Identification and validation of key genes of Xiaozhengtongluo formula in the treatment of diabetic nephropathy by Mendelian randomization

Technical Summary: Identification and Validation of Key Genes of Xiaozhengtongluo Formula in the Treatment of Diabetic Nephropathy by Mendelian Randomization

Problem Statement
Diabetic nephropathy (DN) is a leading cause of chronic kidney disease and hospitalization, yet current pharmacological interventions often fail to halt its progression. The Xiaozhengtongluo (XZTL) formula, a traditional Chinese medicine (TCM) compound based on the theory of "phlegm and blood stasis," has demonstrated clinical efficacy in treating DN. However, its precise molecular targets and mechanisms of action remain poorly defined. The study aims to identify the key genes through which XZTL exerts its therapeutic effects on DN and to validate these findings using a combination of network pharmacology, Mendelian randomization (MR), and experimental verification.

Methodology
The study employed a multi-step integrative approach combining bioinformatics, causal inference, and wet-lab validation:

1. Data Collection and Screening:

   • Differentially Expressed Genes (DEGs): Transcriptomic data from GEO datasets (GSE96804 and GSE30528) were analyzed to identify DEGs between DN and control glomerular tissues.
   • Target Acquisition: Drug target genes for XZTL were retrieved from the TCMSP database based on active ingredients with high oral bioavailability and drug-likeness. Disease target genes for DN were obtained from GeneCards and the Therapeutic Target Database (TTD).
   • Candidate Gene Identification: The intersection of DEGs, drug targets, and disease targets yielded 29 candidate genes.

2. Mendelian Randomization (MR) Analysis:

   • To establish causal relationships, expression quantitative trait loci (eQTL) data were used as instrumental variables (IVs).
   • Five MR algorithms (MR Egger, Weighted median, Inverse variance weighted [IVW], Weighted mode, Simple mode) were applied. The IVW method served as the primary analysis.
   • Sensitivity analyses, including heterogeneity tests, horizontal pleiotropy assessments, and leave-one-out (LOO) analyses, were conducted to ensure result reliability. The Steiger test was used to rule out reverse causality.

3. Bioinformatics and Network Analysis:

   • Diagnostic Value: Receiver Operating Characteristic (ROC) curve analysis was performed on the candidate genes to evaluate their diagnostic potential (AUC > 0.7).
   • Functional Enrichment: Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Gene Set Enrichment Analysis (GSEA) were conducted to identify enriched pathways.
   • Regulatory Networks: Molecular regulatory networks (lncRNA-miRNA-mRNA) and transcription factor (TF) interactions were constructed using databases like DIANA-microT, miRDB, starBase, and miRNet.
   • Molecular Docking: The binding affinity between the identified key genes and the top active components of XZTL was verified using CB-Dock2.

4. Experimental Validation:

   • A DN rat model was established using a high-fat diet and streptozotocin (STZ) injection.
   • Sprague-Dawley rats were divided into control, DN, and XZTL intervention groups.
   • Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) was used to validate the expression levels of the identified key genes in renal glomerular tissues.

Key Contributions and Results

• Identification of Key Genes: From 29 initial candidates, MR analysis narrowed the focus to three key genes: APOD, IGFBP3, and LPL.
  • IGFBP3 was identified as a risk factor for DN (upregulated).
  • LPL and APOD were identified as protective factors (downregulated in DN).
• Causal Validation: MR analysis confirmed a statistically significant causal link between these three genes and DN. Sensitivity analyses (heterogeneity, pleiotropy, LOO) and the Steiger test supported the robustness of these causal inferences.
• Diagnostic Potential: ROC curve analysis in both GSE96804 and GSE30528 datasets showed that all three genes had Area Under the Curve (AUC) values greater than 0.7, suggesting their utility as diagnostic biomarkers.
• Pathway Enrichment: The three key genes were co-enriched in 26 pathways, including oxidative phosphorylation, heme metabolism, fatty acid metabolism, and epithelial-mesenchymal transition.
• Regulatory Mechanisms:
  • IGFBP3 is regulated by the MIR497HG/hsa-miR-19a-3p axis and co-regulated by transcription factors SP1 and SP3.
  • LPL is regulated by the NEAT1/hsa-miR-29a-3p axis and co-regulated by SP1 and SP3.
  • APOD is regulated by the transcription factor NOTCH3.
• Molecular Docking: The study demonstrated strong binding affinities between the key genes and specific XZTL components:
  • IGFBP3 binds to Quercetin (-7.5 kcal/mol).
  • LPL binds to Oleic acid (-6.5 kcal/mol).
  • APOD binds to Baicalein (-6.1 kcal/mol).
• Experimental Confirmation: RT-qPCR results in the rat model confirmed the bioinformatics findings: IGFBP3 was upregulated, while LPL and APOD were downregulated in the DN group. XZTL intervention significantly restored the expression of these genes toward normal levels.

Significance and Claims
The paper claims to provide a comprehensive reference for the mechanism of XZTL in treating DN. By integrating network pharmacology with Mendelian randomization, the study moves beyond correlation to suggest causal relationships between specific genes and DN pathology. The identification of APOD, IGFBP3, and LPL as core targets offers:

1. Mechanistic Insight: A molecular basis for XZTL's renoprotective effects, specifically through the regulation of lipid metabolism, oxidative stress, and energy metabolism pathways.
2. Therapeutic Targets: Validation of these genes as potential targets for DN therapy, with specific active components (Quercetin, Oleic acid, Baicalein) identified as the likely mediators.
3. Diagnostic Value: The potential use of these three genes as novel biomarkers for the early screening and monitoring of DN, given their high diagnostic accuracy (AUC > 0.7).

The authors acknowledge limitations, including the exclusion of CAV1 due to SNP distribution issues and the reliance on database predictions for regulatory networks, which require further cellular and animal elucidation. However, the study concludes that these findings offer a scientific basis for the clinical application of XZTL and suggest directions for future drug development and multi-gene diagnostic models.