Case Study: Temporal Transcriptomic Analysis of Cadmium Chloride Nephrotoxicity

1. Introduction: Context, Motivation, and Project Framework

Nephrotoxicity Challenge: Nephrotoxicity, or kidney injury, is a frequent outcome of exposure to a wide variety of xenobiotics and can result in acute and chronic kidney damage. Traditional toxicological assessment methods, such as animal in vivo studies, are limited by poor predictive capacity and a lack of human relevance.

Advantage of hiPSC Models: A paradigm shift towards in vitro testing offers new opportunities to understand toxicity mechanisms. While existing renal cell lines (e.g., RPTEC/TERT1) are useful, they do not account for genetic and epigenetic variations inherent to the human population. Human induced Pluripotent Stem Cells (iPSCs) serve as a promising tool to overcome these limitations by offering the potential to study genetic variability across individuals. This study specifically utilized hiPSC-derived renal Proximal Tubular Epithelial-Like (PTL) cells.

Project Context (in3 and Edelweiss Connect): This research was conducted to demonstrate the robustness and long-term stability of the PTL model for high-throughput nephrotoxicity screening. The study was performed within the framework of the in3 project (Marie Skłodowska-Curie Innovative Training Networks). To ensure data interoperability and reusability across this network, the raw read count data and corresponding metadata files were specifically uploaded and maintained on an internal instance of the EdelweissData™ data management system. This system, managed by researchers affiliated with Edelweiss Connect GmbH, provides advanced functionality, including an application programming interface (API), to allow filtering and access to data.

2. Methodology and Chemical Exposure

Cell Model and Technology: The study utilized hiPSC-derived PTL cells exposed to the environmental nephrotoxicant Cadmium Chloride (CdCl2). Transcriptomic analysis was performed using the TempO-Seq methodology and the EUTox v2.1 panel (3565 probes). TempO-Seq is a cost-effective, high-throughput transcriptomics technology used to quantify transcript changes and identify genes activated in a sequential, time-dependent manner.

Temporal Exposure Design: PTL cells were repeatedly exposed to a sublethal concentration of 5 µM Cadmium Chloride every 24 hours. Samples were collected over an extended timeline, including early time points (1h, 2h, 4h, 6h, 8h, 12h, 16h, 20h, 24 h) and late time points (72 h and 168 h).

Data Analysis: Normalisation and differential expression analysis were performed using the DESeq2 library. A threshold of adjusted p-value 0.05 and log2 fold changes >2 or <-2 was used to select significantly differentially expressed genes (DEGs).

3. Key Findings: Temporal Gene Expression and Mechanisms

Differential Expression and Early Response: Quality control confirmed good sample quality for analysis. Applying the defined cutoffs resulted in 115 significantly differentially expressed genes (DEGs).

• The genes showing the highest fold changes were Metallothioneins (up to 12), HSPA6 (13), and HMOX1 (10).

• These maximum changes occurred early, specifically between 4h and 8h.

Mechanistic Insights (Pathways and Transcription Factors): Analysis identified specific stress response pathways activated by CdCl2 exposure:

• Metal Response Pathway: This is one of the most well-known pathways activated by Cadmium Chloride exposure. Metallothioneins (MTs) showed very high expression levels and drastic increases starting at the first hour, reaching a maximum around 6 hours, and then levelling off to a plateau that remained significantly above ground level until and beyond 24 hours. This sustained high expression supports the idea that MTs are constantly binding to Cadmium to form inert complexes and detoxify the cells. Metal response was classified as an early and late-responder mechanism.

• Oxidative Stress Pathway (Nrf2): Activation of Nrf2-mediated oxidative stress is a known response to kidney toxicity. Ingenuity Pathway Analysis (IPA) identified the "NRF2-mediated Oxidative Stress Response" as a major pathway with the highest significance. Most oxidative stress-related genes were activated in the early hours. Unlike metallothioneins, many Nrf2-related genes show a maximum followed by a decrease back toward an unperturbed state by 24 hours.

• ER Stress Response: Endoplasmic Reticulum (ER) stress response was also clearly classified as an early-responder mechanism. This pathway includes ATF4 activation of genes in response to endoplasmic reticulum stress.

• Late Effects: Genes associated with apoptosis (e.g., FOSL1, DUSP1, PCNA) and peroxisome proliferative signalling were more related to long-term exposure (72h and 168h) or a more general negative effect of long incubation time on the cell model.

  
  

4. Advanced Temporal Analysis

Limitations of Standard Clustering: Standard data-driven clustering methods, such as Hierarchical Clustering and STEM Clustering, were found to be suboptimal for analyzing the temporal profiles.

• Hierarchical Clustering failed because it concentrates only on the absolute fold changes, leading to the separation of highly expressed genes (like Metallothioneins) into different clusters despite similar profiles, or grouping dissimilar profiles together if they share small expression levels.

• STEM Clustering, while designed for short time series, used profiles that were too general for this application, grouping genes that had continuous activation (plateau) together with genes that quickly returned to ground level.

Development of New Profile Fitting: To overcome these limitations, a new method was developed to fit temporal expression levels to profiles modelled by standard mathematical functions (including Maxwell, Gamma, and cumulative distributions).

• This procedure successfully grouped gene behaviours into distinct patterns, such as genes showing a strong maximum followed by a return to baseline (Behaviour 1), or genes reaching a maximum and then sustaining an elevated plateau until the end of the measurement period (Behaviour 2). This allowed for a more accurate biological interpretation related to the trajectory of stress and repair mechanisms.

  
  

5. Conclusion and Significance

The study successfully demonstrated the utility of hiPSC-derived PTL cells and TempO-Seq for studying the temporal effects of Cadmium Chloride. The time-resolved transcriptomic data provided critical insights into the sequential activation of stress response and repair mechanisms, such as Metal Response, Oxidative Stress (Nrf2), and ER stress. The work also highlights the necessity of developing custom methodologies, like the novel profile fitting procedure, to adequately handle the complexity and noise of time-series omics data derived from in vitro exposure scenarios.