Prognosis model for predicting cervical cancer based on autophagy-related gene and construction method thereof

Abstract

The invention discloses a prognosis model for predicting cervical cancer based on autophagy-related genes and a construction method of the prognosis model, and belongs to the technical field of biomedicine. The model contains four characteristic genes related to prognosis of cervical cancer: BCL2, SPNS1, TM9SF1 and TP73, and the characteristic genes can become biological markers related to cervical cancer; the calculation formula of the prognosis model is as follows: risk score = (-0.411 * BCL2 gene expression quantity) + (0.753 * SPNS1 gene expression quantity) + (0.669 * TM9SF1 gene expression quantity) + (-0.398 * TP73 gene expression quantity). The prognosis model provided by the invention can evaluate the prognosis of the cervical cancer patient, improve the prognosis prediction capability of the cervical cancer patient, effectively identify the high-risk patient, assist in predicting the curative effect of immunotherapy, detect and intervene the high-risk patient earlier in clinic, improve the survival rate and life quality of the patient, and improve the clinical application prospect. A reference is provided for individualized diagnosis and treatment of cervical cancer; and tests and external verification prove that the model is stable and effective.

Description

Technical Field

[0001] This invention belongs to the field of biomedical technology, specifically relating to a prognostic model for cervical cancer based on autophagy-related genes and its construction method. Background Technology

[0002] Cervical cancer is one of the most common malignant tumors in women. Early-stage cervical cancer patients have a relatively good prognosis after receiving standard treatment. However, patients with advanced cervical cancer often have pelvic or distant metastases, and their 5-year overall survival rate is only 15-20%. The high metastatic potential of cervical cancer leads to disease recurrence and progression, which is one of the leading causes of death. Therefore, in-depth research into the mechanisms of malignant progression of cervical cancer is currently a key focus of cervical cancer research.

[0003] Autophagy is a lysosome-dependent catabolistic process in eukaryotic cells. It involves the engulfment of cytoplasmic proteins or organelles into vesicles, where they bind to lysosomes to form autolysosomes, degrading their contents to meet cellular metabolic needs or facilitate organelle self-renewal. This process involves multiple autophagy genes. Autophagy plays a dual role in different types of tumors, both inhibiting tumor growth and promoting tumor progression. In cervical cancer, downregulation of Beclin1, the first discovered human autophagy gene, is an independent risk factor for prognosis. Overexpression of Beclin1 significantly increases the killing effect of chemotherapy drugs on tumor cells. Therefore, the expression of autophagy-related genes is closely related to tumor development and patient prognosis.

[0004] In recent years, several studies have found that autophagy is involved in the development and progression of cervical cancer. Hu et al. examined the expression levels of autophagy markers Beclin1 and LC3 in cervical cancer, high-grade cervical intraepithelial neoplasia, and normal cervical tissue, finding that Beclin1 and LC3 expression was downregulated in cervical cancer, and their expression levels were not significantly correlated with age, FIGO stage, differentiation degree, or lymph node metastasis. Zhang et al. found that MAP7 promotes the proliferation and invasion of cervical cancer by regulating the autophagy pathway. However, the mechanisms and application value of autophagy in predicting the prognosis and guiding immunotherapy in cervical cancer patients remain unclear, and large-scale, systematic, and in-depth studies are lacking.

[0005] Therefore, it is necessary to analyze the expression and prognosis of autophagy-related genes in cervical cancer. This can not only provide a theoretical basis for constructing a prognostic risk model for cervical cancer, but also help patients improve their prognosis. Summary of the Invention

[0006] The purpose of this invention is to overcome the problems existing in the prior art and provide a prognostic model for cervical cancer based on autophagy-related genes, which contains four characteristic genes related to the prognosis of cervical cancer: BCL2, SPNS1, TM9SF1 and TP73. These characteristic genes can serve as biomarkers for cervical cancer and can be applied to basic medical research and genetic engineering development.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: a prognostic model for cervical cancer based on autophagy-related genes, containing four characteristic genes related to the prognosis of cervical cancer: BCL2, SPNS1, TM9SF1 and TP73. These four characteristic genes can serve as biomarkers related to cervical cancer.

[0008] The prognostic model is calculated as follows: Risk score = (-0.411 × BCL2 gene expression level) + (0.753 × SPNS1 gene expression level) + (0.669 × TM9SF1 gene expression level) + (-0.398 × TP73 gene expression level).

[0009] This invention also provides a method for constructing a prognostic model for cervical cancer based on autophagy-related genes, comprising the following steps:

[0010] S1. Obtain transcriptomic data and clinical data of cervical cancer patients from TCGA and GTEx databases, and obtain autophagy-related genes from the human autophagy database. The transcriptomic data includes normal cervical tissue samples and cervical cancer tissue samples.

[0011] S2. From the transcriptome data obtained in step S1 above, differentially expressed autophagy genes are screened according to the criteria of |log2 (fold change)|>0.5 and false discovery rate<0.05.

[0012] S3. Using the expression values ​​of differentially expressed autophagy genes obtained in step S2 above as independent variables, and with death as the key event, univariate Cox regression analysis is used to finally screen out autophagy genes related to prognosis.

[0013] S4. Select autophagy genes related to prognosis from step S3 above, and establish an autophagy-related prognostic model using LASSO regression analysis.

[0014] S5. Based on the correlation coefficient between gene-related expression levels and risk scores, calculate the risk score and divide cervical cancer patients into high-risk and low-risk groups. Plot Kaplan-Meier survival curves and ROC curves. Using age, histological grade, clinical stage, and risk score as independent variables and death as the endpoint event, obtain independent prognostic factors for cervical cancer through univariate and multivariate Cox regression. Externally validate the model's predictive ability using the GEO database GSE52903 as the validation set.

[0015] In step S1, 232 autophagy genes were obtained from the human autophagy database. Transcriptome data included gene expression and expression levels from 258 cervical cancer samples and 13 normal cervical tissue samples. Patient clinical data included age, clinical stage, histological grade, survival time, and survival status. The inclusion criteria for patients were pathological diagnosis of cervical cancer, no missing follow-up time, and complete clinical data.

[0016] In step S2, 32 differentially expressed autophagy genes were obtained through screening, namely APOL1, ATG16L2, BAK1, BCL2, BID, BIRC5, CASP3, CDKN2A, CFLAR, DIRAS3, DLC1, EIF4EBP1, ERN1, GABARAP, GNAI3, GRID1, IFNG, IKBKE, IL24, ITPR1, MAP1LC3A, NRG2, PIK3C3, PINK1, PRKCQ, PTK6, RGS19, SPNS1, TM9SF1, TNFSF10, TP63, and TP73.

[0017] In step S5, the differences in the immune microenvironment, the expression differences of immune checkpoint genes, and the differences in immune epigenetic scores (IPS) between the high-risk and low-risk groups are assessed to help predict the efficacy of immunotherapy.

[0018] In step S5, clinical stage and risk score are determined to be independent prognostic factors for cervical cancer through univariate and multivariate Cox regression.

[0019] The beneficial effects of this invention are:

[0020] 1) This invention constructs a prognostic model for cervical cancer based on autophagy-related genes. The model contains four characteristic genes related to the prognosis of cervical cancer: BCL2, SPNS1, TM9SF1 and TP73. These characteristic genes can serve as biomarkers for cervical cancer. The prognostic model is calculated as follows: Risk score = (-0.411 × BCL2 gene expression level) + (0.753 × SPNS1 gene expression level) + (0.669 × TM9SF1 gene expression level) + (-0.398 × TP73 gene expression level).

[0021] 2) The prognostic model of the present invention can assess the prognosis of cervical cancer patients, improve the predictive ability of cervical cancer patients' prognosis, effectively identify high-risk patients, assist in predicting the efficacy of immunotherapy, enable earlier detection and intervention of high-risk patients in clinical practice, improve patients' survival rate and quality of life, and provide a reference for individualized diagnosis and treatment of cervical cancer; and through testing and external validation, the model has been confirmed to be stable and effective. Attached Figure Description

[0022] Figure 1 This is a flowchart of the prognostic model construction method of the present invention.

[0023] Figure 2 These are the four autophagy genes screened by univariate Cox regression in the prognostic model of this invention;

[0024] Figure 3 The Kaplan-Meier survival curves and the areas under the curves (AUC) at 1, 3, and 5 years are plotted in the prognostic model of this invention.

[0025] Figure 4 This illustrates the differences in the immune microenvironment between the high-risk and low-risk groups in the prognostic model of this invention.

[0026] Figure 5 The expression differences of immune checkpoint genes in the high-risk and low-risk groups in the prognostic model of this invention;

[0027] Figure 6 This refers to the difference in immune epigenetic scores (IPS) between the high-risk and low-risk groups in the prognostic model of this invention. Detailed Implementation

[0028] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments, including the construction and clinical application of a cervical cancer prognostic model based on autophagy-related genes.

[0029] 1. Materials and Methods:

[0030] 1.1 Data Collection:

[0031] Clinical data and gene expression data of 258 cervical cancer tissue samples and 13 normal cervical tissue samples were downloaded from the TCGA database (https: / / portal.gdc.cancer.gov / ). Gene expression data of 10 normal cervical tissue samples were downloaded from the GTEx database (https: / / gtexportal.org / home / ). Clinical information and gene expression data of 55 cervical cancer tissue samples from the GSE52903 dataset were downloaded from the GEO database (https: / / www.ncbi.nlm.nih.gov / geo).

[0032] Inclusion criteria: ① Pathological diagnosis of cervical cancer; ② No missing follow-up time; ③ Complete clinical data including age and clinical stage. Exclusion criteria: ① Incomplete follow-up time; ② Incomplete clinical data. The TCGA and GTEx datasets were merged and corrected using the "normalizeBetweenArrays()" function from the "limma" package in R software (version 4.1.2).

[0033] The clinicopathological characteristics of the samples included in the study are shown in Table 1 below.

[0034] Table 1 shows the clinicopathological features of 313 cervical cancer tissue samples [n(%)].

[0035]

[0036]

[0037] 1.2 Screening for differentially expressed autophagy-related genes:

[0038] 232 autophagy-related gene ARGs were obtained from the human autophagy database (HADBb) (http: / / www.autophagy.lu / ). Using the "limma" package, differentially expressed ARGs were screened from 258 cervical cancer tissues and 13 normal cervical tissues according to the criteria of |log2 fold change (FC)|>0.5 and false discovery rate (FDR)<0.05. Gene ontology (GO) analysis was performed on the differentially expressed ARGs using the "org.Hs.eg.db", "ggplot2", and "enrichplot" packages in R. P<0.05 was considered statistically significant.

[0039] Thirty-two differentially expressed autophagy genes were identified through screening: APOL1, ATG16L2, BAK1, BCL2, BID, BIRC5, CASP3, CDKN2A, CFLAR, DIRAS3, DLC1, EIF4EBP1, ERN1, GABARAP, GNAI3, GRID1, IFNG, IKBKE, IL24, ITPR1, MAP1LC3A, NRG2, PIK3C3, PINK1, PRKCQ, PTK6, RGS19, SPNS1, TM9SF1, TNFSF10, TP63, and TP73.

[0040] 1.3 Establishment of risk scores for autophagy-related genes:

[0041] First, the Cox regression analysis of differentially expressed ARGs was performed using the "survival" package to screen out ARGs associated with cervical cancer prognosis; the four prognostic autophagy genes were BCL2, SPNS1, TM9SF1 and TP73.

[0042] Immunohistochemical staining techniques from the HPA (Human Protein Atlas, HPA) database (https: / / www.proteinatlasorg / ) were used to verify the protein expression of ARGs associated with cervical cancer prognosis in cervical cancer. A risk scoring model was established using prognostic ARGs through LASSO Cox regression, with the risk score calculated as: Risk Score = Gene Expression Level 1 × Regression Coefficient 1 + Gene Expression Level 2 × Regression Coefficient 2 + ... + Gene Expression Level n × Regression Coefficient n.

[0043] Risk scores were calculated for all cervical cancer samples, and patients were divided into high-risk and low-risk groups based on the median. The independent prognostic predictive value of the risk scores was tested using Kaplan-Meier survival curves, time-dependent receiver operating characteristic (ROC) curves, and univariate and multivariate Cox regression. The reliability of the model was validated on the GSE52903 dataset.

[0044] 1.4 Immune infiltration analysis:

[0045] The ESTIMATE method was used to assess the Immune score, stromal score, and Estimate score (i.e., immune component, stromal component, and total component) of each sample in the high- and low-risk groups. The CIBERSORT method was used to calculate the infiltration of 22 types of immune cells in 258 cervical cancer tissue samples. Spearman correlation analysis was used to analyze the correlation between risk scores and immune cells; P < 0.05 was considered statistically significant.

[0046] 1.5 Prediction of Immunotherapy Efficacy:

[0047] Box plots were generated to compare the differences in expression of human leukocyte antigen (HLA) and immune checkpoints between high- and low-risk groups using the "limma" package. Immunophenoscores (IPS) for all cervical cancer samples were downloaded from the Cancer Immunome Atlas (TCIA) database (https: / / tcia.at / ). IPS is a predictive indicator of antibody response to anti-cytotoxic T-lymphocyte antigen 4 (CTLA-4) and / or anti-programmed cell death protein 1 (PD-1). Differences in IPS between high- and low-risk groups were compared; P < 0.05 was considered statistically significant.

[0048] 2. Experimental Results:

[0049] 2.1 Screening for differentially expressed autophagy-related genes:

[0050] From 258 cervical cancer tissue samples and 13 normal cervical tissue samples, 32 differentially expressed ARGs were screened according to the criteria of |log2FC|>0.5 and FDR<0.05. Among them, 17 genes, including APOL1, BAK1, BIRC5, EIF4EBP1, GNAI3, IFNG, and TP73, were upregulated, while 15 genes, including ATG16L2, BCL2, CFLAR, DLC1, ERN1, NRG2, and PIK3C3, were downregulated. GO functional enrichment analysis of the 32 differentially expressed ARGs revealed that biological processes (BP) were mainly enriched in macroautophagy, endogenous apoptosis signaling pathways, and autophagy regulation; cellular components (CC) were mainly enriched in autophagosomes, autophagosome membranes, and autolysosomes; and molecular functions (MF) were mainly enriched in ubiquitin-protein ligase binding, ubiquitin-like protein ligase binding, protease binding, and death receptor binding.

[0051] 2.2 Establishment of a risk scoring model for autophagy-related genes:

[0052] To understand the impact of 32 differentially expressed ARGs on the prognosis of cervical cancer, univariate Cox regression analysis was used to identify four genes associated with the prognosis of cervical cancer. Among them, SPNS1 and TM9SF1 were considered risk factors, with high expression levels associated with poor prognosis, while BCL2 and TP73 were considered protective factors, with high expression levels associated with better prognosis (see Table 2 below).

[0053] Table 2 shows the differentially expressed autophagy-related genes associated with prognosis.

[0054]

[0055] These four prognostic ARGs were used to construct a risk scoring model using LASSO Cox regression, see [link to model]. Figure 2 A and 2B. Risk score = -0.411 × BCL2 expression level + 0.753 × SPNS1 expression level + 0.669 × TM9SF1 expression level - 0.398 × TP73 expression level.

[0056] Based on the risk scoring formula, a risk score was calculated for each cervical cancer patient, dividing them into a high-risk group (n=129) and a low-risk group (n=129). To further determine the expression of these four prognostic ARGs in cervical cancer, immunohistochemical data from the HPA database were used for validation. It was found that in the high-risk group, compared with the low-risk group, the expression of SPNS1 and TM9SF1 was upregulated, while the expression of BCL2 and TP73 was downregulated. Kaplan-Meier survival curves were plotted, revealing a worse prognosis in the high-risk group and a better prognosis in the low-risk group, with a statistically significant difference (P=0.001). Figure 3 A.

[0057] 2.3 Evaluation and validation of the autophagy-related gene risk scoring model:

[0058] Based on clinicopathological features, univariate and multivariate Cox regression analyses were performed on the established risk scoring model. The results showed that clinical stage and risk score were independent prognostic factors for cervical cancer, as shown in Table 3 below.

[0059] Table 3 shows the results of univariate and multivariate Cox regression analyses.

[0060]

[0061] To further validate the reliability of the risk model, ROC curves for 1-year, 3-year, and 5-year survival rates were plotted. The areas under the curves (AUCs) for 1-year, 3-year, and 5-year survival rates were 0.761, 0.722, and 0.757, respectively. (See attached diagram). Figure 3 B. Additionally, external validation was performed using the GSE52903 dataset from the GEO database. Based on risk scores, participants were divided into a high-risk group (n=28) and a low-risk group (n=27). Kaplan-Meier survival curves were plotted, revealing that the high-risk group had a worse prognosis, while the low-risk group had a better prognosis; the difference was statistically significant (P=0.036). (See [link to relevant documentation]). Figure 3C. ROC curves for 1-year, 3-year, and 5-year survival were plotted. The areas under the curve (AUC) for 1-year, 3-year, and 5-year survival were 0.707, 0.664, and 0.716, respectively. See [link to ROC curve]. Figure 3 D. These results indicate that the autophagy-related gene risk scoring model has good predictive value, high reliability of the prediction results, and certain clinical significance.

[0062] 2.4 Immune infiltration analysis using an autophagy-related gene risk scoring model:

[0063] The difference in total immune function (TME) between high- and low-risk groups for autophagy-related genes was assessed using ESTIMATE. The results showed that the Immune score (immune component) and Estimate score (total component) were higher in the low-risk group than in the high-risk group, suggesting that the low-risk group may contain more immune cell components. (See...) Figure 4 The infiltration of 22 immune cells in 258 cervical cancer tissues was calculated using CIBERSORT. Spearman correlation analysis was performed to analyze the correlation between risk score and immune cells. The results showed that the risk score was negatively correlated with CD8+ T cells, helper follicular T cells, M1 macrophages, resting mast cells, and resting dendritic cells, while it was positively correlated with M0 macrophages, activated mast cells, and resting NK cells.

[0064] 2.5 Predicting the efficacy of immunotherapy using an autophagy-related gene risk scoring model:

[0065] The expression differences of HLA and immune checkpoint genes in high- and low-risk groups were compared. The results showed that the expression of HLA and immune checkpoint genes was higher in the low-risk group than in the high-risk group. (See...) Figure 5 The IPS score is a better predictor of response to anti-CTLA-4 and / or anti-PD-1 antibodies. By comparing the differences in IPS scores between high-risk and low-risk groups for PD-1 positivity alone and for double PD-1 and CTLA-4 positivity, the results showed that the IPS score in the low-risk group was significantly higher than that in the high-risk group. (See [link to relevant documentation]). Figure 6 These results suggest that the low-risk group had higher immunogenicity and better efficacy with immune checkpoint inhibitors.

[0066] This invention determined the expression levels of 232 ARGs in cervical cancer by analyzing transcriptomic and clinical data from cervical cancer and normal cervical tissues in the TCGA and GTEx databases, and screened out 32 differentially expressed ARGs. GO functional enrichment analysis revealed that these differentially expressed genes were mainly enriched in autophagy, endogenous apoptosis signaling pathways, death receptor binding, ubiquitin-protein ligase binding, ubiquitin-like protein ligase binding, and protease binding. Univariate Cox regression analysis identified four prognostic-related ARGs (SPNS1, TM9SF1, TP73, and BCL2).

[0067] In this invention, a cervical cancer prognostic prediction model was successfully constructed using LASSO Cox regression and the four prognostic-related ARGs mentioned above. The risk score for each cervical cancer patient was calculated, and patients were divided into high-risk and low-risk groups. The overall survival time of the high-risk group was significantly shorter than that of the low-risk group. Univariate and multivariate Cox regression analyses revealed that clinical stage and risk score were independent prognostic factors for cervical cancer. The area under the ROC curve (AUC) for predicting 1-year, 3-year, and 5-year survival rates of cervical cancer patients was greater than 0.7. External validation using the GSE52903 dataset yielded the same results. These results indicate that the cervical cancer prognostic prediction model based on ARG expression has good predictive value, high reliability, and good clinical application value.

[0068] This invention assessed the differences in TME between high- and low-risk groups and found that low-risk groups may contain more immune cell components. Using CIBERSORT, the infiltration of 22 immune cells in 258 cervical cancer tissues was calculated. It was found that the risk score was negatively correlated with CD8+ T cells, helper follicular T cells, M1 macrophages, resting mast cells, and resting dendritic cells, while it was positively correlated with M0 macrophages, activated mast cells, and resting NK cells.

[0069] This invention establishes a prognostic prediction model for cervical cancer patients based on four autophagy genes related to prognosis. The model has stable predictive performance, independent prognostic value and clinical relevance, and can help predict the efficacy of immunotherapy, providing a reference for individualized diagnosis and treatment of cervical cancer patients.

[0070] The above description is only used to illustrate the technical solution of the present invention and is not intended to limit it. Any other modifications or equivalent substitutions made by those skilled in the art to the technical solution of the present invention, as long as they do not depart from the spirit and scope of the technical solution of the present invention, should be covered within the scope of the claims of the present invention.

Claims

1. A prognostic model for cervical cancer based on autophagy-related genes, characterized in that: It contains four characteristic genes associated with cervical cancer prognosis: BCL2, SPNS1, TM9SF1, and TP73. These four characteristic genes can serve as biomarkers for cervical cancer.

2. The prognostic model for cervical cancer based on autophagy-related genes according to claim 1, characterized in that: The prognostic model is calculated as follows: Risk score = (-0.411 × BCL2 gene expression level) + (0.753 × SPNS1 gene expression level) + (0.669 × TM9SF1 gene expression level) + (-0.398 × TP73 gene expression level).

3. A method for constructing a prognostic model for cervical cancer based on autophagy-related genes as described in any one of claims 1 or 2, characterized in that: Includes the following steps: S1. Obtain transcriptomic data and clinical data of cervical cancer patients from TCGA and GTEx databases, and obtain autophagy-related genes from the human autophagy database. The transcriptomic data includes normal cervical tissue samples and cervical cancer tissue samples. S2. From the transcriptome data obtained in step S1 above, differentially expressed autophagy genes are screened according to the criteria of |log2 (fold change)|>0.5 and false discovery rate<0.

05. S3. Using the expression values ​​of differentially expressed autophagy genes obtained in step S2 above as independent variables, and with death as the key event, univariate Cox regression analysis is used to finally screen out autophagy genes related to prognosis. S4. Select autophagy genes related to prognosis from step S3 above, and establish an autophagy-related prognostic model using LASSO regression analysis. S5. Based on the correlation coefficient between gene-related expression levels and risk scores, calculate the risk score and divide cervical cancer patients into high-risk and low-risk groups. Plot Kaplan-Meier survival curves and ROC curves. Using age, histological grade, clinical stage, and risk score as independent variables and death as the endpoint event, obtain independent prognostic factors for cervical cancer through univariate and multivariate Cox regression. Externally validate the model's predictive ability using the GEO database GSE52903 as the validation set.

4. The method for constructing a prognostic model for cervical cancer based on autophagy-related genes according to claim 3, characterized in that: In step S1, 232 autophagy genes were obtained from the human autophagy database. Transcriptome data included gene expression and expression levels from 258 cervical cancer samples and 13 normal cervical tissue samples. Patient clinical data included age, clinical stage, histological grade, survival time, and survival status. The inclusion criteria for patients were pathological diagnosis of cervical cancer, no missing follow-up time, and complete clinical data.

5. The method for constructing a prognostic model for cervical cancer based on autophagy-related genes according to claim 3, characterized in that: In step S2, 32 differentially expressed autophagy genes were obtained through screening, namely APOL1, ATG16L2, BAK1, BCL2, BID, BIRC5, CASP3, CDKN2A, CFLAR, DIRAS3, DLC1, EIF4EBP1, ERN1, GABARAP, GNAI3, GRID1, IFNG, IKBKE, IL24, ITPR1, MAP1LC3A, NRG2, PIK3C3, PINK1, PRKCQ, PTK6, RGS19, SPNS1, TM9SF1, TNFSF10, TP63, and TP73.

6. The method for constructing a prognostic model for cervical cancer based on autophagy-related genes according to claim 3, characterized in that: In step S5, the differences in the immune microenvironment, the expression differences of immune checkpoint genes, and the differences in immune epigenetic scores (IPS) between the high-risk and low-risk groups are assessed to help predict the efficacy of immunotherapy.

7. The method for constructing a prognostic model for cervical cancer based on autophagy-related genes according to claim 3, characterized in that: In step S5, clinical stage and risk score are determined to be independent prognostic factors for cervical cancer through univariate and multivariate Cox regression.

Images