Case study in3 DrugMatrix: Toxicogenomic Signatures of Kidney Toxicity: Pinpointing 109 Differentially Expressed Genes in Rat DrugMatrix Data

Executive Summary

The case study leveraged rat in vivo transcriptomics data from the DrugMatrix Database to investigate mechanisms of nephrotoxicity. The primary objective was to perform biomarker discovery by identifying differentially expressed genes (DEGs) specific to toxic compounds compared to non-toxic compounds. The study harmonized data for several nephrotoxic and non-nephrotoxic compounds based on dose and exposure duration. Analysis revealed 109 DEGs found exclusively in the toxic compounds group. Two specific genes, ECI1 and ACOT12, were proposed as potential biomarkers for nephrotoxicity, with their upregulation observed in nephrotoxic compounds. These findings contribute to enriching existing AOPs, such as AOP 256, related to kidney toxicity.

Methods:

Data Collection and Selection

The transcriptomics data used originated from the DrugMatrix Database, focusing on rat in vivo data for both nephrotoxins and non-nephrotoxins. The initial step involved compiling a reference list of nephrotoxic and non-nephrotoxic compounds for which sufficient toxicogenomics data was available in the DrugMatrix database and literature.

• Toxic Compounds included Amikacin, Doxorubicin, Diclofenac, Cisplatin, Carbon Tetrachloride, Cadmium Chloride, Gentamicin, Tobramycin, Vancomycin, and Neomycin.

  • Example exposure parameters: Cadmium Chloride was tested at 35 µmol/kg for 1 and 5 days, and Gentamicin at 200 mg/kg for 1 and 5 days.

• Non-Toxic Compounds included Allopurinol, Dexamethasone, Captopril, and Lovastatin.

  • Example exposure parameters: Allopurinol was tested at 50 mg/kg for 1, 3, and 5 days, and Lovastatin at 15 mg/kg for 0.25, 1, 3, and 5 days.

    
    

Data Harmonization and Analysis

The collected in vivo data underwent harmonization based on experimental conditions, particularly dose (MTD) and duration of exposure.

The analysis workflow involved:

1. Limma analysis applied to the rat in vivo DrugMatrix data.

2. Selection criteria for identifying differentially expressed genes (DEGs) required a p-value < 0.05 and a log2 fold change > 1 or < -1.

3. The core goal was to derive common DEGs shared across the toxic compounds and then identify DEGs exclusive to the toxic group (by comparing toxic DEGs against toxic + non-toxic common DEGs).

4. Pathways enrichment analysis was performed using the findings, referencing AOPwiki.

   
   

Results:

Differentially Expressed Genes (DEGs)

The overall Limma analysis of the Affymetrix data resulted in numerous DEGs for individual compounds, such as Amikacin (63 genes), Gentamicin (721 genes), Cisplatin (595 genes), Allopurinol (2462 genes), Lovastatin (4338 genes), and Dexamethasone (1008 genes). The central finding was the identification of 109 differentially expressed genes that were only found in the toxic compounds group.

Pathway Mapping and Biomarkers

Linking these 109 DEGs to biological pathways highlighted several affected areas, including:

• Metabolism.

• DNA repair.

• Cell Cycle.

• Immune system.

• Gene Expression and Signal transduction.

• Transport of small molecules.

• Reproduction and Programmed cell death.

• Neuronal System.

• Specific metabolic pathways included Metabolism of lipids, Fatty acid metabolism, and Mitochondrial Fatty Acid beta oxidation.

Two specific genes were discovered as potential biomarkers for nephrotoxicity: ECI1 and ACOT12. These genes showed upregulation when exposed to nephrotoxic compounds.

Discussion:

The case study successfully demonstrated the use of bioinformatics workflows to mine large toxicogenomics datasets (DrugMatrix) for biomarker discovery in the context of nephrotoxicity.

Linking Findings to Adverse Outcome Pathways (AOPs)

The identified biomarkers, ECI1 and ACOT12, provide potential Key Events (KEs) that can be used to enrich existing AOPs or propose new ones related to kidney toxicity. Biomarker discovery helps make AOPs more understandable, specific, and predictive. The DEG set (including ECI1 and ACOT12) maps to pathways relevant to AOP-256 (mitochondrial dysfunction) and could support biomarker-hypothesis generation and may inform future AOP evidence entries. The biological pathways linked to the 109 toxic-specific DEGs, particularly Mitochondrial Fatty Acid beta oxidation, serve as crucial information for mapping the AOP network.

Future Validation and Comparison

The initial findings based on the in vivo rat data (Case 1) proposed ECI1 and ACOT12 as candidate nephrotoxicity markers derived from DrugMatrix analysis; further in vitro/in vivo validation studies were then planned. (Note: There is emerging work linking peroxisomal fatty-acid metabolism and kidney injury; a 2025 study implicates ACOT12 mechanistically in kidney damage contexts. Clear, direct evidence of ECI1 as a nephrotoxicity biomarker is limited in current literature (it’s a peroxisomal/mitochondrial β-oxidation enzyme; expression can change, but validation as a diagnostic/safety biomarker isn’t established). The logical next step in the research framework (Case 2) is to validate and concretize this hypothesis through in vitro studies.