Machine learning-assisted high-content imaging analysis of 3D MCF7 microtissues for estrogenic effect prediction.

Endocrine-disrupting chemicals (EDCs) pose a significant threat to human well-being and the ecosystem. However, in managing the many thousands of uncharacterized chemical entities, the high-throughput screening of EDCs using relevant biological endpoints remains challenging. Three-dimensional (3D) culture technology enables the development of more physiologically relevant systems in more realistic biochemical microenvironments. The high-content and quantitative imaging techniques enable quantifying endpoints associated with cell morphology, cell-cell interaction, and microtissue organization. In the present study, 3D microtissues formed by MCF-7 breast cancer cells were exposed to the model EDCs estradiol (E2) and propyl pyrazole triol (PPT). A 3D imaging and image analysis pipeline was established to extract quantitative image features from estrogen-exposed microtissues. Moreover, a machine-learning classification model was built using estrogenic-associated differential imaging features. Based on 140 common differential image features found between the E2 and PPT group, the classification model predicted E2 and PPT exposure with AUC-ROC at 0.9528 and 0.9513, respectively. Deep learning-assisted analysis software was developed to characterize microtissue gland lumen formation. The fully automated tool can accurately characterize the number of identified lumens and the total luminal volume of each microtissue. Overall, the current study established an integrated approach by combining non-supervised image feature profiling and supervised luminal volume characterization, which reflected the complexity of functional ER signaling and highlighted a promising conceptual framework for estrogenic EDC risk assessment.

Development of a human liver microphysiological co-culture system for higher throughput chemical safety assessment.

Chemicals in the systemic circulation can undergo hepatic xenobiotic metabolism, generate metabolites and exhibit altered toxicity compared to their parent compounds. This paper describes a two-chamber liver-organ co-culture model in a higher-throughput 96-well format for the determination of toxicity on target tissues in the presence of physiologically relevant human liver metabolism. This two-chamber system is a hydrogel formed within each well consisting of a central well (target tissue) and an outer ring-shaped trough (human liver tissue). The target tissue chamber can be configured to accommodate a three-dimensional (3D) spheroid-shaped microtissue, or a two-dimensional (2D) cell mono-layer. Culture medium and compounds freely diffuse between the two chambers. Human differentiated HepaRGTM liver cells are used to form the 3D human liver microtissues, which displayed robust protein expression of liver biomarkers (albumin, asialoglycoprotein receptor, Phase I cytochrome P450 (CYP3A4) enzyme, multidrug resistance-associated protein 2 transporter, and glycogen), and exhibited Phase I/II enzyme activities over the course of 17 days. Histological and ultrastructural analyses confirmed that the HepaRG microtissues presented a differentiated hepatocyte phenotype, including abundant mitochondria, endoplasmic reticulum and bile canaliculi. Liver microtissue zonation characteristics could be easily modulated by maturation in different media supplements. Furthermore, our proof-of-concept study demonstrated the efficacy of this co-culture model in evaluating testosterone-mediated androgen receptor responses in the presence of human liver metabolism. This liver-organ co-culture system provides a practical, higher-throughput testing platform for metabolism-dependent bioactivity assessment of drugs/chemicals, to better recapitulate the biological effects and potential toxicity of human exposures.

Machine Learning-Assisted High-Content Imaging Analysis of 3D MCF7 Microtissues for Estrogenic Effect Prediction.

Endocrine-disrupting chemicals (EDCs) pose a significant threat to human well-being and the ecosystem. However, in managing the many thousands of uncharacterized chemical entities, the high-throughput screening of EDCs using relevant biological endpoints remains challenging. Three-dimensional (3D) culture technology enables the development of more physiologically relevant systems in more realistic biochemical microenvironments. The high-content and quantitative imaging techniques enable quantifying endpoints associated with cell morphology, cell-cell interaction, and microtissue organization. In the present study, 3D microtissues formed by MCF-7 breast cancer cells were exposed to the model EDCs estradiol (E2) and propyl pyrazole triol (PPT). A 3D imaging and image analysis pipeline was established to extract quantitative image features from estrogen-exposed microtissues. Moreover, a machine-learning classification model was built using estrogenic-associated differential imaging features. Based on 140 common differential image features found between the E2 and PPT group, the classification model predicted E2 and PPT exposure with AUC-ROC at 0.9528 and 0.9513, respectively. Deep learning-assisted analysis software was developed to characterize microtissue gland lumen formation. The fully automated tool can accurately characterize the number of identified lumens and the total luminal volume of each microtissue. Overall, the current study established an integrated approach by combining non-supervised image feature profiling and supervised luminal volume characterization, which reflected the complexity of functional ER signaling and highlighted a promising conceptual framework for estrogenic EDC risk assessment.

A novel co-culture model of human prostate epithelial and stromal cells for androgenic and antiandrogenic screening.

The risk assessment of endocrine-disrupting chemicals (EDCs) greatly relies on in vitro screening. A 3-dimensional (3D) in vitro prostate model that can reflect physiologically-relevant prostate epithelial and stromal crosstalk can significantly advance the current androgen assessment. This study built a prostate epithelial and stromal co-culture microtissue model with BHPrE and BHPrS cells in scaffold-free hydrogels. The optimal 3D co-culture condition was defined, and responses of the microtissue to androgen (dihydrotestosterone, DHT) and anti-androgen (flutamide) exposure were characterized using molecular and image profiling techniques. The co-culture prostate microtissue maintained a stable structure for up to seven days and presented molecular and morphological features of the early developmental stage of the human prostate. The cytokeratin 5/6 (CK5/6) and cytokeratin 18 (CK18) immunohistochemical staining indicated epithelial heterogeneity and differentiation in these microtissues. The prostate-related gene expression profiling did not efficiently differentiate androgen and anti-androgen exposure. However, a cluster of distinctive 3D image features was identified and could be applied in the androgenic and anti-androgenic effect prediction. Overall, the current study established a co-culture prostate model that provided an alternative strategy for (anti-)androgenic EDC safety assessment and highlighted the potential and advantage of utilizing image features to predict endpoints in chemical screening.

A human-derived prostate co-culture microtissue model using epithelial (RWPE-1) and stromal (WPMY-1) cell lines.

The development and normal function of prostate tissue depends on signalling interactions between stromal and epithelial compartments. Development of a prostate microtissue composed of these two components can help identify substance exposures that could cause adverse effects in humans as part of a non-animal risk assessment. In this study, prostate microtissues composed of human derived stromal (WPMY-1) and epithelial (RWPE-1) cell lines grown in scaffold-free hydrogels were developed and characterized using immunohistochemistry, light microscopy, and qRT-PCR. Within 5 days after seeding, the microtissues self-organized into spheroids consisting of a core of stromal WPMY-1 cells surrounded by epithelial RWPE-1 cells. The RWPE-1 layer is reflective of intermediate prostatic epithelium, expressing both characteristics of the luminal (high expression of PSA) and basal (high expression of cytokeratins 5/6 and 14) epithelial cells. The response of the microtissues to an androgen (dihydrotestosterone, DHT) and an anti-androgen (flutamide) was also investigated. Treatment with DHT, flutamide or a mixture of DHT and flutamide indicated that the morphology and self-organization of the microtissues is androgen dependent. qRT-PCR data showed that a saturating concentration of DHT increased the expression of genes coding for the estrogen receptors (ESR1 and ESR2) and decreased the expression of CYP1B1 without affecting the expression of the androgen receptor. With further development and optimization RWPE-1/WPMY-1 microtissues can play an important role in non-animal risk assessments.

Differential toxicity of water versus gavage exposure to trichloroethylene in rats.

Trichloroethylene (TCE) is a persistent environmental contaminant that causes male reproductive toxicity. We investigated whether transient increases in TCE exposure modulated male reproductive toxicity by exposing rats via daily oral to repeated gavage exposures (1000 mg/kg/day) and through drinking water (0.6% TCE) for 14 weeks. The gavage route resulted in reversible reduction of epididymis weight, and reduced body weight that persisted for up to 12-weeks after cessation of exposure. Physiologically-based pharmacokinetic modeling predicted that the gavage route results in higher Cmax and AUC exposure of TCE compared to drinking water exposure, explaining the observed differences in toxicity between dosing regimens.

Effects of continuous bisphenol A exposure from early gestation on 90 day old rat testes function and sperm molecular profiles: A CLARITY-BPA consortium study.

Bisphenol A (BPA) is a ubiquitous industrial chemical that has been identified as an endocrine disrupting compound (EDC). There is growing concern that early life exposures to EDCs, such as BPA, can adversely affect the male reproductive tract and function. This study was conducted as part of the Consortium Linking Academic and Regulatory Insights on BPA Toxicity (CLARITY-BPA) to further delineate the toxicities associated with continuous exposure to BPA from early gestation, and to comprehensively examine the elicited effects on testes and sperm. NCTR Sprague Dawley rat dams were gavaged from gestational day (GD) 6 until parturition, and their pups were directly gavaged daily from postnatal day (PND) 1 to 90 with BPA (2.5, 25, 250, 2500, 25,000, 250,000 µg/kg/d) or vehicle control. At PND 90, the testes and sperm were collected for evaluation. The testes were histologically evaluated for altered germ cell apoptosis, sperm production, and altered spermiation. RNA and DNA isolated from sperm were assessed for elicited changes in global mRNA transcript abundance and altered DNA methylation. Effects of BPA were observed in changes in body, testis and epididymis weights only at the highest administered dose of BPA of 250,000 µg/kg/d. Genome-wide transcriptomic and epigenomic analyses failed to detect robust alterations in sperm mRNA and DNA methylation levels. These data indicate that prolonged exposure starting in utero to BPA over a wide range of levels has little, if any, impact on the testes and sperm molecular profiles of 90 day old rats as assessed by the histopathologic, morphometric, and molecular endpoints evaluated.

Information-dependent enrichment analysis reveals time-dependent transcriptional regulation of the estrogen pathway of toxicity.

The twenty-first century vision for toxicology involves a transition away from high-dose animal studies to in vitro and computational models (NRC in Toxicity testing in the 21st century: a vision and a strategy, The National Academies Press, Washington, DC, 2007). This transition requires mapping pathways of toxicity by understanding how in vitro systems respond to chemical perturbation. Uncovering transcription factors/signaling networks responsible for gene expression patterns is essential for defining pathways of toxicity, and ultimately, for determining the chemical modes of action through which a toxicant acts. Traditionally, transcription factor identification is achieved via chromatin immunoprecipitation studies and summarized by calculating which transcription factors are statistically associated with up- and downregulated genes. These lists are commonly determined via statistical or fold-change cutoffs, a procedure that is sensitive to statistical power and may not be as useful for determining transcription factor associations. To move away from an arbitrary statistical or fold-change-based cutoff, we developed, in the context of the Mapping the Human Toxome project, an enrichment paradigm called information-dependent enrichment analysis (IDEA) to guide identification of the transcription factor network. We used a test case of activation in MCF-7 cells by 17ß estradiol (E2). Using this new approach, we established a time course for transcriptional and functional responses to E2. ERα and ERß were associated with short-term transcriptional changes in response to E2. Sustained exposure led to recruitment of additional transcription factors and alteration of cell cycle machinery. TFAP2C and SOX2 were the transcription factors most highly correlated with dose. E2F7, E2F1, and Foxm1, which are involved in cell proliferation, were enriched only at 24 h. IDEA should be useful for identifying candidate pathways of toxicity. IDEA outperforms gene set enrichment analysis (GSEA) and provides similar results to weighted gene correlation network analysis, a platform that helps to identify genes not annotated to pathways.

Estradiol Exposure Differentially Alters Monolayer versus Microtissue MCF-7 Human Breast Carcinoma Cultures.

The development of three-dimensional (3D) cultures is increasing, as they are able to provide the utility of in vitro models and the strength of testing in physiologically relevant systems. When cultured in a scaffold-free agarose hydrogel system, MCF-7 human breast carcinoma cells organize and develop into microtissues that contain a luminal space, in stark contrast to the flat morphology of MCF-7 two-dimensional (2D) monolayer cultures. Following exposure to 1nM E2, expression of typical estrogen-responsive genes, including progesterone receptor (PGR), PDZ containing domain 1 (PDZK1) and amphiregulin (AREG) is increased in both 2D and 3D cultures. When examining expression of other genes, particularly those involved in cell adhesion, there were large changes in 3D MCF-7 microtissues, with little to no change observed in the MCF-7 monolayer cultures. Together, these results indicate that while the initial estrogen-regulated transcriptional targets respond similarly in 2D and 3D cultures, there are large differences in activation of other pathways related to cell-cell interactions.

Genetic variability in a frozen batch of MCF-7 cells invisible in routine authentication affecting cell function.

Common recommendations for cell line authentication, annotation and quality control fall short addressing genetic heterogeneity. Within the Human Toxome Project, we demonstrate that there can be marked cellular and phenotypic heterogeneity in a single batch of the human breast adenocarcinoma cell line MCF-7 obtained directly from a cell bank that are invisible with the usual cell authentication by short tandem repeat (STR) markers. STR profiling just fulfills the purpose of authentication testing, which is to detect significant cross-contamination and cell line misidentification. Heterogeneity needs to be examined using additional methods. This heterogeneity can have serious consequences for reproducibility of experiments as shown by morphology, estrogenic growth dose-response, whole genome gene expression and untargeted mass-spectroscopy metabolomics for MCF-7 cells. Using Comparative Genomic Hybridization (CGH), differences were traced back to genetic heterogeneity already in the cells from the original frozen vials from the same ATCC lot, however, STR markers did not differ from ATCC reference for any sample. These findings underscore the need for additional quality assurance in Good Cell Culture Practice and cell characterization, especially using other methods such as CGH to reveal possible genomic heterogeneity and genetic drifts within cell lines.

Morphologic effects of estrogen stimulation on 3D MCF-7 microtissues.

In the development of human cell-based assays, 3-dimensional (3D) cell culture models are intriguing as they are able to bridge the gap between animal models and traditional two-dimensional (2D) cell culture. Previous work has demonstrated that MCF-7 human breast carcinoma cells cultured in a 3D scaffold-free culture system self-assemble and develop into differentiated microtissues that possess a luminal space. Exposure to estradiol for 7 days decreased lumen formation in MCF-7 microtissues, altered microtissue morphology and altered expression of genes involved in estrogen signaling, cell adhesion and cell cycle regulation. Exposure to receptor-specific agonists for estrogen receptor alpha, estrogen receptor beta and g-protein coupled estrogen receptor resulted in unique, receptor-specific phenotypes and gene expression signatures. The use of a differentiated scaffold-free 3D culture system offers a unique opportunity to study the phenotypic and molecular changes associated with exposure to estrogenic compounds.

Into the depths: Techniques for in vitro three-dimensional microtissue visualization.

Three-dimensional (3-D) in vitro platforms have been shown to closely recapitulate human physiology when compared with conventional two-dimensional (2-D) in vitro or in vivo animal model systems. This confers a substantial advantage in evaluating disease mechanisms, pharmaceutical drug discovery, and toxicity testing. Despite the benefits of 3-D cell culture, limitations in visualization and imaging of 3-D microtissues present significant challenges. Here we optimized histology and microscopy techniques to overcome the constraints of 3-D imaging. For morphological assessment of 3-D microtissues of several cell types, different time points, and different sizes, a two-step glycol methacrylate embedding protocol for evaluating 3-D microtissues produced using agarose hydrogels improved resolution of nuclear and cellular histopathology characteristic of cell death and proliferation. Additional immunohistochemistry, immunofluorescence, and in situ immunostaining techniques were successfully adapted to these microtissues and enhanced by optical clearing. Utilizing the Clear(T2) protocol greatly increased fluorescence signal intensity, imaging depth, and clarity, allowing for more complete confocal fluorescence microscopy imaging of these 3-D microtissues compared with uncleared samples. The refined techniques presented here address the key challenges associated with 3-D imaging, providing new and alternative methods in evaluating disease pathogenesis, delineating toxicity pathways, and enhancing the versatility of 3-D in vitro testing systems in pharmacological and toxicological applications.

MCF-7 Human Breast Cancer Cells Form Differentiated Microtissues in Scaffold-Free Hydrogels.

Three-dimensional (3D) cultures are increasing in use because of their ability to represent in vivo human physiology when compared to monolayer two-dimensional (2D) cultures. When grown in 3D using scaffold-free agarose hydrogels, MCF-7 human breast cancer cells self-organize to form directionally-oriented microtissues that contain a luminal space, reminiscent of the in vivo structure of the mammary gland. When compared to MCF-7 cells cultured in 2D monolayer culture, MCF-7 microtissues exhibit increased mRNA expression of luminal epithelial markers keratin 8 and keratin 19 and decreased expression of basal marker keratin 14 and the mesenchymal marker vimentin. These 3D MCF-7 microtissues remain responsive to estrogens, as demonstrated by induction of known estrogen target mRNAs following exposure to 17ß-estradiol. Culture of MCF-7 cells in scaffold-free conditions allows for the formation of more differentiated, estrogen-responsive structures that are a more relevant system for evaluation of estrogenic compounds than traditional 2D models.