Experimental method for relation between cytokine level and tumor immune microenvironment in circulation

Abstract

The invention discloses an experimental method for relation between cytokine level and tumor immune microenvironment in circulation, which comprises the following steps of: S1, collecting colorectal cancer patient data and normal data for contrast, and preferably screening a gene expression data set from a Gene Expression Omnibus (GEO) database; according to the method, a plurality of gene expression data sets from the GEO database are collected and analyzed, so that differential expression genes between colorectal cancer and normal tissues can be systematically identified, and a solid foundation is provided for subsequent research. The process not only covers batch correction, normalization and batch effect elimination of data, but also visually presents data distribution changes through principal component analysis, ensures the accuracy and reliability of analysis results, confirms that IL6 and IFN gamma are one of the most critical factors affecting the colorectal cancer immune microenvironment, and participate in formation of immune infiltration.

Description

Technical Field

[0001] The present invention relates to the technical field of tumor immune microenvironment research, and in particular to an experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment. Background Art

[0002] Microsatellite instability (MSI) status and tumor infiltrating lymphocytes (TIL) are recognized prognostic factors for colorectal cancer. The combined analysis of TIL and MSI status has continuously provided direction for clinical treatment in the past. In clinical practice, ICIs-related treatment is recommended for patients with MSI CRC, but 90% of patients with MSS CRC cannot benefit from ICIs treatment. The possible reason is the special tumor immune microenvironment of MSS CRC. In the process of continuous research on the functions of various cells in the tumor immunosuppressive microenvironment, researchers have found that some cytokines are involved in the process of immune infiltration, such as IL6 involved in the immune infiltration process of NK cells, Treg cells and CD8+T cells. However, most of the research is limited to the cytokine levels in cells and their surrounding environment. In the study of solid tumors, few researchers have paid attention to the role of cytokines in serum. Malignant tumors are systemic diseases. Under this concept, the importance of circulating components in tumor occurrence and development is self-evident. Especially during treatment, treatment methods targeting cytokines will directly enter the circulation. In studies of colorectal cancer and pancreatic cancer, researchers have found that the levels of cytokines such as IL6 in serum are closely related to the overall survival of patients. In the detection of serum IL6 in patients with diffuse large B-cell lymphoma, researchers have also found an association between IL6 serum levels and the occurrence of DLBCL. Summary of the invention

[0003] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment.

[0004] In order to achieve the above object, the present invention adopts the following technical solutions:

[0005] An experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment, comprising the following steps:

[0006] S1: Collect colorectal cancer patient data and normal control data, preferably select gene expression data sets from the Gene Expression Omnibus (GEO) database;

[0007] S2: Analyze the differential gene expression in the data, use R software and limma R package to perform batch correction and normalization operations on the data, and identify the differentially expressed genes (DEGs) between control samples and disease samples;

[0008] S3: Perform weighted correlation network analysis (WGCNA) to construct a gene co-expression network and identify colorectal cancer-related modules;

[0009] S4: Intersection operation was performed on the differentially expressed genes (DEGs) and the key genes identified by WGCNA to obtain candidate key genes, and GO / KEGG enrichment analysis was performed;

[0010] S5: construct protein-protein interaction (PPI) network, calculate the degree of nodes, and identify the core genes of PPI network;

[0011] S6: Receiver operating characteristic (ROC) curve was constructed to evaluate the diagnostic value of hub gene;

[0012] S7: Use integrated and standardized gene expression profile data to identify different types of immune cells in tissues and perform immune infiltration analysis;

[0013] S8: Immunohistochemistry (IHC) was used to measure the degree of staining infiltration of specific immune cells in tissue specimens from patients with colorectal cancer.

[0014] Preferably, in step S1, the gene expression datasets screened from the GEO database include GSE35279, GSE87211 and GSE110224.

[0015] Furthermore: In the S2 step, the sva R package is used to eliminate the batch effect, and the ggplot2 R package and the ggpubr R package are used to perform principal component analysis (PCA) to visualize the distribution changes of the data in the dimensionality reduction space.

[0016] Further: In the S3 step, the WGCNA R package was used to construct a gene co-expression network, and the modules were detected by hierarchical clustering and dynamic tree cutting function, and 19 co-expression modules closely related to clinical characteristics were identified and displayed through a clustering tree.

[0017] As a preferred solution of the present invention: in the S4 step, the VennDiagram R package is used to perform an intersection operation on the differentially expressed genes and the key genes identified by WGCNA, and the ClusterProfiler R package is used to perform GO and KEGG enrichment analysis to analyze potential functions and gene pathways.

[0018] As a further solution of the present invention: in the S5 step, the intersection genes after the intersection operation are imported into the STRING online tool to construct a PPI network, and the PPI network is visualized and analyzed using the Cytoscape software. The degree values ​​of the nodes are calculated using the cytoHubba plug-in, and the targets with degree values ​​higher than the median are regarded as the core genes of the PPI network. The genes are sorted according to the degree values, with emphasis on the top 10 key genes in the degree values.

[0019] As a further scheme of the present invention: In the step S6, a ROC curve is constructed to evaluate the diagnostic value of the hub gene, including the area under the curve (AUC). AUC values ​​in the ranges of 0.5-0.7, 0.7-0.9 and >0.9 are respectively judged as low, medium and high accuracy, and a nomogram is used to construct a prediction model between the expression level of the hub gene and the risk of disease. The regression model is calibrated using the rms package, and a calibration curve is drawn to evaluate the predictive accuracy of the model. The hub genes (CXCL10, IL6, PRKACB, IFNG, IL1A) are used to calculate the total score. The value of each of these variables is scored on the point scale axis, and the total score is calculated for estimation.

[0020] Based on the above scheme: In the S7 step, the CIBERSORT algorithm is used to generate an immune cell infiltration matrix to evaluate the infiltration degree of key immune cell types in tumor and normal samples, and heat maps, correlation heat maps and violin plots are drawn for visual presentation.

[0021] On the basis of the above scheme: In the step S8, the tumor tissue was fixed with formalin, embedded in paraffin and cut into slices with a thickness of 3 mm, immunohistochemical staining was performed, and the staining infiltration degree of specific immune cells was calculated by Image J pro.

[0022] On the basis of the above scheme: the CIBERSORT algorithm was used in step S7 to further calculate the differences in the immune microenvironment, and the correlation between the expression of IL-6, IFNG and IL-2 and the infiltration pattern of immune cells was detected compared with the control group.

[0023] The beneficial effects of the present invention are:

[0024] 1. An experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment. By collecting and analyzing multiple gene expression data sets from the GEO database, it can systematically identify differentially expressed genes between colorectal cancer and normal tissues, providing a solid foundation for subsequent research. This process not only covers batch correction, normalization and elimination of batch effects of data, but also intuitively presents data distribution changes through principal component analysis, ensuring the accuracy and reliability of the analysis results, confirming that IL6 and IFNγ are one of the most critical factors affecting the immune microenvironment of colorectal cancer, and are involved in the formation of immune infiltration, and are closely related to the infiltration of M2 macrophages and CD8+T cells; it was found that some inflammatory cells are indeed correlated with some cytokines in serum, and the cytokine levels in serum do affect the infiltration of immune cells in the tumor immune microenvironment, especially IL6 and IFNγ, and these cytokines may be involved in the formation of CRC immune resistance.

[0025] 2. An experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment. Through immunohistochemistry experiments, this method marked and quantitatively analyzed key immune cells, verified the results of bioinformatics analysis, and explored the relationship between cytokine expression levels and patient prognosis and immunotherapy effects, providing a useful reference for clinical diagnosis and treatment. When the IL6 expression level is high, patients have a better prognosis after ICIs treatment.

[0026] 3. An experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment. Through weighted gene co-expression network analysis technology, this method successfully mined co-expression gene modules closely related to colorectal cancer and identified key genes in the core modules. This step is of great significance for revealing the interaction between genes and exploring the relationship between genes and clinical phenotypes.

[0027] 4. An experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment. Through GO / KEGG enrichment analysis, this method deeply explores the functions of key genes and the signal pathways involved, providing valuable clues for understanding the pathogenesis and potential therapeutic targets of colorectal cancer. At the same time, the construction of the protein-protein interaction (PPI) network further reveals the interaction relationship between core genes.

[0028] 5. An experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment, using the CIBERSORT algorithm, which can efficiently evaluate the infiltration degree of key immune cell types in tumor and normal samples, and present the correlation and infiltration differences of immune cells through a variety of visualization methods. This step is crucial to understanding the characteristics of the immune microenvironment of colorectal cancer and its relationship with cytokine levels. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is the result of the three data sets of GSE35279, GSE87211 and GSE110224 in the experimental method of the relationship between circulating cytokine levels and tumor immune microenvironment proposed in the present invention;

[0030] Figure 2 It is a visualization display of DEGs in an experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment proposed in the present invention;

[0031] Figure 3 It is a weighted correlation network analysis result diagram of an experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment proposed by the present invention;

[0032] Figure 4 This is a GO / KEGG analysis result diagram of an experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment proposed by the present invention;

[0033] Figure 5 It is a ROC curve diagram of an experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment proposed by the present invention. DETAILED DESCRIPTION

[0034] The technical solution of this patent is further described in detail below in conjunction with specific implementation methods.

[0035] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0036] Embodiment 1:

[0037] An experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment, such as Figure 1 As shown, the following steps are included:

[0038] S1: Data collection: Data of colorectal cancer patients and normal control data were collected. The preferred gene microarray data of colorectal cancer patients came from Gene Expression Omnibus (GEO). Three gene expression data sets were carefully selected from the GEO database, namely GSE35279 (derived from the GPL6480 platform), GSE87211 (derived from the GPL13497 platform), and GSE110224 (derived from the GPL570 platform), among which:

[0039] The GSE35279 dataset covers 74 microdissected MSS colorectal tumor samples and 5 microdissected normal colon epithelia;

[0040] The GSE87211 dataset contains 203 MSS rectal cancer samples and 160 matched mucosal control samples;

[0041] The GSE110224 dataset consists of tumor tissue samples and normal colon tissue samples from 17 patients with MSS colorectal adenocarcinoma;

[0042] S2: Analysis of differential gene expression in the data: R software (version 4.4.0) and limma R package (version 3.60.4) were used to perform batch correction and normalization operations on the data; to eliminate batch effects, the sva R package (version 3.52.0) was introduced, which can identify and construct alternative variables for high-dimensional data sets;

[0043] The ggplot2 R package (version 3.5.1) and ggpubr R package (version 0.6.0) were used to perform principal component analysis (PCA) on the data sets before and after the batch effect was removed in the R environment, and the distribution changes of the data in the dimensionality reduction space were visually presented with the help of visualization, so as to facilitate the evaluation of the effect of the batch effect removal. The limma package was mainly used to identify the differentially expressed genes (DEGs) between the control samples and the disease samples. Genes that satisfied |log2FC|>1 and adjusted P value <0.05 were identified as differentially expressed genes between disease tissues and control tissues. A total of 1893 differentially expressed genes (DEGs) were found between the tumor group and the control group. Among these DEGs, 880 genes were up-regulated, while 1013 genes were down-regulated.

[0044] The pheatmap R package (version 1.0.12) was used to draw heat maps and volcano maps to visualize the DEGs;

[0045] S3: Weighted correlation network analysis: Weighted gene co-expression network analysis (WGCNA) can be used to mine and explore co-expressed gene modules, with the aim of exploring the association between module genes and clinical phenotypes. By using the database merged in the previous step, the R package WGCNA (version 1.72-5) was used to construct a gene co-expression network to identify modules related to colorectal cancer:

[0046] First, based on the standard deviation > 0.1, genes whose expression value fluctuations in all samples were less than 0.1 were removed. Such genes may have stable expression due to measurement errors or other reasons and usually do not contain valuable biological information;

[0047] Secondly, the goodSamplesGenes function was used to filter the differential gene expression matrix, exclude low-quality samples and genes, and then construct a scale-free co-expression network;

[0048] Third, the contiguity was calculated using softPower (β = 4) derived from co-expression similarity, which was subsequently converted into a topological overlap matrix (TOM) to determine gene ratios and dissimilarity;

[0049] Fourth, modules were detected by hierarchical clustering and dynamic tree cutting function, and average linkage hierarchical clustering was used to classify genes with the same expression profile into modules, setting the minimum genome size of TOM-based dissimilarity metric and gene dendrogram to 60;

[0050] Fifth, by calculating the difference of module characteristic genes, 0.25 was set as the threshold, a cutting line was selected for the module dendrogram, and several modules were combined for further study, and then the characteristic gene network was visualized;

[0051] Sixth, combining the cluster tree branches with clinical phenotype data, using gene significance (GS) and module membership (MM) for comprehensive evaluation, accurately identifying the core modules closely related to clinical characteristics, and finally obtaining 19 co-expression modules, which are displayed by the cluster tree ( Figure 3 B);

[0052] S4: Key gene screening and GO / KEGG analysis: The VennDiagram (version 1.6.0) R package was used to perform intersection operations on differentially expressed genes (DEGs) and key genes identified by weighted gene co-expression network analysis (WGCNA), thereby obtaining 1301 candidate key genes ( Figure 4 A); Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis was performed using the ClusterProfiler (version 4.12.0) R package, and a P value less than 0.05 was set as the statistically significant standard to analyze potential functions and gene pathways;

[0053] The bubble plot of P value ranking shows the top 10 pathway terms ( Figure 4B), these genes mainly play a role in signaling pathways such as “Cytokine-cytokine receptor interaction”, “Neuroactive ligand-receptor interaction”, “Viralprotein interaction with cytokine and cytokine receptor”, “IL-17signalingpathway” and “Calcium signaling pathway”;

[0054] After applying the R package for GO analysis, we obtained a total of 792 GO terms, including 630 BP terms, 41 CC terms, and 121 MF terms. The top 6 terms of BP, CC, and MF are visualized in the circle plot with the minimum P value ( Figure 4 C), BP-enriched terms mainly include “inorganic anion transport”, “hormone metabolic process”, “chloride transport”, “monoatomic anion transport” and “vascular processin circulatory system”; C terms mainly include “apical plasma membrane”, “apical part of cell”, “basal plasma membrane”, “basal part of cell” and “basolateral plasmamembrane”; MF terms mainly include “cytokine activity”, “growth factor activity”, “monoatomic ion channel activity”, “chloride transmembrane transporteractivity” and “inorganic anion transmembrane transporter activity”;

[0055] S5: Construction of protein-protein interaction network: The intersection genes after the intersection operation were imported into the STRING online tool, and the object in the tool was limited to "Homo sapiens", the required minimum interaction score was set to the highest confidence (0.9), and the disconnected nodes were hidden;

[0056] Subsequently, the protein-protein interaction (PPI) network was visualized and analyzed using Cytoscape 3.9.0 software, and the degree values ​​of the nodes were calculated using the cytoHubba plug-in. In this step, the targets with degree values ​​higher than the median were the core genes of the PPI network. The genes were ranked according to their degree values, with an emphasis on the top 10 key genes in degree value.

[0057] S6: Construction of receiver operating characteristic (ROC) curve: The receiver operating characteristic (ROC) curve was further validated for the hub gene to evaluate the diagnostic value of differentially expressed genes (DEGs), including the area under the curve (AUC), sensitivity, and 1-specificity;

[0058] AUC values ​​in the range of 0.5-0.7, 0.7-0.9, and >0.9 were judged as low, medium, and high accuracy, respectively. Sensitivity and 1-specificity were used together to evaluate the authenticity of the model. The closer the two were to 1.0, the higher the authenticity of the model. The nomogram, a statistical prediction tool, was used to construct a prediction model between the expression level of the hub gene and the risk of disease. The rms package was used to calibrate the regression model, and a calibration curve was drawn to evaluate the prediction accuracy of the model.

[0059] The area under the curve (AUC) of all hub genes was greater than 0.7, indicating that the model had a good fitting effect ( Figure 5 A) A total score is calculated using the hub genes (CXCL10, IL6, PRKACB, IFNG, IL1A). The values ​​of each of these variables give a score on a point scale axis. The total score can be easily calculated by adding each individual score and projecting it onto a lower total scale, enabling the estimation of colorectal cancer disease risk ( Figure 5 B), the calibration curve of the nomogram is as follows Figure 5 As shown in C, this shows that the probability of colorectal cancer predicted by the nomogram is highly consistent with the actual probability.

[0060] S7: Immune infiltration analysis: The integrated and standardized gene expression profile data generated in step S6 were used to identify different types of immune cells in the tissue, and the CIBERSORT algorithm was used to generate an immune cell infiltration matrix, which can efficiently evaluate the infiltration degree of 22 key immune cell types contained in tumor and normal samples; pheatmap (version 1.0.12) and corrplot (version 0.94) packages were used to draw heat maps and correlation heat maps, respectively, to visualize the correlation of the 22 infiltrating immune cells; vioplot (version 0.5.0) package was used to draw violin plots of the immune cell infiltration matrix data to visualize the differences in immune cell infiltration;

[0061] Among them, the urquoise module has the strongest correlation with the clinical characteristics of tumors (r = 0.33) ( Figure 3 C), the average gene significance in each module was estimated ( Figure 3 D), and turquoise module was used to construct the correlation diagram between module members and gene significance, and 5980 genes were obtained (cor=0.94, P=1e-200) ( Figure 2 E);

[0062] The immune infiltration analysis of IL6, IL2, and IFNG, common clinical indicators in hub genes, was performed, and the CIBERSORT algorithm was used to further calculate the differences in the immune microenvironment. Compared with the control group, the proportions of Plasmacells, T cells CD8, T cells CD4 naive, T cells CD4 memory resting, T cells CD4memory activated, T cells regulatory (Tregs), NK cells resting, Macrophages M0, Dendritic cells activated, and Mast cells activated in the tumor group were relatively high, while the proportions of B cells naive, B cells memory, T cells follicular helper, Monocytes, Macrophages M2, Mast cells resting, and Neutrophils were relatively low (p<0.05).

[0063] There is a relationship between the expression of IL-6, IFNG and IL-2 and the infiltration pattern of immune cells; IL-6 is positively correlated with Neutrophils, Mast cells activated, Monocytes, and Tcells CD4 naive, which means that with the increase of IL-6 expression level, the infiltration degree of Neutrophils, Mast cells activated, Monocytes and Tcells CD4 naive in the tumor microenvironment may increase. This positive correlation shows that IL-6 plays a certain role in regulating the function and distribution of these immune cells. On the other hand, IL-6 is negatively correlated with Macrophages M2, Tcells CD8, Macrophages M1, and B cells naive, indicating that when the expression level of IL-6 increases, the infiltration degree of Macrophages M2, T cells CD8, Macrophages M1 and B cells naive may decrease; this negative correlation reflects the inhibitory effect of IL-6 on different types of immune cells, and the complex mutual regulation mechanism between these immune cells and IL-6; at the same time, INFG is also related to Tregs, T cells CD4memory resting was negatively correlated, and T cells CD4 memory activated was positively correlated. IL2 was negatively correlated with NK cells resting, and positively correlated with Macrophages M2 and T cells CD8;

[0064] S8: Immunohistochemistry (IHC) was performed on tissue specimens from 55 patients with MSS colorectal cancer, which were fixed in formalin, embedded in paraffin, and cut into sections with a thickness of 3 mm;

[0065] The sections were dewaxed in xylene, dehydrated in ethanol, and then incubated with 0.3% hydrogen peroxide in methanol for 10 min to eliminate endogenous peroxidase;

[0066] Then, rabbit anti-human CD3, CD4, CD8 (purchased from MXB Biotechnology, Fujian, China) and rabbit anti-human CD163, CD206 (purchased from Proteintech, Wuhan, China) monoclonal antibodies were dropped into the monoclonal antibodies and incubated overnight in a refrigerator at 4 °C;

[0067] Rabbit anti-human IgG H&L and rabbit anti-mouse IgG H&L (Beijing Zhongshan Jinggao Biotechnology Co., Ltd., Beijing, China) were washed with PBS, added dropwise, incubated at room temperature for 20 minutes, washed with PBS, developed with DAB for 3 minutes, counterstained with hematoxylin, dehydrated, vitrified with xylene, and sealed with neutral glue. A KFBO digital pathology slide scanner (KF-PRO-020-HI) was used for panoramic scanning and information storage;

[0068] Finally, Image J pro was used to calculate the degree of staining and infiltration of specific immune cells to interpret the results.

[0069] In the above steps, three gene expression data sets were screened from the GEO database, 1893 genes were extracted for differential gene expression analysis, and then 19 co-expression modules were obtained through WGCNA analysis. 1301 genes were further screened for key genes, GO / KEGG analysis, and PPI network construction; then 10 hub genes were obtained;

[0070] Common tumor-associated immune cells CD3, CD4, CD8, CD163, and FOXP3 were labeled by immunohistochemistry, and the levels of common cytokines in serum were obtained by flow cytometry. The statistical results showed that IL6 and IFNγ were one of the most critical factors affecting the immune microenvironment of colorectal cancer, and were involved in the formation of immune infiltration, and were closely related to the infiltration of M2 macrophages and CD8+T cells. It was found that some inflammatory cells were indeed correlated with some cytokines in serum, and the cytokine levels in serum did affect the infiltration of immune cells in the tumor immune microenvironment, especially IL6 and IFNγ, and these cytokines may be involved in the formation of CRC immune resistance. Within the tumor, CD4+T cells were correlated with the levels of IL17, TNFα, and IFNγ in serum. By further grouping the common cytokine levels in serum, it was observed that there was a significant difference between the CD4+T cells in the tumor at different levels of IL4 and IL10 in serum.

[0071] Survival analysis showed that the prognosis of CRC patients was significantly different due to different expression levels of IL6, IFNG, and IL2. According to the existing CRC diagnosis and treatment guidelines, ICIs are recommended for MSI-H / MMR patients. The patients were further divided into two groups, MSS / MSI-L (602 patients) and MSI-H (252 patients), and cytokine-related survival analysis was performed on them. Among MSI-H patients, the prognosis of patients did not show significant differences due to different cytokine expression levels. However, among MSS / MSI-L patients, although no significant differences were observed in the prognosis of patients between the IL2 and IFNG groups, when the IL6 expression level was high, CRC patients with MSS / MSI-L had a better prognosis;

[0072] Further grouping by IL6 level, 520 patients after ICIs treatment were grouped and it was found that patients with high IL6 expression levels had better prognosis after ICIs treatment.

[0073] The above is a preferred specific implementation manner of the present invention, and the protection scope of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by any technician familiar with the field within the technical scope disclosed by the present invention in combination with the prior art or public common sense, within the spirit and principle of the present invention, shall be covered by the protection scope of the present invention.

Claims

1. An experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment, comprising the following steps: S1: Collect colorectal cancer patient data and normal control data, preferably select gene expression data sets from the Gene Expression Omnibus (GEO) database; S2: Analyze the differential gene expression in the data, use R software and limma R package to perform batch correction and normalization operations on the data, and identify the differentially expressed genes (DEGs) between control samples and disease samples; S3: Perform weighted correlation network analysis (WGCNA) to construct a gene co-expression network and identify colorectal cancer-related modules; S4: Intersection operation was performed on the differentially expressed genes (DEGs) and the key genes identified by WGCNA to obtain candidate key genes, and GO / KEGG enrichment analysis was performed; S5: construct protein-protein interaction (PPI) network, calculate the degree of nodes, and identify the core genes of PPI network; S6: Receiver operating characteristic (ROC) curve was constructed to evaluate the diagnostic value of hub gene; S7: Use integrated and standardized gene expression profile data to identify different types of immune cells in tissues and perform immune infiltration analysis; S8: Immunohistochemistry (IHC) was used to measure the degree of staining infiltration of specific immune cells in tissue specimens from patients with colorectal cancer.

2. The experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment according to claim 1, characterized in that: In step S1, the gene expression datasets screened from the GEO database include GSE35279, GSE87211 and GSE110224.

3. The experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment according to claim 2, characterized in that: In the S2 step, the sva R package was used to eliminate the batch effect, and the ggplot2 R package and the ggpubr R package were used to perform principal component analysis (PCA) to visualize the distribution changes of the data in the dimensionality reduction space.

4. The experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment according to claim 2, characterized in that: In the S3 step, the WGCNA R package was used to construct a gene co-expression network, and the modules were detected by hierarchical clustering and dynamic tree cutting functions to identify 19 co-expression modules closely related to clinical characteristics and display them through a clustering tree.

5. The experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment according to claim 2, characterized in that: In the S4 step, the VennDiagram R package was used to perform an intersection operation on the differentially expressed genes and the key genes identified by WGCNA, and the ClusterProfiler R package was used to perform GO and KEGG enrichment analysis to analyze potential functions and gene pathways.

6. The experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment according to claim 2, characterized in that: In the S5 step, the intersection genes after the intersection operation are imported into the STRING online tool to construct a PPI network, and the PPI network is visualized and analyzed using the Cytoscape software. The degree values ​​of the nodes are calculated using the cytoHubba plug-in, and the targets with degree values ​​higher than the median are regarded as the core genes of the PPI network. The genes are sorted according to the degree values, with emphasis on the top 10 key genes in the degree values.

7. The experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment according to claim 2, characterized in that: In the S6 step, an ROC curve is constructed to evaluate the diagnostic value of the hub gene, including the area under the curve (AUC). AUC values ​​in the ranges of 0.5-0.7, 0.7-0.9, and >0.9 are judged as low, medium, and high accuracy, respectively. A nomogram is used to construct a prediction model between the expression level of the hub gene and the risk of disease. The regression model is calibrated using the rms package, and a calibration curve is drawn to evaluate the predictive accuracy of the model. The total score is calculated using the hub genes (CXCL10, IL6, PRKACB, IFNG, IL1A). The value of each of these variables is scored on the point scale axis, and the total score is calculated for estimation.

8. The experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment according to claim 2, characterized in that: In the S7 step, the CIBERSORT algorithm is used to generate an immune cell infiltration matrix to evaluate the infiltration degree of key immune cell types in tumor and normal samples, and heat maps, correlation heat maps and violin plots are drawn for visual presentation.

9. The experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment according to claim 2, characterized in that: In the step S8, the samples were fixed with formalin, embedded in paraffin and cut into slices with a thickness of 3 mm, immunohistochemical staining was performed, and the staining infiltration degree of specific immune cells was calculated by Image J pro.

10. The experimental method for the relationship between circulating cytokine levels and tumor immune microenvironment according to claim 8, characterized in that: In the S7 step, the CIBERSORT algorithm is used to further calculate the differences in the immune microenvironment, and the correlation between the expression of IL-6, IFNG and IL-2 and the infiltration pattern of immune cells is detected compared with the control group.

Images