For a long time, there have been two major challenges in understanding the mechanistic roots of traditional Chinese medicine, which have created a situation where it is "difficult to be recognized".
i) The lack of scientific basis for the classical theories of TCM hinders the understanding of TCM from a modern biomedical perspective;
ii) The complexity of the chemical composition of Chinese herbal medicines and the protein targets of the chemicals are often unknown, so it is not feasible to use the traditional method of "brute force attack" to screen Chinese herbal chemicals.
Understanding traditional and natural medicine can lead to world-changing drug discovery. Despite the best effects of individual herbs, traditional Chinese medicine lacks a scientific basis and is often considered a myth.
Recently, a study published in Science Advances**"Network Medicine Framework Reveals Generic Herbsymptom Effectiveness of Traditional Chinese Medicine", in which researchers established a network medicine framework and revealed the "dialectical treatment" of traditional Chinese medicine** The principle can be explained by the topological relationship between disease symptoms and TCM herbal targets on the human protein interaction network!The researchers found that the proteins associated with the symptoms form a network module, and the network proximity of the herbal target to the symptom module can determine the effectiveness of the herbal medicine symptom and validate the results with patient data from hospitals. The proposed cybermedicine framework reveals the scientific basis of TCM and establishes a paradigm for understanding the molecular basis and disease of natural medicine.
In this study, the investigators developed a network medicine framework that includes the following: (1) symptom-gene association (2) human protein interaction group (3) herb, component, target association and (4) symptom-herb association.
The framework theorizes the scientific basis of traditional Chinese medicine as the topological relationship between symptom-related proteins and herbal targets on the protein-protein interaction network, and designs eight indicators to analyze and evaluate the network.
Findings: Symptom-related proteins form local modules in the protein-protein interacting group (Figure 2A).
The profquencing of the proteins associated with different symptoms indicates that different symptoms perturb different regions of the PPI (Figure 2BC).
The results of Figure 2DE show that the network distance between symptoms is inversely correlated with the number of them co-occurring in the disease, indicating that the closer the network distance, the more likely the symptoms are to coexist in the same disease. The study also found that network distance was inversely correlated with the biosimilarity of symptoms, suggesting that symptoms that were closer in the PPI were more likely to be biologically similar.
Figure 2Symptom patterns of PPI in humans.
Next, Chinese herbal medicine was introduced into it, and the TCM-symptom network was determined to describe the ** effect. A multimodal approach was used to study the proximity of the herb-symptom network, and the protein interaction relationship between the herb target and the symptom was analyzed (Fig. 3A). By using two sets of datasets, 461 herbs containing 915 chemicals and 7518 proteins were obtained (Figure 3b). Through eight network measures, a method was established to assess the effect of herb-symptomatic ** (Figure 3c). The results showed that the network proximity was consistent with the known effective herb-symptomatic pairs, with an AUC value of 065 to 072, indicating that this method has a high ** ability. The best-performing process is to use the HIT dataset and a proximity Z-score with an AUC of 072, significantly superior to the results of the previous drug-disease relationship** (Figure 3D). Using the "fever" symptom as an example to illustrate the role of herb-symptom proximity, the study successfully identified the herbs associated with ** fever through network measurements, validating the usefulness of the framework in guiding herbal medicine selection (Fig. 3e).
Figure 3Herb-symptom network proximity** effect.
The network medicine framework, validated by real patient data, has shown reliability in terms of symptom relationships and herbal-symptomatic proximity. The results of the study based on the electronic medical record data of 1936 inpatients with liver cirrhosis from Hubei Provincial Hospital of Traditional Chinese Medicine in Wuhan showed that symptoms with shorter distances were more likely to coexist in the PPI network, validating the association between distance and symptom coexistence in the PPI network (Fig. 4AB). Network proximity validation for herbal-symptomatic pairs includes physician-prescribed herb-symptom pairs that are more adjacent to the PPI network, and valid herbal-symptomatic pairs tend to be more proximity to the network, which is consistent with expertise (Figure 4C). Validated by the propensity score matching method, network proximity was shown to be a good** indicator of herbal effectiveness in patient data (Figure 4d). In addition, the examples of Atractylodes macrocephalus and anorexia showed that the network proximity successfully demonstrated the effectiveness of herbal medicines, and was verified in patient data, which is consistent with the theory of traditional Chinese medicine.
Figure 4A networked healthcare framework based on hospital inpatient data.
The Cybermedicine Framework reveals opportunities for herbal discovery and repurposing, highlighting the ability of the Cybermedicine Framework to identify potential candidate herbs.
Figure 650 herb-symptomatic pairs (partial).