A composition of biomarkers for liver cancer diagnosis and application thereof

Abstract

The application discloses a composition containing biomarkers, wherein the biomarkers comprise FOXD2-AS1, miR-9-5p and STMN1 genes, preferably, the biomarkers further comprise COL15A1 and CCNE2 genes. The application finds that the biomarkers are differentially expressed in hepatocellular carcinoma, and FOXD2-AS1 and miR-9-5p and miR-9-5p and STMN1 have a target binding relationship, and the three form a ceRNA combination, and the combination further combines COL15A1 and CCNE2 to be used as a biomarker combination for diagnosing hepatocellular carcinoma, and the combination has a higher diagnostic value for hepatocellular carcinoma.

Description

Technical Field

[0001] This invention relates to a composition of biomarkers for the diagnosis of liver cancer and its use. Background Technology

[0002] Hepatocellular carcinoma (HCC), originating from the cancerous transformation of liver parenchymal cells, is one of the leading causes of cancer incidence and mortality worldwide. According to global cancer data reports from 2018, approximately 841,000 people are diagnosed with liver cancer annually, and 782,000 die from the disease. Currently, viral infections (including HBV and HCV), diet (including excessive alcohol consumption and eating moldy food), unhealthy lifestyles (such as obesity and smoking), and other diseases (such as type 2 diabetes and schistosomiasis) are known to be major risk factors for liver cancer. Although hepatitis B vaccination is included in the national immunization program, its effectiveness is limited to young people. While HBV infection rates and the incidence of liver cancer caused by it have decreased significantly, the situation remains severe for the elderly. Therefore, exploring early diagnostic biomarkers is crucial for the prevention and control of liver cancer.

[0003] The cancer transcriptome is characterized by the abnormal expression and regulatory relationships between protein-coding and non-coding transcripts. Competing endogenous RNA (ceRNA) networks are formed by combining the functions of both protein-coding and non-coding RNAs. In recent years, an increasing number of researchers have focused on exploring the functional roles of ceRNAs in cancer. MicroRNAs (miRNAs), long non-coding RNAs (lncRNAs), circular RNAs (circRNAs), pseudogenes, and viral transcripts can all function as ceRNAs in the human body, participating in the regulatory mechanisms of the ceRNA network under physiological and pathological conditions. Among them, miRNAs belong to a class of ncRNAs, composed of 22-25 nucleotides, and mainly act on the expression of target genes. Changes in miRNA expression patterns can be affected by locus mutations, epigenetic methylation, transcription factor regulation, or dysregulation of mature miRNA processing. lncRNAs are longer than 200 bp, containing a 3′ poly(a) tail, a 5′ cap, and a promoter. Although they do not encode proteins, they still play a crucial role in normal development and tumorigenesis. circRNAs are single-stranded, covalently closed circular RNAs formed by the direct linking of the 5′ and 3′ ends of linear RNA. Primarily found in the cytoplasm, they do not encode proteins and exhibit greater stability compared to linear RNAs. These non-coding RNAs constitute a ceRNA network pool within the human body, enabling communication and co-regulation during development and disease progression. For example, lncRNA UCA1, acting as a miR-182-5p sponge, positively regulates the expression of Delta-like ligand 4 (DLL4), promoting the malignant phenotype of renal cell carcinomas.

[0004] Databases such as The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) are indispensable public databases in bioinformatics analysis. Researchers can download tumor-related data from these databases for analysis to uncover the expression trends of certain genes in specific tumors, laying the foundation for further exploration of tumor pathogenesis and greatly promoting research in cancer biology.

[0005] Currently, serum AFP, liver ultrasound, and imaging examinations including CT, MRI, PET, and SPECT each have their advantages in the diagnosis of liver cancer, but pathological biopsy remains the definitive diagnostic method. New diagnostic biomarkers include serum alpha-fetoprotein isoform (lensculinaris agglutinin-reactive fraction of AFP, AFP-L3) and abnormal prothrombin (protein induced by vitamin Kabsence / antagonist-II, PIVKAII), but their clinical applicability is limited. Treatment for liver cancer mainly includes surgical resection, hepatic artery embolization, radiofrequency ablation, and radiotherapy / chemotherapy. However, due to the difficulty in early and accurate diagnosis, treatment outcomes are unsatisfactory, and the prognosis for liver cancer patients remains poor. Therefore, exploring valuable molecular biomarkers and therapeutic targets for liver cancer is crucial for conquering HCC. Summary of the Invention

[0006] This invention provides the following technical solution:

[0007] A composition containing biomarkers, said biomarkers including the FOXD2-AS1, miR-9-5p, and STMN1 genes.

[0008] According to the composition of the present invention, FOXD2-AS1 is a long non-coding RNA (lncRNA), miR-9-5p is a small RNA (miRNA), and STMN1 is a messenger RNA (mRNA).

[0009] According to the composition of the present invention, the biomarker further comprises the COL15A1 and CCNE2 genes.

[0010] According to the composition of the present invention, both COL15A1 and CCNE2 are messenger RNAs (mRNAs).

[0011] The present invention also provides a composition of the above-described biomarkers for the diagnosis and prognostic assessment of hepatocellular carcinoma.

[0012] Specifically, the application of the composition containing biomarkers of the present invention is as follows: compared with the control group, the concentration of the above-mentioned biomarkers from the sample to be diagnosed or evaluated changes, thereby diagnosing whether the source of the sample has liver cancer.

[0013] Therefore, the present invention also provides a composition for the diagnosis of liver cancer, the composition comprising a detection reagent for biomarkers including the FOXD2-AS1, miR-9-5p, and STMN1 genes. Preferably, it further comprises the COL15A1 and CCNE2 genes.

[0014] According to the composition of the present invention, the detection reagent is a reagent capable of quantitatively analyzing the biomarker based on the nucleic acid level.

[0015] According to the composition of the present invention, the nucleic acid level detection assay is a reagent used for nucleic acid amplification reaction, polymerase chain reaction, real-time quantitative polymerase chain reaction, and reverse transcription polymerase chain reaction.

[0016] The nucleic acid level detection reagent contains the nucleic acid sequence of the biomarker and primers for specifically recognizing the nucleic acid sequence.

[0017] Those skilled in the art will understand that, once a biomarker is identified, the reagents and methods for detecting the various biomarkers described above are known to those skilled in the art or can be determined by technical means.

[0018] The present invention also provides the use of the above composition in the preparation of liver cancer diagnostic agents or diagnostic kits.

[0019] The biomarkers described in this invention refer to genes that are differentially regulated between individuals with liver cancer and normal individuals. The diagnosis refers to determining the susceptibility of the examined individual to a specific disease (e.g., liver cancer), whether they have the disease, and monitoring their status during treatment. The prognostic assessment includes determining the recovery status after being diagnosed with liver cancer and receiving treatment. The sample described in this invention is a biopsy specimen of the patient's tumor solid tissue.

[0020] Beneficial effects

[0021] This invention found that the relative expression levels of lncRNA FOXD2-AS1, miRNA miR-9-5p, and mRNAs STMN1, COL15A1, and CCNE2 in hepatocellular carcinoma (HCC) tissues were generally higher than in adjacent normal tissues, indicating that these five genes are upregulated in HCC tissues. This invention also found a targeted regulatory relationship between FOXD2-AS1 and miR-9-5p. miR-9-5p also showed a targeted regulatory relationship with STMN1, while no targeted regulatory relationship was found between miR-9-5p and COL15A1 or CCNE2. However, COL15A1 showed differential expression among groups based on differentiation grade (p = 0.036), lesion size (p = 0.028), and TNM stage (p = 0.005), and CCNE2 showed differential expression in TNM stage (p = 0.005), thus indicating greater potential as molecular markers. Experimental studies in this invention show that the combined diagnostic efficacy of ceRNA with COL15A1 and CCNE2 for HCC is higher than that of individual genes for HCC, and can serve as a biomarker for HCC diagnosis. Attached Figure Description

[0022] Appendix Figure 1 Construction of the Cox regression model and its diagnostic value analysis: A. 10-fold cross-validation plot, B. Lasso regression coefficient distribution of 16 genes, C. Forest plot corresponding to the multivariate Cox regression model, D. ROC curve to evaluate the effectiveness of the diagnostic risk scoring model, E. Kaplan-Meier risk survival curve of HCC patients;

[0023] Appendix Figure 2 Kaplan-Meier survival analysis and identification of ceRNA arrays: Kaplan-Meier survival curves of 8 genes in AH.HCC patients, Venn plot of I.hsa-miR-9-5p target gene, J. screened ceRNA array;

[0024] Appendix Figure 3 Electrophoresis images of lncRNA and mRNA PCR products (M: DNA marker, 1: β-actin, 2: FOXD2-AS1, 3: STMN1, 4: COL15A1, 5: CCNE2);

[0025] Appendix Figure 4 Sequencing peak diagram: A. β-actin sequencing representative fragment, B. FOXD2-AS1 sequencing representative fragment, C. STMN1 sequencing representative fragment, D. COL15A1 sequencing representative fragment, E. CCNE2 sequencing representative fragment;

[0026] Appendix Figure 5 Amplification curves and standard curves of 4 genes and 1 internal reference gene: A. β-actin, B. FOXD2-AS1, C. STMN1, D. COL15A1, E. CCNE2;

[0027] Appendix Figure 6 qRT-PCR results: A. Amplification and melting curves of FOXD2-AS1; B. Amplification and melting curves of has-miR-9-5p; C. Amplification and melting curves of STMN1; D. Amplification and melting curves of COL15A1; E. Amplification and melting curves of CCNE2;

[0028] Appendix Figure 7 Scatter plot of lncRNA, miRNA and mRNA expression in cancer and adjacent normal tissues: A. FOXD2-AS1, B. has-miR-9-5p, C. STMN1, D. COL15A1, E. CCNE2;

[0029] Appendix Figure 8Schematic diagram of binding sites and bar chart of dual-luciferase detection: A. Dual-luciferase reporter assay results between has-miR-9-5p and FOXD2-AS1; B. Dual-luciferase reporter assay results between has-miR-9-5p and STMN1; C. Dual-luciferase reporter assay results between has-miR-9-5p and COL15A1; D. Dual-luciferase reporter assay results between has-miR-9-5p and CCNE2.

[0030] Appendix Figure 9 ROC curves: A. FOXD2-AS1, B. has-miR-9-5p, C. STMN1, D. COL15A1, E. CCNE2, F. ceRNA combination (FOXD2-AS1 / miR-9-5p / STMN1), G. ceRNA network (FOXD2-AS1 / miR-9-5p / STMN1 / COL15A1 / CCNE2) Detailed Implementation

[0031] The biomarker compositions and applications of the present invention will be further described in detail below with reference to specific embodiments. It should be understood that the following embodiments are merely illustrative and explanatory of the present invention and should not be construed as limiting the scope of protection of the present invention. All technologies implemented based on the above content of the present invention are covered within the scope of protection intended by the present invention. Unless otherwise stated, the raw materials and reagents used in the following embodiments are commercially available products or can be prepared by known methods.

[0032] Example 1: Screening of FOXD2-AS1 / miR-9-5p / STMN1 / COL15A1 / CCNE2 markers

[0033] 1.1 Data Download and Processing

[0034] A total of 377 HCC samples and 50 control samples were downloaded from the TCGA (https: / / portal.gdc.cancer.gov / ) database. Patients with incomplete clinical data, missing gene expression profiles, non-primary hepatocellular carcinoma, or other hepatobiliary tumors or tumors of other organs were excluded. Finally, 366 HCC samples were included in the subsequent analysis. Among them, 239 patients were still alive at the follow-up cutoff date, and 138 patients had died.

[0035] The hepatocellular carcinoma lncRNA, miRNA, and mRNA transcript files downloaded from the TCGA database were analyzed for differential expression using the DESeq2 package in R4.0.2 software ((false discovery rate, FDR) < 0.05, |log(fold change)| > 1), yielding differentially expressed lncRNAs, miRNAs, and mRNAs. DElncRNAs-DEmiRNAs and DEmiRNAs-DEmRNAs interaction pairs were searched in the miRcode

[10] and starBase

[11] databases (hypergeometric distribution test p < 0.05, correlation test p < 0.05). A ceRNA network was then constructed using Cytoscape 3.7.2 software. GO annotation analysis and KEGG pathway analysis were performed on the mRNAs in the ceRNA network in the DAVID database to elucidate the role of mRNAs in the ceRNA network; p < 0.01 was statistically significant. Univariate Cox regression and lasso regression analyses were performed on the ceRNA network to prevent overfitting. Based on this, multivariate Cox regression analysis was performed to construct a risk scoring model, and Kaplan-Meier survival curves and ROC curves were used to evaluate the diagnostic value of the model. Batch Kaplan-Meier survival analysis was performed on each gene in the ceRNA network to assess the risk of each gene. The intersection of DEmRNAs matched by key miRNAs in the starBase database with target genes matched in the TargetScan, miRDB, and PicTar databases was obtained.

[0036] 1.2 Screening of ceRNA arrays

[0037] Nine genes were screened using univariate Cox regression and lasso regression analysis. Figure 1 Further multivariate Cox regression analysis (A, B) revealed that the six genes CCNE1, EZH2, CTHRC1, CBX2, COL15A1, and miR-9-5p were optimal for constructing the risk scoring model. Figure 1 C). Survival curves show that high-risk patients have significantly shorter survival times. Figure 1 E). The areas under the curve (AUC) of the ROC curve for 1-, 3-, and 5-year survival were 0.77, 0.73, and 0.764, respectively, indicating that the model has high accuracy in predicting the survival of HCC patients. Figure 1 D).

[0038] Kaplan-Meier survival analysis was performed on all genes in the network in batches, and one miRNA, miR-9-5p, and one lncRNA, FOXD2-AS1, were identified as being associated with HCC prognosis. Figure 2 A, B). To make the mRNAs targeting miR-9-5p more accurate, the intersection of the predicted StarBase (7 target genes), TargetScan (1292 target genes), miRDB (1236 target genes), and PicTar (642 target genes) was taken. The results showed that COL15A1, CCNE2, and STMN1 all have targeting relationships with miR-9-5p in all four databases. Figure 2 I). Constructing a ceRNA network diagram for ceRNA-related genes ( Figure 2 J). In summary, the FOXD2-AS1 / miR-9-5p / STMN1 / COL15A1 / CCNE2 axis was established as a ceRNA array related to HCC diagnosis and prognosis from the TCGA database.

[0039] 1.3 Relationship between ceRNA arrays and clinical features of HCC

[0040] The chi-square test was used to analyze the relationship between the expression levels of FOXD2-AS1, miR-9-5p, STMN1, COL15A1, and CCNE2 and age, sex, grade, TNM stage, and stage in clinical data downloaded from the TCGA database. The results showed that miR-9-5p was significantly different from TNM-T stage (p = 0.040) and stage (p = 0.024), STMN1 from grade (p = 0.001), COL15A1 from TNM-T stage (p = 0.000) and stage (p = 0.003), and CCNE2 from age (p = 0.009) and grade (p = 0.000). All differences were considered statistically significant with p < 0.05. No correlation was found between any of the genes and other clinical information (p > 0.05). Detailed results are shown in the table below.

[0041] Table 1. Relationship between FOXD2-AS1, Has-miR-9-5p, STMN1, COL15A1, CCNE2 and clinical characteristics of liver cancer patients

[0042]

[0043]

[0044] Table 2

[0045]

[0046] 1.4 Conclusion

[0047] This invention constructed a ceRNA network in HCC based on the TCGA database, ultimately identifying one lncRNA (FOXD2-AS1), one miRNA (miR-9-5P), and three mRNAs (STMN1, COL15A1, and CCNE2) with the potential to serve as biomarkers for HCC diagnosis and prognosis. Using this ceRNA distribution-based network as the research object not only narrows down the range of target ncRNAs but also provides specific candidate molecular biomarkers for assessing the prognosis of HCC, which will help expand our understanding of the ceRNA mechanisms involved in the early development of HCC. However, the expression trends of these five genes in clinical HCC tissues and their regulatory relationships require further verification.

[0048] Example 2: Diagnostic value of 5 key genes in liver cancer tissue

[0049] 2.1 Tissue microarray of experimental subjects:

[0050] HCC lncRNA, miRNA, and mRNA tissue cDNA microarrays were all purchased from Shanghai Chipover Biotechnology Co., Ltd. The microarray batch numbers were cDNA-HLivH30PG02 and cDNA-HLivH087Su02. The two microarrays contained a total of 96 cancer samples (experimental group) and 51 adjacent normal tissue samples (control group).

[0051] 2.2 Experimental Reagents and Instruments

[0052] Table 3. Experimental Reagents

[0053]

[0054] Table 4. Experimental Instruments

[0055]

[0056] 2.3 Primer purity of key genes in the ceRNA network

[0057] Primers for four genes (FOXD2-AS1, STMN1, COL15A1, and CCNE2) and a β-actin internal reference gene were designed using BLAST software. These primers were then used to amplify cDNA extracted from HCC tissues via conventional PCR. The PCR products of these four genes and the internal reference gene were then subjected to horizontal nucleic acid electrophoresis at an isobaric voltage of 120V, with RNase-free ddH2O used as a negative control. The PCR products of these four genes and the internal reference gene were then sent to Shanghai Sangon Biotech Co., Ltd. Sangon Biotech performed the sequencing, and the sequencing results were compared with primer sequences designed from BLAST software. Only sequences that matched were used for quantitative real-time PCR (qRT-PCR) detection. The miRNA was reverse transcribed using a tailing method, and the primers were universal; therefore, it was not necessary to verify the primer specificity of miR-9-5p.

[0058] Conclusion: Electrophoretic bands for β-actin (132bp), FOXD2-AS1 (234bp), STMN1 (109bp), COL15A1 (124bp), and CCNE2 (159bp) were obtained, as shown in the figure. Figure 3 As shown in the figure. The results indicate that the electrophoretic bands corresponding to the products amplified by the above genes are single and clear, indicating that the validated primers have high purity and can be used for subsequent qRT-PCR experiments.

[0059] Sequencing of the conventional PCR amplification products and alignment of the sequencing results with the primer sequences corresponding to each gene designed from BLAST software revealed that the alignment results of the primers for β-actin, FOXD2-AS1, STMN1, COL15A1, and CCNE2 were consistent, indicating that the primers have high amplification specificity and can be used for subsequent experiments. Figure 4 The sequencing results are for the above four key genes and β-actin.

[0060] Table 5. PCR primers for 4 key genes and 1 internal control

[0061]

[0062] Table 6. Conventional PCR Reaction System

[0063]

[0064] Table 7. Standard PCR Reaction Procedure

[0065]

[0066] 2.4 Detection of amplification efficiency of key gene primers

[0067] The conventional PCR amplification products were diluted 10,000-fold to serve as the stock solution for qRT-PCR. This stock solution was then serially diluted 10-fold to serve as cDNA for qRT-PCR. Standard curves were plotted for the qRT-PCR results of FOXD2-AS1, STMN1, COL15A1, CCNE2, and β-actin genes, and the amplification efficiency of the corresponding primers was analyzed (qRT-PCR was performed according to Perfect Start). TM(Follow the instructions for use of the Green qPCR SuperMix kit).

[0068] Conclusion: Figure 5 As shown, the amplification curves all exhibited an S-shape. The E values ​​corresponding to the standard curves of β-actin, FOXD2-AS, STMN1, COL15A1, and CCNE2 were 123.5%, 122.9%, 95.7%, 123.7%, and 96.0%, respectively. 2 The values ​​were 0.868, 0.957, 0.994, 0.937, and 0.991, respectively, indicating that the E values ​​were all around 100.0%. 2 The values ​​are all close to 1.000, indicating good amplification efficiency.

[0069] 2.5 qRT-PCR detection

[0070] Strictly follow PerfectStart TM Following the instructions of the Green qPCR SuperMix kit, qRT-PCR was performed on FOXD2-AS1, STMN1, COL15A1, and CCNE2 in HCC tissue cDNA microarrays, with β-actin as an internal control for homogenization. qRT-PCR was performed on miR-9-5p according to the miRcute Plus miRNA qPCR Detection Kit (SYBR Green) instructions, with U6 as an internal control for homogenization. RNase-free ddH2O was used as a negative control, and the qRT-PCR instrument was set to template-free reference (NTC).

[0071] Real-time PCR results showed that the amplification curves of FOXD2-AS1, miR-9-5p, STMN1, COL15A1, and CCNE2 genes were S-shaped, and no amplification was observed in the negative control wells. The Tm values ​​of the melting curves all showed a single peak. Figure 6 Therefore, the Ct value obtained from the intersection of the amplification curve and the threshold line can be used for subsequent statistical analysis.

[0072] The Ct values ​​obtained from qRT-PCR of five key genes in the cancer group and adjacent normal group were used as 2. -△△CtAn algorithm was used to calculate the relative expression levels between the two groups. The Ct values ​​for FOXD2-AS1 in the cancer group and the adjacent normal group were 27.65±1.92 and 29.55±1.78, respectively; for STMN1, the Ct values ​​were 26.93±1.91 and 29.96±1.18, respectively; for COL15A1, the Ct values ​​were 26.01±2.19 and 29.99±1.45, respectively; and for CCNE2, the Ct value was 26.6... The Ct values ​​of lncRNA and mRNA were 1±2.15 and 29.85±1.68, respectively, with β-actin as an internal control for homogenization. The Ct values ​​of miR-9-5p in the cancer group and adjacent normal group were 23.25±1.84 and 24.19±1.30, respectively. The miRNA was homogenized with U6 as an internal control. Furthermore, NTC was not amplified or all Ct values ​​were much smaller than the Ct value of NTC, indicating that the Ct value of the target gene can be used for analysis. Independent samples t-tests were used to compare the differences in gene expression between the two groups. It was found that the expression levels of FOXD2-AS1 (p=0.016), miR-9-5p (p=0.042), STMN1 (p=0.000), COL15A1 (p=0.000), and CCNE2 (p=0.000) were significantly different between the 96 cancer cases and the 51 adjacent normal cases, with corresponding p values ​​less than 0.05. The specific results of the comparison of relative expression levels between the two groups are shown in the table below.

[0073] Table 8. Relative expression levels of key genes in ceRNA network in HCC tissues

[0074]

[0075]

[0076] By plotting a scatter plot of the relative expression levels between the cancer group and the adjacent normal group ( Figure 7 To more intuitively demonstrate the differences in the expression levels of each gene between the two groups, it was found that the expression levels of STMN1, COL15A1, and CCNE2 in the liver cancer group were significantly higher than those in the adjacent normal group, while the expression levels of FOXD2-AS1 and miR-9-5p in the liver cancer group were slightly higher than those in the adjacent normal group, but the differences were still statistically significant.

[0077] One-way ANOVA was performed to compare the expression levels of FOXD2-AS1, miR-9-5p, STMN1, COL15A1, and CCNE2 in HCC tissues with HCC patients' sex, age, differentiation degree, lesion size, and TNM stage to determine whether there were inter-group differences in gene expression levels across different clinical characteristics. As shown in the table below, the results revealed that FOXD2-AS1 showed differential expression across different differentiation degrees (p = 0.021), while no significant differences were observed across the other four clinical characteristics. COL15A1 showed differential expression across different groups in differentiation degree (p = 0.036), lesion size (p = 0.028), and TNM stage (p = 0.005). CCNE2 showed differential expression in TNM stage (p = 0.005) but was not correlated with other clinical factors. The expression of miR-9-5p and STMN1 showed no correlation with clinical characteristics.

[0078] Table 9. Relationship between relative expression levels of five key genes and clinicopathological parameters in HCC tissues

[0079]

[0080]

[0081] Dual-luciferase reporter assay for ceRNA targeting relationship

[0082] The binding sequence fragments of microRNA response elements (MREs) acting on the miR-9-5p seed region of FOXD2-AS1, STMN1, COL15A1, and CCNE2 were predicted using the miRBase, TargetScan, miRanda, PicTar, and MicroT online databases. The 3′-UTRs of one lncRNA and three target genes, as well as the 3′-UTR-miR-9-5p binding site mutant, were cloned downstream of the Luciferase fragment in the pMiR Report vector. The constructed vector was then sequenced to confirm that the target sequence had been cloned to the target location on the vector. The vector and a chemically synthesized miR-9-5p fragment (with NC minics as a control) were then co-transfected into 293T cells, dividing the cells into four groups: NC minics + lncRNA / target gene-wt group, miR-9-5p + lncRNA / target gene-wt group, NC minics + lncRNA / target gene-mu group, and miR-9-5p + lncRNA / target gene-mu group. Experiments were performed according to the instructions of the dual-luciferase reporter gene assay kit. Since FOXD2-AS1 has two binding sites for miR-9-5p, the base fragments from these two binding sites were tandemly constructed to create the FOXD2-AS1 gene 3′-UTR and 3′-UTR mutant, which were then cloned into the pMiR Report vector for analysis.

[0083] Dual-luciferase reporter gene results ( Figure 8 The results showed that FOXD2-AS1 and miR-9-5p, and miR-9-5p and STMN1, have target-binding relationships, while miR-9-5p and COL15A1, and miR-9-5p and CCNE2, have no binding relationships. Therefore, the ceRNA combination that can be differentially expressed in HCC tissues and has target-regulatory relationships is FOXD2-AS1 / miR-9-5p / STMN1. COL15A1 and CCNE2 participate in the miR-9-5p-centered ceRNA network, but have no direct target relationship.

[0084] Analysis of the diagnostic efficacy of ceRNA combinations in HCC tissues

[0085] Finally, based on the results of the previous experiment, ceRNA combinations that are relevant to the survival and prognosis of HCC patients and have a targeted regulatory relationship with each other were determined. Based on the relative expression levels of genes in HCC tissues, ROC curves were plotted for all the key genes screened above to observe the diagnostic accuracy of each gene in HCC tissues. Binary logistic regression was used to calculate the combined predictive factor of the ceRNA combination (FOXD2-AS1 / miR-9-5p / STMN1) and the ceRNA network (FOXD2-AS1 / miR-9-5p / STMN1 / COL15A1 / CCNE2), and ROC curves were plotted for each. The differences in diagnostic efficacy between individual genes and ceRNA combinations, as well as the combined ceRNA network, were evaluated based on the AUC values.

[0086] This study explored the value of FOXD2-AS1, miR-9-5p, STMN1, COL15A1, and CCNE2 as potential biomarkers in the diagnosis and prognosis of HCC by plotting ROC curves. The diagnostic capabilities of ceRNA combinations (FOXD2-AS1 / miR-9-5p / STMN1) and ceRNA networks (FOXD2-AS1 / miR-9-5p / STMN1 / COL15A1 / CCNE2) were analyzed. Figure 9 As shown, the area under the ROC curve (AUC) for FOXD2-AS1 was 0.687 (95% CI: 0.589–0.785, p = 0.002), the AUC for miR-9-5p was 0.633 (95% CI: 0.527–0.739, p = 0.025), the AUC for STMN1 was 0.741 (95% CI: 0.642–0.839, p = 0.000), the AUC for COL15A1 was 0.824 (95% CI: 0.744–0.903, p = 0.000), and the AUC for CCNE2 was 0.835 (95% CI: 0.744–0.926, p = 0.000). All of the above analyses were statistically significant. Binary logistic regression analysis was used to determine the joint predictive factors of the ceRNA combination and its combination with two genes, and ROC curves were plotted. The AUC values ​​were 0.818 (95% CI: 0.732–0.905, p = 0.000) and 0.926 (95% CI: 0.867–0.985, p = 0.000), respectively. This analysis indicates that the AUC value of the ROC curve for the ceRNA combination (FOXD2-AS1 / miR-9-5p / STMN1) is larger than that of its individual components, indicating higher diagnostic efficacy. Furthermore, the AUC value of the ROC curve for the ceRNA combination with two genes is also larger than the AUC value for each gene alone, suggesting even higher diagnostic efficacy.

[0087] The sensitivity and specificity corresponding to the ROC curves of FOXD2-AS1, miR-9-5p, STMN1, COL15A1, CCNE2, and c ceRNA combinations (FOXD2-AS1 / miR-9-5p / STMN1) and ceRNA networks (FOXD2-AS1 / miR-9-5p / STMN1 / COL15A1 / CCNE2) were calculated using the Youden index. As shown in the table below, the cutoff values ​​of FOXD2-AS1, miR-9-5p, STMN1, COL15A1, CCNE2, ceRNA combination, and its combination with two genes are 10.237, 9.616, 9.442, 8.931, 9.336, 0.751, and 0.768, respectively, with corresponding sensitivities of 62.1%, 62.0%, 65.5%, 65.5%, 75.9%, 70.1%, and 84.4%, and specificities of 66.2%, 64.7%, 70.5%, 83.1%, 83.1%, 79.3%, and 93.1%. In summary, using ceRNA combination with COL15A1 and CCNE2 as biomarkers has higher diagnostic value for HCC patients, increasing the likelihood of a confirmed diagnosis.

[0088] Table 10. Diagnostic value analysis of ceRNA arrays in hepatocellular carcinoma tissues.

[0089]

[0090] The embodiments of the present invention have been described above. However, the present invention is not limited to the above embodiments. Any modifications, equivalent substitutions, improvements, etc., within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A composition containing a biomarker, said biomarker being composed of FOXD2-AS1, miR-9-5p and STMN1 genes.

2. A composition containing a biomarker, said biomarker being composed of FOXD2-AS1, miR-9-5p, STMN1, COL15A1 and CCNE2 genes.

3. A composition for the diagnosis of liver cancer, the composition comprising a detection reagent for a biomarker composed of FOXD2-AS1, miR-9-5p and STMN1 genes, or the biomarker composed of FOXD2-AS1, miR-9-5p, STMN1, COL15A1 and CCNE2 genes.

4. The composition according to claim 3, wherein the detection reagent is a reagent capable of quantitatively analyzing the biomarker based on the nucleic acid level.

5. The composition according to claim 4, wherein the detection reagent is a reagent used for nucleic acid amplification reaction, polymerase chain reaction, reverse transcription polymerase chain reaction, or quantitative real-time polymerase chain reaction.

6. The composition according to claim 4, wherein the detection reagent comprises the nucleic acid sequence of the biomarker and primers for specifically recognizing the nucleic acid sequence fragment.

7. The use of the composition according to any one of claims 3-6 in the preparation of a liver cancer diagnostic agent or a liver cancer diagnostic kit.

Images