Document 3ewLRXM4RRwmyO5zRBxvzkVBJ

5 September 2023 ToxProfiler Report Final version Toxys project code: Sponsor reference: Date: 23063 MA-RR-23-5046 2023 09 08 Toxys B.V. Authors: B. ter Braak PhD, L. Loonstra-Wolters MSc., K. Elbertse MSc., R. Derr MSc. Final version 1 5 September 2023 Table of Content Study administration ...................................................................................................................3 Executive summary ....................................................................................................................4 Test protocol ...............................................................................................................................6 Concentration-range finding results ...........................................................................................9 ToxProfiler assay results...........................................................................................................11 Conclusions and discussion ......................................................................................................17 Annex I: Concentration response curves per reporter/chemical ...............................................18 Final version 2 Study administration Test facility: Study director Study coordinator Study operators Sponsor: Sponsor contact: Number of test articles: Date of testing: Number and sequence of tests: Data storage: 5 September 2023 Toxys B.V. Limes 7 2342 DH Oegstgeest The Netherlands Contact: B. ter Braak @toxys.com + 31713322472 B. ter Braak, PhD R. Derr, MSc. L. Loonstra-Wolters, MSc. K. Elbertse, MSc. Honeywell International 115 Tabor Rd. Morris Plains New Jersey 07950 United States Kamlesh Sodani e-mail: @honeywell.com 8 Oxidative stress: CDDO-me Cell cycle stress: Cisplatin ER stress: Tunicamycin Autophagy: Amiodarone Ion stress: Cadmium Chloride Protein stress: Cadmium Chloride Inflammation: human TNF alpha Cytotoxicity: Cadmium Chloride Between August 15th, 2023 and August 25th, 2023 - Concentration range finding with parental HepG2 cells. - ToxProfiler with 7 reporter lines with 3 independent biological replicates. All primary test results are stored at least three years at Toxys and are available upon request. All the quantified data (.txt) per replicate/reporter can be shared upon request. Final version 3 5 September 2023 Executive summary Objective: The goal of this project is to assess the toxic properties and investigate the toxicological mode-of-action of 8 compounds asked to test by Honeywell in the ToxProfiler assay. Methods: The ToxProfiler assay consists of seven human reporter cell lines that were applied to study chemically induced activation of seven distinct stress pathways; oxidative stress (SRXN1-GFP), cell cycle stress (p21-GFP), ER stress (CHOP-GFP), autophagy (LC3-GFP), ion stress (MT1X-GFP), protein stress (HSPA1B-GFP) and inflammation (ICAM1-GFP). The differential induction of the GFP reporters as well as cytotoxicity of the tested compounds were determined by live cell confocal microscopy. Using our automated image segmentation and quantification pipelines, the stress biomarker induction was accurately quantified with a single cell resolution. A custom R-package "modelpod" was used to determine the lowest concentratio40n of the test compounds that resulting in a significant induction of the stress biomarker. Results: The assay control compounds significantly induced their respective stress pathway. Cytotoxicity was observed for 6 out of the 8 test compounds in the tested concentration range. None of the ToxProfiler endpoints was induced by PFPrA and TFA Na in the tested concentration range. In Table 1 an overview of the ToxProfiler reporter results is presented. The POD plotter software was applied to determine if a stress pathway was significantly affected in a concentration dependent fashion. The same PoD approach was applied on the PI (cytotoxicity). Table 1. Summary ToxProfiler assay results. Using the POD plotter software, significant concentration dependent effects are quantified. A red color indicates a significant induction of the corresponding readout, a green color indicates that there was no significant induction of the corresponding ToxProfiler readout in the tested concentration range. Compound name Oxidative Cell cycle ER stress Auto Ion stress Protein Inflam Cyto stress stress phagy stress mation toxicity PFDA PFNA PFOA HFPO-DA (GenX) PfPeA HFBA PFPrA TFA Na Conclusions: The validity of the ToxProfiler assay was confirmed by treatment to various reference compounds and assessing the specificity of the different reporter cell lines. For all significant responses a point of departure (PoD) was determined. Cytotoxicity was observed for 6/8 compounds in the tested concentration range. ER stress was the most common (6/8) stress pathway to be induced. Autophagy was induced by 5/8 compounds. 4/8 compounds significantly induced cell cycle stress and protein stress. Oxidative stress was induced by 3/8 compounds. 2/8 compounds induced ion stress. None of the stress pathways was induced by PFPrA and TFA Na nor was cytotoxicity observed till 10mM. A clear correlation was found between (ToxProfiler) stress pathway activation and carbon length of the tested PFAS. Final version 4 5 September 2023 Background Treatment to chemicals can lead to cell injury. Such cell injury initiates cell stress response signaling pathways. Strong activation of these pathways reflects the onset of toxicity. The ToxProfiler platform consists of a panel of seven reporter cell lines cultured as a 2D monolayer, in which biomarkers of several of such cell stress response pathways have been tagged with a fluorophore (GFP)1,2. With this platform we can dynamically observe activation of these stress pathways (upon chemical treatment) using a high throughput confocal imaging system1,2,3. The in-house developed automated image acquisition and process pipeline allows for a time efficient and high throughput screening1,2,3,4. The data is presented in concentration response plots from which point of departures (PoDs) are derived. The significant stress pathway responses are visualized all together in a hierarchical clustering with a heatmap, the resulting stress signaling fingerprints can then be easily compared, facilitating a chemical read-across based on these features. The analyses provide detailed mechanistic information. In figure 1 an overview of ToxProfiler readouts is presented. Figure 1. Overview of the stress pathways and biomarkers covered by the ToxProfiler reporter system. 1 Wink S, et al. High-content imaging-based BAC-GFP toxicity pathway reporters to assess chemical adversity liabilities. Arch Toxicol. 2017 Mar;91(3):1367-1383. 2 Wink S, et al. Dynamic imaging of adaptive stress response pathway activation for prediction of drug induced liver injury. Arch Toxicol. 2018 May;92(5):1797-1814. 3 Niemeijer M, et al. Systems Microscopy Approaches in Unravelling and Predicting DILI. 2018 In: Drug-Induced Liver Toxicity. Springer New York, pp 611-625. 4 Wink S, et al. Quantitative high content imaging of cellular adaptive stress response pathways... Chem Res Toxicol. 2014 Mar 17;27(3):338-55. Final version 5 5 September 2023 Test protocol The ToxProfiler assay requires a standard cell culture facility, an automated high throughput confocal imager and a dedicated analysis desktop with proprietary software. The ToxProfiler reporter cells were maintained by culturing them in 2D on uncoated cell culture dishes with DMEM medium supplemented with 10% FBS and 100 U/mL penicillin/streptomycin. During the chemical treatment and reporter analysis the ToxProfiler cells were cultured in 384-well imaging plates. Cytotoxicity testing/concentration range finding For chemical testing, first a concentration range finding was performed using parental HepG2 cells (strain HB-8065). Parental HepG2 cells were treated with 14 different concentrations, at a 1/3 log difference between each concentration, of the test substances. The maximum included concentration was 10 mM. However, for most of the compounds, solubility limited the maximal concentration in the test assay. For compounds dissolved in DMSO the maximal concentration was set to 0.4% of the maximal solubility. Cell death was measured as fraction of Propidium Iodide (PI, a stain for late apoptosis/necrosis) positive cells after 24 h treatment. Cadmium Chloride was included as a positive control and used as a validity check for the concentration range finding assay. Based on this data the appropriate concentrations were selected for the ToxProfiler reporter assay. When significant cell death (>15% PI positive cells) was observed, the lowest concentration at which cell death was observed was included as the highest concentration together with the subsequent 6 sub-cytotoxic concentrations. When no cell death was observed the 7 highest concentrations are included for the ToxProfiler assay. ToxProfiler assay The seven independent HepG2 BAC-GFP reporter cell lines were seeded with 8,000-9,000 cells per well in uncoated 384-well imaging plates. Two days after seeding, HOECHST (#33342, a DNA stain) was added. After one hour incubation, the HOECHST medium was removed and the compound treatment was performed in fresh cell culture medium containing the PI stain. Cells were imaged using an Operetta CLS imager at 24 h after treatment. Two images per well were taken (technical replicates) using four different channels (GFP; stress response activation, HOECHST: nuclei, Digital Phase Contrast: cytoplasm, PI: cell death). The GFP levels of the ToxProfiler reporters were quantified in their expected subcellular location (see table 2) and determined on a single cell level using automated segmentation pipelines (see figure 2). This resulted in concentration response graphs, point of departures (PoDs) and hierarchical clustering overviews for easy interpretation of the generated data. Table 2. ToxProfiler reporter lines with thei4r corresponding positive control compounds and timepoints. Reporter Corresponding Pathway + Control compound Subcellular localisation Time point HepG2-SRXN1-GFP Oxidative stress CDDO-me Cytoplasmatic 24 h HepG2-p21-GFP Cell cycle stress Cisplatin Nuclear 24 h HepG2-CHOP-GFP ER stress Tunicamycin Nuclear 24 h HepG2-LC3-GFP Autophagy Amiodarone Cytoplasmatic foci 24 h HepG2-MT1X-GFP Ion stress Cadmium Chloride Cytoplasmatic 24 h HepG2-HSPA1B-GFP Protein stress Cadmium Chloride Cytoplasmatic 24 h HepG2-ICAM1-GFP Inflammation hTNF Cytoplasmatic membrane 24 h Final version 6 5 September 2023 Figure 2. Automated image segmentation and quantification description. A) Example of a concentration response with the raw confocal images of Cadmium Chloride treatment on HepG2-HSPA1B-GFP. B) Example of the automated image segmentation and analysis pipeline. This process is performed on all images. Positive reference treatments with CDDO-me (oxidative stress), cisplatin (cell cycle stress), tunicamycin (ER stress), amiodarone (autophagy), cadmium chloride (ion stress, protein stress and cytotoxicity), human TNF alpha (inflammation) were included in each plate of the respective ToxProfiler reporter, which allowed for plate-wise min-max normalization. Solvent concentrations were kept constant over the concentration range and never exceeded 0.4% for DMSO. All data was corrected for potential auto-fluorescence. For which the parental HepG2s were exposed to the same conditions (compounds/concentrations). The mean fluorescence caused by the compound was then subtracted from the ToxProfiler results of the respective compound/concentration. This approach is only feasible when the auto-fluorescence do not disturb the image segmentation process. This experiment was conducted as a non-GLP study, however general principles to conduct proper scientifically correct in vitro experiments were adhered to, and particular care was taken to proper handling of test compound (stock) solutions to prevent/minimize degradation of the test articles, based on instructions/compound information from the suppliers. Final version 7 5 September 2023 Data analysis For the analysis of the images, a Harmony 5.1 based image segmentation and analysis pipeline was applied that returns the GFP/PI induction in a specific sub-cellular localization and per individual cell (Figure 2). Data quality checks are performed in the harmony software and further processing of the data was performed using a pipeline of custom R-scripts. To be able to compare the GFP reporter activation for different compounds from one screen to another, a min-max plate-wise normalization step was performed using the positive and negative control samples that were included on every plate. Point of departures (PoDs) were calculated using a custom R-package "modelpod". Concentration response curves were fitted with loess regression. The intersect of this fitted curve with the mean of the solvent control plus two times the standard deviation of the concentration-response curve regression gives a value corresponding to the lowest concentration at which we observe a significant (positive or negative) effect (the PoD). Test criteria The ToxProfiler assay is considered to have a positive response when a PoD was calculated. The PoDs were determined not only for reporter responses, but also for PI (cytotoxicity). For each reporter a positive reference control was included, the respective GFP induction of this compound should have a value of at least 2x the standard deviation above the background GFP response induced by the solvent. Final version 8 5 September 2023 Concentrationrange finding results Test parameters A single 24 h continuous exposure to the parental HepG2 cell line. Propidium iodide (PI) dye was used to quantify fraction of dead cells. Chemical information and included concentrations on all the test and reference compounds can be found in table 3 and 4 PFDA and PFNA precipitated in assay medium in the tested concentrations. PFDA, PFNA and PFOA showed auto-fluorescence in the HOECHST channel in only the highest tested concentrations. The observation of compound precipitation and auto-fluorescence did not affect the concentration selection of these compounds. In all cases cytotoxicity determined the selected concentrations Concentration response plots for the cytotoxicity can be found in figure 3. Assay performed following standard protocols. Table 3. Chemical information and concentration ranges for the concentration range finding experiments. Compound name Full name CAS# Cleanrgbtohn MW Provider Product# PFDA Perfluorodecanoic acid 335-76-2 C10 514.08 Sigma-Aldrich 177741-5G PFNA Perfluorononanoic acid 375-95-1 C9 464.08 Sigma-Aldrich 394459-5G PFOA Perfluorooctanoic acid 335-67-1 C8 414.07 Sigma-Aldrich 171468-5G HFPO-DA (GenX) 2(h,3e,p3t,a3f-ltueotrraofplurooproo-x2y-)propanoic acid 13252-13-6 C6 330.05 Enamine EN300-18553200 PfPeA Perfluoropentanoate 2706-90-3 C5 264.05 Sigma-Aldrich 396575-5ML HFBA Heptafluorobutyric acid 375-22-4 C4 214.04 Sigma-Aldrich 164194-5G PFPrA Perfluoropropionic acid 422-64-0 C3 164.03 Sigma-Aldrich 245917-10G TFA Na Sodium trifluoroacetate 2923-18-4 C2 136.01 Sigma-Aldrich 132101-25G Table 4. Chemical information and concentration ranges for the concentration range finding experiments. # Compound reference Solvent Stock conc. (in Lowest conc. Highest conc. # of Difference (name/number) mM) (in M) (in M) Conc. between conc. 1 PFDA DMSO 2500 0.46 10000 14 1/3 log 2 PFNA DMSO 2500 0.46 10000 14 1/3 log 3 PFOA DMSO 2500 0.46 10000 14 1/3 log 4 HFPO-DA (GenX) DMSO 2500 0.46 10000 14 1/3 log 5 PfPeA DMSO 2500 0.46 10000 14 1/3 log 6 HFBA DMSO 2500 0.46 10000 14 1/3 log 7 PFPrA DMSO 2500 0.46 10000 14 1/3 log 8 TFA Na DMSO 2500 0.46 10000 14 1/3 log Internal reference compound* A Cadmium Chloride DMSO 50 0.01 200 14 1/3 log *supplied by Toxys. Reference compound was obtained from SigmaAldrich. Final version 9 5 September 2023 Figure 3. Cytotoxicity measured in 14 concentrations for the concentration range finding experiments. Normalized concentration response plots of cell death data of all compounds included in this project at timepoint 24h as measured with the PI stain in the parental HepG2 (wild type) line. The black dots represent the mean over two technical replicates with error bars representing the standard deviation. The red line indicates the used threshold of cytotoxicity; 15% PI positive cells. The green shaded area represents the 7 selected concentrations for the reporter assay. The pink shaded area shows the concentrations at which either precipitations in assay medium were observed or the compound concentration that causes issues with the HOECHST segmentation. Final version 10 5 September 2023 ToxProfiler assay results Test parameters A single 24 h continuous exposure to the 7 independent ToxProfiler reporter cell lines. Information on the tested concentrations can be found in table 5. Number of independent biological replicates: 3. Number of technical replicates (for each biological replicates): 2. Quantified GFP- and cytotoxicity- concentration response curves can be found in figure 4. A heatmap with a hierarchical clustering with all the ToxProfiler data can be found in figure 5. Points of departures (PoDs) from ToxProfiler endpoints are listed in table 6. Assay performed following standard protocols. Table 5. Concentration ranges for the ToxProfiler reporter assay. # Compound name Stock Lowest conc. Highest conc. # of Conc. Difference between conc. 1 PFDA 2.5 M 2.15 M 215 M 7 1/3 log 2 PFNA 2.5 M 4.64 M 464 M 7 1/3 log 3 PFOA 2.5 M 10 M 1000 M 7 1/3 log 4 HFPO-DA (GenX) 2.5 M 21.54 M 2154 M 7 1/3 log 5 PfPeA 2.5 M 100 M 10000 M 7 1/3 log 6 HFBA 2.5 M 100 M 10000 M 7 1/3 log 7 PFPrA 2.5 M 100 M 10000 M 7 1/3 log 8 TFA Na 2.5 M 100 M 10000 M 7 1/3 log Internal reference compounds* 1 CDDO-me 0.125 mM 0.03 nM 0.5 M 7 5-fold 2 Cisplatin 3.33 mM 0.1 M 10 M 7 1/3 log 3 Tunicamycin 2.5 mM 0.1 M 10 M 7 1/3 log 4 Amiodarone 5 mM 0.2 M 20 M 7 1/3 log 5 Cadmium Chloride 1.08 mM 0.04 M 4.31 M 7 1/3 log 6 human TNF alpha 10 mg/ml 0.6 pg/ml 10 ng/ml 7 5-fold *Supplied by Toxys. All reference compounds are obtained from SigmaAldrich except for human TNF alpha. Human TNF alpha was purchased at R&D systems. Final version 11 Oxidativestress: Cell cyclestress: ERstress: Autophagy: Ionstress: 5 September 2023 Protein stress: Inflammation: Cell death: Final version 12 5 September 2023 Oxidativestress: Cell cyclestress: ERstress: Autophagy: Ionstress: Protein stress: Inflammation: Cell death: Final version 13 Oxidativestress: Cell cyclestress: ERstress: Autophagy: Ionstress: 5 September 2023 Protein stress: Inflammation: Cell death: Final version 14 5 September 2023 Oxidativestress: Cell cyclestress: ERstress: Autophagy: Ionstress: Protein stress: Inflammation: Cell death: Figure 4. Concentration response plots of the 7 ToxProfiler GFP reporters for each compound. The graph on the left represents the quantified and normalized GFP responses in which each color represents one of the 7 ToxProfiler GFP reporters. The curves on the right represents cell death data, in which the black dots represent the average fraction of cell death (PI positive cells) over three biological replicates. The error bars represent the standard deviation over the three biological replicates. Final version 15 5 September 2023 Figure 5. Heatmap of all the quantified and normalized data. Each row represents a compound with 7 concentration levels in which 7 represents the highest concentration. The different columns represent the different readouts: the 7 ToxProfiler reporters and PI. The red gradient indicates an induction of the integrated GFP intensity/PI signal as compared to the solvent control. A white color indicates similar levels as the control. Note that the different chemicals are not screened on an equimolar level but rather on an equitoxic level. The chemicals are ordered based on their carbon length with the C10 on top. Table 6. Point of departures (PoDs) per reporter with cytotoxicity information for all compounds. For each compound the lowest PoD was marked with a bold/italic/underlined font to indicate the primary response. If the difference between the lowest POD and the successive POD of that chemical was <2fold also the second POD was highlighted. PoDs are shown in M. An empty well indicates that the compound at the tested concentration range did not significantly affect that specific readout. Compound Carbon ToxProfiler (TP) stress pathway Cytotoxicity TP Name Length Oxidative Cell cycle ER Autophagy Ion Protein Inflammation Rank SRXN1 P21 CHOP LC3 MT1X HSPA1B ICAM1 PI PFDA C10 134 151 153 151 1 PFNA C9 150 150 150 353 251 2 PFOA C8 215 450 215 300 300 450 300 3 HFPODA C6 650 1000 700 300 1379 4 PfPeA C5 4642 3000 1000 2200 5 HFBA C4 7000 6000 2154 7700 6 PFPrA C3 7-8 TFA Na C2 7-8 Final version 16 5 September 2023 Conclusions and discussion Overall conclusions per compound In table 7 an overall conclusion per compound is presented that is based on cell count. PI and GFP reporter data. Only when the effect was significant (and thus a PoD was determined) the effect was listed in the overall conclusion table. Table 7. Overview of the primary toxic properties of the tested compounds. The conclusions are based only on the ToxProfiler assay results and only on the tested concentration range (table 4). Compound name Brief overall conclusion PFDA Compound induces ER stress in a concentration dependent manner at 134 M. Ion stress, protein stress and cytotoxicity are induced at 150M. PFNA PFOA HFPODA (GenX) PfPeA HFBA PFPrA TFA Na Compound induces autophagy, oxidative stress and ER stress in a concentration dependent manner at 150M. Protein stress is induced at 353M. Cytotoxicity is induced at a concentration of 251M. Compound induces oxidative stress and ER stress in a concentration dependent manner at 215M. Ion stress and autophagy are induced at 300M. Both cell cycle stress and protein stress are observed at 450M. Cytotoxicity is induced at a concentration of 300M. Compound affects the autophagy response in a concentration dependent manner at 300M. Oxidative stress, ER stress and cell cycle stress are induced at 650M, 700M and 1000M, respectively. Cytotoxicity is induced at a concentration of 1379M. Compound affects the autophagy response in a concentration dependent manner at 1000M. ER stress and cell cycle stress are induced at 3000M and 4642M, respectively. Cytotoxicity is induced at a concentration of 2200M. Compound affects the autophagy response in a concentration dependent manner at 2154M. ER stress and cell cycle stress are induced at 6000M and 7000M, respectively. Cytotoxicity is induced at a concentration of 7700M. This compound did not induce any of the ToxProfiler endpoints in the tested concentration range. Also, no cytotoxicity was found in the tested concentrations (up to 10mM). This compound did not induce any of the ToxProfiler endpoints in the tested concentration range. Also, no cytotoxicity was found in the tested concentrations (up to 10mM). Final version 17 5 September 2023 Annex I: Concentration response curves per reporter/chemical Final version 18 5 September 2023 Final version 19 5 September 2023 Final version 20 5 September 2023 Supplemental figure 1. Concentration response plots of the test compounds per ToxProfiler readout. On the X-axis the concentrations are depicted (in M). On the y-axis the plate-wise normalized biomarker-GFP Final version 21 5 September 2023 measurements or the fraction PI positive (dead) cells are shown. Error bars represent the standard deviation over the three biological replicates. Final version 22