Artificial intelligence for high content imaging in drug discovery.

Artificial intelligence (AI) and high-content imaging (HCI) are contributing to advancements in drug discovery, propelled by the recent progress in deep neural networks. This review highlights AI's role in analysis of HCI data from fixed and live-cell imaging, enabling novel label-free and multi-channel fluorescent screening methods, and improving compound profiling. HCI experiments are rapid and cost-effective, facilitating large data set accumulation for AI model training. However, the success of AI in drug discovery also depends on high-quality data, reproducible experiments, and robust validation to ensure model performance. Despite challenges like the need for annotated compounds and managing vast image data, AI's potential in phenotypic screening and drug profiling is significant. Future improvements in AI, including increased interpretability and integration of multiple modalities, are expected to solidify AI and HCI's role in drug discovery.

A Decade in a Systematic Review: The Evolution and Impact of Cell Painting.

High-content image-based assays have fueled significant discoveries in the life sciences in the past decade (2013-2023), including novel insights into disease etiology, mechanism of action, new therapeutics, and toxicology predictions. Here, we systematically review the substantial methodological advancements and applications of Cell Painting. Advancements include improvements in the Cell Painting protocol, assay adaptations for different types of perturbations and applications, and improved methodologies for feature extraction, quality control, and batch effect correction. Moreover, machine learning methods recently surpassed classical approaches in their ability to extract biologically useful information from Cell Painting images. Cell Painting data have been used alone or in combination with other - omics data to decipher the mechanism of action of a compound, its toxicity profile, and many other biological effects. Overall, key methodological advances have expanded Cell Painting's ability to capture cellular responses to various perturbations. Future advances will likely lie in advancing computational and experimental techniques, developing new publicly available datasets, and integrating them with other high-content data types.

The anti-leprosy drug clofazimine reduces polyQ toxicity through activation of PPARγ.

BACKGROUND: PolyQ diseases are autosomal dominant neurodegenerative disorders caused by the expansion of CAG repeats. While of slow progression, these diseases are ultimately fatal and lack effective therapies. METHODS: A high-throughput chemical screen was conducted to identify drugs that lower the toxicity of a protein containing the first exon of Huntington's disease (HD) protein huntingtin (HTT) harbouring 94 glutamines (Htt-Q94). Candidate drugs were tested in a wide range of in vitro and in vivo models of polyQ toxicity. FINDINGS: The chemical screen identified the anti-leprosy drug clofazimine as a hit, which was subsequently validated in several in vitro models. Computational analyses of transcriptional signatures revealed that the effect of clofazimine was due to the stimulation of mitochondrial biogenesis by peroxisome proliferator-activated receptor gamma (PPARγ). In agreement with this, clofazimine rescued mitochondrial dysfunction triggered by Htt-Q94 expression. Importantly, clofazimine also limited polyQ toxicity in developing zebrafish and neuron-specific worm models of polyQ disease. INTERPRETATION: Our results support the potential of repurposing the antimicrobial drug clofazimine for the treatment of polyQ diseases. FUNDING: A full list of funding sources can be found in the acknowledgments section.

A Decade in a Systematic Review: The Evolution and Impact of Cell Painting.

High-content image-based assays have fueled significant discoveries in the life sciences in the past decade (2013-2023), including novel insights into disease etiology, mechanism of action, new therapeutics, and toxicology predictions. Here, we systematically review the substantial methodological advancements and applications of Cell Painting. Advancements include improvements in the Cell Painting protocol, assay adaptations for different types of perturbations and applications, and improved methodologies for feature extraction, quality control, and batch effect correction. Moreover, machine learning methods recently surpassed classical approaches in their ability to extract biologically useful information from Cell Painting images. Cell Painting data have been used alone or in combination with other -omics data to decipher the mechanism of action of a compound, its toxicity profile, and many other biological effects. Overall, key methodological advances have expanded Cell Painting's ability to capture cellular responses to various perturbations. Future advances will likely lie in advancing computational and experimental techniques, developing new publicly available datasets, and integrating them with other high-content data types.

Corrigendum: Development of new approach methods for the identification and characterization of endocrine metabolic disruptors-a PARC project.

[This corrects the article DOI: 10.3389/ftox.2023.1212509.].

From pixels to phenotypes: Integrating image-based profiling with cell health data as BioMorph features improves interpretability.

Cell Painting assays generate morphological profiles that are versatile descriptors of biological systems and have been used to predict in vitro and in vivo drug effects. However, Cell Painting features extracted from classical software such as CellProfiler are based on statistical calculations and often not readily biologically interpretable. In this study, we propose a new feature space, which we call BioMorph, that maps these Cell Painting features with readouts from comprehensive Cell Health assays. We validated that the resulting BioMorph space effectively connected compounds not only with the morphological features associated with their bioactivity but with deeper insights into phenotypic characteristics and cellular processes associated with the given bioactivity. The BioMorph space revealed the mechanism of action for individual compounds, including dual-acting compounds such as emetine, an inhibitor of both protein synthesis and DNA replication. Overall, BioMorph space offers a biologically relevant way to interpret the cell morphological features derived using software such as CellProfiler and to generate hypotheses for experimental validation.

Development of new approach methods for the identification and characterization of endocrine metabolic disruptors-a PARC project.

In past times, the analysis of endocrine disrupting properties of chemicals has mainly been focused on (anti-)estrogenic or (anti-)androgenic properties, as well as on aspects of steroidogenesis and the modulation of thyroid signaling. More recently, disruption of energy metabolism and related signaling pathways by exogenous substances, so-called metabolism-disrupting chemicals (MDCs) have come into focus. While general effects such as body and organ weight changes are routinely monitored in animal studies, there is a clear lack of mechanistic test systems to determine and characterize the metabolism-disrupting potential of chemicals. In order to contribute to filling this gap, one of the project within EU-funded Partnership for the Assessment of Risks of Chemicals (PARC) aims at developing novel in vitro methods for the detection of endocrine metabolic disruptors. Efforts will comprise projects related to specific signaling pathways, for example, involving mTOR or xenobiotic-sensing nuclear receptors, studies on hepatocytes, adipocytes and pancreatic beta cells covering metabolic and morphological endpoints, as well as metabolism-related zebrafish-based tests as an alternative to classic rodent bioassays. This paper provides an overview of the approaches and methods of these PARC projects and how this will contribute to the improvement of the toxicological toolbox to identify substances with endocrine disrupting properties and to decipher their mechanisms of action.

Evaluating the utility of brightfield image data for mechanism of action prediction.

Fluorescence staining techniques, such as Cell Painting, together with fluorescence microscopy have proven invaluable for visualizing and quantifying the effects that drugs and other perturbations have on cultured cells. However, fluorescence microscopy is expensive, time-consuming, labor-intensive, and the stains applied can be cytotoxic, interfering with the activity under study. The simplest form of microscopy, brightfield microscopy, lacks these downsides, but the images produced have low contrast and the cellular compartments are difficult to discern. Nevertheless, by harnessing deep learning, these brightfield images may still be sufficient for various predictive purposes. In this study, we compared the predictive performance of models trained on fluorescence images to those trained on brightfield images for predicting the mechanism of action (MoA) of different drugs. We also extracted CellProfiler features from the fluorescence images and used them to benchmark the performance. Overall, we found comparable and largely correlated predictive performance for the two imaging modalities. This is promising for future studies of MoAs in time-lapse experiments for which using fluorescence images is problematic. Explorations based on explainable AI techniques also provided valuable insights regarding compounds that were better predicted by one modality over the other.

Merging bioactivity predictions from cell morphology and chemical fingerprint models using similarity to training data.

The applicability domain of machine learning models trained on structural fingerprints for the prediction of biological endpoints is often limited by the lack of diversity of chemical space of the training data. In this work, we developed similarity-based merger models which combined the outputs of individual models trained on cell morphology (based on Cell Painting) and chemical structure (based on chemical fingerprints) and the structural and morphological similarities of the compounds in the test dataset to compounds in the training dataset. We applied these similarity-based merger models using logistic regression models on the predictions and similarities as features and predicted assay hit calls of 177 assays from ChEMBL, PubChem and the Broad Institute (where the required Cell Painting annotations were available). We found that the similarity-based merger models outperformed other models with an additional 20% assays (79 out of 177 assays) with an AUC > 0.70 compared with 65 out of 177 assays using structural models and 50 out of 177 assays using Cell Painting models. Our results demonstrated that similarity-based merger models combining structure and cell morphology models can more accurately predict a wide range of biological assay outcomes and further expanded the applicability domain by better extrapolating to new structural and morphology spaces.

New regulators of the tetracycline-inducible gene expression system identified by chemical and genetic screens.

The tetracycline repressor (tetR)-regulated system is a widely used tool to specifically control gene expression in mammalian cells. Based on this system, we generated a human osteosarcoma cell line, which allows for the inducible expression of an EGFP fusion of the TAR DNA-binding protein 43 (TDP-43), which has been linked to neurodegenerative diseases. Consistent with previous findings, TDP-43 overexpression led to the accumulation of aggregates and limited the viability of U2OS. Using this inducible system, we conducted a chemical screen with a library that included FDA-approved drugs. While the primary screen identified several compounds that prevented TDP-43 toxicity, further experiments revealed that these chemicals abrogated the doxycycline-dependent TDP-43 expression. This antagonistic effect was observed with both doxycycline and tetracycline, and in several Tet-On cell lines expressing different genes, confirming the general effect of these compounds as inhibitors of the tetR system. Using the same cell line, a genome-wide CRISPR/Cas9 screen identified epigenetic regulators such as the G9a methyltransferase and TRIM28 as potential modifiers of TDP-43 toxicity. Yet again, further experiments revealed that G9a inhibition or TRIM28 loss prevented doxycycline-dependent expression of TDP-43. In summary, we have identified new chemical and genetic regulators of the tetR system, thereby raising awareness of the limitations of this approach to conduct chemical or genetic screening in mammalian cells.

Integrating cell morphology with gene expression and chemical structure to aid mitochondrial toxicity detection.

Mitochondrial toxicity is an important safety endpoint in drug discovery. Models based solely on chemical structure for predicting mitochondrial toxicity are currently limited in accuracy and applicability domain to the chemical space of the training compounds. In this work, we aimed to utilize both -omics and chemical data to push beyond the state-of-the-art. We combined Cell Painting and Gene Expression data with chemical structural information from Morgan fingerprints for 382 chemical perturbants tested in the Tox21 mitochondrial membrane depolarization assay. We observed that mitochondrial toxicants differ from non-toxic compounds in morphological space and identified compound clusters having similar mechanisms of mitochondrial toxicity, thereby indicating that morphological space provides biological insights related to mechanisms of action of this endpoint. We further showed that models combining Cell Painting, Gene Expression features and Morgan fingerprints improved model performance on an external test set of 244 compounds by 60% (in terms of F1 score) and improved extrapolation to new chemical space. The performance of our combined models was comparable with dedicated in vitro assays for mitochondrial toxicity. Our results suggest that combining chemical descriptors with biological readouts enhances the detection of mitochondrial toxicants, with practical implications in drug discovery.

Morphological profiling of environmental chemicals enables efficient and untargeted exploration of combination effects.

Environmental chemicals are commonly studied one at a time, and there is a need to advance our understanding of the effect of exposure to their combinations. Here we apply high-content microscopy imaging of cells stained with multiplexed dyes (Cell Painting) to profile the effects of Cetyltrimethylammonium bromide (CTAB), Bisphenol A (BPA), and Dibutyltin dilaurate (DBTDL) exposure on four human cell lines; both individually and in all combinations. We show that morphological features can be used with multivariate data analysis to discern between exposures from individual compounds, concentrations, and combinations. CTAB and DBTDL induced concentration-dependent morphological changes across the four cell lines, and BPA exacerbated morphological effects when combined with CTAB and DBTDL. Combined exposure to CTAB and BPA induced changes in the ER, Golgi apparatus, nucleoli and cytoplasmic RNA in one of the cell lines. Different responses between cell lines indicate that multiple cell types are needed when assessing combination effects. The rapid and relatively low-cost experiments combined with high information content make Cell Painting an attractive methodology for future studies of combination effects. All data in the study is made publicly available on Figshare.

NOX4 regulates TGFß-induced proliferation and self-renewal in glioblastoma stem cells.

Glioblastoma (GBM) is the most aggressive and common glioma subtype, with a median survival of 15 months after diagnosis. Current treatments have limited therapeutic efficacy; thus, more effective approaches are needed. The glioblastoma tumoural mass is characterised by a small cellular subpopulation - glioblastoma stem cells (GSCs) - that has been held responsible for glioblastoma initiation, cell invasion, proliferation, relapse and resistance to chemo- and radiotherapy. Targeted therapies against GSCs are crucial, as is understanding the molecular mechanisms that govern the GSCs. Transforming growth factor ß (TGFß) signalling and reactive oxygen species (ROS) production are known to govern and regulate cancer stem cell biology. Among the differentially expressed genes regulated by TGFß in a transcriptomic analysis of two different patient-derived GSCs, we found NADPH oxidase 4 (NOX4) as one of the top upregulated genes. Interestingly, when patient tissues were analysed, NOX4 expression was found to be higher in GSCs versus differentiated cells. A functional analysis of the role of NOX4 downstream of TGFß in several patient-derived GSCs showed that TGFß does indeed induce NOX4 expression and increases ROS production in a NOX4-dependent manner. NOX4 downstream of TGFß regulates GSC proliferation, and NOX4 expression is necessary for TGFß-induced expression of stem cell markers and of the transcription factor nuclear factor erythroid 2-related factor 2 (NRF2), which in turn controls the cell's antioxidant and metabolic responses. Interestingly, overexpression of NOX4 recapitulates the effects induced by TGFß in GSCs: enhanced proliferation, stemness and NRF2 expression. In conclusion, this work functionally establishes NOX4 as a key mediator of GSC biology.

An Open-Source Modular Framework for Automated Pipetting and Imaging Applications.

The number of samples in biological experiments is continuously increasing, but complex protocols and human error in many cases lead to suboptimal data quality and hence difficulties in reproducing scientific findings. Laboratory automation can alleviate many of these problems by precisely reproducing machine-readable protocols. These instruments generally require high up-front investments, and due to the lack of open application programming interfaces (APIs), they are notoriously difficult for scientists to customize and control outside of the vendor-supplied software. Here, automated, high-throughput experiments are demonstrated for interdisciplinary research in life science that can be replicated on a modest budget, using open tools to ensure reproducibility by combining the tools OpenFlexure, Opentrons, ImJoy, and UC2. This automated sample preparation and imaging pipeline can easily be replicated and established in many laboratories as well as in educational contexts through easy-to-understand algorithms and easy-to-build microscopes. Additionally, the creation of feedback loops, with later pipetting or imaging steps depending on the analysis of previously acquired images, enables the realization of fully autonomous "smart" microscopy experiments. All documents and source files are publicly available to prove the concept of smart lab automation using inexpensive, open tools. It is believed this democratizes access to the power and repeatability of automated experiments.

Prolonged estrogen deprivation triggers a broad immunosuppressive phenotype in breast cancer cells.

Among others, expression levels of programmed cell death 1 ligand 1 (PD-L1) have been explored as biomarkers of the response to immune checkpoint inhibitors in cancer therapy. Here, we present the results of a chemical screen that interrogated how medically approved drugs influence PD-L1 expression. As expected, corticosteroids and inhibitors of Janus kinases were among the top PD-L1 downregulators. In addition, we identified that PD-L1 expression is induced by antiestrogenic compounds. Transcriptomic analyses indicate that chronic estrogen receptor alpha (ERα) inhibition triggers a broad immunosuppressive program in ER-positive breast cancer cells, which is subsequent to their growth arrest and involves the activation of multiple immune checkpoints together with the silencing of the antigen-presenting machinery. Accordingly, estrogen-deprived MCF7 cells are resistant to T-cell-mediated cell killing, in a manner that is independent of PD-L1, but which is reverted by estradiol. Our study reveals that while antiestrogen therapies efficiently limit the growth of ER-positive breast cancer cells, they concomitantly trigger a transcriptional program that favors their immune evasion.

A phenomics approach for antiviral drug discovery.

BACKGROUND: The emergence and continued global spread of the current COVID-19 pandemic has highlighted the need for methods to identify novel or repurposed therapeutic drugs in a fast and effective way. Despite the availability of methods for the discovery of antiviral drugs, the majority tend to focus on the effects of such drugs on a given virus, its constituent proteins, or enzymatic activity, often neglecting the consequences on host cells. This may lead to partial assessment of the efficacy of the tested anti-viral compounds, as potential toxicity impacting the overall physiology of host cells may mask the effects of both viral infection and drug candidates. Here we present a method able to assess the general health of host cells based on morphological profiling, for untargeted phenotypic drug screening against viral infections. RESULTS: We combine Cell Painting with antibody-based detection of viral infection in a single assay. We designed an image analysis pipeline for segmentation and classification of virus-infected and non-infected cells, followed by extraction of morphological properties. We show that this methodology can successfully capture virus-induced phenotypic signatures of MRC-5 human lung fibroblasts infected with human coronavirus 229E (CoV-229E). Moreover, we demonstrate that our method can be used in phenotypic drug screening using a panel of nine host- and virus-targeting antivirals. Treatment with effective antiviral compounds reversed the morphological profile of the host cells towards a non-infected state. CONCLUSIONS: The phenomics approach presented here, which makes use of a modified Cell Painting protocol by incorporating an anti-virus antibody stain, can be used for the unbiased morphological profiling of virus infection on host cells. The method can identify antiviral reference compounds, as well as novel antivirals, demonstrating its suitability to be implemented as a strategy for antiviral drug repurposing and drug discovery.

A chemical screen for modulators of mRNA translation identifies a distinct mechanism of toxicity for sphingosine kinase inhibitors.

We here conducted an image-based chemical screen to evaluate how medically approved drugs, as well as drugs that are currently under development, influence overall translation levels. None of the compounds up-regulated translation, which could be due to the screen being performed in cancer cells grown in full media where translation is already present at very high levels. Regarding translation down-regulators, and consistent with current knowledge, inhibitors of the mechanistic target of rapamycin (mTOR) signaling pathway were the most represented class. In addition, we identified that inhibitors of sphingosine kinases (SPHKs) also reduce mRNA translation levels independently of mTOR. Mechanistically, this is explained by an effect of the compounds on the membranes of the endoplasmic reticulum (ER), which activates the integrated stress response (ISR) and contributes to the toxicity of SPHK inhibitors. Surprisingly, the toxicity and activation of the ISR triggered by 2 independent SPHK inhibitors, SKI-II and ABC294640, the latter in clinical trials, are also observed in cells lacking SPHK1 and SPHK2. In summary, our study provides a useful resource on the effects of medically used drugs on translation, identified compounds capable of reducing translation independently of mTOR and has revealed that the cytotoxic properties of SPHK inhibitors being developed as anticancer agents are independent of SPHKs.

High content drug screening for Fanconi anemia therapeutics.

BACKGROUND: Fanconi anemia is a rare disease clinically characterized by malformations, bone marrow failure and an increased risk of solid tumors and hematologic malignancies. The only therapies available are hematopoietic stem cell transplantation for bone marrow failure or leukemia, and surgical resection for solid tumors. Therefore, there is still an urgent need for new therapeutic options. With this aim, we developed a novel high-content cell-based screening assay to identify drugs with therapeutic potential in FA. RESULTS: A TALEN-mediated FANCA-deficient U2OS cell line was stably transfected with YFP-FANCD2 fusion protein. These cells were unable to form fluorescent foci or to monoubiquitinate endogenous or exogenous FANCD2 upon DNA damage and were more sensitive to mitomycin C when compared to the parental wild type counterpart. FANCA correction by retroviral infection restored the cell line's ability to form FANCD2 foci and ubiquitinate FANCD2. The feasibility of this cell-based system was interrogated in a high content screening of 3802 compounds, including a Prestwick library of 1200 FDA-approved drugs. The potential hits identified were then individually tested for their ability to rescue FANCD2 foci and monoubiquitination, and chromosomal stability in the absence of FANCA. CONCLUSIONS: While, unfortunately, none of the compounds tested were able to restore cellular FANCA-deficiency, our study shows the potential capacity to screen large compound libraries in the context of Fanconi anemia therapeutics in an optimized and cost-effective platform.

Gefitinib and Afatinib Show Potential Efficacy for Fanconi Anemia-Related Head and Neck Cancer.

PURPOSE: Fanconi anemia rare disease is characterized by bone marrow failure and a high predisposition to solid tumors, especially head and neck squamous cell carcinoma (HNSCC). Patients with Fanconi anemia with HNSCC are not eligible for conventional therapies due to high toxicity in healthy cells, predominantly hematotoxicity, and the only treatment currently available is surgical resection. In this work, we searched and validated two already approved drugs as new potential therapies for HNSCC in patients with Fanconi anemia. EXPERIMENTAL DESIGN: We conducted a high-content screening of 3,802 drugs in a FANCA-deficient tumor cell line to identify nongenotoxic drugs with cytotoxic/cytostatic activity. The best candidates were further studied in vitro and in vivo for efficacy and safety. RESULTS: Several FDA/European Medicines Agency (EMA)-approved anticancer drugs showed cancer-specific lethality or cell growth inhibition in Fanconi anemia HNSCC cell lines. The two best candidates, gefitinib and afatinib, EGFR inhibitors approved for non-small cell lung cancer (NSCLC), displayed nontumor/tumor IC50 ratios of approximately 400 and approximately 100 times, respectively. Neither gefitinib nor afatinib activated the Fanconi anemia signaling pathway or induced chromosomal fragility in Fanconi anemia cell lines. Importantly, both drugs inhibited tumor growth in xenograft experiments in immunodeficient mice using two Fanconi anemia patient-derived HNSCCs. Finally, in vivo toxicity studies in Fanca-deficient mice showed that administration of gefitinib or afatinib was well-tolerated, displayed manageable side effects, no toxicity to bone marrow progenitors, and did not alter any hematologic parameters. CONCLUSIONS: Our data present a complete preclinical analysis and promising therapeutic line of the first FDA/EMA-approved anticancer drugs exerting cancer-specific toxicity for HNSCC in patients with Fanconi anemia.

The antimalarial drug amodiaquine stabilizes p53 through ribosome biogenesis stress, independently of its autophagy-inhibitory activity.

Pharmacological inhibition of ribosome biogenesis is a promising avenue for cancer therapy. Herein, we report a novel activity of the FDA-approved antimalarial drug amodiaquine which inhibits rRNA transcription, a rate-limiting step for ribosome biogenesis, in a dose-dependent manner. Amodiaquine triggers degradation of the catalytic subunit of RNA polymerase I (Pol I), with ensuing RPL5/RPL11-dependent stabilization of p53. Pol I shutdown occurs in the absence of DNA damage and without the subsequent ATM-dependent inhibition of rRNA transcription. RNAseq analysis revealed mechanistic similarities of amodiaquine with BMH-21, the first-in-class Pol I inhibitor, and with chloroquine, the antimalarial analog of amodiaquine, with well-established autophagy-inhibitory activity. Interestingly, autophagy inhibition caused by amodiaquine is not involved in the inhibition of rRNA transcription, suggesting two independent anticancer mechanisms. In vitro, amodiaquine is more efficient than chloroquine in restraining the proliferation of human cell lines derived from colorectal carcinomas, a cancer type with predicted susceptibility to ribosome biogenesis stress. Taken together, our data reveal an unsuspected activity of a drug approved and used in the clinics for over 30 years, and provide rationale for repurposing amodiaquine in cancer therapy.