Globally, the incidence and death rates of lung cancer are among the highest in the list of malignant tumors. Lung adenocarcinoma (LUAD) is a type of non-small cell lung cancer (NSCLC), which accounts for about 40%–55% of lung cancer patients (Herbst et al., 2018). LUAD tends to occur in female and male non-smokers, and is more common in China, and is characterized by brain, bone and liver metastases (Steinet al., 2019). LUAD usually has no obvious clinical symptoms in the early stage of the disease, and about 75% of patients are in the middle or late stage when they are found, with a low 5-year survival rate, which seriously affects the prognosis of patients (Wanget al., 2017). Although tyrosine kinase inhibitors and immunotherapy have brought some survival benefits to some patients with LUAD, the overall survival rate of patients is still low. Therefore, there is an urgent need to find safer and more effective drugs to treat LUAD.

Traditional Chinese medicine (TCM) has rich clinical experience and a large number of effective drugs in the treatment of tumors. Hedyotis diffusa (H. diffusa) is slightly bitter, sweet, cold in nature, belongs to the stomach, large intestine and small intestine meridians, and has the efficacy of clearing heat and removing toxins, inducing diuresis and clearing diuresis (Corraleset al., 2020). With a variety of active ingredients, H. diffusa can treat a variety of tumors such as digestive system tumors, lung cancer, prostate cancer and cervical cancer, and has anti-tumor effects such as inhibiting angiogenesis, inhibiting proliferation, and promoting apoptosis (Ma et al., 2014; Ye et al., 2019). Relevant studies have shown that the main antitumor components in H. diffusa are ursolic acid and oleanolic acid, and the antitumor mechanism of ursolic acid is to inhibit the growth of breast and colon cancer cell lines by triggering apoptosis, cell cycle arrest, and anti-metastatic and anti-angiogenic properties through various molecular targets and signaling pathways, and oleanolic acid is to induce cell apoptosis by regulating different signaling pathways and biological processes through multiple targets at the same time (He et al., 2018). Oleanolic acid, on the other hand, induces apoptosis by regulating different signaling pathways and biological processes through multiple targets, including intracellular calcium levels, NF-κB signaling pathway, Notch signaling pathway, JAK/STAT3 signaling pathway, and the expression of ribose polymerase (Chan et al., 2019). Based on the current research, there is still a lack of large samples of clinical data analysis and related mechanism studies on the treatment of LUAD with H. diffusa.

Symptom phenotypes (i.e., symptoms and signs), one of the main clinical manifestations of disease conditions, that could be obtained by human natural perception and cognition abilities, play a vital role for medical visiting, clinical diagnosis, and disease treatment. In recent years, researchers have developed a series of systems or network pharmacological strategies to detect the molecular mechanisms of TCM (Schroen, et al., 2015). A study extracted symptom co-occurrences from clinical textbooks to construct phenotype network of symptoms with clinical co-occurrence and incorporated high-quality symptom-gene associations and protein–protein interactions to explore the molecular network patterns of symptom phenotypes (Shu, et al., 2021). The work of Lu et al. showed that the integrated network analysis method could be used for identifying robust symptom clusters (SCs) and investigate the molecular mechanisms of these SCs, which would be valuable for symptom science and precision health (Lu, et al., 2020). By establishing a protein-protein interaction network framework, it was revealed that the general rule in TCM treatment selection is network distance (Gan, et al., 2023). By constructing a symptom-based human disease network, the authors analyzed the relationship between disease-symptom similarity and gene-protein interactions, and explored the relationship between the diversity of clinical manifestations of a single disease and its molecular mechanism (Zhou, et al., 2014). Taking advantage of GPU performance, the algorithm implementation of drug-symptom correlation analysis can be effectively accelerated (Tian, et al., 2020). Herb-Target Interaction Network (HTINet) approach obtained low-dimensional expression of nodes by constructing a heterogeneous network and performing network embedding, and used this expression to predict the relationship between Chinese herbal medicines and targets (Wang, et al., 2019). Using deep learning and heterogeneous network topology, a drug-target association prediction method that calculates the similarity between drug and target has been proposed (Zong, et al., 2019). A method using an iterative algorithm of three-layer heterogeneous networks to achieve drug repositioning is proposed (Wang, et al., 2014). A Network-based Random Walk with Restart (NRWRH) method is proposed to predict unknown drug-target interactions (Chen, et al., 2012).

Based on real-world clinical data, this study reviewed and analyzed the application and symptom improvement of H. diffusa in LUAD, and performed relevant statistical analyses. In addition, the relationship network between symptom improvement and disease genes was also analyzed, and the gene target highly related to the symptom improvement of lung adenocarcinoma was finally screened out, and the molecular and cellular experiments were designed to validate the regulatory effect of H. diffusa on this target. The present study demonstrated the efficacy and possible molecular mechanism of Hedyotis diffusa in the treatment of lung cancer at both clinical and experimental levels, and provided the basis and new ideas for the study of the molecular mechanism of H. diffusa in the treatment of LUAD. The technical flowchart design we studied is shown below (Figure 1).

To assess whether there is a correlation between two categorical variables, we used chi-square and t-tests on the processed data. The chi-square test evaluates the degree of association between variables (gender, symptoms) based on the difference between observed and expected frequencies. When performing the chi-square test, we first construct a contingency table to organize the observed frequency distributions of different categorical variables together. The expected frequency for each cell is then calculated, which is the frequency expected based on the assumed independence. Next, the differences between the observed and expected frequencies are calculated and these differences are squared and normalized. Finally, the chi-square statistic is obtained by summing the standardized differences. It can be used to compare the difference between the observed frequency and the expected frequency to determine whether the difference exceeds the range caused by randomness. The t-test is used to compare whether there are significant differences in the means (age, length of hospitalization) of two samples. When performing a t-test, we first calculate the mean and variance of each sample. Next, using these statistical indicators, we calculate the t-value, which represents the size of the difference between the two sample means relative to their standard errors.

LUAD is the most common subgroup of NSCLC with aggressive and metastatic potential, leading to drug resistance and treatment failure (Wang et al., 2009; Travis et al., 2015). Despite advances in the detection and treatment of NSCLC, unfortunately, the prognosis remains poor when the disease is detected in late clinical stages, and some patients are still plagued by rapid disease recurrence and progression, thus resulting in a 5−year survival rate of only about 15% (Erridge et al., 2007). TCM participation is an important means in the treatment of LUAD. Due to the characteristics of syndrome differentiation and treatment, combined with surgery, chemotherapy and targeted therapy, TCM plays an anti-tumor role from the two aspects of self-healing ability and disease resistance. Recently, an increasing number of studies (Wu et al., 2022; Huang et al., 2022; Qian et al., 2019) showed that the herb has some antitumor effects, including in the treatment of LUAD. It is a kind of Chinese herbal herb family, which is a famous Chinese herbal medicine with thousands of years of clinical practice history. It is an important component of various anticancer formulations; it was reported to inhibit tumor cell proliferation and metastasis and reduce side effects after chemotherapy (Song et al., 2019). Early pharmacological studies confirmed that the herb has medicinal properties such as anti-tumor, anti-inflammatory, immunoregulatory, antioxidant and other biological activities (Shen et al., 2016). The current studies on the mechanism of lung adenocarcinoma focus on the theoretical stage of network pharmacology or only in vitro experimental stage, and the analysis of real world clinical data has not been reported. Based on this, based on real-world clinical data, our study analyzed the key targets of LUAD from the perspective of drug improvement of symptoms, and used biological experiments to verify them, so as to provide certain theoretical basis and support for the mechanism of LUAD in TCM.

H. diffusa is an important component of TCM clinical anticancer prescription (Han et al., 2020; Wazir et al., 2021; Zhang et al., 2021), and the results of this study showed that it is used significantly more frequently in patients with LUAD than in patients with other diseases (p < 0.05). A literature study on the treatment of NSCLC (Qi et al., 2021) showed that the herb is one of the more frequent Chinese medicines in prescription. At the same time, many studies in Taiwan have shown that it is also one of the most frequently prescribed TCM (Hung et al., 2017; Kuo et al., 2018; Ting et al., 2017; Wang et al., 2021) in the treatment of other cancers (such as liver cancer, gastric cancer, nasopharyngeal cancer and pancreatic cancer). An experimental study (Huang et al., 2022) showed that the H. diffusa injection significantly reduced the survival of lung adenocarcinoma cells in vitro, inhibited the growth of BALB/c nude mice in vivo, and induced iron death of VDAC 2/3 regulation by Bax by inhibiting Bcl-2 gene. In addition, the results of this study showed a significant improvement in the clinical symptoms (such as cough, chest tightness, wheezing, sputum, shortness of breath, headache, dry mouth), and improved the quality of life.

Most of the previous studies have focused on the relationship between TCM and disease, for example, TCM diagnosis and herbal therapy are based on the symptom phenotype rather than disease diagnosis. Thus, our study builds on symptoms rather than disease. By focusing on symptoms rather than disease, our approach is consistent with the practice of TCM in diagnosing and treating patients based on the symptom phenotype. The study by Gan et al. (Gan et al., 2023) found that proteins associated with symptoms tend to cluster in a local PPI module, and the network proximity between herbal targets and symptom modules indicates the effectiveness of herbs in treating symptoms. Based on the above improvement symptom gene set and the TCM target set, we combined the above five symptom genes into one group, and put the total symptom gene set and the TCM target set into the gene association analysis platform respectively. The overlapping genes obtained were CTNNB1, STAT3, CASP8, and APC. We picked out the highest weighted target CTNNB1 to unfold the following validation. WANG et al. (Wang et al., 2017) used network pharmacology to analyze the potential targets of H. diffusa-Astragalus membranaceus for the treatment of colorectal cancer, and found that the core target also contained CTNNB1. Zhou et al. (Zhou et al., 2020) included 564 patients with LUAD in the study and found a poor prognosis in patients with primary lung adenocarcinoma with CTNNB1 mutation. In vitro, we combined the H. diffusa with lung adenocarcinoma A549 cells, and the apoptosis experiment showed that the higher concentration of drug groups promoted apoptosis more than the lower concentration. Furthermore, WANG et al. (Wang et al., 2022) showed that H. diffusa significantly reduced the survival rate of lung adenocarcinoma cells in vitro and significantly inhibited the cell adhesion, invasion and migration of lung adenocarcinoma A549 cells. We further verified the effect of H. diffusa on CTNNB1 at the mRNA and protein levels, and found that compared with the control group, the drug group significantly inhibited the mRNA and protein levels of CTNNB1 gene, and the inhibitory effect of the high drug group was more obvious than that of the low concentration of the drug group. It may speculate that it may promote apoptosis of lung cancer A549 cells by inhibiting the CTNNB1 gene. In addition, the shortcomings of our study compared to the study by Peng et al. are that we did not consider prescriptions, but rather looked at the effects of individual drugs (Wang et al., 2022; Peng et al., 2023).

In conclusion, our research based on real world data from clinical LUAD symptoms changes and the relationship between H. diffusa, using network algorithm for key targets, and through the experiment verified the key target, the whole argument method is rigorous, can for clinical treatment of lung adenocarcinoma provide certain theoretical support.