Summarizing the Effective Herbs for the Treatment of Hypertensive Nephropathy by Complex Network and Machine Learning

Jia-Ming Huan, Wen-Ge Su, Wei Li, Chao Gao, Peng Zhou, Fu Chun-sheng, Xiaofeng Wang, Yimin Wang, Yi-Fei WangEvidence-based 2021 · DOI

TL;DR

This paper leverages complex network analysis and machine learning on electronic medical records to identify 14 key traditional Chinese medicine (TCM) herbs effective for treating hypertensive nephropathy (HN). By analyzing herb compatibility, targets, and biological pathways, the study elucidates potential mechanisms for these TCM treatments, offering a data-driven approach to understanding and optimizing HN management with TCM.

What Problem Does This Paper Address?

Hypertensive nephropathy (HN) is a significant complication of hypertension, leading to kidney damage. While traditional Chinese medicine (TCM) has a long history of use in treating HN, the specific effective prescriptions and their underlying therapeutic approaches have not been systematically summarized or analyzed. This lack of comprehensive understanding hinders the integration and optimization of TCM for HN management. The paper aims to address this by using modern computational methods to mine TCM data, identify effective herbs and their combinations, and explore their biological mechanisms.

Methods

The study employed a multi-faceted approach combining data mining, network analysis, and machine learning:

  • Complex Network Analysis: Electronic medical records (EMRs) of 1499 TCM prescriptions for HN were analyzed. A network was constructed where herbs were nodes, and their co-occurrence in prescriptions formed weighted edges. Hierarchical network extraction was used to identify core herb combinations.
  • Apriori Algorithm: This algorithm was used to analyze the compatibility rules (association rules) between herbs, identifying frequently co-prescribed herb pairs.
  • Relative Risk (RR) and Chi-Square Test: These statistical methods were used to identify herbs that were significantly more likely to be used in HN patients compared to non-HN patients, highlighting potentially specific HN-treating herbs.
  • Machine Learning (Clustering): Hierarchical clustering algorithms (Ward's method) were applied at multiple levels:
  • TCM Symptom Level: Herbs were clustered based on their association with specific TCM symptoms related to HN.
  • Target Level: Herbs were clustered based on their known molecular targets.
  • Pathway Level: Herbs were clustered based on their enrichment in specific GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) signaling pathways relevant to HN.
  • Data Mining and Database Integration: Public databases (TCMSP, SymMap, ETCM, PubMed, CNKI, NCBI GEO, OMIM, DisGeNET, Malacards) were used to gather information on herb compounds, targets, and HN-related genes.
  • Random Walk (RWR) Algorithm: Used to evaluate the importance of compounds by measuring their "closeness" to HN-related targets within a biological network.

Datasets

  • Electronic Medical Records (EMRs): Data from 30,695 anonymous patients diagnosed with hypertension at the Affiliated Hospital of Shandong University of Traditional Chinese Medicine (July 2014 - May 2017), specifically focusing on 2055 patients diagnosed with HN. This dataset was used to extract 1499 TCM prescriptions.
  • Public Databases:
    • Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP)
    • SymMap
    • Encyclopedia of Traditional Chinese Medicine (ETCM)
    • PubMed and CNKI databases (for biomedical literature)
    • NCBI GEO database (for gene expression data, specifically GSE99325)
    • OMIM, DisGeNET, and Malacards (for human disease-related genes)
    • STRING database (for protein-protein interaction networks)

Key Insights and Ideas

  • Data-Driven Identification of Core Herbs: The study successfully moved beyond anecdotal evidence to identify a specific set of 14 herbs that are consistently used and likely effective for treating HN, based on real-world prescription data.
  • Multilevel Mechanistic Understanding: By analyzing herbs at the symptom, target, and pathway levels, the research provides a more holistic view of how these TCM components might work. This multi-dimensional approach is crucial for understanding complex diseases like HN.
  • Network Pharmacology as a Tool: The application of complex network analysis and machine learning demonstrates the power of these computational approaches in deciphering the intricate relationships within TCM prescriptions and their biological effects.
  • Synergistic Action of Herbs: The identification of compatibility rules (Apriori algorithm) suggests that the effectiveness of TCM for HN likely stems from the synergistic action of multiple herbs rather than individual components.
  • Targeting Key Inflammatory Pathways: The findings point towards TCM herbs modulating critical inflammatory pathways (e.g., TNF, IL1B, IL6, NF-kB, MAPK) and the Renin-Angiotensin-Aldosterone System (RAAS), which are known contributors to kidney damage in HN.

Key Results

  • 1499 TCM Prescriptions Analyzed: The study analyzed a substantial dataset of TCM prescriptions for HN.
  • 14 Effective Herbs Identified: The core set of herbs identified for treating HN includes: huangqi, danshen, dangshen, fuling, baizhu, danggui, chenpi, banxia, gancao, qumai, cheqianzi, ezhu, qianshi, and niuxi.
  • Significant Herb Compatibility: 41 important association rules were found, highlighting common and effective herb pairings.
  • Modulation of Inflammatory Pathways: The identified herbs were found to downregulate inflammatory cytokines (TNF, IL1B, IL6) and signaling pathways (NF-kB, MAPK), reducing renal inflammatory damage.
  • Hierarchical Clustering Reveals Groupings: At different levels (symptoms, targets, pathways), the 14 herbs could be grouped into 3 or 4 clusters, indicating shared or complementary therapeutic roles.
  • Identification of Core Compounds: 241 effective core compounds were identified, with examples like Quercetin and Scopoletin highlighted.

Limitations and Future Work

  • Reliance on Public Databases: The study acknowledges that compound, target, and gene information was primarily sourced from public databases, and there's a need for more experimental validation (proteomics, metabonomics).
  • Experimental Validation Needed: While the study proposes mechanisms, direct experimental studies are required to confirm the effects of these herbs on specific pathways like NF-kB, cAMP, and mTOR.
  • Limited Scope of EMR Data: The EMR data is from a specific hospital over a defined period, which might limit generalizability.
  • Future Directions: The authors suggest that this data mining approach can provide a direction for future clinical practice and experimental research, emphasizing the need for further studies to validate the proposed mechanisms and optimize TCM treatments for HN.

Summary updated

Abstract

Hypertensive nephropathy is a common complication of hypertension. Traditional Chinese medicine has been used in the clinical treatment of hypertensive nephropathy for a long time, but the commonly used prescriptions have not been summarized, and the basic therapeutic approaches have not been discussed. Based on data from 3 years of electronic medical records of traditional Chinese medicine used at the Affiliated Hospital of Shandong University of Traditional Chinese Medicine, a complex network and machine learning algorithm was used to explore the prescribed herbs of traditional Chinese medicine in the treatment of hypertensive nephropathy (HN). In this study, complex network algorithms were used to describe traditional Chinese medicine prescriptions for HN treatment. The Apriori algorithm was used to analyze the compatibility of these treatments with modern medicine. Data on the targets and regulatory genes related to hypertensive nephropathy and the herbs that affect their expression were obtained from public databases, and then, the signaling pathways enriched with these genes were identified on the basis of their participation in biological processes. A clustering algorithm was used to analyze the therapeutic pathways at multiple levels. A total of 1499 prescriptions of traditional Chinese medicines used for the treatment of hypertensive renal damage were identified. Fourteen herbs used to treat hypertensive nephropathy act through different biological pathways: huangqi, danshen, dangshen, fuling, baizhu, danggui, chenpi, banxia, gancao, qumai, cheqianzi, ezhu, qianshi, and niuxi. We found the formulae of these herbs and observed that they could downregulate the expression of inflammatory cytokines such as TNF, IL1B, and IL6 and the NF-<i>κ</i>B and MAPK signaling pathways to reduce the renal inflammatory damage caused by excessive activation of RAAS. In addition, these herbs could facilitate the deceleration in the decline of renal function and relieve the symptoms of hypertensive nephropathy. In this study, the traditional Chinese medicine approach for treating hypertensive renal damage is summarized and effective treatment prescriptions were identified and analyzed. Data mining technology provided a feasible method for the collation and extraction of traditional Chinese medicine prescription data and provided an objective and reliable tool for use in determining the TCM treatments of hypertensive nephropathy.
BibTeX
@inproceedings{huan2021summarizing,
  author = {Jia-Ming Huan and Wen-Ge Su and Wei Li and Chao Gao and Peng Zhou and FU Chun-sheng and Xiaofeng Wang and Yimin Wang and Yi-Fei Wang},
  title = {{Summarizing the Effective Herbs for the Treatment of Hypertensive Nephropathy by Complex Network and Machine Learning}},
  booktitle = {Evidence-based},
  year = {2021},
  doi = {10.1155/2021/5590743},
}
Figure