CN114277155A - Construction method of postoperative recurrence prediction model of hepatocellular carcinoma patient - Google Patents

Abstract

The invention provides a construction method of a prediction model of postoperative recurrence of a hepatocellular carcinoma patient. The expression and distribution of 8 NKG2D ligands (MICA, MICB and ULBP1-6) in HCC are systematically studied, and the MICA, MICB and ULBP4 are found to be highly expressed in HCC tissues, and ULBP3 possibly plays an important role in HCC progression. Suggesting that high affinity NKG2D-CAR-NK cells can be constructed by targeting MICB in HCC. RFS predictive models based on MICA, ULBP3, and ULBP5 can predict recurrence in patients with liver cancer after hepatectomy, and may have potential clinical value.

Description

Construction method of postoperative recurrence prediction model of hepatocellular carcinoma patient

Technical Field

The invention relates to the technical field of molecular biology and medical diagnosis, in particular to a construction method of a postoperative recurrence prediction model of a hepatocellular carcinoma patient.

Background

Hepatocellular carcinoma is one of the most common malignant tumors worldwide, and its fatality rate and mortality rate are still high. Despite the rapid progress in the treatment of hepatocellular carcinoma in recent years, however, prognostic methods for this fatal disease are still rare. Currently, surgical resection has become the safest treatment for HCC due to its low mortality during surgery. However, high recurrence rates remain a major cause of postoperative death, leading to poor overall prognosis for patients with liver cancer. Therefore, finding new markers for predicting tumor recurrence can guide us in clinical practice to risk stratification, individualized monitoring and targeted intervention of patients.

NK cells are important effector cells in the innate immune system, performing immune clearance and immune surveillance functions, and are the first line of defense of the body against invading pathogens and tumor cells. NK cells are highly enriched in the liver, account for 30% -50% of intrahepatic lymphocytes and account for 2-5 times of peripheral blood. NK cells in the liver have higher cytotoxicity and cytokine-producing ability than NK cells in peripheral blood, and have unique immunophenotyping and functional characteristics in part. NKG2D is a C-type lectin receptor, an important activating receptor of NK cells, and can recognize a series of ligands with different affinities and diverse structures, including MHC-class I molecule-related A/B (MICA/B) and 6 UL16 binding proteins (ULBP 1-6). NKG2DL is not normally expressed in normal cells, and is often expressed in virus-infected cells, and tumor cells such as liver cancer, pancreatic cancer, colon cancer, lung cancer, melanoma, and neuroblastoma. These ligands all bind to NKG2D, activate downstream cascade signaling, and play an important role in the anti-tumor immune response.

It has been shown that NKG2DL is expressed in liver cancer cells and correlates with prognosis. The 2012 study of Kamimura et al found that ULBP1 is mainly expressed in Dysplastic Nodules (DN), liver cancer cells of highly-differentiated and moderately-differentiated HCC, but not in cancer cells of less-differentiated HCC; and HCC patients with negative ULBP1 expression had significantly shorter relapse-free survival times than HCC patients with positive ULBP1 expression. In 2021, Cadoux et al found that ULBP1 and other ligands MICA, MICB and ULBP2 are highly expressed in liver cancer, the expression of the ligands is positively correlated with the grading and low differentiation (low levels of differentiation) of tumors, and patients with high expression of MICA, MICB, ULBP1 and ULBP2 have poor prognosis and are easy to relapse in early stage. This suggests that the relationship between NKG2D ligand expression in liver cancer and early relapse in patients remains controversial.

Disclosure of Invention

The invention aims to discuss the expression of NKG2D ligands such as MICA, MICB, ULBP1-6 and the like in hepatocellular carcinoma, evaluate the relation with prognosis and clinical factors, and further construct a model for predicting early recurrence of hepatocellular carcinoma based on the NKG2D ligand.

To achieve the object of the present invention, in a first aspect, the present invention provides a biomarker for predicting postoperative recurrence in a hepatocellular carcinoma patient, which is a relative expression amount of MICA and/or ULBP3 genes in the hepatocellular carcinoma patient; MICA gene expression level and postoperative recurrence probability are in positive correlation, and ULBP3 gene expression level and postoperative recurrence probability are in negative correlation.

In a second aspect, the invention provides application of a reagent for detecting the expression level of MICA, ULBP3 and/or ULBP5 genes in preparing a reagent for evaluating postoperative recurrence of hepatocellular carcinoma patients.

Preferably, the reagent for detecting the MICA gene expression level comprises a primer shown as SEQ ID NO. 1-2, the reagent for detecting the ULBP3 gene expression level comprises a primer shown as SEQ ID NO. 3-4, and the reagent for detecting the ULBP5 gene expression level comprises a primer shown as SEQ ID NO. 5-6.

In a third aspect, the invention provides a construction method of a prediction model of postoperative recurrence of a hepatocellular carcinoma patient, which comprises the steps of respectively determining the relative expression amounts of MICA, ULBP3 and ULBP5 genes in the hepatocellular carcinoma patient based on a QRT-PCR method, substituting the relative expression amounts into a formula (1) for risk assessment, and if the score is more than or equal to 0.951, determining that the postoperative recurrence risk of the hepatocellular carcinoma patient is high, and if the score is less than 0.951, determining that the postoperative recurrence risk of the hepatocellular carcinoma patient is low;

0.29795X [ MICA ] -0.79165X [ ULBP3] + 1.28571X [ ULBP5] formula (1)

Wherein [ MICA ] is the relative expression level of MICA gene;

[ ULBP3] is the relative expression level of ULBP3 gene;

[ ULBP5] is the relative expression level of ULBP5 gene.

In the present invention, the MICA, MICB and ULBP1-6 genes have reference sequence numbers of CCDS56412.1, CCDS75423.1, CCDS5223.1, CCDS5222.1, CCDS5225.1, CCDS59044.1, CCDS43514.1 and CCDS5224.1 on NCBI, respectively.

In a fourth aspect, the present invention provides the use of MICA, ULBP3 and/or ULBP5 genes for predicting the risk of postoperative recurrence in a hepatocellular carcinoma patient.

The present invention included 374 patients with liver cancer from the TCGA database into subjects and randomly divided into training and validation sets at a 7:3 ratio. Multifactorial Cox regression models screen for NKG2D ligands that could be used to predict RFS (relapse free survival). Predictive efficacy is assessed by the area under the Receiver Operating Characteristic (ROC) curve. At the same time, OEP000321 data sets and HCC patients in our Guilin cohort were included to verify the accuracy of this predictive model. Finally, the expression of NKG2D ligand was verified by immunohistochemical staining on a tissue chip containing 137 patients with liver cancer.

Results MICA, ULBP3 and ULBP5 were found to be significantly associated with relapse in the training set of TCGA, and from this, RFS predictive models of multifactorial Cox regression were constructed with areas under the 1-year and 3-year ROC curves of 0.69 and 0.68, respectively. At the same time, a risk score is constructed using the regression coefficients for each variable for classifying patients into low-risk and high-risk groups. In the TCGA liver cancer dataset and OEP000321 dataset, the high risk group had significant differences from recurrence compared to the low risk group, further demonstrating the accuracy of this prediction model. By analyzing clinical characteristics, the risk score of a patient increases with increasing stage in the TNM stage. Immune infiltration assessment was performed on each patient using CIBERSORT and significant differences were found in the degree of infiltration of M1, M2 and Treg cells in the high risk group and the low risk group. Most importantly, in TCGA liver cancer data sets and tissue chips, we found that ULBP3 expression was significantly higher in the paracarcinoma of liver cancer patients than in cancer tissues, and that patients with high ULBP3 expression had longer relapse-free survival.

The expression and distribution of 8 NKG2D ligands (MICA, MICB and ULBP1-6) in HCC are systematically studied, and the MICA, MICB and ULBP4 are found to be highly expressed in HCC tissues, and ULBP3 possibly plays an important role in HCC progression. Suggesting that high affinity NKG2D-CAR-NK cells can be constructed by targeting MICB in HCC. RFS predictive models based on MICA, ULBP3, and ULBP5 can predict recurrence in patients with liver cancer after hepatectomy, and may have potential clinical value.

Drawings

FIG. 1 is a flow chart of an experiment according to a preferred embodiment of the present invention.

FIG. 2 is a graph showing the expression of NKG2D ligand in liver cancer tissue in a preferred embodiment of the present invention. Wherein, the expression level of NKG2D ligand mRNA in the A.TCGA liver cancer database; B. analyzing the distribution and expression frequency of NKG2D ligand in Guilin liver cancer queue through immunohistochemistry (negative expression; +, weak expression; +, medium expression; + +, high expression); C. statistically analyzing the expression of NKG2D ligand in Guilin liver cancer queue in paracarcinoma and cancer tissues; p <0.01, P <0.001, P < 0.0001.

FIG. 3 is a graph showing the recurrence-free survival (RFS) of NKG2D ligand in liver cancer in accordance with a preferred embodiment of the present invention. Wherein, the A.NKG2D ligand is subjected to RFS analysis in a TCGA liver cancer queue. RFS analysis of NKG2D ligand in Guilin liver cancer cohort.

FIG. 4 shows the screening of NKG2D ligand associated with HCC recurrence in TCGA liver cancer cohort by multifactorial Cox regression analysis in accordance with a preferred embodiment of the present invention.

FIG. 5 is a graph showing the predicted value of 3-NKG2D ligand signature in the training set and validation set in accordance with the preferred embodiment of the present invention. Wherein, the A.TCGA liver cancer data set comprises the survival curves of a training group and a verification group RFS. tcga-LIHC dataset prediction ROC curves for training and validation groups for 1 year and 3 year RFS. Calibration curves for the training and validation groups of the tcga-LIHC data set. OEP000321 data set RFS curves for 3-NKG2D ligand signature. E. RFS curves of 3-NKG2D ligand signature in Guilin cohort.

FIG. 6 is a clinical correlation analysis of 3-NKG2D ligand signature in a preferred embodiment of the invention. Correlation of 3-NKG2D ligand signature with clinical pathological features in A.TCGA-LIHC data set. Correlation between 3-NKG2D ligand signature and clinical pathology features in oep000321 dataset. And C, carrying out immune cell infiltration analysis on low-risk and high-risk populations by a CIBERSORT method in the TCGA-LIHC data set. Oep000321 data set immune cell infiltration analysis was performed by CIBERSORT method on low risk and high risk populations.

FIG. 7 is a graph showing the correlation between NKG2D and its ligand in the preferred embodiment of the present invention. Wherein, the correlation of the NKG2D and the ligand thereof in the TCGA-LIHC queue. Oep000321 data set the correlation of NKG2D and its ligand. C. Expression of ULBP3 mRNA in Guilin liver cancer cohort. D. ULBP3RFS curve based on mRNA expression level. E. Immunohistochemistry for NKG2D ligand in two liver cancer samples.

FIG. 8 is a graph showing the correlation between ULBP3 and immune cell infiltration in hepatocarcinoma tumor tissue in accordance with a preferred embodiment of the present invention.

Detailed Description

The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention. Unless otherwise indicated, the examples follow conventional experimental conditions, such as the Molecular Cloning handbook, Sambrook et al (Sambrook J & Russell DW, Molecular Cloning: a Laboratory Manual,2001), or the conditions as recommended by the manufacturer's instructions.

Example 1

1. Test materials and methods

1.1 data Collection

By 21/2/2021, a total of 374 HCC transcriptome sequencing data and corresponding clinical data downloaded from the TCGA database (TCGA-LIHC) were used as training cohorts (https:// portal.gdc.cancer. gov. /). The OEP000321 dataset downloaded from NODE contained a validation cohort of transcriptome sequencing data and corresponding clinical data for 159 HCC patients (https:// www.biosino.org/NODE). In addition, some liver cancer samples from the affiliated hospital of Guilin medical college (Guilin cohort) from 1/2006 to 31/2016/12/2016 were randomly collected. QRT-PCR was performed using these samples, tissue chips were constructed and immunohistochemistry was performed. Each participant signed a written informed consent. The study was approved by the ethical committee of the national hospital of Guilin medical college, Beijing university, and was in accordance with the declaration of Helsinki principles.

1.2 construction and validation of NKG2D ligand-based prediction models

NKG2D ligand genes (MICA, MICB, ULBP1, ULBP2, ULBP3, ULBP4, ULBP5 and ULBP6) with prognostic value are screened by adopting multifactor Cox regression in a training set by taking a relapse-free survival (RFS) as an observation endpoint. The cutoff value for the NKG2D ligand gene was predicted to be P < 0.05. To further evaluate the performance of NKG2D ligand gene signature, the patients in the training and validation sets were divided into low and high risk groups according to median value. An RFS curve based on a Kaplan-Meier method is drawn by adopting a surfminerR package, a receiver operating characteristic curve (ROC) (RFS in 3 years and 5 years) is drawn by adopting a timeROCR package, and a calibration curve is drawn by adopting an rmsR package.

1.3 clinical relevance analysis and immune cell infiltration analysis

And analyzing and drawing the correlation between the signature of the NKG2D ligand gene and clinical pathological characteristics by using GraphPadprism 8.0.2 software, and evaluating the correlation between the signature of the NKG2D ligand gene and different clinical pathological factors by adopting independent sample t test. Based on transcriptome data we assessed immune cell infiltration using CIBERSORT (https:// ciberstartx. stanford. edu /).

1.4 tissue chips and immunohistochemistry

From 137 HCC specimens from the affiliated hospital of Guilin medical school and the corresponding tissues adjacent to cancer were prepared as tissue chips, which were then subjected to Immunohistochemical (IHC) staining. The antibodies involved in this study were rabbit anti-human MICA/MICB (Abcam, ab224702, 1:50 dilution), rabbit anti-human ULBP1(Abcam, ab238331, 1:1000 dilution), rabbit anti-human ULBP3(Novus, nbp2-31866, 1:1000 dilution), mouse anti-human ULBP4(Santa cruz, sc-390784,1:50 dilution). The specific experimental steps of immunohistochemistry are consistent with the previous method.

2 pathologists scored according to the following criteria: no yellow or brown is negative, point 0; faint yellow is weak positive, 1 point; yellow or dark yellow is moderate positive, 2 points; brown or tan is strongly positive, 3 points. Positive cell proportion score: the proportion of positive cells is 0 min, 1-10% 1 min and 11-50% 2 min; 51-80% of the total weight is 3 min; >80 is 4 min. The staining score of the sample is the value of staining intensity multiplied by the value of positive cell proportion score. Evaluation criteria for NKG2D ligand expression levels: staining scores scored 0 points (-), 1-3 points (+), 4-6 points (++), and 7-12 points (+++).

1.5 real-time quantitative PCR

mRNA from human liver tissue was extracted using RNeasy MiniKit (Qiagen) and quantified using NanoDrop (thermo scientific). And detecting the mRNA expression levels of MICA, ULBP3 and ULBP5 of Guilin people by using real-time fluorescent quantitative polymerase chain reaction (QRT-PCR). The QRT-PCR method is as described previously. QRT-PCR primer sequences are shown in Table 1. MICA, ULBP3 and ULBP5 relative expression levels and GAPDH relative expression levels are normalized.

TABLE 1 primer sequences

1.6 statistics of data

Continuous variables that fit a normal distribution are expressed as mean ± Standard Deviation (SD) and compared using Student's t test. The categorical variables were compared using the chi-square test or the Fisher's exact test. Survival analysis was performed using the Kaplan-Meier method and the Log-rank test. The multi-factor analysis uses Cox regression analysis. Statistical analysis was performed using SPSS18.0(SPSS Inc., Chicago, IL) and R version 4.0.3(https:// www.rproject.org /), with P <0.05 being statistically significant for differences.

2. Results of the experiment

2.1 differential expression of NKG2D ligand in tumor tissue and in paraneoplastic tissue of hepatocellular carcinoma patients

The patient grouping and subsequent experimental procedures in this study are shown in figure 1. In the TCGA-LIHC database, the mRNA expression levels of MICA, MICB, ULBP1 and ULBP4 in liver cancer tissues are obviously higher than those in normal liver tissues, while the expression level of ULBP3 in tumor tissues is lower than that in normal liver tissues. mRNA expression of ULBP2, ULBP5, ULBP6 was not different between normal tissue and tumor tissue (fig. 2A). In addition, immunohistochemistry was used to detect protein expression of NKG2D ligand in the Guilin cohort organization chip. NKG2D ligand is differentially expressed in the cytoplasm of both tumor cells and normal cells. MICA/B and ULBP4 were expressed as moderate (IHC score 4-6) or high (IHC score 7-12) in HCC tumors above 2/3, while ULBP1, ULBP3 and UBP2/5/6 were expressed as non-moderate (IHC score 0) or low (IHC score 1-3) in HCC tumors above 2/3 (fig. 2B). As expected, the protein expression patterns of MICA/B, ULBP3, ULBP4, and ULBP2/5/6 in paired tumor tissues and adjacent non-tumor tissues were consistent with the TCGA-LIHC database. However, there was no significant difference in ULBP1 protein expression between paired liver cancer tumor tissue and adjacent non-tumor liver tissue in the Guilin cohort (fig. 2C).

2.2 relationship between NKG2D ligand expression and HCC clinicopathological features

To assess the significance of NKG2D ligand expression in HCC, we analyzed the relationship of NKG2D ligand expression levels to patient clinical pathological variables in the Guilin cohort (table 2). MICA/B low expression is associated with tumor volume (P0.029) and PVTT (P0.025). High expression of ULBP1 was associated with cirrhosis (P ═ 0.036). Low expression of ULBP3 correlates with high expression of HBsAg (P ═ 0.022) and smaller tumor sizes (P < 0.001). Low expression of ULBP4 was associated with tumor infiltration and metastasis (P ═ 0.011). Low expression of ULBP2/5/6 was associated with late stage TNM staging (P ═ 0.031). The expression of NKG2D ligand has no significant correlation with sex, age, family history, alcohol consumption, AFP, tumor number or lymphatic invasion.

2.3 correlation of expression of NKG2D ligand with prognosis

Furthermore, we confirmed the significance of NKG2D ligand in HCC prognosis in TCGA-LIHC dataset, OEP000321 dataset (HBV-associated HCC cohort) and Guilin cohort. In the TCGA-LIHC dataset, high expression of ULBP4 was associated with poor relapse-free survival (P0.04178) (fig. 3A). In OEP000321 dataset, high MICA expression was associated with poor relapse-free survival (P. 0.02264). In the Guilin cohort, low expression of ULBP3 was associated with poor relapse-free survival (P ═ 0.0075) (FIG. 3B).

These results suggest that a single NKG2D ligand is insufficient to predict HCC recurrence in different HCC cohorts. Therefore, it is assumed that a prediction model constructed from multiple NKG2D ligand expressions can predict recurrence of patients after liver cancer surgery.

2.4 models based on MICA, ULBP3 and ULBP5 expression can predict postoperative RFS in hepatocellular carcinoma patients

374 HCC patients in the TCGA dataset were randomly divided into a training group and a validation group at a 7:3 ratio. Multifactor Cox regression results showed that MICA (HR ═ 1.347; 95% CI ═ 1.012-1.793; P ═ 0.041), ULBP3(HR ═ 0.453; 95% CI ═ 0.231-0.889; P ═ 0.021) and ULBP5(HR ═ 3.617; 95% CI ═ 1.819-7.194; P <0.001) were associated with relapse-free survival (fig. 4). The 3-NKG2D ligand signature risk score for each HCC patient was then calculated from the coefficients of 3NKG2D ligands as: risk score 0.29795 × [ MICA ] -0.79165 × [ ULBP3] +1.28571 × [ ULBP5 ]. HCC patients were then divided into "low risk" and "high risk" groups according to median value (6.413752597). In the TCGA dataset as training cohort, patients in the high risk group had a shorter relapse-free survival time than those in the lower risk group (fig. 5A). The ROC curves show that the area under the curve (AUC) for 3-year and 5-year RFS in the training set is 0.69 and 0.68, respectively, and the validation set is 0.63 and 0.63 (fig. 5B). The calibration curve shows good agreement between the predicted and observed survival probabilities (fig. 5C). Taken together, these results indicate that 3-NKG2D ligand signature has a better predictive RFS performance.

To further validate the model, we evaluated the predicted effect of 3-NKG2D ligand signature on RFS using OEP000321 dataset (fig. 5D) and Guilin cohort (fig. 5E) as validation cohorts. QRT-PCR was used to detect mRNA expression levels of Guilin group MICA, ULBP3 and ULBP 5. Survival curves based on the Kaplan-Meier method showed significantly poorer RFS in high-risk group HCC patients, both in the validation cohort and in the training cohort. This result indicates that 3-NKG2D ligand signature also has good RFS prediction ability in a separate cohort.

2.53-NKG 2D ligand signature correlation with clinical pathological characteristics

To explore the clinical significance of 3-NKG2D ligand signature in HCC, we analyzed the relationship between 3-NKG2D ligand signature and clinical pathology. In the TCGA-LIHC dataset, high risk patients are more prone to lymph node metastasis than low risk patients: however, there was no relationship between risk scores calculated from signature and gender, viral infection or tumor stage (fig. 6A). In the OEP000321 dataset, the results show that with increasing TNM and BCLC stages, the patient risk score increases. However, there was no significant correlation between this signature and either portal vein cancer emboli, tumor numbers, or cirrhosis (fig. 6B).

Given the important role of NKG2D ligands in tumor immunity, we speculate that 3-NKG2D ligand signature may be related to the immune microenvironment. Therefore, we analyzed immune cell subtypes in HCC samples using CIBERSORT and found that this signature correlates with follicular helper T cells or regulatory T cells in the TCGA-LIHC dataset (fig. 6C). In patients with HBV-associated HCC with OEP000321 dataset, the degree of infiltration of M1 macrophages was lower in the high risk group than in the low risk group. In contrast, M2 macrophages were more infiltrated in the high risk group than in the low risk group (fig. 6D).

2.6 correlation between NKG2D and ligand in hepatocellular carcinoma

First, we investigated the relationship between NKG2D ligands in the HCC cohort. In the TCGA-LIHC cohort, two by two of the 8 NKG2D ligands were highly correlated (FIG. 7A). In OEP000321 dataset there was a positive correlation between MICA and MICB, ULBP1 and ULBP2, ULBP1 and ULBP3, ULBP2 and ULBP3, ULBP2 and ULBP5, ULBP4 and ULBP 5. In the Guilin cohort, there was a clear correlation between NKG2D ligands, except for ULBP4 and MICA/B, ULBP4 and ULBP2/5/6 (FIG. 7B).

To investigate the correlation between NKG2D and its ligand, we further evaluated their expression in the TCGA-LIHC dataset and OEP000321 dataset. Correlation analysis showed that NKG2D expression was positively correlated with either ULBP3 or ULBP2 in the TCGA-LIHC dataset (fig. 7A). Interestingly, MICB expression was positively correlated with NKG2D expression in both sets of data (fig. 7A and 7B).

2.7ULBP3 may play an important role in the progression of hepatocellular carcinoma

By comprehensively analyzing the correlation between NKG2D ligands, we found that ULBP3 has significant correlation with other ligands (including NKG2D) in three datasets. Furthermore, protein (fig. 2C) and mRNA expression of ULBP3 in HCC tumor tissues (fig. 7C) was much lower than in paraneoplastic tissues in the TCGA and Guilin cohorts, and RFS were longer in patients with higher ULBP3 expression (fig. 3B and fig. 7D). Suggesting that ULBP3 may affect the progression of HCC.

The TIMER2 tool was used to analyze the relationship between ULBP3 expression and immune infiltration in the TCGA-LIHC database. The results showed that expression of ULBP3 was negatively correlated with the purity of tumor cells in tumor tissues, suggesting that ULBP3 expression was low in tumor cells, consistent with the above studies. At the same time, expression of ULBP3 was associated with NK cells, CD8⁺Infiltration levels of T cells and DC cells were positively correlated (fig. 8), suggesting that ULBP3 may be involved in tumor-mediated immune escape in liver cancer.

3. Discussion of the related Art

Hepatocellular carcinoma has clear etiology and risk factors, such as Hepatitis B Virus (HBV) infection, Hepatitis C Virus (HCV) infection, alcohol (ethanol), non-alcoholic fatty liver disease (NAFLD), carcinogen exposure, diabetes, and the like. Expression of NKG2D ligand is an indicator of cellular stress and may be induced by viral infection or malignant transformation. NKG2D ligand has been shown to be expressed in a number of malignancies, particularly hepatocellular carcinoma. At present, NKG2D ligand becomes a key target for cancer treatment, so that the determination of the expression and distribution of NKG2D ligand in cancer is particularly important. The invention comprehensively discusses the expression of 8 NKG2D ligands in liver cancer tumor tissues, and researches show that the expression of MICA and MICB in TCGA and Guilin queues is obviously higher than that of tissues beside the cancer, and the results are consistent with the previous researches. In addition, the research also finds that ULBP4 has obvious differential expression in liver cancer tumor tissues and paracarcinoma tissues. Compared with other ligands, MICA/B and ULBP4 are moderately expressed or highly expressed in liver cancer tissues higher than 2/3, which provides a theoretical basis for NKG2D ligand as a therapeutic target.

The main treatment mode of hepatocellular carcinoma is surgical treatment, but 50-70% of patients with hepatocellular carcinoma have relapse after the postoperative treatment, and finally the overall prognosis of the patients with hepatocellular carcinoma is poor. Currently, AFP has been shown to be associated with the development and progression of hepatocellular carcinoma and is used as an indicator of diagnosis and prognosis. However, about 30% -40% of liver cancer patients have consistently negative serum AFP, which means that this fraction of patients lacks appropriate prognostic indicators. Therefore, there is a great need to find new biomarkers to predict recurrence after liver cancer surgery. Previous studies have demonstrated a relationship between single NKG2D ligand expression and prognosis, but these prognostic values remain controversial in different cohorts. The invention discovers that a plurality of NKG2D ligands can be simultaneously expressed in the same cancer tissue of the liver cancer, so that the accuracy of constructing the liver cancer recurrence model based on a plurality of NKG2D ligands is probably higher. Subsequently, MICA, ULBP3 and ULBP5 were screened by multifactorial Cox regression models for association with recurrence in patients with liver cancer, and signature was constructed. We also verified the reliability of this predictive model in the TCGA-LIHC dataset, the OEP000321 dataset, and the Guilin dataset. Further, by analyzing the relationship between signature and clinical characteristics, the risk value based on signature is found to be not significantly different between virus-infected liver cancer patients and non-virus-infected liver cancer patients, which suggests that the prediction model is not affected by virus infection. However, we found that high risk values based on signature were significantly associated with lymph node metastasis, BCLC staging and TNM staging, suggesting that liver cancer patients with high risk values have higher tumor malignancy.

Since NKG2D ligand triggers anti-tumor immunity by activating NKG 2D-expressing immune cells, we performed an immune infiltration assessment using the CIBERSORT tool on each patient in the TCGA and OEP000321 datasets, finding that NK cells expressing NKG2D do not differ in both the high risk and low risk groups, which are differentiated based on regression coefficients. However, in the HBV-associated liver cancer cohort (OEP000321), M2 was significantly higher in the high risk group than in the low risk group, whereas M1 was the opposite. This suggests that the immune response elicited by NKG2D ligands in different liver cancer cohorts may be influenced by a variety of factors, requiring further experimental validation.

Previous studies have focused primarily on the mechanisms and roles of MICA, MICB, ULBP1, and ULBP2 in HCC, while less have been studied for other NKG2D ligands. In this study, we found that ULBP3 levels were significantly correlated with other NKG2D ligands in three datasets, suggesting that ULBP3 may play an important role in the progression of HCC by modulating the expression of other NKG2D ligands. Furthermore, expression of ULBP3 was lower in HCC tumor tissues than in non-tumor tissues, and HBsAg levels were high and RFS was short in HCC patients with low expression of ULBP 3. Notably, expression of ULBP3 was positively correlated with NK cell numbers. HBsAg has been shown to inhibit expression of MICA and MICB in HCC cells by inducing cellular miRNAs. The HBsAg is suggested to possibly inhibit the expression of the liver cancer cell ULBP3, and provides an opportunity for tumor immune escape. However, this view requires further investigation to confirm.

NK cells have been shown to mediate the oncolytic effect of liver cancer by binding to ligands through the NKG2D receptor. Namely, the liver cancer cells highly express MICA, MICB and ULBP, trigger NK cells to recognize tumor cells through NKG2D, release various cytokines, secrete granzyme, express suicide related factor ligand (FasL) and tumor necrosis factor related apoptosis inducing ligand (TRAIL), dissolve target cells and play a role in resisting HCC in multiple ways. However, as tumor cells progress, it can significantly down-regulate the expression of NKG2D receptor on NK cells and CD8+ T cells. Therefore, decreased expression of NK cell NKG2D may be one of the important mechanisms of NK dysfunction in tumor patients. Therefore, it has become a core concept of immunotherapy research to maximally activate the expression of NKG2D on immune cells and fully induce the expression of NKG2DL on tumor cells. Currently, researchers have constructed NKG2D-CAR-NK cells based on Chimeric Antigen Receptor (CAR) technology using NKG2D as a signaling molecule that can signal to mediate cytotoxicity intracellularly via CD3 ζ and DAP 10. Compared with NK cells of endogenous NKG2D receptors, the NKG 2D-CAR-induced NK cells can remarkably enhance the activity of the NK cells on the whole, and have effective killing effect on various tumor cells expressing NKG2 DL. According to the invention, high expression of MICB in liver cancer cells is found in both a TCGA database and an OEP000321 database, and the expression of the MICB is obviously and positively correlated with the expression of NKG 2D. This suggests that the construction of high affinity NKG2D-CAR-NK cells against MICB would be expected to be an important means for anti-liver cancer therapy.

In conclusion, the expression and distribution of 8 NKG2D ligands in HCC are researched by the system, and the MICA, the MICB and the ULBP4 are found to be highly expressed in HCC tissues, and the ULBP3 possibly plays an important role in HCC progression. Suggesting that high affinity NKG2D-CAR-NK cells can be constructed by targeting MICB in HCC. RFS predictive models based on MICA, ULBP3, and ULBP5 can predict the recurrence of HCC, potentially with clinical value.

Although the invention has been described in detail hereinabove with respect to a general description and specific embodiments thereof, it will be apparent to those skilled in the art that modifications or improvements may be made thereto based on the invention. Accordingly, such modifications and improvements are intended to be within the scope of the invention as claimed.

Sequence listing

<110> Beijing university Hospital

Construction method of <120> postoperative recurrence prediction model for hepatocellular carcinoma patient

<130> KHP221110523.7

<160> 6

<170> SIPOSequenceListing 1.0

<210> 1

<211> 20

<212> DNA

<213> Artificial Sequence (Artificial Sequence)

<400> 1

gaatccggcg tagtcctgag 20

<210> 2

<211> 20

<212> DNA

<213> Artificial Sequence (Artificial Sequence)

<400> 2

cctgacgcca ggtcagtatg 20

<210> 3

<211> 20

<212> DNA

<213> Artificial Sequence (Artificial Sequence)

<400> 3

tggtctcaat gagagactgc 20

<210> 4

<211> 21

<212> DNA

<213> Artificial Sequence (Artificial Sequence)

<400> 4

tatggctttg ggttgagcta a 21

<210> 5

<211> 21

<212> DNA

<213> Artificial Sequence (Artificial Sequence)

<400> 5

catgtgtctc ctcatatgct c 21

<210> 6

<211> 19

<212> DNA

<213> Artificial Sequence (Artificial Sequence)

<400> 6

taaagtcacg cgagtcacg 19

Claims (4)

1. A biomarker for predicting postoperative recurrence in a hepatocellular carcinoma patient, wherein the biomarker is a relative expression level of MICA and/or ULBP3 genes in the hepatocellular carcinoma patient; MICA gene expression level and postoperative recurrence probability are in positive correlation, and ULBP3 gene expression level and postoperative recurrence probability are in negative correlation.

2. Application of a reagent for detecting MICA, ULBP3 and/or ULBP5 gene expression level in preparing a reagent for evaluating postoperative recurrence of hepatocellular carcinoma patients.

3. The use of claim 2, wherein the reagent for detecting the expression level of MICA gene comprises the primer shown in SEQ ID NO. 1-2, the reagent for detecting the expression level of ULBP3 gene comprises the primer shown in SEQ ID NO. 3-4, and the reagent for detecting the expression level of ULBP5 gene comprises the primer shown in SEQ ID NO. 5-6.

4. The construction method of the model for predicting postoperative recurrence of hepatocellular carcinoma patients is characterized in that relative expression levels of MICA, ULBP3 and ULBP5 genes in hepatocellular carcinoma patients are respectively determined based on a QRT-PCR method, and then are substituted into the formula (1) for risk assessment, if the score is more than or equal to 0.951, the risk of postoperative recurrence of hepatocellular carcinoma patients is high, and if the score is less than 0.951, the risk of postoperative recurrence of hepatocellular carcinoma patients is low;

0.29795X [ MICA ] -0.79165X [ ULBP3] + 1.28571X [ ULBP5] formula (1)

Wherein [ MICA ] is the relative expression level of MICA gene;

[ ULBP3] is the relative expression level of ULBP3 gene;

[ ULBP5] is the relative expression level of ULBP5 gene.

Images