Impact of Qi-Invigorating Traditional Chinese Medicines on Diffuse Large B Cell Lymphoma Based on Network Pharmacology and Experimental Validation

Background: It has been verified that deficiency of Qi, a fundamental substance supporting daily activities according to the Traditional Chinese Medicine theory, is an important symptom of cancer. Qi-invigorating herbs can inhibit cancer development through promoting apoptosis and improving cancer microenvironment. In this study, we explored the potential mechanisms of Qi-invigorating herbs in diffuse large B cell lymphoma (DLBCL) through network pharmacology and in vitro experiment. Methods: Active ingredients of Qi-invigorating herbs were predicted from the Traditional Chinese Medicine Systems Pharmacology Database. Potential targets were obtained via the SwissTargetPrediction and STITCH databases. Target genes of DLBCL were obtained through the PubMed, the gene-disease associations and the Malacards databases. Overlapping genes between DLBCL and each Qi-invigorating herb were collected. Hub genes were subsequently screened via Cytoscape. The Gene Ontology and pathway enrichment analyses were performed using the DAVID database. Molecular docking was performed among active ingredients and hub genes. Hub genes linked with survival and tumor microenvironment were analyzed through the GEPIA 2.0 and TIMER 2.0 databases, respectively. Additionally, in vitro experiment was performed to verify the roles of common hub genes. Results: Through data mining, 14, 4, 22, 22, 35, 2, 36 genes were filtered as targets of Ginseng Radix et Rhizoma, Panacis Quinquefolii Radix, Codonopsis Radix, Pseudostellariae Radix, Astragali Radix, Dioscoreae Rhizoma, Glycyrrhizae Radix et Rhizoma for DLBCL treatment, respectively. Then besides Panacis Quinquefolii Radix and Dioscoreae Rhizoma, 1,14, 10, 14,13 hub genes were selected, respectively. Molecular docking studies indicated that active ingredients could stably bind to the pockets of hub proteins. CASP3, CDK1, AKT1 and MAPK3 were predicted as common hub genes. However, through experimental verification, only CASP3 was considered as the common target of Qi-invigorating herbs on DLBCL apoptosis. Furthermore, the TIMER2.0 database showed that Qi-invigorating herbs might act on DLBCL microenvironment through their target genes. Tumor-associated neutrophils may be main target cells of DLBCL treated by Qi-invigorating herbs. Conclusion: Our results support the effects of Qi-invigorating herbs on DLBCL. Hub genes and immune infiltrating cells provided the molecular basis for each Qi-invigorating herb acting on DLBCL.


INTRODUCTION
Diffuse large B cell lymphoma (DLBCL) is an aggressive lymphoma and the most common subtype of lymphoma, accounting for more than 30% of the lymphoma cases (Siegel et al., 2021). Although 60-70% of DLBCL patients reach longterm remission in response to a combination of rituximab, cyclophosphamide, doxorubicin, and prednisone (R-CHOP) immunochemotherapy, tumor recurrent, coupled with therapeutic resistant figures remain high.
Immune cells are main components of TME surrounding DLBCL cells, including T lymphocytes, macrophages and natural killer (NK) cells. It has been identified that a heterogeneous DLBCL microenvironment differs significantly in the subtype of tumor-infiltrating immune cells. For instance, a high proportion of programmed cell death protein 1 (PD1) + CD8 + T cells and programmed cell death-ligand 1 (PD-L1) + T cells in the TME was found to predict poor survival in DLBCL, whereas high expression of immune checkpoint cytotoxic T-lymphocyteassociated protein 4 (CTLA4) on T cells might be associated with favorable outcome (Xu-Monette et al., 2019). A recent study indicated that high proportions of T-cell immunoglobulin, mucin-domain containing 3 (TIM3) + T cells and TIM3+CD4 + cells have an independent adverse impact on survival (Autio et al., 2021). However, the above targeted therapy is not always feasible in DLBCL (Ansell et al., 2019;Barrueto et al., 2020;Kline et al., 2020). Much is still unknown concerning DLBCL microenvironment. Therefore, there is an urgent need to develop alternative or complementary therapeutics for DLBCL.
Traditional Chinese medicine (TCM) is gradually being accepted as a therapeutic option of cancer for they have provided much promise in vivo experiment and clinical trials (Parekh et al., 2009;Xiang et al., 2019). Based on TCM theory, the human body is recognized as an organic entity with Yin-Yang balance, which is a philosophical concept that represent a unity of opposing forces in the Universe. Generally speaking, Yin refers to stable or negative factors, while Yang represents positive and active features. if the Yin-Yang balance in a single individual is broken, the human body becomes sick. This theory is somewhat similar to TME. Qi is one of the most important products derived from the interaction between Yin and Yang, which stimulates the flow of blood throughout the body, and promotes the absorption and utility of the nutrients of food. The deprivation of Qi can be caused by many factors including malnutrition, fatigue, surgery and chronic diseases, leading to symptoms such as cough, short breath, muscle weakness, and immune-deficiency. Previous studies reported qi deficiency was an important syndrome of cancer (Hou et al., 2021;Qi et al., 2021). Qi-invigorating herbs, including Ginseng Radix et Rhizoma, Astragali Radix, and Glycyrrhizae Radix et Rhizoma, have been verified for their positive roles in cancer immune regulation Wang et al., 2020a). Several Qi-invigorating decoctions can inhibit cancer development through altering cancer immune microenvironment in vitro and vivo (Du et al., 2016;Wu et al., 2019;Xu et al., 2020). Furthermore, some Qiinvigorating herbs could directly promote the apoptosis of tumor cells in vivo and vitro (Xiong et al., 2018;Wang et al., 2020b). In lymphoma, TCMs have been verified to enhance physical function, reduce adverse effects of chemotherapy, and improve long-term survival (Hijikata et al., 1995;Ben-Arye et al., 2010;Guo et al., 2015;Kim et al., 2018). However, the underlying mechanisms and potential targets of Qi-invigorating herbs in DLBCL remain unclear. It is this idea that is the starting point of this article hoping to provide a theoretical basis for the application of Qi-invigorating herbs in DLBCL. In this study, we conducted a network pharmacology strategy for Qi-invigorating herbs. The so-called network pharmacology is a method which combines systems biology with multidirectional pharmacology based on high-throughput omics data analysis and network database retrieval (Hopkins, 2008;Li and Zhang, 2013). Through this discipline and experimental validation, the interaction between Qi-invigorating herbs and DLBCL would be discussed for the first time, which is helpful to elucidate the impact of Qi-invigorating herbs on DLBCL and provide guidance for herb selection, prescription formation, or even new drug design. The detailed flowchart of this study was shown in Figure 1.  Panacis Quinquefolii Radix (XiYangShen), Codonopsis Radix (DangShen), Pseudostellariae Radix (TaiZiShen), Astragali Radix (HuangQi), Glycyrrhizae Radix et Rhizoma (GanCao), Atractylodis Macrocephalae Rhizoma (BaiZhu) and Dioscoreae Rhizoma (ShanYao). All chemical constituents of the above herbs were obtained from the Traditional Chinese Medicine Systems Pharmacology (TCMSP) database (https://tcmsp-e.com/) (Ru et al., 2014). The potential active ingredients of the above Qiinvigorating herbs were selected according to the oral bioavailability (OB)≥30% (Xu et al., 2012), the Drug-Likeness (DL) Index≥0.18 and the drug half-life (HL)≥4 h (Jin et al., 2021). Chemical structures of the active ingredients were then collected from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/ ) (Kim et al., 2016).

Screening Drug Targets
The target genes of the active ingredients were obtained from the SwissTargetPrediction (http://www.swisstargetprediction.ch/) and STITCH (https://stitch.embl.de/) databases (Kuhn et al., 2008;Daina et al., 2019). By uploading chemical structures (SMILE files) to these databases and selecting "Homo sapiens" for the species, results were outputted. In the STITCH database, a target with combined-score ≥ 0.7 was selected, while in the SwissTargetPrediction database, probability ≥ 0.7 was assigned as the criteria for screening target genes of the active ingredients. The gene names of the targets were obtained from the UniProt database (https://www.uniprot.org/) (Consortium, 2019). All the target genes of each active ingredient were obtained after combining the two databases and excluding duplicates. The target genes of each Qi-invigorating herb were collected after combining target genes of its active ingredients and excluding the repetitive targets.

Constructing Protein-Protein Interaction Networks and Screening Key Targets
Venn diagrams via R software 4.1.0 were conducted to show overlapping targets between DLBCL and each Qi-invigorating herb. These targets were considered as potential targets for Qiinvigorating herbs acting on DLBCL.
To illustrate the enrichment of candidate targets in terms of the gene function, the study applied the Database for Annotation Visualization and Integrated Discovery (DAVID, https://david. ncifcrf.gov/) 6.8 to perform Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genome (KEGG) analyses (Huang et al., 2007). GO contains molecular function (MF), cellular component (CC), and biological process (BP) analyses. p-value < 0.05 were set as the criteria for screening (Huang et al., 2009).
PPI networks of candidate targets were constructed using the STRING database (https://string-db.org/) (Szklarczyk et al., 2021). We also used the Cytoscape 3.7.2 software to construct PPI networks based on the STRING results (Shannon et al., 2003). The criteria for considering significant was interactions with a combined score ≥ 0.7. Genes with degrees ≥ 5 were considered as hub genes. The hub genes were also input into the DAVID 6.8 for GO and KEGG enrichment analyses. These hub genes were subsequently prepared for molecular docking.

Molecular Docking
The crystal structures of the above hub genes were downloaded from the Protein Data Bank (https://www.pdb.org/) (Berman et al., 2000). PyMol (version 2.3.2) was used to process proteins, including removing water and ligands. The 2D structures of the active ingredients were obtained from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/) and considered as ligands. AutoDockTools 1.5.6 was run for molecular docking (Trott and Olson, 2010). During the docking process, hydrogen atoms were added to all proteins, and partial atomic charges were calculated. Target proteins were held as rigid and ligands allowed traveling freely. After the flexible-ligand docking, the grid box size was set to 40 × 40 × 40 points with a spacing of 0.375 Å. The Lamarckian Genetic Algorithm was used to search for populations of 150 individuals with a mutation rate of 0.02 over 10 generations. The results were sorted on the basis of the binding energy. The smaller the binding energy value, the more stable the structure after binding.

Analyzing Genes Acting on Diffuse Large B Cell Lymphoma Survival
The GEPIA2 database (http://gepia2.cancer-pku.cn/) was used for validation of hub gene expression in DLBCL and survival analysis .

Screening Common Hub Genes and Active Ingredients
A Venn diagram was constructed to select common hub genes and ingredients. Hub genes and targeted active ingredients existed in at least three groups were considered as common hub genes and ingredients, respectively. for 3 hours Miao et al., 2019). The decoctions were concentrated to 0.2 g/ml. The supernatant was collected as research objects after centrifugation, filtration, and sterilization Miao et al., 2019).

Cell Cultivation
Human DLBCL cell lines SU-DHL-2 (activated B-cell-like lymphoma) and SU-DHL-6 (germinal center B-cell-like lymphoma) were obtained from the American Type Culture Collection (ATCC, Manassas, VA, United States) and cultured in Roswell Park Memorial Institute-1640 (RPMI-1640) medium (HyClone, United States) supplemented with 10% fetal bovine serum (FBS) (CellBox, China) and 1% penicillin-streptomycin. All the cell lines were authenticated using short tandem repeat analysis. Cells were maintained at 37°C in standard culture conditions.
Cell Viability Assay 1 × 10 4 SU-DHL-2 or SU-DHL-6 cells were incubated into the 96well plates in 100 μL serum medium per well. A series of two-fold dilutions of decoctions were carried out to give concentrations of

Cell Proliferation Analysis
Proliferative ability of DLBCL cells was analyzed by CCK-8. Briefly, 1 × 10 4 cells were incubated into the 96-well plates in 100 μL serum medium. Subsequently, the cells were left negative control and experimental groups for 0, 24, 48, 72 h. The plates were read at 450 and 630 nm as reference wavelength. Each experiment was repeated in triplicate.

Flow Cytometry
The Annexin V-APC/propidium iodide (PI) apoptosis detection kit (KeyGEN, China) was used to detect apoptosis by flow cytometry. DLBCL cells were cultured in vitro and divided into nine groups, including untreated (

Statistical Analysis
All data were analyzed by the SPSS 20.0 (IBM, United States) or GraphPad Prism 7 (GraphPad Software Inc., United States) software. p < 0.05 was considered as statistical significance. p < 0.01 was considered as obvious statistical significance.    Figure 2 and the target names of Qi-invigorating herbs were summarized in Supplementary Table S2.

Screening of Common Targets Between Qi-Invigorating Herbs and Diffuse Large B Cell Lymphoma
A total of 1174 target genes associated with DLBCL were retrieved from the PubMed, DisGeNET, and Malacards

Gene Ontology and Pathway Functional Enrichment Analyses of Interacting Targets
To further explore functions of the above interacting targets, GO and pathway functional enrichment analyses were performed with DAVID Bioinformatics Resources 6.8. It was demonstrated that there were some similarities as well as differences among Qi-invigorating herbs. According to p-values, the top 5 BP, MF, CC and KEGG pathways of interacting targets were shown in

Gene Ontology and Pathway Functional Enrichment Analyses of Hub Genes
We also performed GO and KEGG pathway enrichment analyses on the hub genes. Bar charts of GO (BP, CC, MF) and bubble charts of KEGG pathways were obtained. As shown in Figure 3, GO and KEGG enrichment analyses predicted hub genes in Codonopsis Radix mainly involved in response to reactive oxygen species, serine/threonine protein kinase complex, cyclin-dependent protein serine/threonine kinase activity, progesterone-mediated oocyte maturation and so forth; hub genes in Pseudostellariae Radix mainly involved in response to reactive oxygen species, serine/threonine protein kinase complex, histone kinase activity, progesterone-mediated oocyte maturation and so forth; hub genes in Astragali Radix mainly involved in execution phase of apoptosis, spindle, cysteine-type endopeptidase activity involved in apoptotic signaling pathway, colorectal cancer and so forth; while hub genes in Glycyrrhizae Radix et Rhizoma mainly involved in signal transduction in absence of ligand, spindle, protein serine/threonine kinase activity, protein serine/threonine kinase complex, colorectal cancer and so forth. It is shown that these hub genes mainly enriched in cancer-related pathways. From the top ten KEGG pathway results, the significantly enriched hub genes were CASP3, CDK2, AKT1, MAPK3, MAPK1, CCNB1, CDK1, CCNA2, JUN, CDK4, FOS, TNF in Codonopsis Radix; CASP3, AKT1, CDK2, CCNB1, CCNA2, CDK1, JUN, STAT1, VEGFA in Pseudostellariae Radix; CASP3, MAPK3, AKT1, BIRC5, CASP9, AR, CASP8, PLK1, CDK1 in Astragali Radix; CASP3, MAPK3, AKT1, BIRC5, CASP9, CDK1, PLK1 in Glycyrrhizae Radix et Rhizoma.

Correlation Between Hub Genes and Survival Time
In order to discover whether hub genes had correlations with DLBCL prognosis, hub genes were input into the GEPIA 2.0 database. Firstly, we tested their transcription levels in datasets concerning DLBCL vs non-tumor tissues. As shown in Figure 5A, all the hub genes were differently expressed in DLBCL as compared to those in non-tumor tissues. Then comparison of each gene's survival curves was made and p values were calculated by log-rank. Results showed that only alteration in CASP3 and BDNF correlated with disease-free survival, for their p-values of log-rank were less than or close to 0.05, when group cutoff was set to 50% ( Figures 5B,C). In view of the small number of samples, a retrospective study with larger sample is expected to confirm the conclusion.

Common Target Genes and Active Ingredients of Different Qi-Invigorating Herbs
It is noteworthy that different Qi-invigorating herbs exist same active ingredients, i.e. different Qi-invigorating herbs may target same genes in DLBCL. Thus, we first took the intersection of hub gene related ingredients. However, no common hub gene related ingredients were found. Subsequently, additional analysis was performed on hub genes. Through R software 4.1.0, it was detected that four hub genes (CASP3, CDK1, AKT1, MAPK3) corresponded to at least three of five Qi-invigorating herbs ( Figure 6). Therefore, it is suggested that CASP3, CDK1, AKT1 and MAPK3 may be common target genes for different Qi-invigorating herbs acting on DLBCL.

Cytotoxicity Study of Qi-Invigorating Herbs in Diffuse Large B Cell Lymphoma Cells
In order to verify the effect of common hub genes in Qiinvigorating herbs, an in vitro experiment was carried out. Firstly, we evaluated the cytotoxic efficacy of Qi-invigorating herbs in two DLBCL cell lines at concentrations between 3 and 10 mg/ml. Results showed that the IC50 values were 3.91 mg/ml, 4.01 mg/ml for Astragali Radix; 2.4 mg/ml, 1.94 mg/ml for Glycyrrhizae Radix et Rhizoma; 2.05 mg/ml, 3.2 mg/ml for Ginseng Radix et Rhizoma; 1.73 mg/ml, 4.11 mg/ml for Pseudostellariae Radix; 3.23 mg/ml, 4.1 mg/ml for Codonopsis Radix in SUDHL2 and SUDHL6 cells, respectively. Considering the cytotoxicity of all Qi-invigorating herbs, 5 mg/ml concentration of Qi-invigorating herbs was selected as a suitable intervention concentration for subsequent experiments.

The Role of Qi-Invigorating Herbs in the Proliferation of Diffuse Large B Cell Lymphoma Cells
CCK-8 assays were applied to determine the effects of Astragali Radix, Glycyrrhizae Radix et Rhizoma, Ginseng Radix et Rhizoma, Pseudostellariae Radix and Codonopsis Radix on DLBCL cell proliferation. Results showed that the proliferation ability of SU-DHL-2 and SU-DHL-6 cells was decreased via incubating with Qi-invigorating herbs ( Figures 7A,B).

The Role of Qi-Invigorating Herbs in the Apoptosis of Diffuse Large B Cell Lymphoma Cells
To measure the induction of apoptosis, cells were treated with 5 mg/ ml of Astragali Radix, Glycyrrhizae Radix et Rhizoma, Ginseng Radix et Rhizoma, Pseudostellariae Radix and Codonopsis Radix decoctions for 48 h, respectively. Flow cytometry analysis exhibited that incubation with Glycyrrhizae Radix et Rhizoma resulted in a remarkable rise in cell apoptosis [(33.09 ± 0.99)% in SUDHL2 and (31.3 ± 1.3)% in SUDHL6], comparing to negative control groups ( Figures 7C,D). Meanwhile cell apoptosis in DLBCL cells treated with Ginseng Radix et Rhizoma was slightly elevated ( Figures  7C,D). Thus it is suggested that some Qi-invigorating herbs participated in the apoptosis of DLBCL cells.

The Role of Common Hub Genes in Diffuse Large B Cell Lymphoma Cells Treated With Qi-Invigorating Herbs
In order to further elaborate the roles of common hub genes, we tested CASP3, CDK1, AKT1 and MAPK3 expression in DLBCL cells incubated with Qi-invigorating herbs, respectively. As

DISCUSSION
TCM is a very important cancer treatment strategy in China. It has been widely reported that TCM herbs can improve DLBCL survival and provide a complementary therapy for cases that are refractory or have relapsed. According to the TCM theory, disturbance of Yin-Yang is one of the main reasons for carcinogenesis. To alleviate symptoms of disease, TCM aims to restore the balance of Yin and Yang in addition to eliminate pathogenic factors ("Quxie"), which is somewhat similar to TME from the view of theory. Based on preclinical studies, Qiinvigorating herbs have been proved to have immune enhancement and multi-target regulation. Therefore, a systematic study of Qi-invigorating herbs in DLBCL would elucidate their mechanisms and provide guidance for herb selection, or even personalized therapy. In this study, eight commonly used Qi-invigorating herbs were included, which have been widely employed in herbal prescriptions for cancer treatment. Active ingredients were first selected from the TCMSP database. Then potential targets were output via the SwissTargetPrediction and STITCH databases. By intersection with target genes of DLBCL, it was found that besides Atractylodis Macrocephalae Rhizoma, the other seven Qi-invigorating herbs could act on DLBCL. GO and KEGG pathway enrichment results revealed that there were some similarities as well as differences among Qiinvigorating herbs.
Through PPI construction networks, hub genes were output, which implicated Ginseng Radix et Rhizoma, Codonopsis Radix, Pseudostellariae Radix, Astragali Radix, Glycyrrhizae Radix et Rhizoma might be more appropriate for herbal prescriptions in treating DLBCL. The suppressed proliferation by these Qiinvigorating herbs in vitro verified this view. Excellent molecular docking results for hub genes and their corresponding ingredients further confirmed this view.
Interestingly, CASP3, CDK1, AKT1 and MAPK3 were considered as common target genes. Among them, CASP3 is one important member of caspases, characterized by programmed cell death (Galluzzi et al., 2016). It is often used as a marker for efficacy of cancer therapy. In TME, it has been indicated that CASP3 activation could trigger pyroptosis, a form of cell death that is critical for immunity by cleaving gasdermin E (GSDME) in GSDMEexpressing tumors (Wang et al., 2017;Zhang et al., 2020). Similar finding was observed in T-cell lymphoma, where immunogenic cell death depended on CASP3 activity, with reduced antitumor immunity generated by CASP3-deficient EL4 cells (Jaime-Sanchez et al., 2020). Our study detected that DLBCL cells treated with Glycyrrhizae Radix et Rhizoma or Ginseng Radix et Rhizoma promoted their apoptosis accompanied by up-regulated CASP3. Thus it is concluded that Qi-invigorating herbs regulated DLBCL cell FIGURE 8 | relationship between CDK1 and tumor-infiltrating immune cells (CD4 + T cells and neutrophils); (I). The relationship between CCNB1 and tumorinfiltrating immune cells (MDSCs and neutrophils); (J). The relationship between CDK2 and tumor-infiltrating immune cells (CD8 + T cells and neutrophils); (K). The relationship between FOS and tumor-infiltrating immune cells (CD4 + T cells and neutrophils); (L). The relationship between JUN and tumor-infiltrating immune cells (CD8 + T cells and neutrophils); (M). The relationship between BIRC5 and tumor-infiltrating immune cells (neutrophils, myeloid dendritic cells and MDSCs); (N). The relationship between CDK4 and CD4 + T cells; (O). The relationship between MCL1 and neutrophils, CD8 + T cells, macrophage; (P). The relationship between PLK1 and neutrophils; (Q). TOP2A and tumor-infiltrating immune cells (CD8 + T cells and neutrophils); (R). The relationship between MAPK3 and tumor-infiltrating immune cells (CD8 + T cells and neutrophils); (S). The relationship between STAT1 and tumor-infiltrating immune cells (B cells, CD8 + T cells, macrophage, myeloid dendritic cells, neutrophils, MDSCs); (T). The relationship between NEK2 and neutrophils; (U). The relationship between TYMS and MDSCs; (V). The relationship between VEGFA and neutrophils; (W). MAPK1 and tumor-infiltrating immune cells (CD8 + T cells and neutrophils); (X). Although CDK1, AKT1 and MAPK3 were considered as other common hub genes for Qi-invigorating herbs, most CDK1, AKT1, MAPK3 expression did not change significantly in vitro. Therefore, it is concluded that CDK1, AKT1 and MAPK3 are not common hub genes in DLBCL cells treated with Qi-invigorating herbs. In this study we have well described the effect of common hub genes by wet lab experiments. The effects and mechanisms of other hub genes required further verification in the future. The detailed mechanisms obtained by experiments may provide a guidance for herb selection, prescription formation, new drug design, or even personalized therapy.
Molecular docking results revealed that one common hub gene exhibited good binding energy for multiple active ingredients and one ingredient targeted more than one hub gene. Based on binding affinity and specific ligand-receptor residue interactions, there are six active ingredients (luteolin, glycitein, acacetin, quercetin, mairin, formononetin) targeting more than three hub genes in each Qiinvigorating herb, indicating they have potential as DLBCL drugs. So far, some of these ingredients have been well described for their antilymphoma effects in experimental studies. luteolin, an active ingredient extracted from Codonopsis Radix and Pseudostellariae Radix, has been tested to inhibit the proliferation of classical Hodgkin's lymphoma cells via caspase activation (Gharbaran et al., 2020). Quercetin, an active ingredient of Astragali Radix and Glycyrrhizae Radix et Rhizoma, has been detected to inhibit cancer cell growth by modulating AKT and NF-κB pathways in lymphoma models (Soofiyani et al., 2021). The experimental results also displayed the prospect of Qi-invigorating herbs in DLBCL treatment. Thus, more pharmacodynamic studies are required to examine their effects in DLBCL models.
Besides directly targeting cancer cells, previous evidence suggested that Qi-invigorating herbs could exert antitumor effects by upregulating immune responses even in immunosuppressive TME. Therefore, to explore the mechanisms of Qi-invigorating herbs for DLBCL treatment in depth, we conduct a comprehensive TME analysis for hub genes. In DLBCL microenvironment, although much is still unknown concerning the composition of immune cells, CD4 + T cells, CD8 + T cells, neutrophils, macrophages, myeloid DCs and MDSCs have been widely tested as independent predictors of DLBCL outcome (Azzaoui et al., 2016;Chen et al., 2016;Judd et al., 2017;Kusano et al., 2017;Ciavarella et al., 2018;Staiger et al., 2020;Merdan et al., 2021;Xu et al., 2021). Manfroi et al. (Manfroi et al., 2017) further demonstrated an association between neutrophils and A Proliferation-Inducing TNF Ligand (APRIL). They reported that APRIL produced by DLBCL-related neutrophils increased tumor aggressiveness and affected disease prognosis (Manfroi et al., 2017). In our study, the relationship between hub genes and tumor-infiltrating immune cells (B cells, CD4 + T cells, CD8 + T cells, neutrophils, macrophages, myeloid DCs and MDSCs) was tested via the TIMER 2.0 database. Since most hub genes were found to be associated with neutrophils in infiltrating levels, it was reasonable to speculate that Qi-invigorating herbs mainly affected DLBCL TME through acting on tumor-associated neutrophils. Further research is needed to validate our findings.
In a word, this study systematically explored the active ingredients, potential targets, hub genes and the impact of immune cells in Qi-invigorating herbs for DLBCL treatment. Hub genes and immune infiltrating cells provided the molecular basis for each Qi-invigorating herb acting on DLBCL. Study on common mechanisms revealed that CASP3 might be the common target of Qi-invigorating herbs on DLBCL apoptosis. Tumor-associated neutrophils may be main target cells of DLBCL treated by Qi-invigorating herbs. Although this study provided preliminary predictions, it was important for the modernization of TCM. Such researches would undoubtedly increase our understanding of DLBCL pathogenesis. However, the conclusions still need to be verified with subsequent wet lab experiments.

CONCLUSION
Through network pharmacology analysis and experimental validation, Qi-invigorating herbs were proved to be appropriate for DLBCL treatment. Potential targets, hub genes and targeted immune cells of Qi-invigorating herbs helped us to understand the mechanisms. Study on common mechanisms revealed that CASP3 might be the common target of Qi-invigorating herbs on DLBCL apoptosis. Tumor-associated neutrophils may be main target cells of DLBCL treated by Qi-invigorating herbs. However, the mechanisms of hub genes still required verification, which may provide a guidance for herb selection, prescription formation, new drug design, or even personalized therapy.
In all, our findings may provide a theoretical basis for the application of Qi-invigorating herbs in DLBCL. This article may provide a reference and clinical transformation value for TCM.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Materials, further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
QH: Data collection, data mining; Drafting the article; experimental validation; JL: Data collection and qRT-PCR; SH: Data collection and and qRT-PCR; JS: Conception and design of the work; Critical revision of the article; Final approval of the version to be published.