Case Study in3 Paroquat Toxicity

Beyond the Cage: Human iPSC Multi-Organ Models Validate Next-Generation Chemical Risk Assessment

1. Motivation and Context

Need for New Approach Methodologies (NAMs): Current chemical risk assessment, largely based on OECD guidelines, often involves tests performed on animals, raising ethical, financial, and scientific concerns. Governmental agencies worldwide are promoting the development of human-based assays to achieve the goals of the 3Rs (Replacement, Reduction, and Refinement of Animal Experiments).

The in3 Project: This multi-organ hiPSC strategy was developed within the framework of the in3 project (Marie Skłodowska-Curie Action-Innovative Training Network). The in3 project aims toward the synergistic development and utilization of in vitro and in silico tools for human chemical and nanomaterial safety assessment. The organs targeted by the in3 consortium include the kidney, liver, brain, lung, and blood-brain barrier (BBB).

Role of Edelweiss Connect: Researchers affiliated with Edelweiss Connect GmbH were instrumental in this study, particularly in data management and bioinformatics. The raw read count data and corresponding metadata files generated from the toxicity assessment were specifically uploaded and maintained on an internal instance of the EdelweissData™ management system. This system facilitated easier accessibility and ensured data interoperability for reuse within the broader in3 network

Advantage of hiPSCs: Human induced pluripotent stem cells (hiPSCs) offer a valuable tool for NAMs because they are a renewable source of cells that can be differentiated into human-relevant phenotypes, overcoming limitations associated with immortalised cell lines or primary cells (which are hard to obtain in quantity). This approach eliminates interspecies bias in toxicity assessment.

2. Strategy and Methodology

Multi-Organ In Vitro Strategy: The researchers proposed an in vitro multi-organ strategy using hiPSCs to reliably assess chemical toxicity.

Models Used: hiPSCs from two donors were differentiated into models representing various tissues/organs:

• Brain: 2D Neural Cells (NC) and 3D BrainSpheres (BS).

• Blood-Brain Barrier (BBB): Brain-Like Endothelial Cells (BLECs).

• Kidney: Podocytes (PODO) and Proximal Tubular-Like cells (PTL).

• Liver: Hepatocyte-Like Cells (HLC).

• Vasculature: Endothelial Cells (EC) (Note: The EC model was later removed due to unreliable low read counts during TempO-Seq analysis).

Case Study Compound: Paraquat (PQ), a widely used herbicide with known toxic effects on the kidneys and brain, was chosen as the reference chemical.

Assessment Method: Models were exposed to PQ for acute toxicity testing (24 h and/or 48 h). Cytotoxicity curves were established, and sub-cytotoxic concentrations were selected for transcriptomic analysis using TempO-Seq. TempO-Seq utilized a panel of 3565 probes to reveal deregulation of gene expression and pathways.

3. Key Findings

Differential Sensitivity to Paraquat

The models exhibited differential cytotoxic sensitivity to acute PQ exposure, confirming that sensitivity depends on the specific cell type and organ modelled:

• Most Sensitive: 2D Neural Cells (NC) were the most sensitive (IC50: 49 μM).

• Brain Models: The 3D brain model (BS) was less sensitive than the 2D model (NC) (IC50: 200 μM at 24 h).

• Peripheral Organs: Hepatocyte-Like Cells (HLC) appeared to be the least sensitive (IC50: 1093 μM at 48 h).

• Potential Explanations: The differential sensitivity might be related to the higher basal expression of amino acid transporters (SLC3A2, SLC7A11) in the most sensitive models (BS, NC, PODO), which may facilitate PQ uptake. Conversely, higher expression of anti-oxidant defence enzymes (SODs and GSTs) in HLC might explain its relative resistance.

  
  

Shared Mechanisms of Action

Transcriptomic analysis identified pathways deregulated by PQ exposure across multiple models, aligning with existing toxicological knowledge. No pathway was shared by all six remaining models, but several were shared by five or four models:

1. Oxidative Stress via Nrf2: The pathway "Oxidative stress induced gene expression via Nrf2 markers" was found deregulated in five cell models. Key stress response genes belonging to this pathway, such as HMOX1, NQO1, MAFF, and MAFG, showed concentration-dependent increases in expression across all six models tested, confirming oxidative stress as a major mechanism of action for PQ.

2. Unfolded Protein Response (UPR): The UPR pathway was highlighted in four out of the six models. Genes associated with UPR, including ATF4, PPP1R15A, and DDIT3, showed clear concentration-dependent increases in expression in five models (excluding PODO).

3. ESR-mediated signaling: The pathway "ESR-mediated signaling" (Estrogen Receptor-mediated signaling) was revealed in four cell culture systems, including both brain models (NC and BS). This mechanism, recently proposed in PQ toxicity, suggests the strategy can detect both known and novel mechanisms.

   
   

Organ-Specific Responses

While generalized stress pathways were shared, each model also exhibited specific deregulation, suggesting organ-specific reactions to PQ:

• Brain Models (NC, BS): Top deregulated genes included those involved in development (SOX1, INSM1, THRB), suggesting the brain models possess the weakest self-defense capacity compared to peripheral organs.

• Kidney/Brain Models: Upregulation of Vascular Endothelial Growth Factor A (VEGFA) was shared by both brain models and both kidney models, confirming its relevance as a potential marker of chemical injury in these specific organs.

  
  

4. Conclusions and Significance

The proposed multi-organ hiPSC-derived strategy, coupled with cost-effective TempO-Seq transcriptomics, successfully assessed chemical toxicity and elucidated both known and less known mechanisms of PQ toxicity.

Advantages: The strategy allows parallel assessment of chemical toxicity on multiple tissues exclusively using human cells, enabling a better evaluation of differential sensitivity and eliminating interspecies bias.

Future Improvements: To further enhance chemical risk assessment for human health, the strategy should incorporate hiPSC-derived models for other organs (e.g., heart and lungs), strive for standardization of procedures across laboratories, and include an evaluation of the distribution kinetics of chemicals to facilitate in vitro to in vivo extrapolation (IVIVE).