lower the entry barriers to high-throughput profiling experiments by streamlining image-based profiling assays and providing applications for advanced learning technologies such as easy to deploy deep neural networks.