Abstract Gastric cancer (GC) is one of the most prominent malignancies that originate in the epithelial cells of the gastric mucosa and is one of the main causes of cancer-related mortality worldwide. New circulating biomarkers of exosomal RNA might have great potential for non-invasive early prognosis of GC. Sijunzi Decoction (SJZD) is a typical representative formula of the method of benefiting Qi and strengthening the spleen in Traditional Chinese Medicine (TCM). However, the effects and mechanism of SJZD in treating GC remain unclear. This study looked for biomarkers of exosomal RNA for early prognosis of GC, and explored the mechanism of SJZD in treating GC. A gastric cancer model with spleen deficiency syndrome was established in nude mice, and the curative effects of SJZD were investigated. Differentially expressed miRNAs in plasma and saliva exosomes were sequenced and analyzed. Potential target genes of these miRNAs were predicted and applied for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) signaling pathway enrichment annotation. Overlapping miRNAs in saliva and plasma samples were analyzed, and qRT-PCR was performed for verification. miR-151a-3p was selected, and qRT-PCR further determined that miR-151a-3p was downregulated in saliva and plasma exosomes from the SJZD group. The intersected miR-151a-3p target genes were predicted and enriched in the extrinsic apoptotic signaling pathways. SJZD significantly ameliorates gastric cancer with spleen deficiency syndrome in mouse models, and exosomal miRNAs, particularly miR-151-3p, might be modulated by SJZD in plasma and saliva. The exosomal miR-151-3p in saliva may serve as a non-invasive potential marker for gastric cancer diagnosis and prognosis. Keywords: Gastric cancer, Exosomal miRNAs, Sijunzi decoction (SJZD), Spleen deficiency syndrome, Exosomal miR-151-3p 1. Introduction Gastric cancer (GC) originates from the gastric mucosa's epithelial cells and is a leading cause of cancer-related deaths worldwide due to its rapid progression and high recurrence rate [[45]1,[46]2]. Early diagnosis is challenging due to non-specific symptoms [[47][3], [48][4], [49][5]]. While gastroscopy with pathological examination is the most effective screening method, its intrusiveness makes it impractical for populous countries [[50]6,[51]7]. Thus, exploring non-invasive diagnostic methods for early GC detection is crucial. Exosomes are extracellular vesicles, 40–150 nm in diameter, secreted by all cell types [[52][8], [53][9], [54][10]]. Being highly heterogeneous [[55]11] and reflecting the phenotype of their parent cells, they contain proteins, RNA, and DNA [[56][12], [57][13], [58][14]]. While their primary function remains unclear, they are believed to act as cellular waste disposal systems [[59]15] and play roles in intercellular communication and signal transduction [[60]14,[61]16]. In the tumor environment, exosomes can promote carcinogenesis by influencing angiogenesis, immunity, and metastasis [[62][17], [63][18], [64][19]]. Exosomes have been detected in various bodily fluids, including blood [[65]20,[66]21], and elevated exosome levels have been observed in the circulation of patients with several cancers, including gastric carcinoma [[67][22], [68][23], [69][24]]. Furthermore, Baran et al. reported that the mRNA levels of melanoma antigen-1 (MAGE-1) and human epidermal growth factor receptor-2 (HER2) were dramatically elevated in plasma exosomes of patients with gastric carcinoma [[70]25]. Thus, circulating exosomes hold promise as non-invasive biomarkers for early cancer detection and treatment. New circulating exosomal RNA biomarkers might have great potential for GC non-invasive prognosis. Plasma exosomes, with their bilayer vesicle structure, offer enhanced stability and advantages over traditional tumor markers in tumorigenesis and metastasis [[71][26], [72][27], [73][28]]. Saliva, a diagnostic fluid secreted by various glands, contains nucleic acids, proteins, and small molecules similar to serum, reflecting body pathophysiological changes [[74][29], [75][30], [76][31]]. Salivary exosomal RNA has been used to detect oral diseases [[77]32,[78]33] and cancers like breast and pancreatic [[79]29]. Notably, GAPDH mRNA from salivary exosomes was identified in lung cancer-bearing mice [[80]34]. This suggests exosomes from distant tissues might be absorbed by salivary glands and released into saliva. Thus, salivary exosomal mRNA holds potential as a diagnostic marker for gastric cancer. MicroRNAs (miRNAs) are a class of small noncoding RNA molecules, typically consisting of 19–25 nucleotides. They have been shown to significantly impact gene expression regulation, particularly in relation to oncogenes and tumor suppressor genes. As a result, miRNAs are increasingly being recognized as a promising molecular tool for the noninvasive diagnosis and prognosis of cancer [[81]35]. Several miRNAs implicated in the development of GC have been discovered, and several miRNAs are now under assessment for their potential diagnostic and/or prognostic use [[82]36]. For example, miR-21 [[83]37], miR-27a [[84]38], and miR-150 [[85]39] have been reported to regulate GC cell proliferation. miR-151-3p has been implicated in several malignancies, including colorectal cancer [[86]40], osteosarcoma [[87]41], cholangiocarcinoma [[88]42], head and neck cancer [[89]43], and other malignancies, serving as tumor-suppressive or promotive miRNA. Notably, in GC, miR-151-3p derived from gastric cancer exosomes induces M2-phenotype polarization of macrophages and promotes tumor growth [[90]44]. However, whether exosomal miR-151-3p participates in more physiological processes in GC remains unclear. According to Traditional Chinese Medicine (TCM) theory, gastric cancer is a deficiency symptom, with phlegm coagulation, Qi stagnation, blood stagnation, and evil heat as the symptoms and spleen deficiency as the basis [[91]45,[92]46]. The weakness of the spleen and stomach plays a key role in the initiation and progression of gastric cancer, which exists in the precancerous disease - precancerous lesion - gastric cancer [[93]47,[94]48]. In TCM, spleen deficiency is one of the most prevalent fundamental symptoms, and it is also a key pathogenesis and common symptom of gastric cancer. Sijunzi Decoction (SJZD) is a typical representative formula of the method of benefiting Qi and strengthening the spleen, which has been widely used in digestive system diseases in the past generations with clinical practice experience [[95]49,[96]50]. To enrich the diagnostic method of spleen deficiency in gastric cancer and the efficacy and molecular mechanism of SJZD in strengthening the spleen and Qi, in this study, a model of gastric cancer with spleen deficiency in the nude mouse was established, the expression profiles of plasma exosomal RNA and salivary exosomal RNA in model mice were identified by high-throughput sequencing, applied bioinformatics and biostatistics methods to find differentially expressed miRNAs, and biomarkers with clinical application in diagnosing and treating gastric cancer were searched for. This study aims to identify biomarkers and investigate potential molecular pathways involved in the use of SJZD to treat spleen deficiency in gastric cancer. 2. Materials and methods Ethical statement This study was approved by the Experimental Animal Ethics Committee of Hunan University of Chinese Medicine (approval number: LL2020061501). The experiment was conducted strictly as per NIH guidelines for the Care and Use of Laboratory Animals. 2.1. Cell line and cell culture A human gastric carcinoma cell line, NCI–N87 (CRL-5822), was procured from ATCC (Manassas, VA, USA) and cultivated in RPMI-1640 Medium (30–2001, ATCC) containing 10% Fetal Bovine Serum (FBS, 30–2020, ATCC). All cell types were cultivated at 37 °C in a humidified atmosphere containing 5% CO[2]. 2.2. Animals 99 BALB/c (nu/nu) male nude mice (4–6 weeks old, weighing 16 g–20 g) were used in this experiment. Nude mice were provided by the Experimental Animal Center of Hunan University of Traditional Chinese Medicine (license number: SYXK [Xiang] 2019∼0009). All nude mice were housed in an SPF-grade barrier system, where the following experiments were performed. Mice were fed under controlled environmental conditions (22 ± 2 °C, 55 ± 10% humidity), with unrestricted access to water in separate cages. 2.3. TCM herbs and formula SJZD was formulated with reference to the original formula of "Taiping Huimin Hejiju Fang" [[97]51,[98]52], consisting of the Panax ginseng C.A.Mey., Atractylodes macrocephala Koidz, Poria cocos (Schw.) Wolf., and Glycyrrhiza uralensis Fisch. (in the ratio of 2:2:2:1) in the doses of 12 g, 12 g, 12 g, and 6 g. All the herbs were procured from Anhui Huizhong TCM Tablets Co., Ltd. in accordance with the original formula's proportions (20021002). Experts from the College of Pharmacy at Hunan University of Chinese Medicine certified their compliance with the 2020 edition of the People's Republic of China Pharmacopeia standards. For the consistency of the TCM formulation, SJZD was extracted according to the standard method described in the pharmacopeia. The obtained liquid was concentrated by a rotary evaporator at 75 °C to make the raw drug concentration 2 g/mL; the liquid was cooled and sealed in a dry sterilized container and stored at 4 °C. Sennae Folium decoction was derived from the leaf of Cassia Angustifolia. Vahl, procuring from the Bozhou Yonggang medicinal slices factory (China). Sennae Folium (500 g) was added with 5000 mL of boiling water to seal and soak for 10 min. The drugs were filtered out, and the liquid was evaporated and concentrated into 1 kg/L of liquid by rotary evaporator at 75 °C. After cooling, the decoction was sealed in a dry sterilized container and stored at 4 °C. 2.4. Quality analysis of SJZD by ultra-performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UPLC-QTOF-MS) UPLC was performed on a Waters ACQUITY UPLC system (Milford MA), consisting of a binary solvent delivery system, an autosampler, and a PDA detector system. An ACQUITY UPLC BEH C18 column (1.7 μm, 2.1 mm × 100 mm) was used for all analyses at 35 °C. The mobile phase was composed of 0.1% formic acid (v/v) (A) and methanol (B). The flow rate was 0.3 mL/min with 60 min of gradient elution: 0–10 min, 10-10% B; 10–40 min, 10–100% B; 40–45 min, 100-100% B; 45–47.5 min, 100-10% B; 47.5–60 min, 10-10%B. The MS analysis was performed on a Waters ACQUITY Synapt Q-TOF/MS spectrometer connected to the Waters ACQUITY UPLC system with an electrosprayionization (ESI) source in both positive and negative ionization modes. The detailed parameters were as follows: multiple ion scanning in IDA mode for each compound, scanning range m/z 100–1000, spray voltage 5500 and -4500 V, temperature 550 °C, collision energy 35 ± 15 eV. TOF/MS primary scanning time 100 ms, TOF/MS/MS secondary scanning time 1150 ms. The compounds were screened automatically in the TCM MS/MS Library and TMK TCM Library provided by SCIEX Company (Massachusetts, USA), followed by manual checking in order to remove the false positive signals. 2.5. Establishment of the orthotopic xenotransplantation model of gastric cancer with spleen deficiency syndrome For spleen deficiency syndrome, 48 BALB/c (nu/nu) nude mice were given limited feeding and gavage with Sennae Folium decoction (1 mL/d for 10 days), which has a bitter-cold cathartic effect. The success of the spleen deficiency state was determined by observing the stool condition, body mass, and food and water intake [[99]53]. Referring to the Chinese Medicine Clinical Terminology and the Clinical Research Guidelines for the Treatment of Spleen Deficiency in Chinese Medicine, the success of modeling was determined by scoring the symptoms of spleen deficiency in nude mice: 1 point for tiredness, 1 point for squinting, 1 point for bowed back, 1 point for reduced food intake, 1 point for reduced water intake, 1 point for reduced body mass, 1 point for soft stool, and 2 points for loose stool. An increase in the sum of the scores indicates an increase in symptoms, and a score of 5 or more indicates successful modeling. Three BALB/c (nu/nu) 4-6w male nude mice were also taken, and NCI–N87 cell suspension (2 × 10^6) was inoculated in the axilla on the first day of the experiment to cultivate gastric cancer tissue blocks [[100]54]. Ten days later, when the subcutaneous graft tumors had grown to around 10 mm in diameter, they were peeled off. The small tumor blocks were implanted into the stomachs of spleen-deficient mice that had been successfully bred. 2.6. Treatment of SJZD 72 BALB/c (nu/nu) male nude mice were randomly assigned into 3 groups (24 mice per group), namely control, model, and SJZD groups. Mice in the model and SJZD groups were prepared as gastric cancer models with spleen deficiency syndrome. Then, mice in the model and SJZD groups were given free ingestion of a sterile full-value pellet diet. At the same time, mice in the SJZD group were given SJZD by gavage once daily (10.92 g/kg, the dose was computed according to the drug dose conversion formula of body surface area, equivalent to the twice dose given to a 70 kg adult) for 2 weeks. SJZD was dissolved in saline before oral administration. The control and model groups were administered the same dose of normal saline for two weeks by gavage. The mental status, eye color, behavioral activity, and stool status of the nude mice in each group were observed and recorded from the 11th to the 24th day of the experiment. On the 24th day of the experiment, nude mice were anesthetized and sacrificed after saliva and plasma collection. Gastric cancer tissues were isolated and exfoliated; tumor weight, size, and inhibition rate were measured and calculated. 2.7. Collection of saliva and plasma Saliva was collected from nude mice for 5–10 min on the 22nd and 24th days of the experiment. Pilocarpine (0.05 mg/100 g) was injected subcutaneously between the ears of nude mice to induce saliva secretion. After collection, saliva samples were subjected to a 15-min centrifugation (2600 g, 4 °C) in a high-speed freezing centrifuge and kept at −80 °C. Blood was obtained from the eyes on the 24th day of the experiment, and whole blood was harvested from nude mice using 5 mL of K3 EDTA anticoagulated vacuum blood collection tubes. The plasma was separated by centrifugation as soon as possible. After centrifugation, the supernatant was divided into new sterile centrifuge tubes and stored at −80 °C. 2.8. Exosomal RNA extraction and quality control Exosomal RNAs were extracted from saliva and plasma samples with an exoRNeasy Midi/Maxi Kit (QIAGEN), and all operations were performed as stipulated by the protocol. The eluted RNA was kept at −80 °C. Agarose gel electrophoresis was carried out to analyze RNA degradation and contamination. Nanodrop was employed to determine RNA purity. Qubit® 2.0 Fluorometer (IMPLEN, CA, USA) was applied to quantify RNA concentration. The Agilent Bioanalyzer 2100 system (Agilent Technologies, CA, USA) was used to determine RNA integrity. 2.9. cDNA library construction and quality control A Small RNA Sample Pre-kit was employed to construct the cDNA library. Adding splices at both ends of Small RNA and then reverse transcription were used to make cDNA constructs based on the Small RNA and total RNA ligated with 5′ and 3’ adapters. PAGE gel electrophoresis was employed to separate the target DNA fragments, and gel cutting was employed to recover the cDNA library. Qubit 2.0 was subsequently applied to perform initial quantification, and the library was diluted to 1 ng/ul. Subsequently, Agilent 2100 was employed to examine the insert size of the library. qRT-PCR (effective library concentration higher than 2 nM) was conducted to quantify the effective concentration of the library, thereby validating the library quality. 2.10. Sequencing The libraries were pooled based on the effective concentration, and then the Illumina SE50 sequencing was carried out. DNA polymerase, splice primers, and four dNTPs with base-specific fluorescent labeling (one at a time) were introduced to the reaction system in the sequencing-while-synthesizing method. Unused free dNTPs and DNA polymerase were eluted after adding all four dNTPs. The fluorescence excitation buffer was then introduced, a laser was utilized to excite the fluorescence signal, the fluorescence signal was captured, and then the optical signal was converted into sequencing bases using computer analysis. Chemical reagents were ultimately applied for fluorescence quenching and removal of the dNTP 3′-OH protecting group for the next sequencing cycle. 2.11. Differentially expressed miRNA analysis Differentially expressed miRNAs obtained from the sequencing were analyzed using the R language (Limma package). Differential miRNAs were obtained by t-test using the thresholds of P < 0.01 and Log2(FC) > 1. 2.12. qRT-PCR analysis To further investigate whether SJZD regulated exosomal miRNA expression in a dose-dependent manner, 24 BALB/c (nu/nu) mice were prepared as gastric cancer models with spleen deficiency syndrome. Then, mice in the low-dose, medium-dose, and high-dose SJZD groups were given SJZD by gavage once a day (5.48 g/kg, 10.92 g/kg or 21.84 g/kg, respectively) for 2 weeks. The model groups were administered the same dose of saline by gavage for two weeks. Exosomal RNAs were extracted from saliva and plasma samples from different groups with an exoRNeasy Midi/Maxi Kit (QIAGEN). The reverse transcription of RNA to cDNA was performed by StarScript II First-strand cDNA Synthesis Mix (GenStar, China). The expression of miR-151a-3p was examined using a SYBRGreen Realtime PCR Master Mix (GenStar). The real-time thermal cycler (Jena, German) was used for qRT-PCR. U6 expression was used as an endogenous control. The primer sequence used for qRT-PCR included the following: miR-151a-3p (reverse transcription primer: 5′-GTCGTATCCAGTGCGTGTCGTGGAGTCGGCAATTGCACTGGATACGACCCTCAA-3′, forward: 5′GCGCTAGACTGAAGCTCC-3’; reverse: 5′-CAGTGCGTGTCGTGGA-3′). U6 (forward: 5′-CTCGCTTCGGCAGCACA-3′, reverse: 5′-AACGCTTCACGAATTTGCGT-3′). 2.13. Integrated target prediction of miR-151a-3p and signaling and functional enrichment annotation The putative miR-151a-3p target genes (with score>0.95 and binding sites 3′UTR) were analyzed using MiRWalk 3.0 ([101]http://mirwalk.umm.uni-heidelberg.de/) and miRNA-target prediction tool Targetscan7.2 ([102]http://www.targetscan.org/vert_72/). For the obtained 22 overlapping genes, functional enrichment of Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways analyses were performed using Metascape (metascape.org/gp/index.html). Metascape is a web-based portal designed to provide a comprehensive gene list annotation and analysis resource for experimental biologists [[103]55]. 2.14. Statistical analysis All data were presented in terms of [MATH: −x :MATH] ±s, and data analysis was performed using SPSS 24.0 software. If the data conformed to the normal distribution and chi-square test, the comparison among groups was conducted using one-way ANOVA. Dunnett-t3 method test was used for those who did not conform to the normal distribution and chi-square test. A P-value of less than 0.01 was considered statistically significant. 3. Results 3.1. Quality analysis of SJZD The main compounds of SJZD were analyzed using UPLC-QTOF-MS for validation. Totally 106 compounds were identified under negative mode, whereas 101 compounds were identified under positive mode ([104]Supplemental Tables 1 and 2). The Elementary particle flow graph chromatogram in positive and negative ion mode for SJZD is shown in [105]Fig. 1. Top 40 effective active components according to relative peak area were marked in [106]Fig. 1 and [107]Table 1. Fig. 1. [108]Fig. 1 [109]Open in a new tab The Elementary particle flow graph chromatogram in positive and negative ion mode for SJZD. (A) Negative ion Mode; (B) Positive ion Mode. The top 40 effective active components according to relative peak area are marked. 1, Quinic acid; 2, Proline; 3, l-Malic acid; 4, Adenine; 5, Pipecolinic acid; 6, Citric acid; 7, Nicotinamide; 8, Adenosine; 9, Isoleucine; 10, Phenprobamate; 11, L-Tryptophan; 12, p-hydroxyl phenylpropanol; 13, Isoscopoletin; 14, Vicenin-2; 15, Schaftoside; 16, Isoliquiritin apioside; 17, Liquiritin; 18, Liquiritin apioside; 19, Isoliquiritigenin; 20, Liquiritigenin; 21, Isoliquiritin; 22, Ononin; 23, Hydroxygenkwanin; 24, Naringenin; 25, Ginsenoside Re; 26, Ginsenoside Rg1; 27, Ganoderol B; 28, Formononetin; 29, Ginsenoside Rf; 30, Glabrene; 31, Atractylenolide Ⅲ; 32, Dehydrocostus lactone; 33, Parthenolide; 34, Ginsenoside Rg2; 35, 20(R)-Ginsenoside Rh1; 36, Germacrone; 37, Isoalantolactone; 38, Ginsenoside Rd; 39, Heterophyllin B; 40, Diammonium glycyrrhizinate. Table 1. Top 40 effective active components of SJZD according to relative peak area. n Retention Time Library Hit Molecular Mass Negative/Positive ion Relative Peak area(%) 1 0.83 Quinic acid 192.06294 Negative ion 0.66% 2 0.9 Proline 115.06299 Positive ion 1.37% 3 0.91 l-Malic acid 134.02021 Negative ion 1.31% 4 0.96 Adenine 135.05389 Positive ion 1.13% 5 1.11 Pipecolinic acid 129.04227 Positive ion 2.37% 6 1.14 Citric acid 192.02687 Negative ion 10.32% 7 1.14 Nicotinamide 122.04787 Positive ion 0.66% 8 1.35 Adenosine 267.0963 Positive ion 1.24% 9 1.52 Isoleucine 131.09443 Positive ion 1.15% 10 2.41 Phenprobamate 165.07875 Positive ion 2.33% 11 4.24 L-Tryptophan 204.08901 Negative ion 0.76% 12 6.53 p-hydroxyl phenylpropanol 166.06183 Negative ion 8.06% 13 16.77 Isoscopoletin 192.04188 Positive ion 0.95% 14 17.41 Vicenin-2 594.16033 Negative ion 0.45% 15 18.63 Schaftoside 564.14925 Negative ion, Positive ion 0.64% 16 18.84 Isoliquiritin apioside 550.16963 Negative ion, Positive ion 2.07% 17 18.99 Liquiritin 418.12528 Negative ion, Positive ion 13.95% 18 19.45 Liquiritin apioside 550.16804 Negative ion, Positive ion 11.29% 19 22.43 Isoliquiritigenin 256.0725 Negative ion, Positive ion 6.17% 20 22.47 Liquiritigenin 256.07261 Positive ion 6.81% 21 22.57 Isoliquiritin 418.12606 Negative ion, Positive ion 3.16% 22 22.61 Ononin 476.13246 Negative ion, Positive ion 1.64% 23 23.87 Hydroxygenkwanin 300.06319 Negative ion 0.44% 24 24.43 Naringenin 272.06766 Negative ion, Positive ion 1.10% 25 26.23 Ginsenoside Re 946.55135 Negative ion 2.82% 26 26.36 Ginsenoside Rg1 846.49815 Negative ion 5.44% 27 26.39 Ganoderol B 440.36455 Positive ion 4.12% 28 27.36 Formononetin 268.0727 Negative ion, Positive ion 3.48% 29 29.32 Ginsenoside Rf 800.49498 Negative ion 3.60% 30 29.91 Glabrene 321.11247 Negative ion 0.52% 31 29.97 Atractylenolide Ⅲ 248.14078 Negative ion 0.55% 32 30.02 Dehydrocostus lactone 230.12949 Positive ion 6.52% 33 30.02 Parthenolide 248.14098 Positive ion 3.28% 34 30.07 Ginsenoside Rg2 784.50122 Negative ion 1.13% 35 30.2 20(R)-Ginsenoside Rh1 684.44603 Negative ion 0.83% 36 30.84 Germacrone 218.13143 Positive ion 0.69% 37 31.26 Isoalantolactone 232.14524 Positive ion 6.35% 38 33.65 Ginsenoside Rd 992.55758 Negative ion 1.21% 39 33.9 Heterophyllin B 778.41608 Negative ion 0.71% 40 36.81 Diammonium glycyrrhizinate 822.40564 Negative ion 6.16% [110]Open in a new tab 3.2. Establishment and assessment of the spleen deficiency syndrome model in nude mice The gastric cancer model with spleen deficiency syndrome was established as described and assessed. On days 3, 6, and 9 of the experiment, the body weight, food intake, and water consumption of nude mice in the model and SJZD groups decreased significantly than the control group (P < 0.01; [111]Fig. 2A–C). Furthermore, compared to the control group, the model construction assignment scores of nude mice in both the model and SJZD group were above 5, with significant differences (P < 0.01; [112]Fig. 2D), indicating the successful construction of the spleen deficiency syndrome model. Fig. 2. [113]Fig. 2 [114]Open in a new tab Assessment of the spleen deficiency syndrome model in nude mice. (A–C) The mice's body weight, food intake, and water intake were monitored on days 3, 6, and 9 of the experiments. (D) Scoring the symptoms of spleen deficiency in nude mice.**P < 0.01, compared to the control group. 3.3. SJZD suppresses tumor growth in gastric cancer mice with spleen deficiency syndrome The orthotopic xenograft tumor model was established in mice with spleen deficiency syndrome, and SJZD treatment was applied ([115]Fig. 3A). The body weight of mice in the model and SJZD groups was significantly lower than the control group (P < 0.01); however, SJZD treatment partially increased the body weight of model mice compared with the model group ([116]Fig. 3B). Evaluation of tumor size and weight revealed similar evidence that, compared to the model group, SJZD therapy significantly reduced the size and weight of tumors in model mice (P < 0.01; [117]Fig. 3C). These data indicate that SJZD suppresses tumor growth in gastric cancer mice with spleen deficiency syndrome. Fig. 3. [118]Fig. 3 [119]Open in a new tab Sijunzi decoction (SJZD) suppresses tumor growth in gastric cancer mice with spleen deficiency syndrome. (A) A schematic diagram showing the grouping, modeling, and treatment methods of model mice. (B) The body weight of mice in three groups was monitored on days 10, 12, 15, 18, 21, and 24 of the experiments. (C) Tumor size and weight of mice in the model and SJZD groups. n = 24. **P < 0.01 compared with the control group; ^Δ^ΔP<0.01 compared with the model group. 3.4. Plasma exosomal miRNA expression in mice from different groups Exosomal miRNAs from plasma and saliva were sequenced in order to find miRNAs with differential expression among different groups. Differentially expressed murine exosomal miRNAs in mice plasma from the control, model, and SJZD groups were displayed in hierarchical clustering heatmap and volcano plots ([120]Fig. 4A–C). Among 507 detected murine miRNAs in the control and model groups, 87 murine miRNAs were significantly differentially expressed between the control and model groups ([121]Fig. 4B–[122]Table S3). Among 406 detected murine miRNAs in the model and SJZD groups, 14 murine miRNAs showing significant differential expression between the model group and the SJZD group were identified ([123]Fig. 4C–[124]Table S4). After sequence comparison, differentially expressed human exosomal miRNAs in mice plasma from the control, model, and SJZD groups were displayed in hierarchical clustering heatmap and volcano plots ([125]Fig. 4D–F). Among 326 detected human miRNAs in the control and model groups, 71 human miRNAs showing significant differential expression between the control and model groups were identified ([126]Fig. 4E–[127]Table S5). Among 293 detected human miRNAs in the model and SJZD groups, 11 human miRNAs showing significant differential expression between the model group and the SJZD group were identified ([128]Fig. 4F–[129]Table S6). Fig. 4. [130]Fig. 4 [131]Open in a new tab Plasma exosomal miRNA expression in mice from different groups. (A–C) Hierarchical clustering heatmap and volcano plots for differentially expressed murine exosomal miRNAs in mice plasma from the control, model, and SJZD groups. (D–F) Hierarchical clustering heatmap and volcano plots for differentially expressed human exosomal miRNAs in mice plasma from the control, model, and SJZD groups. Regarding the murine-derived genomic sequences, miR-128-3p (Log2(FC) = −4.346), miR-328-3p (Log2(FC) = −2.595), and miR-339-5p (Log2(FC) = −2.274) were downregulated within the model group in comparison with the control group, and their expression upregulated after SJZD treatment. miR-184-3p (Log2(FC) = 1.929), miR-1a-3p (Log2(FC) = 1.507), miR-202-5p (Log2(FC) = 2.563), miR-3470a (Log2(FC) = 3.001), miR-3470b (Log2(FC) = 3.609), miR-34c-5p (Log2(FC) = 3.015), and miR-466i-5p (Log2(FC) = 4.990) were increased within the model group in comparison with the control group, and their expression decreased after SJZD treatment. Regarding human-derived genomic sequences, miR-128-3p (Log2(FC) = −4.418), miR-328-3p (Log2(FC) = −2.637), and miR-339-5p (Log2(FC) = −2.338) were downregulated within the model group in comparison with the control group, and their expression upregulated after SJZD treatment. miR-1-3p (Log2(FC) = 1.450), miR-151a-3p (Log2(FC) = 3.267), miR-184 (Log2(FC) = 1.880), miR-202-5p (Log2(FC) = 2.511), and miR-34c-5p (Log2(FC) = 2.960) were increased within the model group in comparison with the control group, and their expression decreased after SJZD treatment. These data suggest that SJZD might alleviate gastric cancer mice with spleen deficiency syndrome in a manner related to miR-328-3p, miR-339-5p, miR-128-3p, miR-1-3p, miR-151a-3p, miR-184, miR-202-5p, miR-34c-5p, miR-184-3p, miR-1a-3p, miR-3470a, miR-3470b, miR-466i-5p, and their target gene functions. 3.5. Functional and signaling pathway enrichment annotation of plasma exosomal miRNA targets For further analysis of the functions and signaling pathways associated with these plasma exosomal miRNAs, TargetScan was used to predict the potential target genes of differentially expressed exosomal miRNAs in mouse plasma, and GO and KEGG enrichment annotation was then performed. Regarding biological process (BP) analysis, the main enrichment items were transcriptional positive regulation, DNA templating, transcribed DNA template, actin cytoskeleton organization, transcription from RNA polymerase II promoter, negative modulation of transcription initiation from RNA polymerase II promoter, DNA templating, phosphatidylinositol-mediated signaling, The main enrichment items were cellular junctional assembly, morphological regulation, and regulation of epithelial cell differentiation ([132]Fig. 5A). Regarding cellular component (CC) analysis, it is mainly enriched in cell membrane, cytoplasm, cytoplasm, nucleus and other cytosolic structures, neuronal projections, Golgi membrane, synapses, presynaptic membrane, cell junctions, laminae, folds, early endosomes, endocytic vesicles, synaptic vesicles, axons and other neuronal sites, mainly for cell-matrix adhesion connections ([133]Fig. 5A). Regarding molecular function (MF) analysis, they are mainly enriched in transcription factor activity, sequence-specific DNA binding, transcriptional coactivator activity, transcriptional activation activity, sequence-specific binding in RNA polymerase II core promoter proximal region, DNA binding in the transcriptional regulatory region, transmembrane receptor protein serine/threonine kinase activity, zinc ion binding, SMAD binding, and transcription factor binding ([134]Fig. 5A). Fig. 5. [135]Fig. 5 [136]Open in a new tab Possible target genes of differentially expressed exosomal miRNAs in mice plasma were predicted by TargetScan and applied for GO functional enrichment analysis (A) and KEGG signaling pathway enrichment annotation (B). KEGG pathway enrichment analysis demonstrated that it was primarily enriched in 23 signaling pathways: cholinergic synapse, melanogenesis, phosphatidylinositol signaling system, insulin secretion, MAPK signaling pathway, PI3K-Akt signaling pathway, estrogen signaling pathway, Adherens junction, colorectal cancer thyroid hormone signaling, TGF -β signaling, RAS signaling, cancer choline metabolism, Rap1 signaling, and other related pathways, prostate cancer, pancreatic cancer and other tumor signaling pathways. The aforementioned findings show differential miRNA gene expression levels in apoptosis, growth, differentiation, and cancer-related pathways. ([137]Fig. 5B). 3.6. Saliva exosomal miRNA expression in mice from different groups Next, hierarchical clustering heatmaps and volcano plots were used to reveal the differential expression of murine exosomal miRNAs in the saliva of mice from the control group, model group, and SJZD group ([138]Fig. 6A–C). Among 274 detected murine miRNAs in the control and model groups, 7 murine miRNAs showing significant differential expression between the control and model groups were identified ([139]Fig. 5B–[140]Table S7). Five murine miRNAs exhibited substantial differential expression between the model and SJZD groups among the 300 detected murine miRNAs ([141]Fig. 6C–[142]Table S8). After sequence comparison, differentially expressed human exosomal miRNAs in mice saliva from the control, model, and SJZD groups were displayed in hierarchical clustering heatmap and volcano plots ([143]Fig. 6D–F). Among 248 detected human miRNAs in the control and model groups, 5 human miRNAs showing significant differential expression between the control and model groups were identified ([144]Fig. 6E–[145]Table S9). Among 278 detected human miRNAs in the model and SJZD groups, 7 human miRNAs showing significant differential expression between the model group and the SJZD group were identified ([146]Fig. 6F–[147]Table S10). Fig. 6. [148]Fig. 6 [149]Open in a new tab Saliva exosomal miRNA expression in mice from different groups. (A–C) Hierarchical clustering heatmap and volcano plots for differentially expressed murine exosomal miRNAs in mice saliva from the control, model, and SJZD groups. (D–F) Hierarchical clustering heatmap and volcano plots for differentially expressed human exosomal miRNAs in mice saliva from the control, model, and SJZD groups. Compared with the murine-derived genomic sequences, miR-184-3p (Log2(FC) = −2.377) was lower expressed within the model group in comparison with the control group, and its expression increased after SJZD treatment. Regarding human-derived genomic sequences, miR-184 (Log2(FC) = −2.129) was lower expressed within the model group in comparison with the control group, and its expression increased after SJZD treatment. miR-151a-3pexpression (Log2(FC) = 2.299) was upregulated within the model group in comparison with the control group, and its expression was downregulated after SJZD treatment. These results indicated that SJZD might exert the effect of benefiting Qi and strengthening the spleen by acting on miR-151a-3p, miR-184, and miR-184-3p. 3.7. Functional and signaling pathway enrichment annotation of saliva exosomal miRNA targets Regarding BP analysis, the main enrichment items are RNA polymerase II promoter transcriptional positive and negative regulation, post-transcriptional RNA gene silencing, miRNA-mediated translation inhibition, glycolytic process modulation, epithelial cell migratory and apoptotic modulation, cellular response to insulin stimulation, etc. ([150]Fig. 7A). Regarding CC analysis, the targets are mainly enriched in the nucleus, cytoplasm, cytoplasmic mRNA processing body, and extra-nucleoplasmic region ([151]Fig. 7A). Regarding MF analysis, it is mainly enriched in transcription factor binding, zinc ion binding, ATP binding, transcriptional activation activity, sequence-specific binding of RNA polymerase II core promoter proximal region, sequence-specific binding of RNA polymerase II transcriptional regulatory region, insulin binding, etc. binding, insulin binding, etc. ([152]Fig. 7A). Fig. 7. [153]Fig. 7 [154]Open in a new tab Possible target genes of differentially expressed exosomal miRNAs in mice saliva were predicted by TargetScan and applied for GO functional enrichment analysis (A) and KEGG signaling pathway enrichment annotation (B). KEGG pathway enrichment analysis demonstrated that it was primarily enriched in 12 signaling pathways: central carbon metabolism, insulin signaling, HIF-1 signaling, neurotrophin signaling, renal cell carcinoma, Rap1 signaling, Foxo signaling, cancer choline metabolism, and cGMP-PKG in cancer ([155]Fig. 7B). These analyses suggest the target genes of miRNAs are enriched in pathways associated with the regulation of apoptosis, growth, and differentiation, as well as tumor-related pathways. 3.8. Selection of promising prognostic miRNA markers for gastric cancer with spleen deficiency syndrome For identifying potential prognostic miRNA markers for gastric cancer with spleen deficiency syndrome, overlapping miRNAs in saliva and plasma samples between the model and SJZD groups were analyzed, and hsa-miR-151a-3p and hsa-miR-184 were found ([156]Fig. 8A). According to the sequencing results, the hsa-miR-151a-3p level decreased both in plasma and salivary exosomes after SZJD therapy. While the hsa-miR-184 level exhibited an inconsistent trend in plasma and salivary exosomes (decreased in plasma exosomes and increased in salivary exosomes). Therefore, miR-151a-3p was selected for further validation in SJZD therapy groups with varying doses. The expression levels of hsa-miR-151a-3p were examined in mice saliva and plasma samples from the model, low-dose, medium-dose, and high-dose SJZD groups using qRT-PCR. [157]Fig. 8B shows that miR-151a-3p was downregulated in both plasma and salivary exosomes in a dose-dependent manner. Next, according to TCGA-STAD and gastric cancer-related miRNA expression microarray [158]GSE93415, miR-151a-3p was upregulated in cancer tissues than control group samples ([159]Fig. 8C–E). The target genes of miR-151a-3p were subsequently predicted using MirWalk 3.0 and TargetScan, and the intersection was taken ([160]Fig. 8F). Extrinsic apoptotic signaling pathways were identified by applying functional and signaling pathway enrichment annotation in Metascope to the intersecting target genes of miR-151a-3p.([161]Fig. 8G). Therefore, exosomal miR-151a-3p might be an underlying biomarker in gastric cancer. Fig. 8. [162]Fig. 8 [163]Open in a new tab Selection of promising prognostic miRNA markers for gastric cancer with spleen deficiency syndrome. (A) Overlapping miRNAs in saliva and plasma samples between the model and SJZD groups were analyzed, and hsa-miR-151a-3p and hsa-miR-184 were found. (B) The expression levels of hsa-miR-151a-3p were examined in mice saliva and plasma samples from the model, low-dose (5.48 g/kg), medium-dose (10.92 g/kg), and high-dose (21.84 g/kg) of SJZD groups using qRT-PCR. The expression levels of miR-151a-3p in control and stomach adenocarcinoma (STAD) samples according to TCGA-STAD (C–D) and [164]GSE93415 (E). (F) Norwalk 3.0 and TargetScan were used to predict the target genes of miR-151a-3p, and the intersection was taken. (G) The intersected target genes of miR-151a-3p were applied for functional and signaling pathway enrichment annotation in Metascope. *P < 0.05, **P < 0.01 compared with the model group or control. 4. Discussion In this study, a gastric cancer model with spleen deficiency syndrome was established in nude mice, and the curative effects of SJZD were investigated. Differentially expressed miRNAs in plasma and saliva exosomes were sequenced and analyzed. Potential target genes of these miRNAs were predicted and applied for GO and KEGG signaling pathway enrichment annotation. Overlapping miRNAs in saliva and plasma samples were analyzed, and miR-151a-3p was selected; intersected miR-151a-3p target genes were predicted and enriched in the extrinsic apoptotic signaling pathways. The common clinical TCM syndromes of gastric cancer include Qi deficiency and blood stagnation, liver Qi affecting the stomach, phlegm and dampness blocking the stomach, stomach heat injuring Yin, spleen and stomach deficiency, stomach Yin deficiency, and Qi and blood deficiency [[165]56]. Jiang et al. [[166]57] investigated 500 patients with gastric cancer and found that spleen deficiency syndrome accounted for the largest proportion of the preoperative single-factor syndromes. Zhao et al. [[167]58] divided 102 gastric cancer patients into two groups according to TCM identification criteria: gastric cancer without and with spleen deficiency syndrome, and showed that gastric cancer patients with the syndrome of spleen deficiency had a poorer clinical prognosis compared with those without spleen deficiency. In this study, in the Sennae Folium decoction-induced spleen deficiency syndrome model in nude mice, weight loss, reduced food and water intake, and tiredness and curling up were observed in the nude mice, and the symptoms were similar to those recorded in medical practice. SJZD, which originated from the third volume of the Song Dynasty's Taiping Huimin Hejiju Fang, is a typical representative formula of the method of benefiting Qi and strengthening the spleen. It was found to act mainly by regulating gastrointestinal functions and enhancing the body's immunity, with anti-tumor and metabolic effects [[168]59,[169]60]. Tian et al. [[170]61] found that SJZD could restore the CaM/CaMK II signaling pathway to basic normal levels in mice with spleen deficiency syndrome. Some studies have also reported the restorative effects of polysaccharides and nonpolysaccharides from SJZD on intestinal microorganisms in a model of spleen deficiency [[171]50]. In addition, SJZD can also promote the repair of gastrointestinal mucosal damage [[172]62]. In this study, SJZD treatment effectively suppressed tumor growth in mice with spleen deficiency syndrome and gastric cancer, giving experimental support for the use of SJZD in the treatment of spleen deficiency syndrome and gastric cancer. Among compounds identified in this study, several have been reported to serve as anti-tumor active ingredients against GC. For example, vicenin-2 inhibits Helicobacterium pylori infection associated gastric carcinogensis in vitro via modulating Akt and Nrf2 signaling [[173]63]. Liquiritin [[174]64] and isoliquiritin effectively suppress proliferation and promote ROS accumulation and apoptosis of GC cell lines [[175]65]. Ginsenoside Re, Ginsenoside Rd and Ginsenoside Rh1 have been reported to exert anti-tumor effects in GC by inhibiting GC metastasis [[176]66] and induce GC cell apoptosis [[177]67,[178]68]. SJZD might exert its functions through these active ingredients, which needs to be investigated in future studies. Numerous studies have demonstrated that exosomal microRNAs can influence a variety of cancer-related characteristics, including cancer's onset, progression, and metastasis, garnering significant attention for exosomal microRNAs [[179]69,[180]70], and additional evidence has indicated that exosomal miRNAs are excellent candidates as biomarkers in the era of precision medicine [[181]71,[182]72]. It was shown that the differential expression of serum exosomes miR-19b-3p and miR-106a-5p in gastric cancer patients correlated with lymph node metastasis and the TNM stage of gastric cancer [[183]73]. Serum exosomal miR-423-5p showed to be more differentially expressed than CEA and CA-199 and could be used to diagnose gastric cancer [[184]74]. In addition to plasma, as a component of human bodily fluids, saliva is affected by pathophysiological changes in the body due to variations in its constituent content [[185]75]. In this experiment, plasma and saliva exosomal miRNAs from mice with gastric cancer and spleen deficiency syndrome were detected using high-throughput sequencing techniques, and differential exosomal miRNAs were screened by comparing gene libraries derived from humans and mice. Interaction of saliva and plasma human-derived differential miRNAs revealed that 2 human-specific differential miRNAs in plasma samples, miR-151a-3p, and miR-184, were also present in salivary exosomes. The presence of human-derived miRNAs in the saliva of nude mice, which could only originate from transplanted human cancer cell-derived tumors, suggests that changes in miRNAs in saliva can directly reflect molecular changes in gastric tumor cells, providing direct evidence for non-invasive early diagnosis of gastric cancer with spleen deficiency syndrome using saliva diagnosis. In addition, differential salivary miRNA expression analysis revealed that miR-184 expression was lower in the model group compared to the control group when human genomic sequences were compared, and increased following SJZD treatment. miR-151a-3p was highly expressed within the model group than the control group and showed to be downregulated following SJZD treatment. miR-151a-3p showed consistent changes in expression in saliva and plasma samples, while miR-184 was inconsistently expressed in saliva and plasma samples. hsa-miR-151a-3p has been previously reported to be upregulated in the serum exosomes from gastric cancer patients [[186]76]. Small extracellular vesicles from gastric cancer that are high in miR-151a-3p promote liver metastasis by generating a niche that increases hepatic stemness [[187]77]. Furthermore, other studies also indicated the upregulation of circulating miR-151a-3p in esophageal adenocarcinoma [[188]78] and cervical cancer [[189]79] patients. Moreover, the intersected target genes of miR-151a-3p were applied for functional and signaling pathway enrichment annotation, and extrinsic apoptotic signaling pathways were found. Although the specific functions of miR-151-3p on SJZD for treating gastric cancer in patients with spleen deficiency syndrome remain unknown, targeting tumor genes with miR-151a-3p will be an important area of future research. Regarding the potential mechanisms, miR-151-5p has been reported to target p53 mRNA and down-regulate p53 levels in SC-M1 gastric cancer cells. The Notch1/miR-151-5p axis contributed to the progression of SC-M1 cells through down-regulation of p53 which in turn repressed FAK promoter activity [[190]80]. Another study found that miR-151-3p induces M2-phenotype polarization in macrophages and increases gastric cancer tumor growth [[191]44]. Therefore, underlying mechanisms might be multifaceted. Notably, several active components of SJZD have been reported to alter miRNA expression in GC and other diseases. Isoliquiritin ameliorates depression in mouse models via upregulating miR-27a levels [[192]81]. Ginsenoside Rg1 administration increased miR-152 to reduce hepatic stellate cell activation [[193]82]. Formonoetin plays an anti-GC function by regulating miR-542-5p [[194]83]. Integrative analyses involving these compounds are necessary. 5. Conclusions In a mouse model with spleen deficiency syndrome, SJZD significantly alleviates gastric cancer. Furthermore, exosomal miRNAs in plasma and saliva, especially miR-151-3p, might be modulated by SJZD. The presence of exosomal miR-151-3p in saliva has the potential to serve as a noninvasive diagnostic and prognostic biomarker for gastric cancer. 6. Limitation There were some limitations in the present study. Firstly, a positive control for gastric cancer therapy, such as 5-FU and cisplatin, was not included in the treatment. A positive control could have offered a clearer baseline for evaluating the effects of SJZD interventions. Secondly, we have identified 207 compounds of SJZD extracts. However, the bioavailability and functions of compounds from SJZD extracts require further investigation. Ethics statement This study was approved by the Experimental Animal Ethics Committee of Hunan University of Chinese Medicine (approval number: LL2020061501). Funding This study was supported by grants from the National Natural Science Foundation of China (No. 82374323), the Natural Science Foundation of Hunan Province (No. 2022JJ30435, 2019JJ40221), and the Key Research and Development Program of Hunan Province (No. 2022SK2124, 2022SK2163). Data availability statement The authors confirm that the data supporting the findings of this study are available within the article and supplementary materials. CRediT authorship contribution statement Ping Yang: Validation, Investigation, Funding acquisition, Data curation. Huijun Lei: Investigation. Yue Fu: Investigation. Cheng Chen: Investigation. Li Tang: Investigation. Shuaishuai Xia: Writing – original draft, Investigation. Yan Guo: Validation, Software. Guangyu Chen: Investigation. Mengzhou Xie: Validation, Software. Jingjing Yang: Writing – review & editing, Validation. Feng Li: Writing – review & editing, Validation, Project administration, Conceptualization. Liang Li: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments