Catgut implantation at acupoints improves anti-PD-1 inhibitor efficacy in lung cancer by inducing immune responses and remodeling the tumor microenvironment.

Introduction
Lung cancer ranks as the third most commonly diagnosed malignancy and represents the second leading cause of cancer-related mortality worldwide [1]. The advent of immune checkpoint inhibitors (ICIs) targeting the programmed death-l/ programmed death ligand-1 (PD-1/PD-L1) has markedly revolutionized the therapeutic paradigm for lung cancer [2–4]. However, the efficacy of single-agent ICIs remains suboptimal in a substantial proportion of patients, primarily due to challenges such as inadequate immune cell infiltration and the immunosuppressive tumor microenvironment (TME), which constitute significant barriers to successful immunotherapy in lung cancer [5–7].
The immune system plays a pivotal role in lung cancer pathogenesis by detecting and eliminating malignant cells. Nevertheless, the phenomenon of “cold tumors”, characterized by a lack of tumor-infiltrating lymphocytes (TILs) and an immune-desert phenotype, fosters cancer progression under conditions of chronic inflammation or immunosuppression [8–10]. This immunologically inert state severely impedes anti-tumor immune responses. Key factors within the TME, including immune cell inhibition, the presence of immunosuppressive cell populations, and chronic inflammation, collectively promote tumor cell proliferation, immune evasion, metastasis, and resistance to immunotherapy [11, 12]. Although combinatorial approaches incorporating chemotherapy, radiotherapy, Vascular Endothelial Growth Factor/Vascular Endothelial Growth Factor receptor (VEGF/VEGFR) inhibitors, and oncolytic viruses have been explored to overcome the low response rates, these strategies have not yielded substantial clinical benefits[13, 14]. Thus, there is an urgent need for research into immunomodulatory strategies to enhance the efficacy of ICIs and bolster anti-tumor immunity in lung cancer.
Acupuncture stimulation, as a complementary and alternative therapeutic modality, offers notable advantages including safety, cost-effectiveness, and minimal adverse effects [15, 16]. In oncology, acupuncture stimulation is widely employed in clinical practice to mitigate cancer-related symptoms and treatment-associated side effects [17–19]. Recent investigations into acupuncture stimulation, particularly electroacupuncture, suggest that it exerts direct anti-tumor effects and potent immunomodulatory activity through the neural–endocrine–immune network [17, 20]. Emerging preclinical evidence indicates that acupuncture stimulation reduces tumor burden and restores immune function in patients with malignancies by augmenting lymphocyte levels and modulating inflammatory cytokines [17, 19–22]. These findings suggest that acupuncture stimulation may enhance the response to immunotherapy in lung cancer. Catgut embedding at acupoints (CIAA) involves the implantation of absorbable sutures at specific acupoints, which provides sustained and prolonged stimulation as the sutures are gradually absorbed. In this study, we hypothesized that CIAA could remodel the TME, enhance anti-tumor immunity, and potentially synergize with anti-PD-1 therapy in lung cancer. The present study aimed to preliminarily explore this hypothesis and provide correlative experimental data through multi-omics and multi-parametric immune profiling.
Our experimental findings demonstrate that CIAA was associated with reduced tumor growth and correlated with enhanced efficacy of anti-PD-1 therapy in murine models of lung cancer. Mechanistically, our exploratory analysis suggests that the combination of CIAA and anti-PD-1 therapy correlated with remodeling of the TME and marked alterations in anti-tumor immune responses. These results provide preliminary, hypothesis-generating evidence for the potential of CIAA as an adjunct to immunotherapy in lung cancer treatment.

Results

CIAA enhances tumor growth inhibition by anti-PD-1 inhibitor in the lung cancer mice model
Previous studies demonstrated that acupoint stimulation improves immune function in cancer patients [17]. We investigated the role of CIAA in lung cancer mice treated with anti-PD-1 inhibitor. After 2 weeks of intervention (Fig. 1A), the tumor volume and weight in the C, P, and C-P groups were significantly reduced compared to those in the M group (C vs. M P = 0.0009; P vs. M P < 0.0001; C-P vs. M P < 0.0001, Fig. 1B and C). Meanwhile, the tumor growth rate in each treatment group was also significantly lower than that in the M group (Fig. 1D). The tumor growth inhibition effect in the combined treatment group was more pronounced than in the single treatment groups. These data collectively suggest an association between CIAA and tumor growth inhibition in mouse models, particularly when combined with anti-PD-1 inhibitor.
A: Summary of experimental design to study the therapeutic effect of combination treatment in LCC tumor-bearing mice.
B-C: The image and tumor weight of LCC tumors in C57BL/6 mice after 14 days of treatment (n = 6).
D: Tumor growth curves with mean tumor volumes ± SEM (n = 8).

Effect of CIAA on peripheral blood immune cells in lung cancer mice
The immune function is an important factor affecting tumor growth. We assessed the effects of treatments on immune function by measuring the percentage of immune cells in the peripheral blood of mice. NK, NKT, and B cells are key components of innate immunity, and an increase in their proportions reflects the restoration of innate immune function [23, 24].
Compared with the model group, the percentage of NK cells was significantly higher in the C-P group (P = 0.037) compared to the P group (P > 0.05, Fig. 2H); NKT cells were significantly higher in the C-P group (P = 0.034, Fig. 2J); and there was no statistically significant difference in the percentage of B cells among the groups (P > 0.05, Fig. 2F).
Adaptive immune function is a critical component of immunity and plays an anti-tumor role in lung cancer, primarily mediated by T cells [25]. The percentage of total T cells was the lowest in the model group, while the percentages in the S-C group, the C group, and the C-P group were all elevated, with the C group showing the highest percentage among all groups (P = 0.001, Fig. 2B). Compared with the model group, the percentage of CD4 + T cells in the S-C group (P = 0.006), the C group (P = 0.003), and the C-P group was significantly lower ((P = 0.034, Fig. 2C). Similarly, the percentage of CD8 + T cells in the S-C group (P = 0.003) and the C group (P = 0.003) was decreased compared to that in the M group (Fig. 2D).
Regulatory T cells (Tregs) are immunosuppressive cells that promote tumor growth, metastasis, and immune evasion by suppressing anti-tumor immune responses [26]. The percentage of Treg cells in the P group was lower than that in the M group (P > 0.05), whereas the percentages in the C group (P = 0.038) and the C-P group were significantly lower (P = 0.043, Fig. 2E). These data indicate that CIAA combined with anti-PD-1 inhibitor is associated with alterations in adaptive and innate anti-tumor immune parameters in peripheral blood.
A. Representative flow cytometry plots.
B-H. Quantification of the percentages of total T cells, CD8 + T cells, CD4 + T cells, Tregs, B cells, NK cells, and NKT cells (n = 5).
I. Cluster heatmap of immune cell subsets. The following symbols were used in figure legends to denote p values: ns, non-significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001.

CIAA is associated with remodeling of the TME

Immune cell infiltration in the TME
The TME is the internal environment where tumor cells arise and survive. It consists of tumor cells, immune cells, and cytokines secreted by various cell types. Based on the extent of immune cell infiltration, tumors can be classified into three immunophenotypes: immune-inflamed, immune desert, and immune rejection [27]. Studies on cold and hot tumors have demonstrated that T cell infiltration within the TME is closely associated with patient prognosis [28]. In cold tumors, such as immune-desert and immune rejection types, T cells struggle to infiltrate the tumor and complete the immune response, leading to persistent tumor growth and metastasis [29]. Since cold tumors exhibit poor responses to anti-PD-1 therapy, converting them into hot tumors represents a key strategy to enhance the efficacy of anti-PD-1 inhibitors.
We analyzed the infiltration of multiple immune cells in mouse tumor tissues using flow cytometry, including total T cells, CD4 + T cells, CD8 + T cells, NK cells, NKT cells, B cells, and Tregs. In the TME, T cells function to kill and destroy tumor cells. However, T cells become dysfunctional due to nutrient limitations and the accumulation of harmful metabolites in the TME [30, 31]. CD4 + T cells and CD8 + T cells can effectively combat tumor cells and inhibit tumor growth [32].
As shown in Fig. 3A, the infiltration of total T cells was the lowest in the M group, while it was significantly higher in the C group (P = 0.005), P group (P < 0.0001), and C-P group (P = 0.005) compared to the M group. CD8 + T cell infiltration was the lowest in the M group, with higher levels observed in the P group and C-P group, though the difference was not significant (P > 0.05) (Fig. 3B). Immunohistochemical analysis demonstrated that the C-P group exhibited a significantly higher infiltration of CD8 + T cells compared to the other groups (P < 0.0001, Fig. 3M). Concurrently, macrophage infiltration in the C-P group was also significantly elevated relative to the other groups (P < 0.0001). Furthermore, the CIAA group alone showed a higher level of infiltration than the model group (P = 0.006, Fig. 3N). The percentage of CD4 + T cells in tumor tissues was significantly higher in the C-P group compared to the M group (P = 0.043, Fig. 3C).
Tregs are immunosuppressive cells that promote tumor growth and metastasis within the TME [33, 34]. The percentage of Tregs in tumor tissues was highest in the M group, while it was significantly lower in the C group (P = 0.024), P group (P = 0.007), and C-P group (P = 0.002), with the lowest percentage observed in the combined treatment group (Fig. 3D).
NK cells can eliminate tumor cells by releasing cytotoxic particles, such as perforin and granzyme, to directly lyse and kill target cells [35]. The percentage of NK cells in tumor tissues was higher in the C group compared to the M group, though the difference was not significant (P = 0.357, Fig. 3F). Compared to the P group, the percentage of NK cells was significantly higher in the C group (P = 0.041) and elevated in the C-P group, but the difference was not significant (P = 0.785).
As antigen-presenting cells, B cells activate T cells and macrophages, produce cytokines and antibodies, and enhance cytotoxicity and phagocytosis to promote tumor immunity [36, 37]. Compared to the M group, the percentage of B cells in tumor tissues was significantly increased in the C group (P = 0.030) and C-P group (P = 0.049, Fig. 3E).
In the M group, the number of immune cells capable of killing tumor cells through phagocytosis was reduced, while the number of Tregs was increased. Following anti-treatment, the infiltration of T cells and CD8 + T cells in the TME increased. CIAA enhanced the infiltration and upregulation of innate immune cells. After combination therapy, the infiltration of total T cells, CD4 + T cells, CD8 + T cells, B cells, and NK cells in tumor tissues increased, while the infiltration of Tregs decreased (Fig. 3G). These data collectively suggest that CIAA and its combination with anti-PD-1 are associated with significant remodeling of the immune cell composition within the tumor microenvironment.

Cytokine release of TME
As a pro-inflammatory cytokine, interleukin-6 (IL-6) is closely associated with the growth, progression, and survival of tumor cells in the TME [38]. Immunohistochemistry revealed the strongest IL-6 positivity in the M group, while weaker positivity was observed in the C group and C-P group compared to other groups. There was no significant difference in IL-6 levels between the M group and the S-C group (P > 0.05), whereas significant differences were observed between the P group (P < 0.001), the C group (P < 0.001), and the C-P group (P < 0.001, Fig. 3H, I). The Elisa assay showed that IL-6 levels were the lowest in the C group, which was significantly different from the M group. Similarly, IL-6 levels in the C-P group were significantly lower compared to the M group (P < 0.05, Fig. 3R).
VEGF stimulates angiogenesis, increases vascular permeability in tumor tissues, promotes tumor cell invasion, and provides nutrients for tumor growth [39]. In this study, VEGF expression was reduced in all intervention groups, indicating suppression of VEGF in tumors. The most significant reduction was observed in the C-P group (P < 0.001, Fig. 3H, J).
Transforming growth factor-ß (TGF-ß), a member of the transforming growth factor superfamily, regulates immune responses [38]. TGF-ß inhibits tumor growth by suppressing tumor cell proliferation and promoting apoptosis. Treatment with CIAA, Sham-CIAA, anti-PD-1 inhibitor, and CIAA combined with anti-PD-1 inhibitor significantly reduced TGF-ß levels, with the most pronounced effect observed in the C group (P < 0.0001, Fig. 3Q).
For Granzyme-B (GZM-B) expression, statistical analysis showed that the level in the C-P group was the highest, with statistically significant differences compared to the M group (P = 0.012), the S-C group (P = 0.041), and the C group (P = 0.029, Fig. 3N). Interferon-alpha (IFN-α) showed trends similar to TGF-ß, with the lowest levels observed in the CIAA group (P < 0.0001, Fig. 3P).
Together, these results indicate inhibitory changes in the TME of the C-P group. These findings collectively demonstrate that CIAA and anti-PD-1 inhibitor combination therapy is associated with complex changes in a spectrum of immunosuppressive and pro-tumorigenic factors within the TME, suggesting a comprehensive remodeling of the tumor microenvironment.
A-F: Quantification of the percentages of total T cells, CD8 + T cells, CD4 + T cells, Tregs, B cells, NK cells, and NKT cells in tumor tissues (n = 5).
G: Graph showing the relative proportions of immune cell subpopulations.
H: Representative immunohistochemistry (IHC) images of VEGF, IL-6, CD8, and F4/80.
I-M: Quantitative analysis of VEGF, IL-6, CD8, and F4/80 staining intensity in IHC images.
N-R: Levels of GZM-B, INF-γ, IFN-α, TGF-ß, and IL-6 analyzed by enzyme-linked immunosorbent assay (Elisa) in tumor tissues.

Effects of CIAA combined with anti-PD-1 inhibitor on the mechanisms in lung cancer mouse models

CIAA altered the expression of differential genes in lung cancer mouse models
To elucidate the molecular mechanisms underlying the effects of CIAA and anti-PD-1 inhibitor on tumor growth, transcriptome sequencing was performed on tumor tissues from lung cancer model mice [40]. We investigated the regulatory effects of CIAA combined with anti-PD-1 inhibitor on gene expression in lung cancer mice by conducting analyses of differentially expressed genes. It was found that CIAA, CIAA combined with anti-PD-1 inhibitor and anti-PD-1 inhibitor could alter the expression of differential genes in mice, as shown in the graph of significant differential gene levels between groups (Fig. 4A). Among them, M and C significant differential genes were Dmbt1, Reg3b, Lypd8; M and P significant differential genes were Fga, Fgb, Gc; M and C-P significant differential genes were Car1, Reg3b, Lypd8; and M and S-C significant differential genes were Xirp2, Mrgprb1, Mrgbpc1.

CIAA changed metabolism differentials in lung cancer model mice
For tumor tissue samples from experimental mice, the ring diagram of metabolite category composition (Fig. 4D) revealed that the main metabolites were concentrated in amino acids and their metabolites (16.83%), benzene and substituted derivatives (16.06%), organic acids and their derivatives (13.97%), and heterocyclic compounds (12.9%).The analysis of differential metabolite clustering revealed significant differences in metabolites across treatment groups (Fig. 4I)[41]. Volcano plots (VP) were used to visualize differences in metabolite levels between sample groups and their statistical significance [42]. Metabolites and differential metabolites in each comparison group are displayed below (Fig. 4E-H).
These data indicate that CIAA and anti-PD-1 inhibitor, both individually and in combination, are associated with significant and context-specific transcriptional and metabolic remodeling in the TME [43, 44], providing clues for further mechanistic investigation.
A. Volcano Plot of Differentially Expressed Genes. Red dots denote significantly differentially expressed genes, while black dots represent genes with no significant differential expression. The top three genes with the largest fold changes in each comparison group are labeled.
B. Sample Correlation Heatmap. The x-axis and y-axis represent different samples, and the color of each square indicates the strength of correlation. Purple denotes positive correlations, while green represents negative correlations.
C. KEGG Enrichment Analysis Bubble Plot for Genes (C) and Metabolites (J). The bubble plot visualizes the enrichment of KEGG pathways, with bubble size representing the number of enriched genes or metabolites and color indicating the significance level.
D. Metabolite Category Composition Ring Diagram. Each color represents a metabolite category, and the area of the color block corresponds to the relative percentage of the category. Larger areas indicate higher proportions of the respective metabolite category.
E–H. Differential Metabolite Volcano Plot. Green points denote down-regulated metabolites, red points represent up-regulated metabolites, and gray points indicate metabolites with no significant differential expression.
I. Differential Metabolite Clustering Heatmap. The horizontal axis represents sample information, and the vertical axis represents metabolite information.
K. KEGG Pathway Enrichment Bar Plot. Red and green bars denote pathways enriched in the metabolomics and transcriptomics datasets, respectively.
M. Correlation Clustering Heatmap. Each row represents a gene, and each column represents a metabolite. Red indicates positive correlations between genes and metabolites, while blue denotes negative correlations.
N. Multi-omics KEGG Enrichment Analysis Bubble Plot. The red–yellow–blue color gradient reflects the significance level of enrichment, represented by the p value. Bubble shape corresponds to different omics layers, and bubble size represents the number of differentially expressed metabolites or genes.

CIAA is associated with alterations in signaling pathways in lung cancer model mice
Metabolic and cellular signaling pathways form the basis of therapeutic targets, making it essential to study changes in these pathways [45]. Differential gene enrichment analysis showed that CIAA combined with anti-PD-1 inhibitor altered pathways such as the PI3K-Akt signaling pathway, ECM-receptor interaction, protein digestion and absorption, and focal adhesion. CIAA alone altered pathways including oxidative phosphorylation, mineral absorption, fat digestion and absorption, and bile secretion. The anti-PD-1 inhibitor significantly altered pathways such as Th17 cell differentiation, Th1 and Th2 cell differentiation, intestinal immune network for IgA production, human T cell leukemia virus 1 infection, and others (Fig. 4C).
Clustering analysis of differential metabolites revealed significant differences in metabolic signaling pathways across treatment groups [46]. Differential metabolites between the P and M groups were primarily associated with dopaminergic synapse, melanogenesis, and neutrophil extracellular trap formation. Differential metabolites between the CIAA and M groups were mainly enriched in metabolic pathways, glycine, serine and threonine metabolism, D-amino acid metabolism, central carbon metabolism in cancer, and aminoacyl-tRNA biosynthesis. Differential metabolites between the C-P and M groups were mainly enriched in tryptophan metabolism, beta-alanine metabolism, central carbon metabolism in cancer, D-amino acid metabolism, carbon metabolism, and biosynthesis of cofactors. Among the metabolic pathways shared by the CIAA and M groups and the C-P and M groups, protein digestion and absorption, phenylalanine metabolism, glycine, serine and threonine metabolism, and central carbon metabolism in cancer were significantly enriched (Fig. 4J).
To better understand the effects of CIAA on lung cancer mice, we analyzed the combination of gene and metabolic differentials to further investigate phenotype and biological process regulation mechanisms. First, KEGG pathways enriched by the two omics were visualized using column charts and enrichment bubble maps. We found that multiple transcriptional and metabolic signaling pathways were altered in lung cancer model mice following combination therapy with CIAA and an anti-PD-1 inhibitor. These pathways included the regulation of TRP channels by inflammatory mediators, neuroactive ligand–receptor interaction, the cAMP signaling pathway, and pathways related to neurodegeneration in multiple diseases (Fig. 4K). Additionally, in the enrichment bubble map, two signaling pathways, ABC transporters and protein digestion and absorption, showed significant alterations in transcription and metabolism (Fig. 4N). Furthermore, all correlation calculations for differentially expressed genes and differential metabolites were performed. A heat map of the associated clusters was plotted, and significant clustering was observed, such as for amino acids and their metabolites, alcohols and amines, benzene and substituted derivatives, and organic acids and their derivatives (Fig. 4M). Our research on lung cancer model mice demonstrated that CIAA combined with anti-PD-1 inhibitor was associated with alterations in multiple pathways closely tied to tumor immunity, though these findings require validation in future functional studies.

Discussion
In this study, we demonstrate that combining catgut embedding at the Zusanli acupoint (CIAA) with anti-PD-1 therapy is associated with a profound and coordinated remodeling of the tumor microenvironment (TME) in lung cancer models, characterized by multi-layered transformations across cellular, cytokine, and molecular axes that collectively suggest a conversion from an immunologically “cold” TME toward an active and potentially treatment-susceptible “hot” state [27, 47]. However, we emphasize that all findings in this study are correlational, not proof of causation.
The tumor microenvironment (TME) in the untreated model (M group) displayed a hallmark immune-desert phenotype [48], characterized by scant infiltration of effector cells (total T, CD8⁺ T, CD4⁺ T, NK, and B cells) and a dominant immunosuppressive landscape featuring abundant Tregs [49] and elevated pro-tumorigenic factors (IL-6[50], VEGF[51]). All therapeutic interventions reversed this profile. Notably, the CIAA + anti-PD-1 (C-P) combination therapy elicited the most robust and coordinated immunologic remodeling. This shift was marked by a significant augmentation of key anti-tumor effectors—including CD8⁺ cytotoxic T cells, CD4⁺ helper T cells, macrophages, and B cells—coupled with the most pronounced depletion of Tregs [52]. Moreover, the combination therapy most effectively suppressed critical pro-tumorigenic pathways, notably inhibiting VEGF-driven angiogenesis [51] and reducing IL-6 levels [50], thereby attenuating both pro-inflammatory and immunosuppressive signaling axes.
Integrated multi-omics analyses delineated the molecular mechanisms underlying this phenotypic transformation. Transcriptomic sequencing demonstrated that each therapeutic regimen elicited a distinct transcriptional signature. The combination therapy specifically and significantly modulated key genes and pathways involved in tumor–stroma crosstalk and cell survival, including PI3K-Akt signaling [53], ECM-receptor interaction [54], and focal adhesion [55].
Complementary metabolomic profiling revealed profound, treatment-specific metabolic reprogramming. Notably, the combination therapy markedly altered immunomodulatory metabolic pathways—particularly tryptophan metabolism [56], beta-alanine metabolism [41], and central carbon metabolism in cancer [57]—which are fundamentally linked to immune cell functionality and nutrient sensing within the TME [43]. These omics findings provide a rich, hypothesis-generating dataset for further investigation into the mechanisms of action of CIAA combined with anti-PD-1 therapy.
In conclusion, this preclinical study in a murine lung cancer model demonstrates that CIAA combined with anti-PD-1 inhibitor treatment is associated with significant changes in immune cell infiltration in the tumor microenvironment and correlates with distinct transcriptomic and metabolomic remodeling features. Our integrated analysis suggests that the anti-tumor effects associated with CIAA plus anti-PD-1 therapy may arise from two complementary axes: anti-PD-1 blockade is predominantly associated with reactivation of adaptive T cell immunity, while CIAA may augment this response by recruiting innate immune cells and instigating broad transcriptional–metabolic reprogramming.

Study limitations
This study has several critical limitations. First and foremost, the primary limitation of this study is the lack of further mechanistic investigation. Although our data reveal significant associations between CIAA treatment and specific immunological changes as well as transcriptomic and metabolomic profiles, these findings constitute a working model based on correlational data rather than functionally validated conclusions. Future studies employing in vivo blocking experiments are necessary to further test these hypotheses. Second, we recognize the lack of an appropriate control group as a key design flaw, which limits our ability to definitively attribute the observed therapeutic effects to acupoint-specific stimulation. Notably, the sham-CIAA group also exhibited immunomodulatory effects on certain peripheral blood immune parameters. However, for key tumor microenvironment parameters (e.g., intratumoral CD8 + T cell infiltration and cytokine levels, as shown in Fig. 3), the CIAA group and its combination with anti-PD-1 therapy demonstrated more favorable trends, although more rigorous controls are needed to confirm this. Third, despite the application of rigorous statistical corrections, the sample size in this study is relatively limited, and the multi-omics findings require validation in independent cohorts.
These limitations imply that the results of this study and any discussion of their translational potential remain entirely speculative at present and merely provide preliminary directions for future research.

Materials and methods

Lung cancer mice models
LCC cells (2 × 106) were subcutaneously injected into the right flanks of C57BL/6 mice. Mice were randomly assigned to treatment groups when their tumors reached approximately 5 mm in diameter. The C57BL/6 mice were purchased from Zhuhai BesTest Bio-Tech Company. These mice were maintained under a 14-h light/10-h dark cycle, with housing temperature and humidity maintained at 22–25 °C and 50%, respectively. Animal handling and experimental procedures conformed to institutional guidelines and were approved by the Animal Research Committee of Guangdong Pharmaceutical University (GZR2024030). Male mice aged 6 to 7 weeks were used in all experiments. Tumor volume was calculated using the formula: Volume = π × length × width × height/6.
Mouse models of lung cancer were randomly divided into five groups: the M group (received no treatment, n = 8), the S-C group (received sham-CIAA treatment, n = 8), the C group (received CIAA treatment, n = 8), the P group (received anti-PD-1 treatment, n = 8), and the C-P group (received combined CIAA and anti-PD-1 treatment, n = 8).

Anti-PD-1 treatment
Mice were administered 10 mg/kg anti-PD-1 (Bio X Cell) intraperitoneally three times per week.

CIAA and sham-CIAA Treatment
We selected “Zusanli” (ST36) on both sides for catgut embedding at acupoints. For CIAA, first, we disinfected the acupoints using 75% alcohol. A 2–3 mm catgut thread was inserted into the acupuncture point to a depth of approximately 4 mm using a 7-gauge burying needle. The catgut thread was then pushed into the tissues, and the burying needle was slowly withdrawn. Finally, the needle hole was pressed with a cotton swab. For sham-CIAA, the steps were similar to those of CIAA: the points and insertion depths were identical, but without the catgut thread.

Flow cytometry
Antibodies included FITC Hamster Anti-Mouse CD3e, APC Rat Anti-Mouse CD4, PE Rat Anti-Mouse CD8a, BV421 Rat Anti-Mouse Foxp3, PE-Cy7 Mouse Anti-Mouse NK-1.1 from BD and PerCP/Cvanine5.anti-mouse/human CD4RB220 from BioLegend. Staining was performed with Cytofix/Cytoperm kits (BD) following the manufacturer’s instructions. For staining of transcription factors, the eBioscience Foxp3/Transcription Factor Staining Buffer Set was used following the manufacturer’s instructions. Flow cytometric analyses were performed using BD Canto2 cell analyzers.

Immunohistochemistry (IHC)
Tumor samples were immediately fixed in 10% formalin and c. Representative Sectons (4 µm in thickness) were obtained. IHC was performed with appropriately diluted primary antibodies (IL-6 and VEGFA polyclonal antibody from BioLegend) at 37 °C for 1 h. These antibodies were used at a 1:500 dilution. After 52-min washes in PBS buffer, the slides were incubated with the secondary antibody (IgG, ab6721, Abcam, UK) at 37 °C for 30 min. After another series of PBS washes, the sections were stained with DAB (Dako), counterstained with hematoxylin, dehydrated, and mounted for viewing. The presence of brownish granules in the cell membrane, cytoplasm, or nucleus was defined as positively stained cells. Changes in tumors were observed and assessed under a microscope.

RNA-seq and metabolomics sequence methods and statistical analysis
The cDNA libraries and untargeted metabolomics were sequenced on the Illumina sequencing platform by Metware Biotechnology Co., Ltd. (Wuhan, China). The RNA-seq and metabolomics sequence methods and analytical approaches are provided in Supplementay Appendices 1.

Statistical analysis
Statistical analyses were performed using GraphPad Prism 8 (GraphPad Software), and results were reported as mean ± SD. Categorical variables were analyzed using the χ2 test, whereas continuous variables were analyzed using analysis of variance (ANOVA) or a two-tailed unpaired t-test. The following symbols were used in figure legends to denote p values: ns, non-significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Supplementary Information
Below is the link to the electronic supplementary material.