Using ToxCast™ Data to Reconstruct Dynamic Cell State Trajectories and Estimate Toxicological Points of Departure

Background: High-content imaging (HCI) allows simultaneous measurement of multiple cellular phenotypic changes and is an important tool for evaluating the biological activity of chemicals. Objectives: Our goal was to analyze dynamic cellular changes using HCI to identify the “tipping point” at which the cells did not show recovery towards a normal phenotypic state. Methods: HCI was used to evaluate the effects of 967 chemicals (in concentrations ranging from 0.4 to 200 μM) on HepG2 cells over a 72-hr exposure period. The HCI end points included p53, c-Jun, histone H2A.x, α-tubulin, histone H3, alpha tubulin, mitochondrial membrane potential, mitochondrial mass, cell cycle arrest, nuclear size, and cell number. A computational model was developed to interpret HCI responses as cell-state trajectories. Results: Analysis of cell-state trajectories showed that 336 chemicals produced tipping points and that HepG2 cells were resilient to the effects of 334 chemicals up to the highest concentration (200 μM) and duration (72 hr) tested. Tipping points were identified as concentration-dependent transitions in system recovery, and the corresponding critical concentrations were generally between 5 and 15 times (25th and 75th percentiles, respectively) lower than the concentration that produced any significant effect on HepG2 cells. The remaining 297 chemicals require more data before they can be placed in either of these categories. Conclusions: These findings show the utility of HCI data for reconstructing cell state trajectories and provide insight into the adaptation and resilience of in vitro cellular systems based on tipping points. Cellular tipping points could be used to define a point of departure for risk-based prioritization of environmental chemicals. Citation: Shah I, Setzer RW, Jack J, Houck KA, Judson RS, Knudsen TB, Liu J, Martin MT, Reif DM, Richard AM, Thomas RS, Crofton KM, Dix DJ, Kavlock RJ. 2016. Using ToxCast™ data to reconstruct dynamic cell state trajectories and estimate toxicological points of departure. Environ Health Perspect 124:910–919; http://dx.doi.org/10.1289/ehp.1409029


Table of Contents
. Sample high-content images for select chemicals: (a) CCCP, (b) taxol, (c) butachlor, (d) fludioxonil and (e) fluazinam. Each panel shows raw images for each chemical as a hierarchical grid in which the stains/fluorochromes are organized as rows, followed by time (sub-rows), and concentrations are given in the columns. The first column from the left shows the negative control (DMSO) followed by increasing concentrations of the treatment chemical. In order to provide a more detailed view of the cell-level changes for concentrations and time points, each image shows four fields. The magnification of all images was 20x unless labeled otherwise. For some chemicals we could not obtain the raw images for all stains/fluorochromes, concentrations and times. Finally, because all figures were generated automatically using the database (and not manually) some of the images may contain artifacts. Figure S2. Raw concentration and time response data for reference chemicals, taxol and CCCP. Panels (a) and (b) show concentration (x-axis) dependent raw responses (y-axis) in cell number (CN), nuclear size (NS), mitochondrial membrane potential (MMP), mitotic arrest (MA) and mitochondrial mass (MM) for CCCP and taxol, respectively. The Responses at 1, 24 and 72 h are shown in different colors (see legend). Raw data (labeled as, "+") and smoothed data are shown as curves (see text for additional information). Panels (c) and (d) show the time (x-axis) dependent raw responses (y-axis) across CN, NS, MMP, MA and MM for CCCP and taxol, respectively. The responses at different concentrations are signified by colors (see legend). The mean and standard deviation of responses for DMSO are shown in green.

Figure S3.
Concentration and time-dependent responses across microtubules (Mt), p53 activity (p53), oxidative stress (OS), mitotic arrest (MA), mitochondrial membrane, potential (MMP), cell cycle arrest (CCA), cell number (CN), stress kinase (SK), mitochondrial mass (MM) and nuclear size (NS) for (a) fludioxonil, (b) fluazinam, and (c) butachlor. Each graph shows the smoothed raw responses in the yaxis, and the time in hours along the x-axis. The responses for 10 treatment concentrations are shown in different colors (see legend). Finally, the mean and standard deviation of raw responses for each endpoint for the DMSO controls are shown in green. Table S1. The algorithms used for extracting extracting cell level features from the raw images. From left to right, each row shows the cell level feature(s), the cell stain or fluorochrome used by the algorithm, the name of the algorithm, and the name of the BioApplication software.

Supplemental Code and Data Zip File
Excel Table S1. The lowest effect concentration (LEC) for 967 chemicals. Each row shows the LECs for one chemical at one time point across the ten high-content imaging endpoints. From left to right the first three columns show the ToxCast chemical identifier (hchem_id), the CAS number (chem_casrn), and the time (timeh) in hours (h). The remaining columns show the micromolar (µM) LEC value for p53, SK (stress kinase), OS (oxidative stress), Mt (microtubules), MM (mitochondrial mass), mitochondrial membrane potential (MMP), mitotic arrest (MA), cell cycle arrest (CCA), nuclear size (NS), and cell number (CN). Table S2. Perturbations of the HepG2 system produced by 967 chemicals across 10 high-content imaging (HCI) endpoints, 10 concentrations and different time points. Each row shows the concentration and time dependent perturbations produced by each chemical across the 10 HCI endpoints. From left to right the first four columns show the ToxCast chemical identifier (hchem_id), the chemical name (chem_name), the time (timeh) in hours (h) and the micromolar (µM) concentration (conc). The remaining columns show perturbations as log 2 (fold change) values for each HCI endpoint including: p53, SK (stress kinase), OS (oxidative stress), Mt (microtubules), MM (mitochondrial mass), mitochondrial membrane potential (MMP), mitotic arrest (MA), cell cycle arrest (CCA), nuclear size (NS), and cell number (CN).

Excel
Excel Table S3. The HepG2 system tipping points and critical concentration (C cr ) analysis for all 967 chemicals at 72 h. Each row shows the mean and standard deviation of C cr for each chemical along with results of the uncertainty analysis. From left to right the columns show the ToxCast chemical identifier (hchem_id), the chemical name (chem_name), the CAS number (chem_cas), the confidence in the tipping point (tip_pt_pcnt), the mean value of C cr (C_crit_mn) and the standard deviation of C cr (C_crit_sd). The last column on the right shows the overall result of the trajectory analysis (traj_analysis), which has one of three possible values: recovered, tipping point and uncertain. Figure S1. Sample high-content images for select chemicals: (a) CCCP, (b) taxol, (c) butachlor, (d) fludioxonil and (e) fluazinam. Each panel shows raw images for each chemical as a hierarchical grid in which the stains/fluorochromes are organized as rows, followed by time (sub-rows), and concentrations are given in the columns. The first column from the left shows the negative control (DMSO) followed by increasing concentrations of the treatment chemical. In order to provide a more detailed view of the cell-level changes for concentrations and time points, each image shows four fields. The magnification of all images was 20x unless labeled otherwise. For some chemicals we could not obtain the raw images for all stains/fluorochromes, concentrations and times. Finally, because all figures were generated automatically using the database (and not manually) some of the images may contain artifacts.    S1. The algorithms used for extracting extracting cell level features from the raw images. From left to right, each row shows the cell level feature(s), the cell stain or fluorochrome used by the algorithm, the name of the algorithm, and the name of the BioApplication software. Excel Table S1. The lowest effect concentration (LEC) for 967 chemicals. Each row shows the LECs for one chemical at one time point across the ten high-content imaging endpoints. From left to right the first three columns show the ToxCast chemical identifier (hchem_id), the CAS number (chem_casrn), and the time (timeh) in hours (h). The remaining columns show the micromolar (µM) LEC value for p53, SK (stress kinase), OS (oxidative stress), Mt (microtubules), MM (mitochondrial mass), mitochondrial membrane potential (MMP), mitotic arrest (MA), cell cycle arrest (CCA), nuclear size (NS), and cell number (CN).

Endpoint
Excel Table S2. Perturbations of the HepG2 system produced by 967 chemicals across 10 high-content imaging (HCI) endpoints, 10 concentrations and different time points. Each row shows the concentration and time dependent perturbations produced by each chemical across the 10 HCI endpoints. From left to right the first four columns show the ToxCast chemical identifier (hchem_id), the chemical name (chem_name), the time (timeh) in hours (h) and the micromolar (µM) concentration (conc). The remaining columns show perturbations as log 2 (fold change) values for each HCI endpoint including: p53, SK (stress kinase), OS (oxidative stress), Mt (microtubules), MM (mitochondrial mass), mitochondrial membrane potential (MMP), mitotic arrest (MA), cell cycle arrest (CCA), nuclear size (NS), and cell number (CN).
Excel Table S3. The HepG2 system tipping points and critical concentration (C cr ) analysis for all 967 chemicals at 72 h. Each row shows the mean and standard deviation of C cr for each chemical along with results of the uncertainty analysis. From left to right the columns show the ToxCast chemical identifier (hchem_id), the chemical name (chem_name), the CAS number (chem_cas), the confidence in the tipping point (tip_pt_pcnt), the mean value of C cr (C_crit_mn) and the standard deviation of C cr (C_crit_sd). The last column on the right shows the overall result of the trajectory analysis (traj_analysis), which has one of three possible values: recovered, tipping point and uncertain.