Segmentation of Image Data from Complex Organotypic 3D Models of Cancer Tissues with Markov Random Fields

Organoid intelligence: Integration of organoid technology and artificial intelligence in the new era of in vitro models

Organoid Intelligence ushers in a new era by seamlessly integrating cutting-edge organoid technology with the power of artificial intelligence. Organoids, three-dimensional miniature organ-like structures cultivated from stem cells, offer an unparalleled opportunity to simulate complex human organ systems in vitro. Through the convergence of organoid technology and AI, researchers gain the means to accelerate discoveries and insights across various disciplines. Artificial intelligence algorithms enable the comprehensive analysis of intricate organoid behaviors, intricate cellular interactions, and dynamic responses to stimuli. This synergy empowers the development of predictive models, precise disease simulations, and personalized medicine approaches, revolutionizing our understanding of human development, disease mechanisms, and therapeutic interventions. Organoid Intelligence holds the promise of reshaping how we perceive in vitro modeling, propelling us toward a future where these advanced systems play a pivotal role in biomedical research and drug development.

Spatiotemporal Tracking of Different Cell Populations in Cancer Organoid Models for Investigations on Photodynamic Therapy

Three-dimensional (3D) in vitro models of tumors are gaining interest as versatile platforms for treatment screening. In this context, heterocellular cultures in which various cell types are co-cultured are being explored to investigate whether partner cells can influence the treatment efficacies. However, when the cells are co-cultured, it is challenging to distinguish them and it becomes impossible to identify whether the treatment affects each cell line in a similar way or if there is a certain selectivity. Here, we propose a protocol in which different cell types are pre-labeled with fluorescent reporters prior to 3D culture initiation. Subsequently, the internal architecture of the 3D cancer models can be longitudinally monitored for model characterization, and to potentially detect architectural and treatment selectivity in response to therapy. This protocol employs quantum dots as non-photobleaching dyes and two-photon excited microscopy as a widely accessible imaging modality. In combination with an appropriate image analysis workflow, this protocol will help to investigate the architectural development of heterotypic microtumor/spheroid/organoid models and possibly identify treatment efficacies on individual cell populations represented within the models.

Machine learning-aided quantification of antibody-based cancer immunotherapy by natural killer cells in microfluidic droplets

Natural killer (NK) cells have emerged as an effective alternative option to T cell-based immunotherapies, particularly against liquid (hematologic) tumors. However, the effectiveness of NK cell therapy has been less than optimal for solid tumors, partly due to the heterogeneity in target interaction leading to variable anti-tumor cytotoxicity. This paper describes a microfluidic droplet-based cytotoxicity assay for quantitative comparison of immunotherapeutic NK-92 cell interaction with various types of target cells. Machine learning algorithms were developed to assess the dynamics of individual effector-target cell pair conjugation and target death in droplets in a semi-automated manner. Our results showed that while short contacts were sufficient to induce potent killing of hematological cancer cells, long-lasting stable conjugation with NK-92 cells was unable to kill HER2+ solid tumor cells (SKOV3, SKBR3) significantly. NK-92 cells that were engineered to express FcγRIII (CD16) mediated antibody-dependent cellular cytotoxicity (ADCC) selectively against HER2+ cells upon addition of Herceptin (trastuzumab). The requirement of CD16, Herceptin and specific pre-incubation temperature served as three inputs to generate a molecular logic function with HER2+ cell death as the output. Mass proteomic analysis of the two effector cell lines suggested differential changes in adhesion, exocytosis, metabolism, transport and activation of upstream regulators and cytotoxicity mediators, which can be utilized to regulate specific functionalities of NK-92 cells in future. These results suggest that this semi-automated single cell assay can reveal the variability and functional potency of NK cells and may be used to optimize immunotherapeutic efficacy for preclinical analyses.

Software tools for 3D nuclei segmentation and quantitative analysis in multicellular aggregates

Today, we are fully immersed into the era of 3D biology. It has been extensively demonstrated that 3D models: (a) better mimic the physiology of human tissues; (b) can effectively replace animal models; (c) often provide more reliable results than 2D ones. Accordingly, anti-cancer drug screenings and toxicology studies based on multicellular 3D biological models, the so-called "-oids" (e.g. spheroids, tumoroids, organoids), are blooming in the literature. However, the complex nature of these systems limit the manual quantitative analyses of single cells' behaviour in the culture. Accordingly, the demand for advanced software tools that are able to perform phenotypic analysis is fundamental. In this work, we describe the freely accessible tools that are currently available for biologists and researchers interested in analysing the effects of drugs/treatments on 3D multicellular -oids at a single-cell resolution level. In addition, using publicly available nuclear stained datasets we quantitatively compare the segmentation performance of 9 specific tools.

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.

Incorporating interaction networks into the determination of functionally related hit genes in genomic experiments with Markov random fields

Motivation

Incorporating gene interaction data into the identification of ‘hit’ genes in genomic experiments is a well-established approach leveraging the ‘guilt by association’ assumption to obtain a network based hit list of functionally related genes. We aim to develop a method to allow for multivariate gene scores and multiple hit labels in order to extend the analysis of genomic screening data within such an approach.

Results

We propose a Markov random field-based method to achieve our aim and show that the particular advantages of our method compared with those currently used lead to new insights in previously analysed data as well as for our own motivating data. Our method additionally achieves the best performance in an independent simulation experiment. The real data applications we consider comprise of a survival analysis and differential expression experiment and a cell-based RNA interference functional screen.

Availability and implementation

We provide all of the data and code related to the results in the paper.

Supplementary information

Supplementary data are available at Bioinformatics online.

Comprehensive high-throughput image analysis for therapeutic efficacy of architecturally complex heterotypic organoids

Bioengineered three-dimensional (3D) tumor models that incorporate heterotypic cellular communication are gaining interest as they can recapitulate key features regarding the intrinsic heterogeneity of cancer tissues. However, the architectural complexity and heterogeneous contents associated with these models pose a challenge for toxicological assays to accurately report treatment outcomes. To address this issue, we describe a comprehensive image analysis procedure for structurally complex organotypic cultures (CALYPSO) applied to fluorescence-based assays to extract multiparametric readouts of treatment effects for heterotypic tumor cultures that enables advanced analyses. The capacity of this approach is exemplified on various 3D models including adherent/suspension, mono-/heterocellular cultures and several disease types. The subsequent analysis revealed specific morphological effects of oxaliplatin chemotherapy, radiotherapy, and photodynamic therapy. The procedure can be readily implemented in most laboratories to facilitate high-throughput toxicological screening of pharmaceutical agents and treatment regimens on organotypic cultures of human disease to expedite drug and therapy development.

A high-content image analysis approach for quantitative measurements of chemosensitivity in patient-derived tumor microtissues

Organotypic, three-dimensional (3D) cancer models have enabled investigations of complex microtissues in increasingly realistic conditions. However, a drawback of these advanced models remains the poor biological relevance of cancer cell lines, while higher clinical significance would be obtainable with patient-derived cell cultures. Here, we describe the generation and data analysis of 3D microtissue models from patient-derived xenografts (PDX) of non-small cell lung carcinoma (NSCLC). Standard of care anti-cancer drugs were applied and the altered multicellular morphologies were captured by confocal microscopy, followed by automated image analyses to quantitatively measure phenotypic features for high-content chemosensitivity tests. The obtained image data were thresholded using a local entropy filter after which the image foreground was split into local regions, for a supervised classification into tumor or fibroblast cell types. Robust statistical methods were applied to evaluate treatment effects on growth and morphology. Both novel and existing computational approaches were compared at each step, while prioritizing high experimental throughput. Docetaxel was found to be the most effective drug that blocked both tumor growth and invasion. These effects were also validated in PDX tumors in vivo. Our research opens new avenues for high-content drug screening based on patient-derived cell cultures, and for personalized chemosensitivity testing.

Development of a 3D Tissue Culture-Based High-Content Screening Platform That Uses Phenotypic Profiling to Discriminate Selective Inhibitors of Receptor Tyrosine Kinases

3D tissue cultures provide a more physiologically relevant context for the screening of compounds, compared with 2D cell cultures. Cells cultured in 3D hydrogels also show complex phenotypes, increasing the scope for phenotypic profiling. Here we describe a high-content screening platform that uses invasive human prostate cancer cells cultured in 3D in standard 384-well assay plates to study the activity of potential therapeutic small molecules and antibody biologics. Image analysis tools were developed to process 3D image data to measure over 800 phenotypic parameters. Multiparametric analysis was used to evaluate the effect of compounds on tissue morphology. We applied this screening platform to measure the activity and selectivity of inhibitors of the c-Met and epidermal growth factor (EGF) receptor (EGFR) tyrosine kinases in 3D cultured prostate carcinoma cells. c-Met and EGFR activity was quantified based on the phenotypic profiles induced by their respective ligands, hepatocyte growth factor and EGF. The screening method was applied to a novel collection of 80 putative inhibitors of c-Met and EGFR. Compounds were identified that induced phenotypic profiles indicative of selective inhibition of c-Met, EGFR, or bispecific inhibition of both targets. In conclusion, we describe a fully scalable high-content screening platform that uses phenotypic profiling to discriminate selective and nonselective (off-target) inhibitors in a physiologically relevant 3D cell culture setting.