A biomarker for predicting immunotherapy benefit of hepatocellular carcinoma, an RNA model and application thereof

Abstract

The application discloses a biomarker for predicting the benefit of hepatocellular carcinoma immunotherapy, an RNA model and application. The application screens 49 markers related to the efficacy of hepatocellular carcinoma immunotherapy, and uses a machine learning algorithm to establish an RNA model, namely an HIRE score model, for predicting the efficacy of immunotherapy based on a tumor microenvironment. The HIRE score model of the application is constructed according to characteristic genes related to the prognosis of immunotherapy of hepatocellular carcinoma patients, and to some extent, reflects the activation of the tumor microenvironment. In combination with the expression levels of tumor cell characteristics and immune checkpoint genes, the HIRE score model can more comprehensively evaluate the tumor and the tumor microenvironment, and then more accurately screen out potential populations benefiting from immunotherapy, provide more individualized evaluation of the immunological benefit for hepatocellular carcinoma patients, and greatly promote the precise process of hepatocellular carcinoma immunotherapy.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of biological medicine, and particularly relates to a biomarker for predicting immunotherapy benefit of hepatocellular carcinoma, an RNA model and application. BACKGROUND

[0002] Hepatocellular carcinoma (HCC) is a global health problem with increasing incidence and mortality. Although surgical treatment and local treatment are widely used worldwide, it is estimated that about 50-60% of HCC patients will eventually receive systemic treatment. There are many prognostic factors for hepatocellular carcinoma. Patients with hepatocellular carcinoma with the same clinical stage and histological grade may have different prognoses after receiving the same treatment regimen.

[0003] In fact, for more than a decade, systemic molecular targeted therapy has been the mainstay of treatment for advanced HCC; some drugs have been shown to provide significant survival benefits as single agents, with first-line sorafenib and lenvatinib (including the Chinese self-developed drug, domafenib) providing a median overall survival (OS) of 11-14 months, and second-line regorafenib, cabozantinib, and ramucirumab providing a median OS of 8-11 months. In recent years, immunotherapy, mainly immune checkpoint blockades (ICBs), has made breakthrough progress in advanced hepatocellular carcinoma. Compared with sorafenib, ICI atezolizumab (an anti-PD-L1 antibody) and anti-VEGFA antibody bevacizumab produced better results than sorafenib in patients with advanced HCC, setting a new first-line benchmark. The median OS duration in this treatment regimen was 19 months, providing more survival benefits for patients and changing the overall treatment landscape for hepatocellular carcinoma. Similarly, compared with sorafenib, anti-PD-1 antibody sintilimab combined with bevacizumab biosimilar (IBI305) has been reported to improve OS in Chinese patients with advanced hepatitis B virus (HBV)-related HCC. In addition to the approval of atezolizumab-bevacizumab for first-line treatment, the FDA has accelerated the approval of anti-PD-1 antibody pembrolizumab as a single agent and nivolumab (another anti-PD-1 antibody) combined with anti-CTLA4 antibody ipilimumab for second-line treatment of advanced HCC based on the efficacy data of early trials. However, these data have not been confirmed in phase III studies. However, the biggest challenge of this therapy is the wide variation in patient response, with multiple studies showing that only a portion of hepatocellular patients can benefit from this therapy. Therefore, it is necessary to understand the determinants of response and resistance to these different drugs and / or combinations in HCC patients, to find biomarkers that can accurately predict the response to this therapy, and to develop corresponding individualized treatment plans to avoid harm and burden to patients caused by over-treatment and inappropriate treatment.

[0004] According to the TNM staging system, there are significant differences in clinical outcomes among different patients with the same type of tumor and the same stage after the same or similar treatment, which indicates that the staging system has limitations in providing prognostic information and guiding postoperative chemotherapy decisions. Currently, imaging has become the standard technique for HCC diagnosis and follow-up. However, imaging techniques have their limitations, especially in the early detection of HCC. Therefore, there is an urgent need for reliable, non / minimally invasive biomarkers. So far, alpha-fetoprotein (AFP) is the only serum biomarker used in clinical practice for HCC management. However, the specificity and sensitivity of AFP are relatively low. Current research on biomarkers for predicting the efficacy of immune checkpoint inhibitors in advanced hepatocellular carcinoma has mainly focused on the field of liquid biopsy. PD-L1 expression was initially considered a potential predictive biomarker of immunotherapy in phase I trials, and subsequently validated in larger-scale clinical trials, paving the way for immunohistochemical evaluation of PD-L1 in clinical practice for various cancer types, including NCSLC, breast cancer, and advanced esophageal cancer. However, the utility of PD-L1 as a biomarker of immunotherapy response in HCC is unclear.

[0005] A meta-analysis exploring the clinical significance of PD-L1 expression in HCC found that PD-L1 expression was not significantly associated with improved survival or remission rates. In the phase I Checkmate-040 trial, designed to evaluate the efficacy and safety of nivolumab alone or in combination with ipilimumab in treating patients with advanced HCC, the amount of PD-L1 expression did not differ between responders, and both PD-L1 positive and negative patients showed comparable response rates. In addition, a phase 1b study evaluating the efficacy and safety of atezolizumab in combination with bevacizumab retrospectively analyzed PD-L1 expression in 86 tumor samples using immunohistochemical staining, and found that patients' clinical response was not related to the amount of PD-L1 expression. The KEYNOTE-224 study analyzed the expression of PD-L1 in immune and tumor cells in 52 patients using immunohistochemical evaluation, and no correlation between response rate and PD-L1 expression in tumor cells was observed.

[0006] Tumor mutation burden (TMB) is defined as the total number of somatic genetic coding errors, base substitutions, gene insertions, or deletion errors detected per million bases. Tumors with high TMB are associated with higher neoantigen expression, which will be more likely to elicit T lymphocyte recognition to activate immune response. Therefore, TMB-high (TMB-H) status has been explored as a predictive biomarker of response to PD-1 / PD-L1 inhibitors. The KEYNOTE-158 phase 2 study showed that pembrolizumab significantly improved in a variety of advanced TMB-H solid tumors, with an ORR of 29% in TMB-H patients, while the ORR of the TMB-low subgroup was 6%. These findings led to the FDA approval of pembrolizumab for the treatment of unresectable or metastatic TMB-H cancer. However, since TMB-H status is rarely found in HCC, this trial did not include any HCC patients. Evidence so far suggests that TMB-H corresponds to a small number of HCC patients, thus limiting the clinical applicability of TMB as a predictive biomarker for HCC.

[0007] The probability of mutations occurring during DNA synthesis is limited by immediate proofreading of replication errors and post-replication mismatch repair (MMR) mechanisms. Microsatellites are a class of short tandem repeat DNA sequences in the genome, which are inherently more susceptible to impaired DNA repair. Defects in the MMR pathway (dMMR) lead to microsatellite instability (MSI) status. MSI status is considered a potential biomarker of response to ICPi therapy. In 2015, the FDA first accelerated the approval of pembrolizumab for the treatment of any MSI-H or dMMR unresectable tumor type based on the observed response rate in the KEYNOTE-016 phase I study. However, MSI-H is a rare feature in HCC patients, resulting in its limited utility as a predictor of response to ICPi, with only <3% of HCCs having MSI status.

[0008] In summary, the existing biomarkers including the expression level of PD-L1, TMB, and MSI-H, etc. all face challenges in the clinical application of liver cancer, suggesting that there is still a need to develop more precise and comprehensive markers to predict the efficacy of immunotherapy for hepatocellular carcinoma. Developing new robust biomarkers and exploring effective biomarker combination detection strategies may be the key to addressing this challenge.

[0009] With the advancement of technology, the understanding of the complexity and diversity of the tumor microenvironment and its impact on treatment response is also advancing. In-depth analysis of the complexity of the tumor microenvironment is likely to reveal advanced biomarkers that will make a difference in identifying patient populations that respond to current ICB therapy and will facilitate the search for new therapeutic modulation targets. Therefore, predicting the responsiveness of immune checkpoint inhibitors PD-1 / PD-L1 mAbs based on the characteristics of tumor immune infiltration is an important initiative to improve the success rate of current hepatocellular carcinoma immune checkpoint inhibitor therapy and develop the next generation of immunotherapy.

[0010] In order to more accurately and comprehensively predict the efficacy of immunotherapy, some international scholars have predicted the efficacy from the perspective of tumor immune microenvironment (including TIDE, IPS, Pan-F-TBRs, etc. score). Among them, TIDE score integrates the expression characteristics of T cell dysfunction and T cell exclusion to simulate tumor immune escape. TIDE score is trained from untreated tumor data and can predict the clinical response of ICB based on tumor characteristics before treatment. However, in some cancer types such as renal cell carcinoma, there are a large number of CD8+ T cell infiltrations, and higher CTLs may not be associated with survival benefits. In addition, according to the sample size or characteristics of a specific data set, there may not be any statistically significant genes interacting with CTLs to affect survival. Therefore, it is urgent to design and develop a multi-gene expression profile and prognosis scoring system for hepatocellular carcinoma based on existing research and technology. SUMMARY

[0011] In view of the above technical problems, the purpose of the present application is to provide a biomarker, RNA model and application for predicting the degree of immunotherapy benefit and prognosis of hepatocellular carcinoma. The specific scheme is as follows.

[0012] In the first aspect of the present application, a marker for predicting the degree of immunotherapy benefit and / or prognosis of hepatocellular carcinoma is provided, and the marker comprises the following genes:

[0013] SNORA74B, GBP1, CXCL9, CXCL10, TRBC2, UBE2D3P2, HNRNPA1P21, CD24, SSR2, TRGV2, GBP4, GBP5, PSMB9, TTYH3, PRF1, TRBC1, LAPTM4B, CD8A, SLC39A4, ACVR1B, TRGC2, GPR171, CEP41, ZC3HAV1L, DDR2, XPR1, PLCE1, GAL3ST4, PWWP2B, HSPB7, TMEM267, SLC12A8, IDO1, B3GNT5, SLC6A9, FCRL3, IFNG, GDPD5, LGI2, ZFAT, AGPAT4, DTHD1, ERFE, ARHGAP39, SPATA12, KLRD1, EPGN, ROR1, TLL2.

[0014] In a second aspect of the present application, an RNA model for predicting the degree of immunotherapy benefit and / or prognosis of liver cancer is provided, wherein the model generates a risk score based on the expression levels of the marker genes described above by the Gscore() function of the R package "HIRE".

[0015] In some embodiments of the present application, the RNA model is constructed by a method comprising the following steps:

[0016] (1) Collect sequencing data of hepatocellular carcinoma patients and divide them into a training group and an internal validation group;

[0017] (2) Filter genes related to progression-free survival and overall survival and having statistical differences by the batch_sur() function of the R package "IOBR";

[0018] (3) Based on the expression levels of the genes obtained in step (2), construct a Random Survival Forest model by the rfsrc() function of the R package "randomForestSRC".

[0019] In some embodiments of the present application, the cut-off value of the risk score in the RNA model is the average of the risk score.

[0020] In some embodiments of the present application, the relevant code for generating a risk score based on the RNA model is as follows:

[0021] # Install the HIRE R package

[0022] devtools::install_github("LiaoWJLab / HIRE")

[0023] # Load R package

[0024] Library("HIRE")

[0025] # Input gene expression matrix - and perform HIRE model risk score calculation

[0026] HIRE <- Gscore(eset = eset, scale = TRUE)

[0027] Detailed code and parameters refer to the link: https: / / github.com / LiaoWJLab / HIRE

[0028] The third aspect of the present application also provides a computer readable storage medium, which stores a computer program, wherein the computer program is executed by a processor to realize the function of the RNA model of the second aspect of the present application.

[0029] The fourth aspect of the present application also provides application of the marker in preparation of a product for predicting the degree of survival benefit and / or prognosis of liver cancer immunotherapy.

[0030] The fifth aspect of the present application also provides application of a substance for detecting the marker in preparation of a product for predicting the degree of survival benefit and / or prognosis of liver cancer immunotherapy.

[0031] The sixth aspect of the present application also provides a product, which comprises the above-mentioned substance for detecting the marker.

[0032] In some embodiments of the present application, the product in the third to sixth aspects of the present application comprises reagents, instruments, software, etc., such as a kit, a diagnostic reagent, a detection reagent, a test paper, a gene chip or a protein chip, etc.

[0033] In some embodiments of the present application, the substance in the third to sixth aspects of the present application comprises a substance for detecting the expression level of the marker gene by a sequencing technology, a nucleic acid hybridization technology, a nucleic acid amplification technology, etc.

[0034] In some embodiments of the present application, the substance in the third to sixth aspects of the present application comprises a primer, an antibody, a probe, etc. For example, in a specific embodiment of the present application, the primer comprises a sequence as shown in SEQ ID NO. 1-98.

[0035] In some embodiments of the present application, the primer in the third to sixth aspects of the present application further comprises an internal reference primer.

[0036] In some embodiments of the present application, the internal reference genes include TUBB and ACTB. In a specific embodiment of the present application, the primer sequences of the internal reference genes TUBB and ACTB are shown in SEQ ID NO. 99-102.

[0037] In some embodiments of the present application, the liver cancer is hepatocellular carcinoma.

[0038] In some embodiments of the present application, the immunotherapy includes immune checkpoint inhibitor therapy.

[0039] In some embodiments of the present application, the immune checkpoint inhibitor includes PD-1 / PD-L1 antibody.

[0040] Compared with the prior art, the present application has the following beneficial effects:

[0041] Currently, the biomarkers for predicting the efficacy of immunotherapy in the field of hepatocellular carcinoma mainly include the expression levels of AFP and PD-L1. However, these two indicators mainly depend on the characteristics of tumor cells, ignoring the key role of tumor microenvironment in mediating tumor immunotherapy response. The present application screens 49 related markers by analyzing the transcriptome profile of hepatocellular carcinoma patients receiving immunotherapy, and establishes a molecular scoring RNA model (HIRE scoring model) based on tumor microenvironment by using machine learning algorithm. The HIRE scoring model of the present application is constructed according to the characteristic genes related to the prognosis of hepatocellular carcinoma patients receiving immunotherapy, which reflects the activation of tumor microenvironment to some extent. In combination with the expression levels of tumor cell characteristics and immune checkpoint genes, the HIRE scoring model can more comprehensively evaluate the tumor and tumor microenvironment, and thus more accurately screen out the potential population benefiting from immunotherapy.

[0042] Compared with existing molecular markers, the model is not only an independent prognostic factor for predicting the progression and death risk of hepatocellular carcinoma patients receiving immunotherapy, but also an effective tool for predicting the response of patients to immunotherapy, providing more individualized evaluation of immunotherapy benefits for hepatocellular carcinoma patients, and having certain clinical transformation significance. BRIEF DESCRIPTION OF DRAWINGS

[0043] Figure 1 To train the HIRE score to predict the efficacy of immunotherapy for hepatocellular carcinoma. Figure 1 A OS survival curve of hepatocellular carcinoma patients with high and low HIRE scores receiving immunotherapy. Figure 1 B PFS survival curve of hepatocellular carcinoma patients with high and low HIRE scores receiving immunotherapy. Figure 1 C ROC curve to evaluate the accuracy of HIRE score in predicting the efficacy of immunotherapy for hepatocellular carcinoma in the training group.

[0044] Figure 2To validate the efficacy of HIRE score in predicting the efficacy of immunotherapy in HCC patients. Figure 2 A. OS survival curve of HCC patients with high and low HIRE score receiving immunotherapy. Figure 2 B. PFS survival curve of HCC patients with high and low HIRE score receiving immunotherapy. Figure 2 C. ROC curve to evaluate the accuracy of HIRE score in predicting the efficacy of immunotherapy in HCC patients in the validation set.

[0045] Figure 3 D. Comparison of the efficacy of HIRE score and other common immune biomarkers in predicting the efficacy of immunotherapy.

[0046] Figure 4 E. To validate the efficacy of HIRE score in predicting the efficacy of immunotherapy in HCC patients in the external validation set. DETAILED DESCRIPTION

[0047] The concept and technical effects of the present application will be described below in conjunction with the embodiments to fully understand the purpose, features and effects of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments of the present application, other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0048] Unless otherwise specified, the reagents, methods and devices used in the present application are conventional reagents, methods and devices in the technical field. Unless otherwise specified, the reagents and materials used in the following examples are commercially available.

[0049] Example 1 Construction of HCC patient cohort

[0050] 1. Construction of training set and validation set patient cohort

[0051] The applicant first applied for data from the U.S. Genentech research team that owns two major immune therapy cohort data of hepatocellular carcinoma: the IMbrave150 patient cohort published in the New England Journal of Medicine and the GO30140A / F group patient cohort published in the Lancet-Oncology. The patient enrollment criteria for the above two studies are as follows: ① the patient is 18 years old or older; ② locally advanced metastatic or unresectable hepatocellular carcinoma (or coexist); ③ no previous systemic treatment for liver cancer or not applicable, or progression after treatment; ④ measurable lesions; ⑤ ECOG score 0 or 1; ⑥ Child-Pugh liver function A level; ⑦ adequate hematological and organ function. ⑧ Among them, the two groups of patients in the GO30140 cohort have additional enrollment criteria. GO30140 A group: Child-Pugh score reaches B7, platelet count is at least 100,000 platelets per μL. GO30140 F group: Child-Pugh liver function A level, life expectancy is at least 3 months, platelet count is at least 75,000 platelets per μL. After combining the enrollment criteria and related screening, a total of 246 patients who received atezolizumab combined with bevacizumab treatment were obtained, and the data were randomly divided into training set and internal validation set (7:3), the training set contains 173 patients, and the validation set contains 73 patients.

[0052] Example 2: Screening of candidate genes for modeling

[0053] The tumor microenvironment, metabolism and tumor intrinsic pathway related gene signatures of the hepatocellular carcinoma patient immune therapy cohort transcriptome data provided by the U.S. Genentech research team were analyzed, and the target genes were screened by analyzing the gene differences and treatment response related gene signatures.

[0054] The present study first uses the batch_surv() function of the R package "IOBR" to obtain genes with a p-value < 0.01 related to progression-free survival and overall survival, and divides the obtained genes into high expression groups and low expression groups according to the average value. From the high group of genes, 49 genes with a p-value < 0.001 are selected as input genes. The input genes include: SNORA74B, GBP1, CXCL9, CXCL10, TRBC2, UBE2D3P2, HNRNPA1P21, CD24, SSR2, TRGV2, GBP4, GBP5, PSMB9, TTYH3, PRF1, TRBC1, LAPTM4B, CD8A, SLC39A4, ACVR1B, TRGC2, GPR171, CEP41, ZC3HAV1L, DDR2, XPR1, PLCE1, GAL3ST4, PWWP2B, HSPB7, TMEM267, SLC12A8, IDO1, B3GNT5, SLC6A9, FCRL3, IFNG, GDPD5, LGI2, ZFAT, AGPAT4, DTHD1, ERFE, ARHGAP39, SPATA12, KLRD1, EPGN, ROR1, TLL2. Finally, based on the expression of the input genes, a random survival forest model is constructed by the rfsrc() function of the R package "randomForestSRC", which generates a risk score based on the expression of these genes.

[0055] The relevant code for the model to generate a risk score based on gene expression is as follows:

[0056] # Install HIRE R package

[0057] devtools::install_github("LiaoWJLab / HIRE")

[0058] # Load R package

[0059] Library("HIRE")

[0060] # Input gene expression matrix - and perform HIRE model score calculation

[0061] HIRE <- Gscore(eset = eset, id_pdata = "ID")

[0062] Example 3: Prediction of HIRE score model in training group for efficacy and prognosis of patients with hepatocellular carcinoma receiving immunotherapy

[0063] According to the method of Example 2, the HIRE scores of 173 patients in the training group were calculated, and the effect of using the HIRE score was explored in the training group. The average HIRE score was set as the cut-off value, and the patients were divided into a high HIRE group and a low HIRE group based on this. The results of survival analysis of patients with OS data showed that patients with low HIRE had significantly longer overall survival when receiving immunotherapy for hepatocellular carcinoma ( Figure 1 A). Patients with high HIRE had a significantly higher survival risk (HR=5.78, 95% CI=2.97-11.22) than those with low HIRE. Subsequently, survival analysis of patients with PFS data showed that patients with low HIRE had a significantly longer progression-free survival (PFS) when receiving immunotherapy. Figure 1 B). Patients with high HIRE had a significantly higher risk of progression (HR=7.26, 95%CI=4.65-11.34) than those with low HIRE. The ROC curve was used to evaluate the efficacy of the HIRE score in predicting the response to immunotherapy in patients with hepatocellular carcinoma. The analysis showed that the AUC value of the HIRE score was 0.9455, indicating that the accuracy of the HIRE score in predicting the efficacy of immunotherapy was 94.55% ( Figure 1 C).

[0064] Example 4: Predictive Effect of the HIRE Scoring Model on the Efficacy of Immunotherapy in Patients with Hepatocellular Carcinoma in the Validation Group

[0065] The effect of using HIRE score was explored in the validation group. The internal validation group (n=73) was selected as the validation cohort. The mean HIRE score was set as the cut-off value, and the patients were divided into high HIRE group and low HIRE group based on this. The results of survival analysis of patients with OS data showed that patients with low HIRE had significantly longer overall survival ( Figure 2 A). Patients with high HIRE had a significantly higher survival risk (HR=3.17, 95% CI=1.14-8.8) than those with low HIRE. Subsequently, survival analysis of patients with PFS data showed that patients with low HIRE had significantly longer progression-free survival (PFS) after immunotherapy. Figure 2 B). Patients with high HIRE had a significantly higher risk of progression (HR=3.8, 95%CI=1.91-7.58) than those with low HIRE. The ROC curve was used to evaluate the efficacy of the HIRE score in predicting the response to immunotherapy in patients with hepatocellular carcinoma. The analysis showed that the AUC value of the HIRE score was 0.8307, indicating that the accuracy of the HIRE score in predicting the efficacy of immunotherapy was 83.07% ( Figure 2C). Next, the HIRE score was compared with other common immune biomarkers for predicting the efficacy of immunotherapy. It was found that the AUC value of the HIRE score in predicting the efficacy of immunotherapy for hepatocellular carcinoma was higher than that of other biomarkers Figure 3

[0066] The accuracy of the HIRE score in predicting the efficacy of immunotherapy for hepatocellular carcinoma was further verified in the external validation cohort (GSE14520, TCGA-LIHC). Similarly, the average of the HIRE score was set as the cut-off value, and the patients were divided into high HIRE group and low HIRE group. In hepatocellular carcinoma patients receiving immunotherapy, the HIRE score distribution of the non-responding group was significantly higher than that of the responding group Figure 4

[0067] In summary, the HIRE score has good predictive effect in the training group and internal and external validation groups, and its accuracy in predicting the efficacy of immunotherapy for hepatocellular carcinoma is better than that of other common immune biomarkers.

[0068] Example 5: Construction of PCR kit and detection implementation method

[0069] (1) Design of PCR primers

[0070] ① In order to realize the clinical transformation of the gene set identified in the above study, so as to realize the real-time detection of individual patients, the primer sequence of the above gene set was predicted using the Primer designing tool method, and the primer sequence corresponding to each marker gene was SEQ ID NO. 1-98 (see Table 1 for details);

[0071] ② In order to realize the standardization of the PCR detection results of the above tag genes, TUBB and ACTB were selected as internal reference genes, and the corresponding primer sequences were SEQ ID NO. 99-102 (see Table 1 for details). The expression level of each tag gene is calculated based on the average expression level of the two internal reference genes.

[0072] Note: Internal reference genes are expressed in all cells and are basic genes for normal life activities. The internal reference genes are added to calculate the standardization parameter of each sample, and the collection of multiple internal reference genes can better standardize the expression level of tumor microenvironment genes. In qPCR detection technology, internal reference genes are needed to perform relative quantification of target genes.

[0073] Table 1: qRT-PCR primer sequences of the gene set

[0074]

[0075] ​​

[0076]

[0077] (2) PCR reagent kit detection and implementation method (RNA extraction, standard value acquisition)

[0078] 1. Extraction of RNA from tumor tissue

[0079] 1) Extraction of RNA from fresh or frozen tissue samples

[0080] ① First, put the tumor sample tissue directly into a mortar, add a small amount of liquid nitrogen, grind quickly, and then add a small amount of liquid nitrogen, grind again, repeat 3 times;

[0081] ② Add 50-100 mg of tissue sample to 1 ml of Trizol and transfer it to a centrifuge tube. The volume of the tissue should not exceed 10% of the volume of Trizol, and then use an electric homogenizer to homogenize for 1-2 min;

[0082] ③ Then, centrifuge at 12000 r / min for 5 min, discard the precipitate, add 200 μL of chloroform per ml of Trizol, tightly cap the centrifuge tube, shake by hand for 15 s, and let it stand at room temperature for 10 min;

[0083] ④ Then, centrifuge at 4°C, 12000 r / min for 15 min, and transfer the upper aqueous phase to a new centrifuge tube. Add 0.6 mL of isopropyl alcohol per ml of Trizol, mix well and let stand at room temperature for 5-10 min;

[0084] ⑤ Then, centrifuge at 4°C, 12000 r / min for 10 min, discard the supernatant, and add 1 mL of 75% ethanol per ml of Trizol, gently shake and suspend the precipitate;

[0085] ⑥ Then, dry at room temperature or vacuum dry for 5-10 min, and then measure the absorbance at 260 nm to determine the concentration of RNA;

[0086] ⑦ Finally, the RNA can be separated into mRNA or stored in 70% ethanol at -70°C. The homogenate before adding chloroform can be stored for more than 1 month;

[0087] 2) Extraction of RNA from formalin-fixed paraffin-embedded (FFPE) tissue samples

[0088] Notes: For FFPE samples, the storage time of the sample is recommended to be less than 2 years;

[0089] ① After removing the surface layer of the wax block in contact with the air, cut 5-8 paraffin sections (5-10 μm thick);

[0090] ② Add PBS vortex mix, room temperature centrifugal, discard supernatant, repeat 3 times;

[0091] ③ The sample is loaded into a sterile centrifuge tube, and dimethylbenzene is added, and vortexed vigorously;

[0092] ④ Use Roche's High Pure FFPET RNA Isolation Kit to extract high-purity RNA.

[0093] 2. RNA quantitative detection

[0094] ① Reverse transcription: according to the steps of the reverse transcription kit, the extracted total RNA of the cells is reverse transcribed into corresponding cDNA. The reverse transcription system is 10 μL, which specifically includes: buffer 2 μL, reverse transcriptase 0.5 μL, oligonucleotide primer 0.5 μL, random primer 0.5 μL, total RNA of cells with a concentration of about 1000 ng 1 μL (adjust the volume of the added RNA sample according to the detected RNA concentration, and ensure the amount of RNA), and the rest of the system is supplemented with RNase-free water. After preparing the reverse transcription system, reverse transcription is carried out.

[0095] ② Configuration of real-time quantitative PCR reaction system: according to the instructions of Roche fluorescent quantitative PCR kit, real-time quantitative PCR experiment is carried out. The reaction system is 10 μL, which specifically includes: DEPC water 3.6 μL, upstream primer 0.2 μL, downstream primer 0.2 μL, SYBR Green I fluorescent reaction liquid 5 μL, cDNA sample 1 μL. Use Roche quantitative fluorescent PCR instrument to amplify cDNA.

[0096] ③ Calculate the expression of the gene by comparison method (ΔΔCt), and compare the cycle threshold (Ct) of the target gene and the housekeeping gene.

[0097] 3. Selection of internal reference genes and tag genes

[0098] ① In order to evaluate the consistency of PCR detection results and high-throughput sequencing data, 65 patients with remaining specimens were randomly selected from the above queue data, and the tissue specimens were detected for the above tag genes and internal reference genes;

[0099] (2) By evaluating the expression abundance and stability of the reference genes, ACTB and TUBB were selected as the reference genes for subsequent normalization; at the same time, the signature genes were used as the components of the PCR detection kit. Signature genes: SNORA74B, GBP1, CXCL9, CXCL10, TRBC2, UBE2D3P2, HNRNPA1P21, CD24, SSR2, TRGV2, GBP4, GBP5, PSMB9, TTYH3, PRF1, TRBC1, LAPTM4B, CD8A, SLC39A4, ACVR1B, TRGC2, GPR171, CEP41, ZC3HAV1L, DDR2, XPR1, PLCE1, GAL3ST4, PWWP2B, HSPB7, TMEM267, SLC12A8, IDO1, B3GNT5, SLC6A9, FCRL3, IFNG, GDPD5, LGI2, ZFAT, AGPAT4, DTHD1, ERFE, ARHGAP39, SPATA12, KLRD1, EPGN, ROR1, TLL2; Reference genes: ACTB and TUBB.

[0100] 4. Calculation of HIRE score

[0101] The expression level of each gene was calculated using the following formula: Nrel = N01 / N02 = N x 2-Ct1 / N x 2-Ct2 = 2-(Ct1-Ct2) = 2-ACt (1 is the detected gene, and 2 is the average value of the reference gene). Then the score of each tumor patient was calculated using the following method: input the patient gene expression level matrix into the Gscore() function of the R package HIRE to calculate the HIRE score of the patient.

[0102] The above detailed description has described the present application in detail, but the present application is not limited to the above embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the purpose of the present application. In addition, the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

Claims

1. A marker for predicting the prognosis of hepatocellular carcinoma after treatment with PD-1 or PD-L1 antibodies, characterized in that: The markers are the following genes: SNORA74B, GBP1, CXCL9, CXCL10, TRBC2, UBE2D3P2, HNRNPA1P21, CD24, SSR2, TRGV2, GBP4, GBP5, PSMB9, TTYH3, PRF1, TRBC1, LAPTM4B, CD8A, SLC39A4, ACVR1B, TRGC2, GPR171, CEP41, ZC3HAV1L, DDR2, XPR1, PLCE1, GAL3ST4, PWWP2B, HSPB7, TMEM267, SLC12A8, IDO1, B3GNT5, SLC6A9, FCRL3, IFNG, GDPD5, LGI2, ZFAT, AGPAT4, DTHD1, ERFE, ARHGAP39, SPATA12, KLRD1, EPGN, ROR1 and TLL2 .

2. A risk assessment model for predicting the prognosis of hepatocellular carcinoma after treatment with PD-1 or PD-L1 antibodies, characterized in that: The model was constructed using a random survival forest algorithm and a risk score was generated based on the marker expression levels as described in claim 1 using the Gscore function of the R package "HIRE".

3. The risk assessment model according to claim 2, characterized in that: The cut-off value of the risk score of the model is the average of the risk scores.

4. Use of a substance for detecting the marker according to claim 1 in the preparation of a product for predicting the prognosis of hepatocellular carcinoma treated with PD-1 or PD-L1 antibodies.

5. The use according to claim 4, characterized in that The substances include primers or probes.

Images