Abstract

   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.

   Keywords: qi-invigorating herbs, traditional Chinese medicine, diffuse
   large B cell lymphoma, network pharmacology, tumor microenvironment

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 ([32]Siegel et al., 2021). Although 60–70% of DLBCL
   patients reach long-term remission in response to a combination of
   rituximab, cyclophosphamide, doxorubicin, and prednisone (R-CHOP)
   immunochemotherapy, tumor recurrent, coupled with therapeutic resistant
   figures remain high.

   In recent years, several agents constitute the new landscape of
   treatments for DLBCL, including brentuximab vedotin ([33]Svoboda et
   al., 2021), ibrutinib ([34]Younes et al., 2019) and polatuzumab vedotin
   ([35]Sehn et al., 2020). Besides, targeting tumor microenvironment
   (TME) is one promising option for increasing evidence strongly
   highlights that lymphomagenesis and DLBCL progression mainly rely on
   its microenvironment ([36]Scott and Gascoyne, 2014).

   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-lymphocyte-associated
   protein 4 (CTLA4) on T cells might be associated with favorable outcome
   ([37]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 ([38]Autio et al., 2021). However, the above targeted therapy
   is not always feasible in DLBCL ([39]Ansell et al., 2019; [40]Barrueto
   et al., 2020; [41]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 ([42]Parekh et al., 2009; [43]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 ([44]Hou et al., 2021;
   [45]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
   ([46]Jiang et al., 2019; [47]Wang et al., 2020a). Several
   Qi-invigorating decoctions can inhibit cancer development through
   altering cancer immune microenvironment in vitro and vivo ([48]Du et
   al., 2016; [49]Wu et al., 2019; [50]Xu et al., 2020). Furthermore, some
   Qi-invigorating herbs could directly promote the apoptosis of tumor
   cells in vivo and vitro ([51]Xiong et al., 2018; [52]Wang et al.,
   2020b). In lymphoma, TCMs have been verified to enhance physical
   function, reduce adverse effects of chemotherapy, and improve long-term
   survival ([53]Hijikata et al., 1995; [54]Ben-Arye et al., 2010; [55]Guo
   et al., 2015; [56]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
   ([57]Hopkins, 2008; [58]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 [59]Figure 1.

FIGURE 1.

   [60]FIGURE 1
   [61]Open in a new tab

   The detailed flowchart of this network pharmacology based study.

Materials and Methods

Screening Active Constituents of Qi-Invigorating Herbs

   Eight common Qi-invigorating herbs were discussed in this study,
   including Ginseng Radix et Rhizoma (RenShen), 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 ([62]https://tcmsp-e.com/) ([63]Ru et al., 2014). The
   potential active ingredients of the above Qi-invigorating herbs were
   selected according to the oral bioavailability (OB)≥30% ([64]Xu et al.,
   2012), the Drug-Likeness (DL) Index≥0.18 and the drug half-life
   (HL)≥4 h ([65]Jin et al., 2021). Chemical structures of the active
   ingredients were then collected from the PubChem database
   ([66]https://pubchem.ncbi.nlm.nih.gov/) ([67]Kim et al., 2016).

Screening Drug Targets

   The target genes of the active ingredients were obtained from the
   SwissTargetPrediction ([68]http://www.swisstargetprediction.ch/) and
   STITCH ([69]https://stitch.embl.de/) databases ([70]Kuhn et al., 2008;
   [71]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 ([72]https://www.uniprot.org/)
   ([73]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.

Obtaining Potential Target Genes for Diffuse Large B Cell Lymphoma

   We used “diffuse large B cell lymphoma,” “diffuse large-B cell
   lymphoma” and “DLBCL” as keywords to identify DLBCL-related genes from
   the PubMed database (update time: September 15, 2021,
   [74]https://www.ncbi.nlm.nih.gov/gene), the gene-disease associations
   database (DisGeNET, [75]https://www.disgenet.org/) and the Malacards
   ([76]https://www.malacards.org/) database ([77]Piñero et al., 2017;
   [78]Rappaport et al., 2017; [79]Piñero et al., 2020). Only those genes
   from “Homo sapiens” were included. In the end, a total of 1174 target
   genes of DLBCL were obtained after removing duplicates.

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 Qi-invigorating 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, [80]https://david.ncifcrf.gov/) 6.8 to
   perform Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genome
   (KEGG) analyses ([81]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
   ([82]Huang et al., 2009).

   PPI networks of candidate targets were constructed using the STRING
   database ([83]https://string-db.org/) ([84]Szklarczyk et al., 2021). We
   also used the Cytoscape 3.7.2 software to construct PPI networks based
   on the STRING results ([85]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 ([86]https://www.pdb.org/) ([87]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
   ([88]https://pubchem.ncbi.nlm.nih.gov/) and considered as ligands.
   AutoDockTools 1.5.6 was run for molecular docking ([89]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 ([90]http://gepia2.cancer-pku.cn/) was used for
   validation of hub gene expression in DLBCL and survival analysis
   ([91]Tang et al., 2019).

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.

Performing Experimental Verification

Preparation of Qi-Invigorating Herbs

   All the Qi-invigorating herbs, including Ginseng Radix et Rhizoma,
   Codonopsis Radix, Pseudostellariae Radix, Astragali Radix, Glycyrrhizae
   Radix et Rhizoma, were purchased from the TCM pharmacy of Fujian
   Medical University Union Hospital, Fujian, China. Each herb (10 g) was
   placed in a beaker, then boiled water (100 ml) was added and the
   solution was kept boiling for 3 hours ([92]Chen et al., 2019; [93]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 ([94]Chen et al., 2019; [95]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 96‐well
   plates in 100 μL serum medium per well. A series of two-fold dilutions
   of decoctions were carried out to give concentrations of 10, 5, 2.5,
   1.25, 0.625 and 0.313 mg/ml. Untreated tumor cells and complete medium
   (blank control) were left as control groups. After 48 h incubation,
   10 µL of CCK8 (AbMole Bioscience) was added into all wells and the
   plates were further incubated for 2 h. Then, the plates were read at
   450 and 630 nm as reference wavelength. After background subtraction,
   half maximal inhibitory concentration (IC50) was calculated using the
   SPSS 20.0 (IBM, United States). The experiments were repeated in
   triplicate for each inhibitor concentration.

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 (blank control), negative control, Astragali Radix group,
   Glycyrrhizae Radix et Rhizoma group, Ginseng Radix et Rhizoma group,
   Pseudostellariae Radix group and Codonopsis Radix group. After
   cultivation for 48 h, cells were washed with PBS, resuspended in 100 µL
   binding buffer at a concentration of 2 × 10^6 cells/ml, and
   anti-Annexin V APC-conjugated antibody and PI were added. The mixtures
   were incubated for 15 min at room temperature, supplemented with
   binding buffer to 500 µL and processed by BD Accuri C6 Plus Flow
   Cytometry (BD Biosciences, United States). Data were analyzed through
   flow cytometry (BD PharmingenTM, United States).

Quantitative Real-Time Polymerase Chain Reaction

   Total RNA of cells was extracted by TRIzol (CWBIO, China) and
   sequentially purified via the chloroform, isopropanol, and 70% ethanol.
   Reverse transcription for mRNA relied on a kit named Hifair III 1^st
   Strand cDNA Synthesis Supermix for qPCR (YEASEN, China). qRT-PCR was
   carried out on Applied Biosystems 7500/7500 Fast Real-time PCR System
   (Thermo Fisher Scientific, United States) and QuantStudio™ 5 Real-Time
   PCR Instrument (384-well block, Thermo Fisher Scientific, United
   States). The GADPH was used to normalize mRNA samples. Relative
   quantification of target primers was calculated by the 2-ΔΔCT method.
   The experiment was repeated three times. Primers used for qRT-PCR
   analysis are listed as follow: GADPH Forward (5′-3′)
   TTG​GTA​TCG​TGG​AAG​GAC​TCA, Reverse (5′-3′)
   AGT​AGA​GGC​AGG​GAT​GAT​GTT; CASP3 Forward (5′-3′)
   CAT​GGA​AGC​GAA​TCA​ATG​GAC​T, Reverse (5′-3′)
   CTG​TAC​CAG​ACC​GAG​ATG​TCA; CDK1 Forward (5′-3′)
   GTC​AGT​CTT​CAG​GAT​GTG​CT, Reverse (5′-3′) ATG​TAC​TGA​CCA​GGA​GGG​AT;
   AKT1 Forward (5′-3′) GTC​ATC​GAA​CGC​ACC​TTC​CAT, Reverse (5′-3′)
   AGC​TTC​AGG​TAC​TCA​AAC​TCG​T; MAPK3 Forward (5′-3′)
   GCT​TCC​GCC​ATG​AGA​ATG​TC; Reverse (5′-3′) GGC​GGA​GTG​GAT​GTA​CTT​GA.

Analyzing Genes Acting on Diffuse Large B Cell Lymphoma Tumor
Microenvironment

   The TIMER database ([96]http://timer.cistrome.org/) is a database for
   comprehensive analysis of immune infiltrating cells across multiple
   cancer types. We applied TIMER2.0 to study the relationship between hub
   genes and tumor-infiltrating immune cells [B cells, CD4^+ T cells,
   CD8^+ T cells, neutrophils, macrophages, and myeloid dendritic cells
   (DCs), myeloid derived suppressor cells (MDSC)] in DLBCL ([97]Li et
   al., 2020).

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.

Results

Filtering of Active Constituents of Qi-Invigorating Herbs

   Among Qi-invigorating herbs, 20 active ingredients were identified from
   Ginseng Radix et Rhizoma, 9 from Panacis Quinquefolii Radix, 19 from
   Codonopsis Radix, 6 from Pseudostellariae Radix, 16 from Astragali
   Radix, 5 from Atractylodis Macrocephalae Rhizoma, 16 from Dioscoreae
   Rhizoma and 76 from Glycyrrhizae Radix et Rhizoma. The details of
   active ingredients were summarized in [98]Supplementary Table S1.

Screening of Drug Targets

   Through the SwissTargetPrediction and Stitch platforms, 63 genes were
   selected as candidate targets of Ginseng Radix et Rhizoma, 27 genes as
   candidate targets of Panacis Quinquefolii Radix, 62 genes as candidate
   targets of Codonopsis Radix, 63 genes as candidate targets of
   Pseudostellariae Radix, 121 genes as candidate targets of Astragali
   Radix, 0 genes as candidate targets of Atractylodis Macrocephalae
   Rhizoma, 23 genes as candidate targets of Dioscoreae Rhizoma, and 136
   genes as candidate targets of Glycyrrhizae Radix et Rhizoma. The
   drug-active compounds-targets network diagrams were shown in [99]Figure
   2 and the target names of Qi-invigorating herbs were summarized in
   [100]Supplementary Table S2.

FIGURE 2.

   [101]FIGURE 2
   [102]Open in a new tab

   Network diagrams of drug-active ingredients-targets. The red nodes
   represent target genes, the yellow nodes represent active ingredients,
   while the blue nodes mean herb. (A). The network diagram of Ginseng
   Radix et Rhizoma-active ingredients-targets; (B). The network diagram
   of Panacis Quinquefolii Radix-active ingredients-targets; (C). The
   network diagram of Pseudostellariae Radix-active ingredients-targets;
   (D). The network diagram of Codonopsis Radix-active
   ingredients-targets; (E). The network diagram of Astragali Radix-active
   ingredients-targets; (F). The network diagram of Dioscoreae
   Rhizoma-active ingredients-targets; (G). The network diagram of
   Glycyrrhizae Radix et Rhizoma-active ingredients-targets.

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 databases. Common targets of DLBCL
   and each Qi-invigorating herb were considered as interacting targets
   between herbs and DLBCL. As shown in [103]Table 1, there existed 14
   interacting targets between Ginseng Radix et Rhizoma and DLBCL, 4
   targets between Panacis Quinquefolii Radix and DLBCL, 22 targets
   between Codonopsis Radix and DLBCL, 22 targets between Pseudostellariae
   Radix and DLBCL, 35 targets between Astragali Radix and DLBCL, 2
   targets between Dioscoreae Rhizoma and DLBCL, and 36 targets between
   Glycyrrhizae Radix et Rhizoma and DLBCL.

TABLE 1.

   Interacting targets between DLBCL and Qi-invigorating herbs.
   Herbs Interacting Targets
   Ginseng Radix et Rhizoma CDK1 ABCB1 AHR TNF PTGS2 BCL2A1 CASP3 ABCG2
   IL10 ICAM1 ABCC1 AR PRKCA RPS6KA3
   Panacis Quinquefolii Radix BCL2A1 CASP3 ICAM1 AR
   Codonopsis Radix CCNB1 CD38 CDK1 TNF CDK4 CCNA2 CASP3 PARP1 ABCG2 IL10
   MAPK1 SYK FOS MAPK3 CDK5R1 CDK2 CDK5 AKT1 JUN AR APP TOP1
   Pseudostellariae Radix CCNB1 CD38 CDK1 CYP1A1 IL5 VEGFA CCNA2 IL13
   STAT1 CASP3 PARP1 ABCG2 SYK FOS ICAM1 CDK5R1 CDK2 CDK5 AKT1 JUN APP
   TOP1
   Astragali Radix NOS2 MET UGT1A1 BIRC5 PLK1 ALK CDK1 CYP1A1 CASP8 ABCB1
   AHR PIM1 HCK AURKB TYMS MCL1 CASP3 KDR ABCG2 HMOX1 LMNB1 IL2 TOP2A
   IGF1R MAPK3 CASP9 AKT1 ABCC1 AR DAPK1 NEK2 PDE4A RPS6KA3 HMGB1 TOP1
   Dioscoreae Rhizoma TNF IL10
   Glycyrrhizae Radix et Rhizoma NOS2 MET UGT1A1 BIRC5 PLK1 ALK BDNF CDK1
   CYP1A1 CASP8 ABCB1 AHR PIM1 HCK AURKB MCL1 CCL2 CASP3 KDR ABCG2 HMOX1
   LMNB1 IL2 TOP2A IGF1R ICAM1 MAPK3 CASP9 AKT1 ABCC1 DAPK1 NEK2 PDE4A
   RPS6KA3 HMGB1 TOP1
   [104]Open in a new tab

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 [105]Table 2. For instance,
   interacting targets between Ginseng Radix et Rhizoma and DLBCL had main
   KEGG pathways in African trypanosomiasis, NF-kappa B signaling pathway,
   amoebiasis, TNF signaling pathway, natural killer cell mediated
   cytotoxicity; while interacting targets between Panacis Quinquefolii
   Radix and DLBCL had main pathways in viral myocarditis, NF-kappa B
   signaling pathway, TNF signaling pathway; interacting targets between
   Codonopsis Radix and DLBCL had main pathways in hepatitis B, T cell
   receptor signaling pathway, pertussis, viral carcinogenesis,
   progesterone-mediated oocyte maturation; interacting targets between
   Pseudostellariae Radix and DLBCL had main pathways in hepatitis B,
   Epstein-Barr virus infection, TNF signaling pathway, pathways in
   cancer, osteoclast differentiation; interacting targets between
   Astragali Radix and DLBCL had main pathways in pathways in cancer,
   progesterone-mediated oocyte maturation, toxoplasmosis, oocyte meiosis,
   colorectal cancer; interacting targets between Dioscoreae Rhizoma and
   DLBCL had main pathways in asthma, African trypanosomiasis, allograft
   rejection, malaria, inflammatory bowel disease; interacting targets
   between Glycyrrhizae Radix et Rhizoma and DLBCL had main pathways in
   pathways in cancer, progesterone-mediated oocyte maturation, chagas
   disease, TNF signaling pathway, toxoplasmosis.

TABLE 2.

   GO and signaling pathway analysis of common targets between
   Qi-invigorating herbs and DLBCL.
   Herbs Type Term Function Count p value
   Ginseng Radix et Rhizoma Biological Process GO:0042493 response to drug
   8 9.30E-10
   GO:0051384 response to glucocorticoid 4 1.54E-05
   GO:0043066 negative regulation of apoptotic process 5 3.13E-04
   GO:0045429 positive regulation of nitric oxide biosynthetic process 3
   4.91E-04
   GO:0006810 transport 4 0.002163
   Cellular Component GO:0045121 membrane raft 3 0.009137
   GO:0005886 plasma membrane 8 0.014002
   GO:0005654 nucleoplasm 6 0.036623
   GO:0005737 cytoplasm 8 0.049495
   Molecular Function GO:0042626 ATPase activity, coupled to transmembrane
   movement of substances 3 5.09E-04
   GO:0005515 protein binding 13 0.002657
   GO:0005524 ATP binding 6 0.0038
   GO:0008559 xenobiotic-transporting ATPase activity 2 0.003845
   GO:0005215 transporter activity 3 0.01019
   KEGG pathway hsa05143 African trypanosomiasis 4 2.15E-05
   hsa04064 NF-kappa B signaling pathway 4 3.96E-04
   hsa05146 Amoebiasis 4 7.07E-04
   hsa04668 TNF signaling pathway 4 7.27E-04
   hsa04650 Natural killer cell mediated cytotoxicity 4 0.001065
   Panacis Quinquefolii Radix Biological Process GO:0043200 response to
   amino acid 2 0.005528
   GO:0097192 extrinsic apoptotic signaling pathway in absence of ligand 2
   0.006062
   GO:0051092 positive regulation of NF-kappaB transcription factor activi
   2 0.023575
   Cellular Component GO:0045121 membrane raft 2 0.033531
   KEGG pathway hsa05416 Viral myocarditis 2 0.024656
   hsa04064 NF-kappa B signaling pathway 2 0.037469
   hsa04668 TNF signaling pathway 2 0.045948
   Codonopsis Radix Biological Process GO:0042493 response to drug 10
   4.51E-11
   GO:0018105 peptidyl-serine phosphorylation 8 1.14E-10
   GO:0045893 positive regulation of transcription, DNA-templated 10
   4.74E-09
   GO:0051090 regulation of sequence-specific DNA binding transcription
   factor activity 5 2.25E-08
   GO:0018107 peptidyl-threonine phosphorylation 5 1.30E-07
   Cellular Component GO:0005654 nucleoplasm 16 1.20E-08
   GO:0005634 nucleus 19 1.58E-07
   GO:0005829 cytosol 15 1.35E-06
   GO:0005667 transcription factor complex 5 6.34E-05
   GO:0043234 protein complex 6 8.70E-05
   Molecular Function GO:0004693 cyclin-dependent protein serine/threonine
   kinase activity 6 4.85E-10
   GO:0004674 protein serine/threonine kinase activity 8 2.29E-07
   GO:0004672 protein kinase activity 7 3.67E-06
   GO:0016301 kinase activity 6 9.60E-06
   GO:0005515 protein binding 21 2.23E-05
   KEGG pathway hsa05161 Hepatitis B 10 8.84E-11
   hsa04660 T cell receptor signaling pathway 8 7.37E-09
   hsa05133 Pertussis 7 4.71E-08
   hsa05203 Viral carcinogenesis 9 5.03E-08
   hsa04914 Progesterone-mediated oocyte maturation 7 1.16E-07
   Pseudostellariae radix Biological Process GO:0042493 response to drug 9
   2.22E-10
   GO:0045893 positive regulation of transcription, DNA-templated 7
   7.50E-06
   GO:0032496 response to lipopolysaccharide 5 1.89E-05
   GO:0045944 positive regulation of transcription from RNA polymerase II
   promoter 7 2.77E-04
   GO:0006954 inflammatory response 5 4.82E-04
   Cellular Component GO:0005667 transcription factor complex 4 7.13E-04
   GO:0005654 nucleoplasm 9 0.001951
   GO:0005634 nucleus 12 0.002961
   GO:0005829 cytosol 8 0.023221
   GO:0005615 extracellular space 5 0.032528
   Molecular Function GO:0019899 enzyme binding 6 1.48E-05
   GO:0070412 R-SMAD binding 3 1.98E-04
   GO:0005515 protein binding 17 2.50E-04
   GO:0042802 identical protein binding 6 6.72E-04
   GO:0003677 DNA binding 6 0.021361
   KEGG pathway hsa05161 Hepatitis B 7 5.32E-07
   hsa05169 Epstein-Barr virus infection 6 6.04E-06
   hsa04668 TNF signaling pathway 5 8.72E-05
   hsa05200 Pathways in cancer 7 1.64E-04
   hsa04380 Osteoclast differentiation 5 1.91E-04
   Astragali Radix Biological Process GO:0043066 negative regulation of
   apoptotic process 12 5.79E-10
   GO:0046777 protein autophosphorylation 9 1.15E-09
   GO:0008283 cell proliferation 10 2.45E-08
   GO:0042493 response to drug 9 9.86E-08
   GO:0006468 protein phosphorylation 10 1.60E-07
   Cellular Component GO:0005654 nucleoplasm 17 1.18E-05
   GO:0005829 cytosol 18 2.47E-05
   GO:0005634 nucleus 23 2.53E-05
   GO:0030496 midbody 5 9.43E-05
   GO:0043234 protein complex 6 9.34E-04
   Molecular Function GO:0005524 ATP binding 18 6.29E-10
   GO:0004674 protein serine/threonine kinase activity 9 6.15E-07
   GO:0004672 protein kinase activity 8 6.09E-06
   GO:0005515 protein binding 31 8.61E-06
   GO:0019899 enzyme binding 7 4.75E-05
   KEGG pathway hsa05200 Pathways in cancer 11 1.71E-06
   hsa04914 Progesterone-mediated oocyte maturation 6 2.25E-05
   hsa05145 Toxoplasmosis 6 7.00E-05
   hsa04114 Oocyte meiosis 6 7.30E-05
   hsa05210 Colorectal cancer 5 1.04E-04
   Dioscoreae Rhizoma Biological Process GO:0032800 receptor biosynthetic
   process 2 1.79E-04
   GO:0002740 negative regulation of cytokine secretion involved in immune
   response 2 3.57E-04
   GO:0034116 positive regulation of heterotypic cell-cell adhesion 2
   6.55E-04
   GO:0044130 negative regulation of growth of symbiont in host 2 9.53E-04
   GO:0050715 positive regulation of cytokine secretion 2 0.001489
   Molecular Function GO:0005125 cytokine activity 2 0.010426
   KEGG pathway hsa05310 Asthma 2 0.004361
   hsa05143 African trypanosomiasis 2 0.004797
   hsa05330 Allograft rejection 2 0.005379
   hsa05144 Malaria 2 0.007123
   hsa05321 Inflammatory bowel disease 2 0.009304
   Glycyrrhizae Radix et Rhizoma Biological Process GO:0043066 negative
   regulation of apoptotic process 12 8.36E-10
   GO:0046777 protein autophosphorylation 9 1.49E-09
   GO:0006468 protein phosphorylation 11 1.45E-08
   GO:0042493 response to drug 9 1.27E-07
   GO:0046677 response to antibiotic 5 4.84E-07
   Cellular Component GO:0030496 midbody 5 1.06E-04
   GO:0005654 nucleoplasm 15 3.51E-04
   GO:0005829 cytosol 16 6.24E-04
   GO:0005634 nucleus 21 6.56E-04
   GO:0043234 protein complex 5 0.007748
   Molecular Function GO:0005524 ATP binding 18 1.12E-09
   GO:0004672 protein kinase activity 9 5.52E-07
   GO:0004674 protein serine/threonine kinase activity 9 7.82E-07
   GO:0005515 protein binding 31 3.01E-05
   GO:0016301 kinase activity 6 1.30E-04
   KEGG pathway hsa05200 Pathways in cancer 10 2.13E-05
   hsa04914 Progesterone-mediated oocyte maturation 6 2.69E-05
   hsa05142 Chagas disease (American trypanosomiasis) 6 6.38E-05
   hsa04668 TNF signaling pathway 6 7.31E-05
   hsa05145 Toxoplasmosis 6 8.35E-05
   [106]Open in a new tab

   If there were more than five terms enriched in this category, top five
   terms were selected according to P value and gene count≧ 5;

   Count: the number of enriched genes in each term.

Construction of PPI Networks and Screening of Hub Genes

   Using the STRING online database and Cytoscape software, interacting
   targets between Ginseng Radix et Rhizoma and DLBCL were filtered into
   the PPI network complex, containing 14 nodes and 38 edges. Among the 14
   nodes, 11 nodes were considered significant, with a combined score ≧
   0.7. CASP3 was identified as the hub gene with the filtering of degree
   ≧ 5. Similarly, 20 nodes were considered significant among 22
   interacting targets between Codonopsis Radix and DLBCL. 14 genes were
   identified as hub genes with the filtering of degree ≥ 5 criteria,
   including JUN, CASP3, CDK2, AKT1, MAPK3, MAPK1, CDK5, CCNB1, CDK1,
   CDK4, CCNA2, FOS, TNF, APP.19 nodes were considered significant among
   22 interacting targets between Pseudostellariae Radix and DLBCL. 10 hub
   genes were identified with the filtering of degree ≥ 5 criteria,
   including JUN, CASP3, AKT1, CDK5, CDK2, CCNB1, CCNA2, STAT1, VEGFA,
   CDK1. Besides, 31 nodes were considered significant among 35
   interacting targets between Astragali Radix and DLBCL. 14 hub genes
   were identified with the filtering of degree ≥ 5 criteria, including
   CASP3, MAPK3, CDK1, AKT1, BIRC5, TOP2A, AURKB, MCL1, PLK1, NEK2, AR,
   CASP9, CASP8, TYMS. Furthermore, 32 nodes were considered significant
   among 36 interacting targets between Glycyrrhizae Radix et Rhizoma and
   DLBCL. 13 genes were identified as hub genes with the filtering of
   degree ≥ 5 criteria, including CASP3, MAPK3, AKT1, BIRC5, CDK1, TOP2A,
   MCL1, AURKB, NEK2, IL2, CASP9, BDNF, PLK1. No hub genes were found in
   Panacis Quinquefolii Radix and Dioscoreae Rhizoma.

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 [107]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.

FIGURE 3.

   [108]FIGURE 3
   [109]Open in a new tab

   Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analyses of
   hub genes. Each figure contains top ten pathways of Gene Ontology or
   Kyoto Encyclopedia of Genes and Genomes. In Gene Ontology enrichment
   analysis, Y-axis is biological process (BP), cellular component (CC)
   and molecular function (MF), X-axis is enrichment score. In Kyoto
   Encyclopedia of Genes and Genomes analysis, Y-axis is pathway name,
   X-axis is enrichment score. Bubble size represents gene numbers
   involved in the pathway enrichment. Color represents p-value, the
   redder the color, the smaller the p-value. (A). Gene Ontology analysis
   of hub genes for Codonopsis Radix; (B). Pathway enrichment analysis of
   hub genes for Codonopsis Radix; (C). Gene Ontology analysis of hub
   genes for Pseudostellariae Radix; (D). Pathway enrichment analysis of
   hub genes for Pseudostellariae Radix; (E). Gene Ontology analysis of
   hub genes for Astragali Radix; (F). Pathway enrichment analysis of hub
   genes for Astragali Radix; (G). Gene Ontology analysis of hub genes for
   Glycyrrhizae Radix et Rhizoma; (H). Pathway enrichment analysis of hub
   genes for Glycyrrhizae Radix et Rhizoma.

   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.

Molecular Docking Verification

   Through AutoDock Vina, all the hub genes and their related active
   ingredients were run for molecular docking. As shown in [110]Figure 4,
   16 active ingredients and 28 hub genes were run via AutoDock Vina for
   molecular docking. The spatial coordinates and the lowest binding
   affinity are shown in [111]Table 3. Results in [112]Figure 4 showed
   that in Ginseng Radix et Rhizoma, beta-sitosterol had a good affinity
   to CASP3 through hydrogen bonding. In Codonopsis Radix, luteolin might
   suppress cancer cell growth through hydrogen-bonding with AKT1, APP,
   CASP3, CCNA2, CCNB1, CDK1, CDK2, CDK5, FOS, JUN; glycitein could
   hydrogen-bond with CDK4, JUN, MAPK1, MAPK3; while stigmasterol only
   bound with TNF by van der Waals forces. In Pseudostellariae Radix, good
   affinities were also detected between luteolin and hub genes, including
   AKT1, CASP3, CCNA2, CCNB1, CDK1, CDK2, CDK5, JUN. Besides, acacetin and
   JUN, STAT1, VEGFA were bound together by hydrogen bonds.
   Beta-sitosterol took effect through CASP3. In Astragali Radix,
   quercetin might hold together with AKT1, AUKRB, CDK1, NEK2, PLK1
   through hydrogen bonds, with MCL1 by van der Waals forces; mairin was
   bound with AKT1, BIRC5, TOP2A by hydrogen bonds, with CASP3 by van der
   Waals forces. Formononetin might induce cancer apoptosis through
   hydrogen-bonding with CASP3, CASP8, CASP9. Isorhamnetin, kaempferol,
   FA, calycosin, hederagenin might play an anti-tumor role through
   targeting AKT1, CDK1 TYMS, MAPK3, AR, respectively. As for Glycyrrhizae
   Radix et Rhizoma, similar results were found in quercetin,
   formononetin, isorhamnetin, kaempferol, calycosin and mairin.
   Meanwhile, CASP3, AKT1, IL2, BDNF could also bond with sitosterol,
   formononetin, naringenin,
   2-(3,4-dihydroxyphenyl)-5,7-dihydroxy-6-(3-methylbut-2-enyl) chromone,
   respectively.

FIGURE 4.

   [113]FIGURE 4
   [114]Open in a new tab

   Molecular docking results: A1-A5. CASP3 and beta-sitosterol, luteolin,
   formononetin, mairin, sitosterol, respectively; B1-B5. AKT1 and
   luteolin, isorhamnetin, mairin,
   quercetin,2-(3,4-dihydroxyphenyl)-5,7-dihydroxy-6-(3-methylbut-2-enyl)c
   hromone, respectively; C1-C3. CDK1 and luteolin, kaempferol, quercetin,
   respectively; D1-D2. MAPK3 and glycitein, calycosin, respectively;
   E1-E7. luteolin and JUN, CDK2, CDK5, CCNB1, CCNA2, APP, FOS,
   respectively; F1-F3. glycitein and JUN, MAPK1, CDK4, respectively;
   G1-G3. acacetin and JUN, STAT1, VEGFA, respectively; H1-H2. mairin and
   BIRC5, TOP2A, respectively; I1-I4. quercetin and AUKRB, MCL1, NEK2,
   PLK1, respectively; J. hederagenin and AR; K1-K3. formononetin and IL2,
   CASP8, CASP9, respectively; L. stigmasterol and TNF; M. naringenin and
   BDNF; N. FA and TYMS.

TABLE 3.

   Docking results of active ingredients with target proteins.
   Targets PDB-ID Ligands Affinity (kcal/mol) Center coordinates(x,y,z)/nm
   CASP3 3KJF beta-sitosterol −5.0 10.319 6.913 2.268
   CASP3 3KJF luteolin −7.5 10.319 6.913 2.268
   CASP3 3KJF formononetin −7.3 10.319 6.913 2.268
   CASP3 3KJF mairin −8.2 10.319 6.913 2.268
   CASP3 3KJF sitosterol −3.5 10.319 6.913 2.268
   AKT1 6S9W luteolin −10.1 −13.631 −12.55 15.018
   AKT1 6S9W isorhamnetin −9.9 −13.631 −12.55 15.018
   AKT1 6S9W mairin −7.6 −13.631 −12.55 15.018
   AKT1 6S9W quercetin −7.7 −13.631 −12.55 15.018
   AKT1 6S9W
   2-(3,4-dihydroxyphenyl)-5,7-dihydroxy-6-(3-methylbut-2-enyl)chromone
   −10.6 −13.631 −12.55 15.018
   CDK1 4YC6 luteolin −8.4 −14.325 12.348 −22.28
   CDK1 4YC6 kaempferol −7.7 −14.325 12.348 −22.28
   CDK1 4YC6 quercetin −7.9 −14.325 12.348 −22.28
   MAPK3 6GES glycitein −8.1 46.838 4.48 −8.117
   MAPK3 6GES calycosin −8.1 46.838 4.48 −8.117
   JUN 1A02 luteolin −8.4 28.223 28.762 60.016
   CDK2 6GUE luteolin −9.6 −29.87 −7.262 18.182
   CDK5 3O0G luteolin −9.3 16.519 34.169 52.349
   CCNB1 2B9R luteolin −7.0 −65.873 41.523 −11.34
   CCNA2 5IF1 luteolin −7.8 −61.902 -39.304 −11.927
   APP 1OWT luteolin −6.3 −0.013 -0.096 0.004
   FOS 1A02 luteolin −8.4 28.223 28.762 60.016
   JUN 1A02 glycitein −7.4 28.223 28.762 60.016
   MAPK1 1PME glycitein −7.1 −4.477 8.779 47.428
   CDK4 4YC6 glycitein −8.4 6.503 5.666 29.291
   JUN 1A02 acacetin −7.9 28.223 28.762 60.016
   VEGFA 3QTK acacetin −7.8 31.497 18.909 7.02
   STAT1 1YVL acacetin −8.0 −15.938 −29.561 171.162
   BIRC5 2RAX mairin −7.1 26.692 −61.209 4.957
   TOP2A 1ZXN mairin −7.5 41.327 23.743 40.984
   AUKRB 4AF3 quercetin −6.1 15.553 −16.602 −3.528
   MCL1 6B4U quercetin −7.6 −8.597 15.908 10.321
   NEK2 2XKD quercetin −6.9 15.251 −11.912 17.204
   PLK1 2OGQ quercetin −7.1 55.303 16.8 27.793
   AR 2PIV hederagenin −5.3 21.023 5.462 11.667
   IL2 1M47 formononetin −5.6 6.751 25.891 14.188
   CASP8 5JQE formononetin −7.0 169.678 29.358 -0.457
   CASP9 2AR9 formononetin −8.5 19.147 38.329 27.142
   TNF 3WD5 stigmasterol −5.8 −20.065 21.641 22.187
   BDNF 1B8M naringenin −6.8 2.715 23.339 14.231
   TYMS 3N5E FA −7.1 −48.904 0.049 11.859
   [115]Open in a new tab

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 [116]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% ([117]Figures 5B,C). In view of the small number of
   samples, a retrospective study with larger sample is expected to
   confirm the conclusion.

FIGURE 5.

   [118]FIGURE 5
   [119]Open in a new tab

   Verification of hub genes in the GEPIA 2.0 database. (A). Expression of
   hub genes in DLBCL as compared to those in non-tumor tissues; (B).
   Correlation between CASP3 and disease-free survival in DLBCL; (C).
   Correlation between BDNF and disease-free survival in DLBCL.

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 ([120]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.

FIGURE 6.

   FIGURE 6
   [121]Open in a new tab

   A Venn diagram for common hub genes. The red part represents Codonopsis
   Radix; the green part represents Ginseng Radix et Rhizoma; the yellow
   part represents Glycyrrhizae Radix et Rhizoma; the brown part
   represents Pseudostellariae Radix; the blue part represents Astragali
   Radix.

Experimental Verification

Cytotoxicity Study of Qi-Invigorating Herbs in Diffuse Large B Cell Lymphoma
Cells

   In order to verify the effect of common hub genes in Qi-invigorating
   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 ([122]Figures 7A,B).

FIGURE 7.

   [123]FIGURE 7
   [124]Open in a new tab

   Experimental validation in vitro. (A). The proliferative ability of
   SUDHL2 cells incubating with different Qi-invigorating herbs; (B). The
   proliferative ability of SUDHL6 cells incubating with different
   Qi-invigorating herbs; (C). The cell apoptosis of SU-DHL-2 cells
   incubating with different Qi-invigorating herbs was measured through
   flow cytometry analysis (From upper left to right: Negative control,
   SUDHL2 cells treated with Astragali Radix, SUDHL2 cells treated with
   Glycyrrhizae Radix et Rhizoma; From lower left to right: SUDHL2 cells
   treated with Ginseng Radix et Rhizoma, SUDHL2 cells treated with
   Pseudostellariae Radix, SUDHL2 cells treated with Codonopsis Radix);
   (D). The cell apoptosis of SU-DHL-6 cells incubating with different
   Qi-invigorating herbs was measured through flow cytometry analysis
   (From upper left to right: Negative control, SUDHL6 cells treated with
   Astragali Radix, SUDHL6 cells treated with Glycyrrhizae Radix et
   Rhizoma; from lower left to right: SUDHL6 cells treated with Ginseng
   Radix et Rhizoma, SUDHL6 cells treated with Pseudostellariae Radix,
   SUDHL6 cells treated with Codonopsis Radix); (E). mRNA expression of
   CASP3, AKT1, CDK1, MAPK3 in SUDHL2 cells incubating with different
   Qi-invigorating herbs; (F). mRNA expression of CASP3, AKT1, CDK1, MAPK3
   in SUDHL6 cells incubating with different Qi-invigorating herbs. *p <
   0.05 and **p < 0.01.

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 ([125]Figures 7C,D). Meanwhile
   cell apoptosis in DLBCL cells treated with Ginseng Radix et Rhizoma was
   slightly elevated ([126]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 shown in [127]Figures 7E,F,
   DLBCL cells treated with Glycyrrhizae Radix et Rhizoma or Ginseng Radix
   et Rhizoma had an effectively increase in CASP3 at mRNA levels compared
   with negative control groups, suggesting that Qi-invigorating herbs
   regulated DLBCL cell apoptosis mainly through targeting CASP3. It was
   noteworthy that most CDK1, AKT1, MAPK3 expression did not change
   significantly in DLBCL cells treated with Qi-invigorating herbs
   ([128]Figures 7E,F). Therefore, it is concluded that CDK1, AKT1 and
   MAPK3 are not common hub genes in DLBCL cells treated with
   Qi-invigorating herbs.

Tumor Microenvironment Analysis

   To further explore the mechanisms of Qi-invigorating herbs acting on
   DLBCL TME, we conducted the analysis of the relationship between hub
   genes and tumor-infiltrating immune cells (B cells, CD4^+ T cells,
   CD8^+ T cells, neutrophils, macrophages, myeloid DCs, MDSCs) from the
   TIMER 2.0 database. As shown in [129]Figure 8, the expression of AKT1
   was negatively associated with tumor purity (r = −0.386, p = 1.15e-02),
   positively correlated with CD8^+ T cells (r = 0.523, p = 4.50e-04) and
   neutrophils (r = 0.583, p = 6.44e-05) in infiltrating levels. Similar
   relationships were found in CASP8, JUN, APP and MCL1 (p < 0.05). CASP3,
   CDK2, MAPK1, TOP2A, MAPK3 only had positive significant correlation
   with CD8^+ T cells and neutrophils (r > 0, p < 0.01). CDK1, CCNA2, FOS
   had statistical correlation with CD4^+ T cells and neutrophils (p <
   0.05), while BIRC5 and CCNB1 correlated with both neutrophils and MDSCs
   (p < 0.05). PLK1, VEGFA, AR only correlated with neutrophils (p <
   0.05), TYMS, AURKB, TNF mainly correlated with MDSCs (p < 0.05). The
   expression of STAT1 was negatively associated with tumor purity (r =
   −0.451, p = 2.69e-03), B cells (r = −0.364, p = 1.94e-02), macrophage
   (r = −0.499, p = 8.89e-04) and MDSCs (r = −0.531, p = 3.57e-04) in
   infiltrating levels, but positively correlated with CD8^+ T cells (r =
   0.641, p = 6.20e-06), myeloid dendritic cells (r = 0.62, p = 1.55e-05)
   and neutrophils (r = 0.809, p = 1.55e-10) in infiltrating levels.
   However, no correlation was found among BDNF, CDK5, IL2 and
   tumor-infiltrating immune cells in DLBCL. Since most hub genes were
   associated with neutrophils in infiltrating levels, we speculated that
   Qi-invigorating herbs mainly affects DLBCL microenvironment through
   acting on neutrophils.

FIGURE 8.

   FIGURE 8
   [130]Open in a new tab

   The relationship between hub genes and tumor-infiltrating immune cells.
   (A). The relationship between AKT1 and tumor-infiltrating immune cells
   (CD8^+ T cells and neutrophils); (B). The relationship between AURKB
   and tumor-infiltrating immune cells (CD4^+ T cells and MDSCs); (C). The
   relationship between APP and tumor-infiltrating immune cells (CD4^+ T
   cells, CD8^+ T cells and neutrophils); (D). The relationship between AR
   and neutrophils; (E). The relationship between CASP3 and
   tumor-infiltrating immune cells (CD8^+ T cells and neutrophils); (F).
   The relationship between CASP8 and tumor-infiltrating immune cells
   (CD8^+ T cells and neutrophils); (G). The relationship between CCNA2
   and tumor-infiltrating immune cells (CD4^+ T cells and neutrophils);
   (H). The relationship between CDK1 and tumor-infiltrating immune cells
   (CD4^+ T cells and neutrophils); (I). The relationship between CCNB1
   and tumor-infiltrating 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). The relationship between TNF and
   myeloid dendritic cells, MDSCs. Myeloid derived suppressor cells: MDSC.

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, Qi-invigorating 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 Qi-invigorating 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 Qi-invigorating 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 ([131]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 GSDME-expressing tumors ([132]Wang et al., 2017;
   [133]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
   ([134]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 apoptosis
   mainly through targeting CASP3. In view of the correlation between
   CASP3 and DLBCL prognosis, it is of great significance to further
   explore the mechanisms of Glycyrrhizae Radix et Rhizoma and Ginseng
   Radix et Rhizoma acing on CASP3 in DLBCL.

   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 Qi-invigorating herb,
   indicating they have potential as DLBCL drugs. So far, some of these
   ingredients have been well described for their anti-lymphoma 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 ([135]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 ([136]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 ([137]Azzaoui et al., 2016;
   [138]Chen et al., 2016; [139]Judd et al., 2017; [140]Kusano et al.,
   2017; [141]Ciavarella et al., 2018; [142]Staiger et al., 2020;
   [143]Merdan et al., 2021; [144]Xu et al., 2021). Manfroi et al.
   ([145]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 ([146]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.

Acknowledgments