Risk model of nucleotide metabolism and immunity signature genes in predicting of hepatocellular carcinoma prognosis and immune microenvironment

Risk model of nucleotide metabolism and immunity signature genes in predicting of hepatocellular carcinoma prognosis and immune microenvironment | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Risk model of nucleotide metabolism and immunity signature genes in predicting of hepatocellular carcinoma prognosis and immune microenvironment Xiaofang Wang, Qinghua Cui, Yuan Zhou This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6048868/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Given the persistent poor prognosis of hepatocellular carcinoma (HCC), it is imperative to establish a multi-feature prognostic prediction model. Methods: LIHC patient data were obtained from TCGA, and LIHC patients were clustered by NMF and a nucleotide metabolism and immune-related genes (NMIRGs) prognostic risk model was built byby cox regression and Lasso regression, the gene expression was detected by immunohistochemistry and qPCR. The GSE14520, GSE10141 and GSE27150 datasets were used as external validation sets. The prognostic value of the model was evaluated by immune checkpoint genes, immune infiltration, functional enrichment analysis (GSEA) and tumor mutation load, and its potential mechanism was explored. The accuracy of our model was evaluated by the C-index value, compared to the reported prediction model. Results: The HCC patients were grouped into two subtypes based on nucleotide metabolism and immune differentially expressed genes (DEGs), with cluster 1 (C1) having more DEGs and worse prognosis than cluster 2 (C2). A NMIRGs signature gene prognostic model was established, with HSP90AA1, HDAC1, STC1, MAPT and CHGA genes as high-risk genes and GHR genes as low-risk genes. We found that the stage and risk score of HCC patients were independent prognostic factors. The model also assess the prognostic risk of patients in relation to immune checkpoint genes and immune infiltration. Further analysis showed that our model's risk score from our model and tumor mutation burden (TMB) were largely independent in predicting the efficacy of immunotherapy and were closely related to immune-related processes. Finally, the model was found to have a higher C-index and accuracy in predicting the prognosis of HCC patients. Conclusions: Our study provides a helpful predictive model for the prognosis of HCC and its immune microenvironment. In addition, it identifies potential new biomarkers associated with HCC prognosis. hepatocellular carcinoma nucleotide metabolism immune microenvironment prognostic risk biomarker Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Hepatocellular carcinoma (HCC) is one of the most common solid tumors, with a high mortality rate and a poor prognosis [1]. In recent years, many novel therapies have been proposed and developed, individual survival rates remain low due to delays in diagnosis, frequent metastasis, and high recurrence rates [2]. Early identification of high-risk HCC patients can significantly improve patient outcomes and reduce the burden of cancer. Therefore, the current priority is to develop and validate new prognostic signals, improve the accuracy of predicting HCC prognostic risk, aid clinical decision-making, and discover new biomarkers for HCC diagnosis and treatment. Recent studies have shown that changes in nucleotide metabolism can impact the biological behavior of tumor cells, including proliferation, immune evasion, metastasis, and therapeutic resistance [3]. The liver is the primary site of nucleotide synthesis and catabolism, and nucleotides play a crucial role in maintaining normal liver function. Altering the concentration of nucleotides in the liver can affect its function [4]. Numerous studies have demonstrated a correlation between cancer initiation and up-regulated nucleotide biosynthesis capacity by oncogenic drivers, including HCC. Enhanced nucleotide metabolism is critical for cancer initiation and progression downstream of oncogene activation [5,6]. Therefore, developing and validating prognostic prediction models of HCC involving nucleotide metabolism-related genes would be a valuable strategy. As the liver is an important immune organ that orchestrates the innate immune response through by activating various immune cells [7], Immunotherapy holds promise for HCC patients [1]. However, the differences in the tumor microenvironment of individual HCC patients restrict the efficacy of immunotherapy [8]. In recent years, the relationship between nucleotide metabolism and immune cells in tumor cells has begun to emerge [3,6,9]. Aberrant nucleotide metabolism promotes tumor growth and suppresses the normal immune response in the tumor microenvironment [3]. Currently, studies have attempted to establish prognostic model of immune-related genes [10,11] or the nucleotide-related genes [12,13] for HCC patients. However, the robustness and effectiveness of the single-feature model was relatively poor, whereas the multiple-feature model will effectively improve the accuracy of predicting the prognostic risk of HCC. Therefore, we attempted to develop a prognostic model for HCC patients by integrating the differential genes of nucleotide metabolism and immunity based on the existing clinical features. According to the nucleotide metabolism and immune signatures DEGs identify the subtype of HCC patients the model was evaluated and verified by external datasets through survival analysis, ROC curve and decision curve analysis. TMB and GSEA were used to explore the underlying mechanism, as well as the immune microenvironment and the effect of immunotherapy are analyzed. The study will help clinicians to more accurately assess the prognosis of HCC patients and provide personalized treatment options for HCC patients, while discovering new biomarkers and providing new ideas and strategies for developing more effective HCC diagnosis and treatment strategies. Materials and Methods Data acquisition and processing RNA sequencing (RNA-seq) transcriptome, clinical information and somatic mutation data were obtained from The Cancer Genome Atlas database (TCGA; https://portal.gdc.cancer.gov/), including 374 LIHC and 50 normal. External validation datasets of HCC patients were obtained from the TCGA database and the Gene Expression Omnibus (GEO; https://www.ncbi.nlm.nih.gov/geo/) database, including The GSE14520 (GPL3721 subset), GSE10141, and GSE27150. The data were normalized, and differentially expressed genes (DEGs) of the samples were analyzed using the R package "limma". The DEGs extraction criteria were as follows: |log2Fold-Change|> 1 and P<0.01 . All gene expression values were treated as log2-scale. When multiple probes corresponded to the same gene symbol, the mean expression was calculated as the final expression value. DEG Subset Extraction and Non-Negative Matrix Factorization (NMF) Clustering From the MSigDB database (https://www.gsea-msigdb.org/gsea/msigdb/) "c2.cp.kegg.v7.4" 947 metabolism-related genes were downloaded. After analysis of the DAVID database [14](https://david.ncifcrf.gov/tools.jsp), 72 nucleotide metabolism-related genes were obtained. Immune-related genes (1793 genes after removing duplicates) were obtained from the ImmPort database [15](https://www.immport.org/home ). The transcriptome data of TCGA database were intersected to obtain Differential expressed genes (DEGs) related to nucleotide metabolism and immunity. Then, the expression of DEGs in each tumor sample in the TCGA database was combined with clinical data by univariate Cox analysis to obtain the expression of prognostic related NMIRGs. The "nsNMF" algorithm of the "NMF" package (v0.23.0) was used to cluster HCC patient samples according to the expression of NMIRGs, and the cluster number K was set between 2 and 10. With "survival" package for HCC patients survival analysis, the progression-free survival (PFS) of HCC patients was validated using the XENA database (https://pancanatlas.xenahubs.net ) pan-cancer data. Analysis of the tumor microenvironment and the immune infiltration Stromal cell and immune cell scores in HCC were calculated using the Estimation of STromal and Immune cells in MAlignant Tumour tissues using Expression (ESTIMATE) [16] software (version 1.0.13), which uses the unique properties of transcriptional profiles to infer the properties of tumor cells as well as the purity of tumors, thereby predicting the levels of immune infiltration. The MCPcounter [17] algorithm estimated the abundance of each immune microenvironment-related cell type in the different HCC subtypes. The "MCPcounter" package provides abundance scores for T cells, CD8 T cells, cytotoxic lymphocytes, B lineage, NK cells, monocytic lineage, myeloid dendritic cells, neutrophils, endothelial cells and fibroblasts for each sample. Prognostic model construction and validation Following reference [18], the model was constructed according to the sample 7:3 within the group. Samples without complete survival data were excluded, and 259 samples (70%) out of 371 TCGA-LIHC datasets were used as training set (TCGAtrain set), 112 samples (30%) were used as internal test set (TCGAtest set). Lasso regression model was constructed by "glmnet" and "caret" R package, 9 NMIRGs signature genes were determined to construct prognostic model, and high-risk and low-risk groups were divided according to the median risk score of each sample. HCC samples in GSE14520, GSE10141 were used as external validation set (GEOtest set). The expression of DEGs in each tumor sample of GSE14520 dataset was compared with survival time and survival state by univariate Cox analysis, and the expression of prognostic related NMIRGs was obtained. The model was verified by survival analysis, ROC curve at 1, 3, 5 years (using the "rms" R package), decision curve analysis (DCA) and risk curve. Survival differences in Kaplan-Meier curves were compared using the log-rank test. The ROC curve and the area under the curve (AUC) were obtained by using the "timeROC" package. Evaluation and comparison of the prognostic models Differences in clinical characteristics between the TCGAtest and TCGAtrain groups were analyzed using the chi-squared test. Correlation analysis was performed using the Pearson's method, and the Mann-Whitney U test was used to compare differences between the two groups. Clinical data metrics were integrated using logistic regression to generate a survival prediction nomogram. Independent prognostic factors were analyzed by univariate and multivariate Cox risk regression analyses using the "survival" package. The model performance was evaluated by correlation analysis of immune checkpoint genes (31 genes of HCC [19,20]) and immune infiltration in high-risk and low-risk groups of HCC patients. The TCIA database [21](https://tcia.at/home) was used to analyze the immunotherapeutic effect of PD-1 and CTLA-4 in HCC. Gene set enrichment analysis (GSEA) of the active function or pathway to explore the underlying mechanisms. Correlation analysis of risk score, microsatellite instability (MSI), TMB and tumor immune microenvironment, to evaluate the prognostic value of this model. Finally, we compared our model with the previous models reported in the references, and the accuracy of our model was evaluated by the C-index value. All statistical analyses of the study were performed using R software (version 4.3.1), P<0.05 indicated that the difference was statistically significant. Detection of the expression of NMRIGs signature The expression of NMIRGs signature genes were detected by Real Time PCR. After 24 hours of culture of HUH7 and MHCC97-H, total RNA of HCC cells was extracted using total RNA extraction kit (Thermo, US). Then, the total RNA was reversely transcribed into cDNA using reverse transcription kit (Thermo, US), and the relative gene expression was detected by fluorescent quantitative PCR (Bioer, China). The primer sequences were shown in Supplementary Table 1. Protein expression level of these genes in HCC tissue were retrieval by human protein atlas online database (HPA, https://www.proteinatlas.org/). In the immunohistochemical (IHC) test, the samples of HCC patients were obtained from the TCGA database, where the degree of antibody staining was classified as Not detected, Low, Medium and High. The positive result is tissue dyed dark brown. Results NMIRGs clustered HCC patients into two subtypes, and they showed significant differences in tumor microenvironment and immune infiltration According to our procedure, a total of 412 DEGs were found by combining the NMIRGs with the TCGA transcriptome data (|log2 fold-change| >1, P < 0.01 ), of which 362 were up-regulated DEGs and 50 were down-regulated DEGs (Supplementary Fig. 1A,B). Among them, there were 382 immune-related genes, 34 metabolism-related genes, and 3 crossover genes(Supplementary Fig. 1C), these DEGs can distinguish HCC samples from normal controls (Supplementary Fig. 2). Using univariate Cox regression analysis and NMF clustering were used to divide HCC into C1 and C2 subtypes (Supplementary Fig. 1D, Supplementary Fig. 3). The Kaplan-Meier curve showed that the overall survival (OS) and PFS of C1 were significantly lower than that of C2 ( P < 0.05 , Supplementary Fig. 1E, F), indicating that nucleotide metabolism and immune DEGs (NMIRGs) are able to discriminate the different types of HCC. Studies have shown that nucleotide metabolism is associated with the immune microenvironment [6], so we speculate that NMIRGs may also discriminate immune infiltrates with different HCC subtypes. The unsupervised clustering diagram shows that nucleotide metabolism and immune DEGs showed higher expressions in C1 than in C2 (Fig. 1 A). Using the MCPcounter for eight immune microenvironment-related cell types for analysis, we found that except the neutrophils and endothelial cells, all other cells were significantly increased in subtype C1 ( P < 0.05 ), whereas the neutrophils were significantly increased in subtype C2 (Fig. 1 B), indicating that the immune cells were mainly infiltrated in the C1 tumor samples. Similarly, the results of tumor microenvironment analysis also showed that immune cells scored higher in C1 (Fig. 1 C, D). Both the score and the composite score of stromal cells showed that C1 was higher than C2, indicating that the tumor purity of C1 was lower than C2. Combined with previous result, we believe that C1 represents a highly expressed subtype of nucleotide metabolism and immune DEGs in HCC, with more obvious immune infiltration, lower tumor purity, and poor prognosis. Prognosis prediction model based on nucleotide metabolism and immune-related genes We combined the survival data and DEGs of TCGA HCC patients, and randomly divided the samples into TCGAtrain group and TCGAtest group according to the 7:3 fraction. Through univariate Cox analysis and LASSO regression with cross-validation, nine signature genes (RAC3, HSP90AA1, BTC, MAPT, HDAC1, STC1, CHGA, GAL, GHR) were identified to construct the prognostic model (Fig. 2 A-C, Supplementary Table 2, Supplementary Table 3). The high-risk and low-risk groups were divided according to the median risk score of each sample. Through the analysis of clinical characteristics by chi-squared test, we found that there was no difference in age, sex, grade, stage, T, N and M stage between the high-risk and low-risk groups ( P > 0.05 , Supplementary Table 4), indicating that the prediction model had no statistical deviation in the clinical characteristics of the sample grouping. Using the GSE14520 dataset as an external validation set and comparing the survival rate of TCGAtrain, TCGAtest, and TCGAall groups between the high-risk and low-risk HCC patients, it was found that the survival rate of the high-risk group was significantly lower than that of the low-risk group ( P < 0.05 , Fig. 2 D-G). ROC curves at 1,3, and 5 years showed that the area under the curve of each group was greater than 0.5 (Fig. 2 H-K), it indicating that the model has predictive power to discriminate the prognostic risk of HCC patients in the high-risk and low-risk groups. Characterizing the clinical relevance of the prediction model The model characteristics were verified using internal and external datasets, and it was found that as the risk increased, the curve became progressively steeper and the number of patients who died gradually increased (Fig. 3 A-D), which was consistent with the expected prediction. In the TCGAtrain dataset, the expression of the GHR gene decreased with higher risk, while the expression of the HSP90AA1 , HDAC1 , RAC3 , STC1 , MAPT and BTC genes increased (Fig. 3 A-D), but the CHGA and GAL genes were only upregulated in some very-high-risk samples. This suggests that the high-risk genes are the HSP90AA1 , HDAC1 , RAC3 , STC1 , MAPT and BTC genes, and the low-risk genes are the GHR genes. For different clinical characteristics of the high-risk and low-risk groups, the survival rate of the high-risk group was lower in grade (G1-4), stage (I-Ⅳ), T stage (T1-4) compared with the low-risk group ( P < 0.05 , Fig. 3 E-J), showing that our model is suitable for predicting of HCC in different periods. Looking at the risk scores of patients among clinical characteristics, there were no significant differences in age, sex, N stage and M stage in HCC patients ( P > 0.05 , Supplementary Fig. 4), but significant differences in grade, stage and T stage ( P < 0.05 , Fig. 3 K-M). Among them, the risk scores of the patients gradually increased during the progression from G1 to G4, stage I to stage Ⅲ, T1 to T4 (Fig. 3 K-M, P < 0.05 ). The result suggests that the model is able to predict differences in clinical characteristics and prognosis of HCC patients with different progression in the high-risk and low-risk groups. Nomogram derived from the prediction model predict survival partly independent to the tumor stage To further demonstrate the prediction ability of our model in comparison with basic clinical characteristics, logistic regression was performed on the entire TCGA-LIHC cohort and a survival prediction nomogram was constructed. As shown in the figure, the 1-year, 3-year, and 5-year survival rates predicted by the model were 0.891, 0.786, and 0.711, respectively (Fig. 4 A, Supplementary Table 5). The C-index value of the calibration curve was 0.759, indicating that the nomogram predicted survival and actual survival well (Fig. 4 B). There are more prognostic features included in the model, so we further tested if the nomogram can be used as independent prognostic factors. By univariate Cox regression analysis, we found that the risk score 95%CI was 1.054(1.029–1.079) and P < 0.001 (Fig. 4 C). Multivariate Cox regression analysis showed that the risk score of 95%CI was 1.042(1.016–1.069), and P = 0.002 (Fig. 4 D). In addition, the stage of HCC was P < 0.001 in univariate and multivariate Cox regression analysis, suggesting that the stage and risk score of HCC patients can be used as independent risk factors to predict the prognosis of HCC patients. In addition, the results of the DCA decision curve further suggested that stage was closely related with the poor prognosis of HCC patients (Supplementary Fig. 5). As shown in Fig. 4 E, the AUC of the nomogram was 0.803, indicating that the prediction of survival of HCC patients by the nucleotide metabolism and immune-related gene signatures model is accurate, indicating its potential clinical value. Links between prognostic model results and immunotherapy efficacy Next, we explored the role of our model in assessing immune status and immunotherapy efficacy. The intersection of 79 immune checkpoint genes discovered by Hu et al. [19] and 47 HCC-related immune checkpoint genes discovered by Long et al. [20] was used to obtain 31 immune checkpoint genes of HCC for analysis. Through the correlation analysis of immune checkpoint related genes, it was found that most of the immune checkpoint related genes were positively correlated with the high-risk and low-risk of HCC patients, and only ADORA2A gene was negatively correlated with the risk of HCC patients (Fig. 5 A-B, P 0.05 ). This suggests that our model can be used to assess the association of risk differences in HCC patients with immune checkpoint genes, and our model may serve as a new indicator for the efficacy of immunotherapy. To check this idea, LIHC immunotherapy data of HCC were obtained through the TCIA database, and compared the effect of PD-1 and CTLA-4 treatment on patients with HCC in the high-low-risk group, we found that the high-low-risk group in the GSE10141 dataset had a significant difference in immunotherapy ( P < 0.05 ). Compared to the high-risk group, the low-risk group fared better when treated with CTLA4 alone, PD-1 alone, or both in combination (Fig. 5 C-F). However, the response of HCC patients in the GSE14520 and GSE27150 datasets to immunotherapy is different from that in the GSE10141 dataset. Only the high-risk and low-risk groups of HCC patients in the double negative group in the GSE14520 dataset are statistically different. There was no difference in immunotherapy efficacy between high-risk and low-risk HCC patients in the GSE27150 dataset (Supplementary Fig. 7). In conclusion, our model can predict the differences in immune status and immunotherapy efficacy between high-risk and low-risk groups of HCC patients for most of the datasets investigated here. Tumor immunity-related biological features are associated with the prediction results Similarly, our model is advantageous for assessing the association between risk differences and immune infiltration in HCC patients. Compared to the low-risk group, the high-risk group had more immune cells, such as NK cells, monocytic lineage, endothelial cell, T cells, myeloid dendritic cells, but the low-risk group had more neutrophils ( P < 0.05 , Fig. 6 A, Supplementary Fig. 8). Furthermore, neutrophils were negatively correlated with the risk of HCC patients, and other immune cells were positively correlated with the risk of HCC patients ( P < 0.05 , Fig. 6 B). While tumor mutation burden (TMB) has been considered as a biomarker for predicting cancer progression and treatment outcome [22], physiological mismatch repair (MMR) proteins aim to detect and correct errors in mismatched nucleotides, the defects of which can lead to microsatellite instability (MSI) in DNA [23]. Therefore, we also analyzed the role of our predictive model in TMB. The TMB of HCC patients was divided into high-TMB and low-TMB groups, and we found that the survival rate of the high-TMB group was significantly lower than that of the low-TMB group ( P < 0.05 , Fig. 6 C, Supplementary Table 6). When the patient risk score was combined TMB, we found that HCC patients with high-TMB in the high-risk group had a lower survival rate, while HCC patients with low-TMB in the high-risk group had a lower survival rate than those in the low-risk group ( P < 0.05 , Fig. 6 D), suggesting that the survival risk of HCC patients and TMB are independent of each other. We analyzed the correlation between HCC risk score, TMB, MSI and immune cells in HCC patients. According to Fig. 6 E, MSI was positively correlated with TMB and patient risk score, except for neutrophils, HCC patients risk scores were positively correlated with most immune cells. In fact, the survival risk of HCC patients and TMB are indeed independent of each other. These results indicate that NMIRGs can be used to predict the prognostic risk of HCC patients, and the risk is positively correlated with TMB and MSI, and the high risk is closely correlated with immune-related processes. We further analyzed the functions or pathways of enrichment among the high-risk and low-risk groups by GSEA enrichment method. As shown from in Fig. 6 F, five biological pathways including cell cycle, cytokine-cytokine receptor interaction, ECM receptor interaction, hematopoietic cell lineage and neuroactive ligand-receptor interaction were enriched in the high-risk group Five biological processes including cytochrome p450, fatty acid metabolism, glycine, serine, and threonine metabolism, limonene and degradation, and primary bile acid biosynthesis were enriched in the low-risk group (Fig. 6 G). This suggests that immune-related biological processes are mainly enriched in the high-risk group of HCC patients, whereas metabolism-related processes are mainly enriched in the low-risk group of HCC patients. Comparison with previous HCC prognostic models Finally, we compared the prediction score of the external validation set GSE14520 with the models already reported in the references, including Zhang_signature [24](Chemokines and chemokine receptors related three-gene model), Cui_signature [25] (Immune index related ten-gene model), Su_signature [26] (Immunogenic cell death related three-gene model), Lu_signature[27](inflammatory response and immunity status related seven-gene model). Using the same calculation method, HCC samples were divided into high-risk and low-risk groups, and the survival analysis showed that our model is better than Zhang_signature and Su_signature, and is at least comparable to Cui_signature and Lu_signature (Fig. 7 A-E). Furthermore, the ROC curve analysis showed that our NMIRGs_signature indeed showed better performance, where the area under ROC curve for 1-year, 3-year and 5-year survival rates in our model were 0.780, 0.695 and 0.710, respectively (Fig. 7 F-J). In comparison with other models, the C-index of our model was 0.697, which is also higher than Zhang_signature (C-index = 0.595), Cui_signature (C-index = 0.587), Su_signature (C-index = 0.621), Lu_signature (C-index = 0.622) (Fig. 7 K). These results indicate that our model has a high accuracy in predicting the prognosis of HCC patients. NMRIGs signature were differentially expressed in HCC patients’ tissues and cells The expression of these nine NMRIGs signature genes used in the construction of the model was searched in the HPA online database. We found that the protein of GHR, HDAC1, MAPT, HSP90AA1, CHGA and STC1 were expressed in the tissues of HCC patients (Fig. 8 A-E), while RAC3 was expressed in cholangiocarcinoma, but was not detected in HCC. And GAL was not detected in HCC patients (Supplementary Table 7). What’s more, the protein of GHR, HDAC1, MAPT, HSP90AA1, CHGA were significantly correlated with the prognosis of HCC patients (Fig. 8 F-J). We analyzed liver cancer cell lines in the GDSC database and found that HuH7 cells were more resistant to HCC therapeutics than other HCC cell lines (Fig. 8 K, Supplementary Table 8). The relative expression levels of NMRIGs signature genes mRNA were expressed in different degrees in Huh7 cell lines, among which the relative expression level of GHR gene was the highest ( P < 0.05 , Fig. 8 L), which was consistent with the results of immunohistochemistry. Discussion HCC remains a challenge to global health. Although immunotherapy has made a breakthrough, the treatment outcome is not the same due to the different of tumor microenvironment, and there is still a significant proportion of HCC patients still with poor prognosis [12]. Therefore, it is urgent to investigate its effective prognostic indicators. Given the role of nucleotide metabolism in HCC progression and the immune microenvironment [3,5,6,9]. The combination of both predictors has a higher predictability[28]. Therefore, this study integrates nucleotide metabolism and immune-related differential genes and HCC patients clinical features, established the prognosis prediction model of HCC patients, in order to help clinicians more accurately evaluate the prognosis of HCC patients and provide personalized treatment and find new prognostic markers. We first classified HCC patients into C1 and C2 subtypes by NMF clustering, and found that C1 cluster had high expression of nucleotide metabolism and immune-related genes, low tumor purity, more immune cell infiltration, and overall survival (OS) and progression-free survival were significantly lower than C2. This suggests that high expression of NMIRGs is associated with poor prognosis and immune infiltration in HCC patients, and has predictive potential. Therefore, next, we screened out 9 nucleotide metabolism and immune signature genes to construct a prognostic model of HCC. HSP90AA1 , HDAC1 , RAC3 , STC1 , MAPT and BTC genes are high-risk genes, and GHR gene is a low-risk gene. Receptor-associated coactivator 3 ( RAC3 ) is an oncogene that is highly expressed in many malignant tumors and regulates cell cycle progression and cell proliferation, invasion, and migration [29]. In fact, RAC3 only showed strong staining in cholangiocarcinoma. Betacellulin ( BTC ) proteins are ligands for EGF receptors, and binding to EGFR, ERBB 4, and other members of the EGF receptor family members can effectively promote angiogenesis [30]. RT-PCR showed that BTC gene was expressed in HuH7 cell line. Chromogranin A ( CHGA ) may play a role in early tumor progression [31], it is mentioned in most liver cancer prediction models, and we found that it is indeed differentially expressed in HCC patients and is associated with the prognosis of patients with liver cancer. Human Heat Shock Protein 90 Alpha Family Class A Member 1 ( HSP90AA1) [32], microtubule-associated protein tau ( MAPT ) [33], and histone deacetylase 1 ( HDAC1 ) [34] are associated with a poor prognosis in HCC. Knockdown of growth hormone receptor ( GHR ) increases the sensitivity of HCC cells to sorafenib [35]. We also detected differential expression of these genes in HCC cells and tissues. In addition, the AUC of the nomogram reached 0.803, indicating that our model has good predictive performance, and these genes are expected to become potential targets of HCC. According to the expression of NMRIGs signature genes, personalized guidance for HCC patients can be targeted to reduce the burden of HCC patients. We also evaluated the predictive performance of the model by immune invasion and the effect of immunotherapy. Because of studies have shown that tumor cells evade immune killing of the immune system by altering the local immune microenvironment [36]. Therefore, changes in the immune microenvironment are critical for assessing tumor progression. We found that our model could well evaluate the correlation between the difference in prognostic risk of HCC patients and immune checkpoint genes and immune infiltration, and the changes in neutrophils were also consistent with the negative correlation between the number of tumor- infiltrating neutrophils and the number of T cells reported in the reference [37]. In addition, when evaluating the effect of PD-1 and CTLA-4 immunotherapy in high-risk and low-risk HCC patients, the results of the external validation dataset were also in line with expectations. There are differences in the effectiveness of immunotherapy in different HCC patients, so the efficacy of immunotherapy is not the same in different datasets. For example, the GSE10141 dataset included more HCC patients with high alcohol consumption than the GSE14520 dataset, and our model showed that they had better immunotherapy outcomes for PD-1 and CTLA-4 in the GSE10141 dataset. In order to explore the potential regulatory mechanism of NMRIGs signature on HCC, we performed GSEA enrichment analysis in the high-risk and low-risk groups, and found that the enrichment of immune-related processes was mainly in the high-risk group of HCC patients, while the enrichment of metabolism-related processes was mainly in the low-risk group. This is an interesting result, which shows that the analysis of HCC progression cannot only consider a single feature, and the combined analysis of multiple features can enable the more comprehensive interpretation of the HCC prognosis. TMB and MSI, as indicators of somatic mutation [23], have been used as biomarkers to predict cancer progression and treatment outcome. High TMB indicates increased production of neoantigens, which can affect the efficacy of immunotherapy. Our analysis showed that the risk score of HCC patients was positively correlated not only with MSI and TMB, but also with most immune cell abundances (except neutrophils). Nonetheless, the NMIRGs risk score and TMB are largely independent to each other in predicting HCC prognosis. In summary, we found that NMIRGs can classify HCC and thus established a predictive model with advantages in predicting survival, assessing clinical relevance, prognostic risk, and immune microenvironment in HCC patients. Potential targets associated with HCC prognosis were also discovered. Most importantly, our predictive model has higher accuracy in predicting the prognosis of HCC patients than some single-feature predictive models. Conclusion In conclusion, we have constructed a clinical model of 9 nucleotide metabolic and immune signature genes that predicting HCC prognosis, which has advantages in predicting the risk and immune microenvironment of HCC patients. It is expected that this model will help clinicians to more accurately assess the prognosis of HCC patients and provide personalized treatment options. In addition, the differential expression of NMIRGs signature genes suggests their potential as biomarkers, providing a new perspective for the diagnosis and treatment of HCC. However, it should be noted that our study is based on re-analysis of public data, and therefore further analysis and validation through experimental studies and prospective clinical investigations are needed. Abbreviations HCC: hepatocellular carcinoma LIHC: Liver hepatocellular carcinoma NMIRGs: nucleotide metabolism and immune-related genes C1: cluster 1 DEGs: Differential Expressed Genes TMB: Tumor mutation load MSI: mitochondrial microsatellite instability TCGA: The Cancer Genome Atlas GEO: Gene Expression Omnibus DAVID: The Database for Annotation, Visualization and Integrated Discovery LASSO: Least Absolute Shrinkage and Selection Operator MCPcounter: Microenvironment Cell Populations counter ESTIMATE: Estimation of STromal and Immune cells in MAlignant Tumour tissues using Expression GSEA: Gene Set EnrichmentAnalysis NMF: Non-Negative Matrix Factorization PFS: progression-free survival DCA: Decision Curve Analysis ROC: Receiver Operating Characteristic RAC3: Receptor-associated coactivator 3 HSP90AA1: Heat Shock Protein 90 Alpha Family Class A Member 1 BTC: Betacellulin MAPT: microtubule-associated protein tau HDAC1: histone deacetylase 1 STC1: stanniocalcin-1 CHGA: Chromogranin A GAL: galactoside GHR: growth hormone receptor PD-1: programmed cell death protein 1 CTLA-4: Cytotoxic T lymphocyte associate protein-4 ADORA2A: Adenosine Receptor A2a Declarations Ethical Statement: Not applicable, due to it is a re-analysis of public data and no new human and animal subjects are involved. Author Contributions: Conceptualization, X.W.; Methodology, X.W., Y.Z.; Formal Analysis, X.W.; Data Curation, X.W.; Writing – Original Draft Preparation, X.W.; Writing – Review & Editing, Y.Z., Q.C.; Supervision, Y.Z., Q.C.; Funding Acquisition, Y.Z. Funding : This work was supported by the National Natural Science Foundation of China (32222020 to Yuan Zhou). Informed Consent Statement: Not applicable, due to it is a re-analysis of public data and no new human and animal subjects are involved. Data availability statement : The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors. Conflict of interest: There is no conflicts of interest in this study. References Vogel, A.; Meyer, T.; Sapisochin, G.; Salem, R.; Saborowski, A. Hepatocellular carcinoma. Lancet 2022 , 400 , 1345-1362, doi:10.1016/S0140-6736(22)01200-4. Anwanwan, D.; Singh, S.K.; Singh, S.; Saikam, V.; Singh, R. Challenges in liver cancer and possible treatment approaches. Biochim Biophys Acta Rev Cancer 2020 , 1873 , 188314, doi:10.1016/j.bbcan.2019.188314. Ma, J.; Zhong, M.; Xiong, Y.; Gao, Z.; Wu, Z.; Liu, Y.; Hong, X. Emerging roles of nucleotide metabolism in cancer development: progress and prospect. Aging (Albany NY) 2021 , 13 , 13349-13358, doi:10.18632/aging.202962. Wu, H.L.; Gong, Y.; Ji, P.; Xie, Y.F.; Jiang, Y.Z.; Liu, G.Y. Targeting nucleotide metabolism: a promising approach to enhance cancer immunotherapy. J Hematol Oncol 2022 , 15 , 45, doi:10.1186/s13045-022-01263-x. Jain, S.; Jacobson, K.A. Purinergic Signaling in Liver Pathophysiology. Front Endocrinol (Lausanne) 2021 , 12 , 718429, doi:10.3389/fendo.2021.718429. Mullen, N.J.; Singh, P.K. Nucleotide metabolism: a pan-cancer metabolic dependency. Nat Rev Cancer 2023 , 23 , 275-294, doi:10.1038/s41568-023-00557-7. Xu, D.; Tian, Y.; Xia, Q.; Ke, B. The cGAS-STING Pathway: Novel Perspectives in Liver Diseases. Front Immunol 2021 , 12 , 682736, doi:10.3389/fimmu.2021.682736. Donne, R.; Lujambio, A. The liver cancer immune microenvironment: Therapeutic implications for hepatocellular carcinoma. Hepatology 2023 , 77 , 1773-1796, doi:10.1002/hep.32740. Han, T.; Kang, D.; Ji, D.; Wang, X.; Zhan, W.; Fu, M.; Xin, H.B.; Wang, J.B. How does cancer cell metabolism affect tumor migration and invasion? Cell Adh Migr 2013 , 7 , 395-403, doi:10.4161/cam.26345. Shen, B.; Zhang, G.; Liu, Y.; Wang, J.; Jiang, J. Identification and Analysis of Immune-Related Gene Signature in Hepatocellular Carcinoma. Genes (Basel) 2022 , 13 , 1834, doi:10.3390/genes13101834. Chen, Y.; Shen, C.; Wu, J.; Yan, X.; Huang, Q. Role of immune related genes in predicting prognosis and immune response in patients with hepatocellular carcinoma. J Biochem Mol Toxicol 2023 , e23519, doi:10.1002/jbt.23519. Xu, Y.J.; He, M.K.; Liu, S.; Huang, L.C.; Bu, X.Y.; Kan, A.; Shi, M. Construction of a single nucleotide variant score-related gene-based prognostic model in hepatocellular carcinoma: analysis of multi-independent databases and validation in vitro. Cancer Cell Int 2021 , 21 , 610, doi:10.1186/s12935-021-02321-z. Wei, T.; Liu, J.; Ma, S.; Wang, M.; Yuan, Q.; Huang, A.; Wu, Z.; Shang, D.; Yin, P. A Nucleotide Metabolism-Related Gene Signature for Risk Stratification and Prognosis Prediction in Hepatocellular Carcinoma Based on an Integrated Transcriptomics and Metabolomics Approach. Metabolites 2023 , 13 , 1116, doi:10.3390/metabo13111116. Sherman, B.T.; Hao, M.; Qiu, J.; Jiao, X.; Baseler, M.W.; Lane, H.C.; Imamichi, T.; Chang, W. DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update). Nucleic Acids Res 2022 , 50 , W216-W221, doi:10.1093/nar/gkac194. Bhattacharya, S.; Dunn, P.; Thomas, C.G.; Smith, B.; Schaefer, H.; Chen, J.; Hu, Z.; Zalocusky, K.A.; Shankar, R.D.; Shen-Orr, S.S.; et al. ImmPort, toward repurposing of open access immunological assay data for translational and clinical research. Sci Data 2018 , 5 , 180015, doi:10.1038/sdata.2018.15. Yoshihara, K.; Shahmoradgoli, M.; Martinez, E.; Vegesna, R.; Kim, H.; Torres-Garcia, W.; Trevino, V.; Shen, H.; Laird, P.W.; Levine, D.A.; et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun 2013 , 4 , 2612, doi:10.1038/ncomms3612. Becht, E.; Giraldo, N.A.; Lacroix, L.; Buttard, B.; Elarouci, N.; Petitprez, F.; Selves, J.; Laurent-Puig, P.; Sautes-Fridman, C.; Fridman, W.H.; et al. Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression. Genome Biol 2016 , 17 , 218, doi:10.1186/s13059-016-1070-5. Moons, K.G.M.; Altman, D.G.; Reitsma, J.B.; Ioannidis, J.P.A.; Macaskill, P.; Steyerberg, E.W.; Vickers, A.J.; Ransohoff, D.F.; Collins, G.S. Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD): Explanation and Elaboration. Annals of Internal Medicine 2015 , 162 , W1-W73, doi:10.7326/m14-0698. Hu, F.F.; Liu, C.J.; Liu, L.L.; Zhang, Q.; Guo, A.Y. Expression profile of immune checkpoint genes and their roles in predicting immunotherapy response. Brief Bioinform 2021 , 22 , bbaa176, doi:10.1093/bib/bbaa176. Long, M.; Zhou, Z.; Wei, X.; Lin, Q.; Qiu, M.; Zhou, Y.; Chen, P.; Jiang, Y.; Wen, Q.; Liu, Y.; et al. A novel risk score based on immune-related genes for hepatocellular carcinoma as a reliable prognostic biomarker and correlated with immune infiltration. Front Immunol 2022 , 13 , 1023349, doi:10.3389/fimmu.2022.1023349. Charoentong, P.; Finotello, F.; Angelova, M.; Mayer, C.; Efremova, M.; Rieder, D.; Hackl, H.; Trajanoski, Z. Pan-cancer Immunogenomic Analyses Reveal Genotype-Immunophenotype Relationships and Predictors of Response to Checkpoint Blockade. Cell Rep 2017 , 18 , 248-262, doi:10.1016/j.celrep.2016.12.019. Choucair, K.; Morand, S.; Stanbery, L.; Edelman, G.; Dworkin, L.; Nemunaitis, J. TMB: a promising immune-response biomarker, and potential spearhead in advancing targeted therapy trials. Cancer Gene Ther 2020 , 27 , 841-853, doi:10.1038/s41417-020-0174-y. Gabbia, D.; De Martin, S. Tumor Mutational Burden for Predicting Prognosis and Therapy Outcome of Hepatocellular Carcinoma. Int J Mol Sci 2023 , 24 , 3441, doi:10.3390/ijms24043441. Zhang, Z.; Mao, M.; Wang, F.; Zhang, Y.; Shi, J.; Chang, L.; Wu, X.; Zhang, Z.; Xu, P.; Lu, S. Comprehensive analysis and immune landscape of chemokines- and chemokine receptors-based signature in hepatocellular carcinoma. Front Immunol 2023 , 14 , 1164669, doi:10.3389/fimmu.2023.1164669. Cui, X.; Han, L.; Cui, L.; Fu, G.; Liu, E.; Wang, D.; Song, B.; Zhang, Y.; Zhou, W.; Wang, H.; et al. Immune index: A gene and cell prognostic signature for immunotherapy response prediction in hepatocellular carcinoma. Pharmacol Res 2023 , 187 , 106583, doi:10.1016/j.phrs.2022.106583. Su, D.; Zhang, Z.; Xia, F.; Liang, Q.; Liu, Y.; Liu, W.; Xu, Z. ICD-related risk model predicts the prognosis and immunotherapy response of patients with liver cancer. Front Pharmacol 2023 , 14 , 1202823, doi:10.3389/fphar.2023.1202823. Lu, G.; Du, R.; Feng, B.; Wang, J.; Zhang, F.; Pei, J.; Wang, Y.; Shang, Y. A Novel Gene Signature Associated with Inflammatory Responses and Immune Status Assists in Prognosis and Intervention for Patients with HCC. J Inflamm Res 2022 , 15 , 6729-6743, doi:10.2147/JIR.S390113. Xu, Y.Y.; Qin, Z.; Gao, J.M.; Yang, Y.; Lu, Y.Z.; Yu, F.Y.; Lv, Y.; Sun, Z.W.; Zhang, J.Y.; Tang, J.N. Combination of Neutrophil Count and Gensini Score as a Prognostic Marker in Patients with ACS and Uncontrolled T2DM Undergoing PCI. Cardiovasc Innov App 2023 , 8 , doi:10.15212/Cvia.2023.0051. de Curtis, I. The Rac3 GTPase in Neuronal Development, Neurodevelopmental Disorders, and Cancer. Cells 2019 , 8 , 1063, doi:10.3390/cells8091063. Dahlhoff, M.; Wolf, E.; Schneider, M.R. The ABC of BTC: structural properties and biological roles of betacellulin. Semin Cell Dev Biol 2014 , 28 , 42-48, doi:10.1016/j.semcdb.2014.01.002. Mahata, S.K.; Corti, A. Chromogranin A and its fragments in cardiovascular, immunometabolic, and cancer regulation. Ann N Y Acad Sci 2019 , 1455 , 34-58, doi:10.1111/nyas.14249. Zuehlke, A.D.; Beebe, K.; Neckers, L.; Prince, T. Regulation and function of the human HSP90AA1 gene. Gene 2015 , 570 , 8-16, doi:10.1016/j.gene.2015.06.018. Chang, C.W.; Shao, E.; Mucke, L. Tau: Enabler of diverse brain disorders and target of rapidly evolving therapeutic strategies. Science 2021 , 371 , eabb8255, doi:10.1126/science.abb8255. Cai, D.; Yuan, X.; Cai, D.Q.; Li, A.; Yang, S.; Yang, W.; Duan, J.; Zhuo, W.; Min, J.; Peng, L.; et al. Integrative analysis of lactylation-related genes and establishment of a novel prognostic signature for hepatocellular carcinoma. J Cancer Res Clin Oncol 2023 , 149 , 11517-11530, doi:10.1007/s00432-023-04947-0. Gao, S.; Ni, Q.; Wu, X.; Cao, T. GHR knockdown enhances the sensitivity of HCC cells to sorafenib. Aging (Albany NY) 2020 , 12 , 18127-18136, doi:10.18632/aging.103625. Giannini, E.G.; Aglitti, A.; Borzio, M.; Gambato, M.; Guarino, M.; Iavarone, M.; Lai, Q.; Levi Sandri, G.B.; Melandro, F.; Morisco, F.; et al. Overview of Immune Checkpoint Inhibitors Therapy for Hepatocellular Carcinoma, and The ITA.LI.CA Cohort Derived Estimate of Amenability Rate to Immune Checkpoint Inhibitors in Clinical Practice. Cancers (Basel) 2019 , 11 , 1689, doi:10.3390/cancers11111689. He, G.; Zhang, H.; Zhou, J.; Wang, B.; Chen, Y.; Kong, Y.; Xie, X.; Wang, X.; Fei, R.; Wei, L.; et al. Peritumoural neutrophils negatively regulate adaptive immunity via the PD-L1/PD-1 signalling pathway in hepatocellular carcinoma. J Exp Clin Cancer Res 2015 , 34 , 141, doi:10.1186/s13046-015-0256-0. Supplementary Files Supplementarymaterials.pdf Supplementary materials Supplementary Figure 1. NMF clustering of HCC based on genes related to nucleotide metabolism and immunity. Supplementary Figure 2. The heatmap of nucleotide metabolism and immune-related DEGs. Supplementary Figure 3. Results of the nsNMF decomposition level. Supplementary Figure 4. Comparison of different gender, different age, different N stages, and M stages. Supplementary Figure 5. The DCA decision curve of each clinical feature. Supplementary Figure 6. Comparison of immune checkpoint-related genes in the high and low-risk groups of TCGA-LIHC patients (Exclude all other immune checkpoints listed in the article). Supplementary Figure 7. Comparison of the immunotherapy effects in TCGA-LIHC patients in the high-risk and low-risk groups (GES14520 dataset and GSE27150 dataset). Supplementary Figure 8. Comparison of immune microenvironment-related cells (B lineage, myeloid dendritic cells, CD8 T cells, fibroblasts) between high-risk and low-risk groups in patients with HCC. Supplementary Table 1. The sequence of primers. Supplementary Table 2. The results of HCC samples from the TCGA database that multivariate Cox regression analysis. Supplementary Table 3. The results of HCC samples from the TCGA database that underwent LASSO regression analysis. Supplementary Table 4. Chi-square test results of clinical features of LIHC patients in TCGA database. Supplementary Table 5. The HCC patients’ high-risk and low-risk results of the nomogram. Supplementary Table 6. The HCC patients’ tumor mutation burden (TMB) results. Supplementary Table 7. HPA database Immunohistochemical patient information. Supplementary Table 8. IC50 of liver cancer cell lines relative to drug sensitivity in GDSC database. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6048868","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":430047038,"identity":"3e90e169-7262-4851-abc2-076945e8cc7d","order_by":0,"name":"Xiaofang Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABCUlEQVRIiWNgGAWjYBACfv7GhgMJBhI8/OzNB8AiEiCCB48WyRmHGw88KLCRk+w5lsBwgBgtBgfSmw8++JBmbHDDx4A4LQwHDoIcdjhx5gyez58/tjHkSc5IYHzwto1B3hyHDsbmRoiWfunebRIH2xiKpSUSmA3ntjEY7mzAroWZAWbLnLPbGIBaEudJJLBJ87YxJBgcwK6FjSERomXDjZzHH6Ba2H/j08ID0QLyfg4DyGGJs4G2MOPTIiEBdhg4kM0kzpyTKJbsedgsOeechOEGHFrsz7c//vjjDzgqH3+oKLPJkziefPDDmzIbeVy2YNiaAAzFBgZo9BAHEohXOgpGwSgYBSMFAADhb2ijxzxo0AAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0003-3450-6408","institution":"Peking University Health Science Center","correspondingAuthor":true,"prefix":"","firstName":"Xiaofang","middleName":"","lastName":"Wang","suffix":""},{"id":430047039,"identity":"6ca95dbe-dbf6-415b-a201-919fcaeca030","order_by":1,"name":"Qinghua Cui","email":"","orcid":"","institution":"Wuhan Institute of Physical Education: Wuhan Sports University","correspondingAuthor":false,"prefix":"","firstName":"Qinghua","middleName":"","lastName":"Cui","suffix":""},{"id":430047040,"identity":"dd2804c5-d437-452c-b69e-420f56484650","order_by":2,"name":"Yuan Zhou","email":"","orcid":"https://orcid.org/0000-0001-5685-066X","institution":"Peking University Health Science Library: Peking University Health Science Center","correspondingAuthor":false,"prefix":"","firstName":"Yuan","middleName":"","lastName":"Zhou","suffix":""}],"badges":[],"createdAt":"2025-02-17 14:32:18","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6048868/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6048868/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":79172734,"identity":"7fa6ff69-1162-493b-9011-8b8803681731","added_by":"auto","created_at":"2025-03-25 09:43:56","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":179987,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAnalysis of the immune microenvironment in different HCC subtypes.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(A) \u003c/strong\u003eComparison of DEGs of two HCC subtypes by unsupervised clustering. \u003cstrong\u003e(B)\u003c/strong\u003e Comparison of the results for the two HCC subtypes using the MCPcounter algorithm. \u003cstrong\u003e(C)\u003c/strong\u003e Heatmap of the immune cell scoring using the MCPcounter algorithm. \u003cstrong\u003e(D)\u003c/strong\u003e Comparison of the tumor microenvironment of the two HCC subtypes using the ESTIMATE algorithm. * \u003cem\u003eP \u0026lt; 0.05\u003c/em\u003e, ** \u003cem\u003eP \u0026lt; 0.01\u003c/em\u003e, *** \u003cem\u003eP \u0026lt; 0.001\u003c/em\u003e.\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6048868/v1/19404d61b9074018434de638.jpeg"},{"id":79174407,"identity":"4aacce7f-e011-443c-9cb9-26b0050734f1","added_by":"auto","created_at":"2025-03-25 09:51:56","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":137506,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConstruction and overall performance evaluation of the prognosis prediction model.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(A)\u003c/strong\u003e The LASSO regression coefficient distribution map, one line represents one gene. \u003cstrong\u003e(B) \u003c/strong\u003eDistribution plot of the LASSO regression parameters. \u003cstrong\u003e(C)\u003c/strong\u003e Bar chart showing the model coefficients of the selected signature genes. \u003cstrong\u003e(D-G)\u003c/strong\u003e Kaplan-Meier curve analysis of high-risk and low-risk HCC patients in each dataset, including internal datasets TCGAtrain, TCGAtest, TCGAall and external GEO test dataset GSE14520. \u003cstrong\u003e(H-K)\u003c/strong\u003e ROC curve of at 1,3, and 5 years in each dataset.\u003c/p\u003e","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6048868/v1/a3dc2d689a0fc2abcb161f7a.jpeg"},{"id":79174405,"identity":"7cb844b9-7747-4cb8-b95a-0474d288c8de","added_by":"auto","created_at":"2025-03-25 09:51:56","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":206076,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAnalysis of basic clinical features associated with the high- and low-risk groups.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(A-D) \u003c/strong\u003eRisk curves, survival status and prognostic signatures heatmap in TCGAtrain, TCGAtest, GEOtest (GSE14520) and TCGAall datasets. \u003cstrong\u003e(E-F) \u003c/strong\u003eKaplan-Meier curve analysis of pairwise comparisons of G1 to G2, G3 to G4 between high-risk and low-risk groups. \u003cstrong\u003e(G-H) \u003c/strong\u003eKaplan-Meier curve analysis of pairwise comparisons of stage I to stage Ⅱ, stage Ⅲ to stage Ⅳ between high-risk and low-risk groups. \u003cstrong\u003e(I-J) \u003c/strong\u003eKaplan-Meier curve analysis of pairwise comparisons of T1 to T2, T3 to T4 between high-risk and low-risk groups. \u003cstrong\u003e(K-M) \u003c/strong\u003eComparison of different grades, different stages, and different T stages. * \u003cem\u003eP\u0026lt;0.05\u003c/em\u003e, **\u003cem\u003e P\u0026lt;0.01\u003c/em\u003e, *** \u003cem\u003eP\u0026lt;0.001\u003c/em\u003e.\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6048868/v1/9a6aa0e3835e93bbaac5aeed.jpeg"},{"id":79178039,"identity":"9f59d9b0-f976-43e7-b930-82b78b50f02c","added_by":"auto","created_at":"2025-03-25 10:07:56","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":66040,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConstruction and evaluation of the prediction model-derived nomogram.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(A) \u003c/strong\u003eNomogram prediction of OS in TCGA-LIHC patients based on prognostic features of nucleotide metabolism and immune DEGs. For each patient, the total score was calculated by adding the scores obtained from the subscales for each variable. On the basis of the total score, the underlying scale was used to predict the probability of survival at 1, 3 or 5 years. The red line indicates the calculation process and the principle of the nomogram. \u003cstrong\u003e(B-C)\u003c/strong\u003eUnivariate and multivariate Cox regression analysis of risk scores and various clinical characteristics. \u003cstrong\u003e(D)\u003c/strong\u003e Correction curve between predicted and actual survival at 1,3 and 5 years.\u003cstrong\u003e (E) \u003c/strong\u003eROC curves of the 1-year, 3-year, and 5-year survival rates of TCGA-LIHC patients, including the nomogram.\u003c/p\u003e","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6048868/v1/de5ba759a971fe031b6e5c72.jpeg"},{"id":79174403,"identity":"585d8242-c0c3-454c-9951-963397219e5c","added_by":"auto","created_at":"2025-03-25 09:51:56","extension":"jpeg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":112660,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe effectiveness of the prognostic model assessing immune status andimmunotherapy efficacy.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(A) \u003c/strong\u003eComparison of the expression of immune checkpoint-related genes in TCGA-LIHC patients between high-risk and low-risk groups (only 9 genes were listed). \u003cstrong\u003e(B)\u003c/strong\u003e Correlation analysis of risk scores and immune checkpoint-related genes in TCGA-LIHC patients.\u003cstrong\u003e (C-F) \u003c/strong\u003eComparison of the immunotherapy effects of TCGA-LIHC patients in the high-risk and low-risk groups when receiving PD-1 and CTLA-4 alone or in combination (GES10141 dataset).\u003c/p\u003e","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6048868/v1/3fcfb9c4537d41d4e0e0e4ef.jpeg"},{"id":79172730,"identity":"d6047198-d930-4b0b-bcad-a9455474ddeb","added_by":"auto","created_at":"2025-03-25 09:43:56","extension":"jpeg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":135820,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExploration of tumor immunity-related biological factors.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(A) \u003c/strong\u003eComparison of immune cells between the high and low-risk groups of TCGA-LIHC patients. \u003cstrong\u003e(B)\u003c/strong\u003e Correlations between the relative abundance of immune microenvironment-related cell types and the risk score. \u003cstrong\u003e(C-D) \u003c/strong\u003eSurvival rate and Kaplan-Meier curves of high-TMB and low-TMB, and high-risk and low-risk in TCGA-LIHC patients. \u003cstrong\u003e(E) \u003c/strong\u003eCorrelation analysis of TMB, MSI, and immune cell infiltration in patients with TCGA-LIHC. \u003cstrong\u003e(F-G) \u003c/strong\u003eKEGG enrichment analysis in the high-risk and low-risk groups of TCGA-LIHC patients.\u003c/p\u003e","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6048868/v1/9db60aa94ac9b23c7f2be7d5.jpeg"},{"id":79176387,"identity":"db167e58-cc0e-440b-ba4d-dea1001822ab","added_by":"auto","created_at":"2025-03-25 09:59:56","extension":"jpeg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":130446,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of model predictive performance with the previous prognostic signatures.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(A-E) \u003c/strong\u003eKaplan-Meier curve analysis of all prediction models, including the prediction models published in the references. Figures A to E are NMIRGs_signature (our model), Zhang_signature, Cui_signature, Su_signature, Lu_signature, respectively. \u003cstrong\u003e(F-J) \u003c/strong\u003eROC curves of the 1-year, 3-year, and 5-year survival in all prediction models. Figures F to J are NMIRGs_signature (our model), Zhang_signature, Cui_signature, Su_signature, Lu_signature, respectively. \u003cstrong\u003e(K) \u003c/strong\u003eC-index analysis of all prediction models.\u003c/p\u003e","description":"","filename":"floatimage7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6048868/v1/814d8106dde31e58771bbd5d.jpeg"},{"id":79172735,"identity":"9f99647a-e02e-46ca-8c3f-749a814d7f2f","added_by":"auto","created_at":"2025-03-25 09:43:56","extension":"jpeg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":115943,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eNMRIGs signature genes are differentially expressed in liver cancer.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(A-E)\u003c/strong\u003e The protein expression levels of GHR, HDAC1, MAPT, HSP90AA1, CHGA in HCC patients and normal tissues by immunohistochemistry (IHC) obtained from HPA online database. \u003cstrong\u003e(F-J)\u003c/strong\u003eThe survival analysis of them by TCGA database. \u003cstrong\u003e(K)\u003c/strong\u003eIC50 comparison of the relative sensitivity of HuH7 cell lines with other HCC cell lines to the HCC therapeutics Oxaliplatin, Cisplatin, Dabrafenib, Gefitinib, and Afatinib obtained from GDSC. \u003cstrong\u003e(L)\u003c/strong\u003e The gene expression level of NMRIGs signature in Huh7 cell lines relative to MHCC97H cell lines were detected by real-time PCR.\u003c/p\u003e","description":"","filename":"floatimage8.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6048868/v1/8bfaa08add801eba5b28e4f0.jpeg"},{"id":82397517,"identity":"1827ba1f-f2e0-4180-bdd4-c8d041e4308a","added_by":"auto","created_at":"2025-05-09 21:00:55","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2437678,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6048868/v1/7a5739bd-9fd3-4a4a-8dee-3e04fcfcd94d.pdf"},{"id":79172763,"identity":"63d71f45-37dc-4444-adc2-9101063e6cb4","added_by":"auto","created_at":"2025-03-25 09:43:57","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":2420501,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplementary materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Figure 1\u003c/strong\u003e. NMF clustering of HCC based on genes related to nucleotide metabolism and immunity.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Figure 2\u003c/strong\u003e. The heatmap of nucleotide metabolism and immune-related DEGs.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Figure 3\u003c/strong\u003e. Results of the nsNMF decomposition level.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Figure 4\u003c/strong\u003e. Comparison of different gender, different age, different N stages, and M stages.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Figure 5\u003c/strong\u003e. The DCA decision curve of each clinical feature.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Figure 6\u003c/strong\u003e. Comparison of immune checkpoint-related genes in the high and low-risk groups of TCGA-LIHC patients (Exclude all other immune checkpoints listed in the article).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Figure 7\u003c/strong\u003e. Comparison of the immunotherapy effects in TCGA-LIHC patients in the high-risk and low-risk groups (GES14520 dataset and GSE27150 dataset).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Figure 8\u003c/strong\u003e. Comparison of immune microenvironment-related cells (B lineage, myeloid dendritic cells, CD8 T cells, fibroblasts) between high-risk and low-risk groups in patients with HCC.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Table 1.\u003c/strong\u003e The sequence of primers.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Table 2.\u003c/strong\u003e The results of HCC samples from the TCGA database that multivariate Cox regression analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Table 3.\u003c/strong\u003e The results of HCC samples from the TCGA database that underwent LASSO regression analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Table 4.\u003c/strong\u003e Chi-square test results of clinical features of LIHC patients in TCGA database.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Table 5. \u003c/strong\u003eThe HCC patients’ high-risk and low-risk results of the nomogram.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Table 6.\u003c/strong\u003e The HCC patients’ tumor mutation burden (TMB) results.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Table 7. \u003c/strong\u003eHPA database Immunohistochemical patient information.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Table 8. \u003c/strong\u003eIC50 of liver cancer cell lines relative to drug sensitivity in GDSC database.\u003c/p\u003e","description":"","filename":"Supplementarymaterials.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6048868/v1/531200b78acb4d50bfc416e2.pdf"}],"financialInterests":"","formattedTitle":"Risk model of nucleotide metabolism and immunity signature genes in predicting of hepatocellular carcinoma prognosis and immune microenvironment","fulltext":[{"header":"Introduction","content":"\u003cp\u003eHepatocellular carcinoma (HCC) is one of the most common solid tumors, with a high mortality rate and a poor prognosis [1]. In recent years, many novel therapies have been proposed and developed, individual survival rates remain low due to delays in diagnosis, frequent metastasis, and high recurrence rates [2]. Early identification of high-risk HCC patients can significantly improve patient outcomes and reduce the burden of cancer. Therefore, the current priority is to develop and validate new prognostic signals, improve the accuracy of predicting HCC prognostic risk, aid clinical decision-making, and discover new biomarkers for HCC diagnosis and treatment.\u003c/p\u003e \u003cp\u003eRecent studies have shown that changes in nucleotide metabolism can impact the biological behavior of tumor cells, including proliferation, immune evasion, metastasis, and therapeutic resistance [3]. The liver is the primary site of nucleotide synthesis and catabolism, and nucleotides play a crucial role in maintaining normal liver function. Altering the concentration of nucleotides in the liver can affect its function [4]. Numerous studies have demonstrated a correlation between cancer initiation and up-regulated nucleotide biosynthesis capacity by oncogenic drivers, including HCC. Enhanced nucleotide metabolism is critical for cancer initiation and progression downstream of oncogene activation [5,6]. Therefore, developing and validating prognostic prediction models of HCC involving nucleotide metabolism-related genes would be a valuable strategy.\u003c/p\u003e \u003cp\u003eAs the liver is an important immune organ that orchestrates the innate immune response through by activating various immune cells [7], Immunotherapy holds promise for HCC patients [1]. However, the differences in the tumor microenvironment of individual HCC patients restrict the efficacy of immunotherapy [8]. In recent years, the relationship between nucleotide metabolism and immune cells in tumor cells has begun to emerge [3,6,9]. Aberrant nucleotide metabolism promotes tumor growth and suppresses the normal immune response in the tumor microenvironment [3]. Currently, studies have attempted to establish prognostic model of immune-related genes [10,11] or the nucleotide-related genes [12,13] for HCC patients. However, the robustness and effectiveness of the single-feature model was relatively poor, whereas the multiple-feature model will effectively improve the accuracy of predicting the prognostic risk of HCC. Therefore, we attempted to develop a prognostic model for HCC patients by integrating the differential genes of nucleotide metabolism and immunity based on the existing clinical features. According to the nucleotide metabolism and immune signatures DEGs identify the subtype of HCC patients the model was evaluated and verified by external datasets through survival analysis, ROC curve and decision curve analysis. TMB and GSEA were used to explore the underlying mechanism, as well as the immune microenvironment and the effect of immunotherapy are analyzed. The study will help clinicians to more accurately assess the prognosis of HCC patients and provide personalized treatment options for HCC patients, while discovering new biomarkers and providing new ideas and strategies for developing more effective HCC diagnosis and treatment strategies.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003e\u003cstrong\u003eData acquisition and processing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRNA sequencing (RNA-seq) transcriptome, clinical information and somatic mutation data were obtained from The Cancer Genome Atlas database (TCGA; https://portal.gdc.cancer.gov/), including 374 LIHC and 50 normal. External validation datasets of HCC patients were obtained from the TCGA database and the Gene Expression Omnibus (GEO; https://www.ncbi.nlm.nih.gov/geo/) database, including The GSE14520 (GPL3721 subset), GSE10141, and GSE27150. The data were normalized, and differentially expressed genes (DEGs) of the samples were analyzed using the R package \u0026quot;limma\u0026quot;. The DEGs extraction criteria were as follows: |log2Fold-Change|\u0026gt; 1 and \u003cem\u003eP\u0026lt;0.01\u003c/em\u003e. All gene expression values were treated as log2-scale. When multiple probes corresponded to the same gene symbol, the mean expression was calculated as the final expression value.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDEG Subset Extraction and Non-Negative Matrix Factorization (NMF) Clustering\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFrom the MSigDB database (https://www.gsea-msigdb.org/gsea/msigdb/) \u0026quot;c2.cp.kegg.v7.4\u0026quot; 947 metabolism-related genes were downloaded. After analysis of the DAVID database [14](https://david.ncifcrf.gov/tools.jsp), 72 nucleotide metabolism-related genes were obtained. Immune-related genes (1793 genes after removing duplicates) were obtained from the ImmPort database [15](https://www.immport.org/home ). The transcriptome data of TCGA database were intersected to obtain Differential expressed genes (DEGs) related to nucleotide metabolism and immunity. Then, the expression of DEGs in each tumor sample in the TCGA database was combined with clinical data by univariate Cox analysis to obtain the expression of prognostic related NMIRGs. The \u0026quot;nsNMF\u0026quot; algorithm of the \u0026quot;NMF\u0026quot; package (v0.23.0) was used to cluster HCC patient samples according to the expression of NMIRGs, and the cluster number K was set between 2 and 10. With \u0026quot;survival\u0026quot; package for HCC patients survival analysis, the progression-free survival (PFS) of HCC patients was validated using the XENA database (https://pancanatlas.xenahubs.net ) pan-cancer data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAnalysis of the tumor microenvironment and the immune infiltration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eStromal cell and immune cell scores in HCC were calculated using the Estimation of STromal and Immune cells in MAlignant Tumour tissues using Expression (ESTIMATE) [16] software (version 1.0.13), which uses the unique properties of transcriptional profiles to infer the properties of tumor cells as well as the purity of tumors, thereby predicting the levels of immune infiltration. The MCPcounter [17] algorithm estimated the abundance of each immune microenvironment-related cell type in the different HCC subtypes. The \u0026quot;MCPcounter\u0026quot; package provides abundance scores for T cells, CD8 T cells, cytotoxic lymphocytes, B lineage, NK cells, monocytic lineage, myeloid dendritic cells, neutrophils, endothelial cells and fibroblasts for each sample.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePrognostic model construction and validation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFollowing reference [18], the model was constructed according to the sample 7:3 within the group. Samples without complete survival data were excluded, and 259 samples (70%) out of 371 TCGA-LIHC datasets were used as training set (TCGAtrain set), 112 samples (30%) were used as internal test set (TCGAtest set). Lasso regression model was constructed by \u0026quot;glmnet\u0026quot; and \u0026quot;caret\u0026quot; R package, 9 NMIRGs signature genes were determined to construct prognostic model, and high-risk and low-risk groups were divided according to the median risk score of each sample. HCC samples in GSE14520, GSE10141 were used as external validation set (GEOtest set). The expression of DEGs in each tumor sample of GSE14520 dataset was compared with survival time and survival state by univariate Cox analysis, and the expression of prognostic related NMIRGs was obtained. The model was verified by survival analysis, ROC curve at 1, 3, 5 years (using the \u0026quot;rms\u0026quot; R package), decision curve analysis (DCA) and risk curve. Survival differences in Kaplan-Meier curves were compared using the log-rank test. The ROC curve and the area under the curve (AUC) were obtained by using the \u0026quot;timeROC\u0026quot; package.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEvaluation and comparison of the prognostic models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDifferences in clinical characteristics between the TCGAtest and TCGAtrain groups were analyzed using the chi-squared test. Correlation analysis was performed using the Pearson\u0026apos;s method, and the Mann-Whitney U test was used to compare differences between the two groups. Clinical data metrics were integrated using logistic regression to generate a survival prediction nomogram. Independent prognostic factors were analyzed by univariate and multivariate Cox risk regression analyses using the \u0026quot;survival\u0026quot; package. The model performance was evaluated by correlation analysis of immune checkpoint genes (31 genes of HCC [19,20]) and immune infiltration in high-risk and low-risk groups of HCC patients. The TCIA database [21](https://tcia.at/home) was used to analyze the immunotherapeutic effect of PD-1 and CTLA-4 in HCC. Gene set enrichment analysis (GSEA) of the active function or pathway to explore the underlying mechanisms. Correlation analysis of risk score, microsatellite instability (MSI), TMB and tumor immune microenvironment, to evaluate the prognostic value of this model. Finally, we compared our model with the previous models reported in the references, and the accuracy of our model was evaluated by the C-index value. All statistical analyses of the study were performed using R software (version 4.3.1), \u003cem\u003eP\u0026lt;0.05\u003c/em\u003e indicated that the difference was statistically significant.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDetection of the expression of NMRIGs signature\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe expression of NMIRGs signature genes were detected by Real Time PCR. After 24 hours of culture of HUH7 and MHCC97-H, total RNA of HCC cells was extracted using total RNA extraction kit (Thermo, US). Then, the total RNA was reversely transcribed into cDNA using reverse transcription kit (Thermo, US), and the relative gene expression was detected by fluorescent quantitative PCR (Bioer, China). The primer sequences were shown in Supplementary Table 1. Protein expression level of these genes in HCC tissue were retrieval by human protein atlas online database (HPA, https://www.proteinatlas.org/). In the immunohistochemical (IHC) test, the samples of HCC patients were obtained from the TCGA database, where the degree of antibody staining was classified as Not detected, Low, Medium and High. The positive result is tissue dyed dark brown.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e \u003cb\u003eNMIRGs clustered HCC patients into two subtypes, and they showed significant differences in tumor microenvironment and immune infiltration\u003c/b\u003e \u003c/p\u003e \u003cp\u003eAccording to our procedure, a total of 412 DEGs were found by combining the NMIRGs with the TCGA transcriptome data (|log2 fold-change| \u0026gt;1, \u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.01\u003c/em\u003e), of which 362 were up-regulated DEGs and 50 were down-regulated DEGs (Supplementary Fig.\u0026nbsp;1A,B). Among them, there were 382 immune-related genes, 34 metabolism-related genes, and 3 crossover genes(Supplementary Fig.\u0026nbsp;1C), these DEGs can distinguish HCC samples from normal controls (Supplementary Fig.\u0026nbsp;2). Using univariate Cox regression analysis and NMF clustering were used to divide HCC into C1 and C2 subtypes (Supplementary Fig.\u0026nbsp;1D, Supplementary Fig.\u0026nbsp;3). The Kaplan-Meier curve showed that the overall survival (OS) and PFS of C1 were significantly lower than that of C2 (\u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/em\u003e, Supplementary Fig.\u0026nbsp;1E, F), indicating that nucleotide metabolism and immune DEGs (NMIRGs) are able to discriminate the different types of HCC.\u003c/p\u003e \u003cp\u003eStudies have shown that nucleotide metabolism is associated with the immune microenvironment [6], so we speculate that NMIRGs may also discriminate immune infiltrates with different HCC subtypes. The unsupervised clustering diagram shows that nucleotide metabolism and immune DEGs showed higher expressions in C1 than in C2 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). Using the MCPcounter for eight immune microenvironment-related cell types for analysis, we found that except the neutrophils and endothelial cells, all other cells were significantly increased in subtype C1 (\u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/em\u003e), whereas the neutrophils were significantly increased in subtype C2 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB), indicating that the immune cells were mainly infiltrated in the C1 tumor samples. Similarly, the results of tumor microenvironment analysis also showed that immune cells scored higher in C1 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC, D). Both the score and the composite score of stromal cells showed that C1 was higher than C2, indicating that the tumor purity of C1 was lower than C2. Combined with previous result, we believe that C1 represents a highly expressed subtype of nucleotide metabolism and immune DEGs in HCC, with more obvious immune infiltration, lower tumor purity, and poor prognosis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003ePrognosis prediction model based on nucleotide metabolism and immune-related genes\u003c/h3\u003e\n\u003cp\u003eWe combined the survival data and DEGs of TCGA HCC patients, and randomly divided the samples into TCGAtrain group and TCGAtest group according to the 7:3 fraction. Through univariate Cox analysis and LASSO regression with cross-validation, nine signature genes (RAC3, HSP90AA1, BTC, MAPT, HDAC1, STC1, CHGA, GAL, GHR) were identified to construct the prognostic model (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA-C, Supplementary Table\u0026nbsp;2, Supplementary Table\u0026nbsp;3). The high-risk and low-risk groups were divided according to the median risk score of each sample. Through the analysis of clinical characteristics by chi-squared test, we found that there was no difference in age, sex, grade, stage, T, N and M stage between the high-risk and low-risk groups (\u003cem\u003eP\u0026thinsp;\u0026gt;\u0026thinsp;0.05\u003c/em\u003e, Supplementary Table\u0026nbsp;4), indicating that the prediction model had no statistical deviation in the clinical characteristics of the sample grouping. Using the GSE14520 dataset as an external validation set and comparing the survival rate of TCGAtrain, TCGAtest, and TCGAall groups between the high-risk and low-risk HCC patients, it was found that the survival rate of the high-risk group was significantly lower than that of the low-risk group (\u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/em\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD-G). ROC curves at 1,3, and 5 years showed that the area under the curve of each group was greater than 0.5 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eH-K), it indicating that the model has predictive power to discriminate the prognostic risk of HCC patients in the high-risk and low-risk groups.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eCharacterizing the clinical relevance of the prediction model\u003c/h2\u003e \u003cp\u003eThe model characteristics were verified using internal and external datasets, and it was found that as the risk increased, the curve became progressively steeper and the number of patients who died gradually increased (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA-D), which was consistent with the expected prediction. In the TCGAtrain dataset, the expression of the GHR gene decreased with higher risk, while the expression of the \u003cem\u003eHSP90AA1\u003c/em\u003e, \u003cem\u003eHDAC1\u003c/em\u003e, \u003cem\u003eRAC3\u003c/em\u003e, \u003cem\u003eSTC1\u003c/em\u003e, \u003cem\u003eMAPT\u003c/em\u003e and \u003cem\u003eBTC\u003c/em\u003e genes increased (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA-D), but the \u003cem\u003eCHGA\u003c/em\u003e and \u003cem\u003eGAL\u003c/em\u003e genes were only upregulated in some very-high-risk samples. This suggests that the high-risk genes are the \u003cem\u003eHSP90AA1\u003c/em\u003e, \u003cem\u003eHDAC1\u003c/em\u003e, \u003cem\u003eRAC3\u003c/em\u003e, \u003cem\u003eSTC1\u003c/em\u003e, \u003cem\u003eMAPT\u003c/em\u003e and \u003cem\u003eBTC\u003c/em\u003e genes, and the low-risk genes are the \u003cem\u003eGHR\u003c/em\u003e genes. For different clinical characteristics of the high-risk and low-risk groups, the survival rate of the high-risk group was lower in grade (G1-4), stage (I-Ⅳ), T stage (T1-4) compared with the low-risk group (\u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/em\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE-J), showing that our model is suitable for predicting of HCC in different periods. Looking at the risk scores of patients among clinical characteristics, there were no significant differences in age, sex, N stage and M stage in HCC patients (\u003cem\u003eP\u0026thinsp;\u0026gt;\u0026thinsp;0.05\u003c/em\u003e, Supplementary Fig.\u0026nbsp;4), but significant differences in grade, stage and T stage (\u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/em\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eK-M). Among them, the risk scores of the patients gradually increased during the progression from G1 to G4, stage I to stage Ⅲ, T1 to T4 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eK-M, P\u0026thinsp;\u003cem\u003e\u0026lt;\u0026thinsp;0.05\u003c/em\u003e). The result suggests that the model is able to predict differences in clinical characteristics and prognosis of HCC patients with different progression in the high-risk and low-risk groups.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eNomogram derived from the prediction model predict survival partly independent to the tumor stage\u003c/h2\u003e \u003cp\u003eTo further demonstrate the prediction ability of our model in comparison with basic clinical characteristics, logistic regression was performed on the entire TCGA-LIHC cohort and a survival prediction nomogram was constructed. As shown in the figure, the 1-year, 3-year, and 5-year survival rates predicted by the model were 0.891, 0.786, and 0.711, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, Supplementary Table\u0026nbsp;5). The C-index value of the calibration curve was 0.759, indicating that the nomogram predicted survival and actual survival well (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). There are more prognostic features included in the model, so we further tested if the nomogram can be used as independent prognostic factors. By univariate Cox regression analysis, we found that the risk score 95%CI was 1.054(1.029\u0026ndash;1.079) and \u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). Multivariate Cox regression analysis showed that the risk score of 95%CI was 1.042(1.016\u0026ndash;1.069), and P\u0026thinsp;=\u0026thinsp;0.002 (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD). In addition, the stage of HCC was \u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.001\u003c/em\u003e in univariate and multivariate Cox regression analysis, suggesting that the stage and risk score of HCC patients can be used as independent risk factors to predict the prognosis of HCC patients. In addition, the results of the DCA decision curve further suggested that stage was closely related with the poor prognosis of HCC patients (Supplementary Fig.\u0026nbsp;5). As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE, the AUC of the nomogram was 0.803, indicating that the prediction of survival of HCC patients by the nucleotide metabolism and immune-related gene signatures model is accurate, indicating its potential clinical value.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eLinks between prognostic model results and immunotherapy efficacy\u003c/h2\u003e \u003cp\u003eNext, we explored the role of our model in assessing immune status and immunotherapy efficacy. The intersection of 79 immune checkpoint genes discovered by Hu et al. [19] and 47 HCC-related immune checkpoint genes discovered by Long et al. [20] was used to obtain 31 immune checkpoint genes of HCC for analysis. Through the correlation analysis of immune checkpoint related genes, it was found that most of the immune checkpoint related genes were positively correlated with the high-risk and low-risk of HCC patients, and only \u003cem\u003eADORA2A\u003c/em\u003e gene was negatively correlated with the risk of HCC patients (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-B, P\u0026thinsp;\u003cem\u003e\u0026lt;\u0026thinsp;0.05\u003c/em\u003e). However, there was no statistical difference in the expression of \u003cem\u003eADORA2A\u003c/em\u003e gene in high-risk and low-risk group (Supplementary Fig.\u0026nbsp;6, P\u0026thinsp;\u003cem\u003e\u0026gt;\u0026thinsp;0.05\u003c/em\u003e). This suggests that our model can be used to assess the association of risk differences in HCC patients with immune checkpoint genes, and our model may serve as a new indicator for the efficacy of immunotherapy. To check this idea, LIHC immunotherapy data of HCC were obtained through the TCIA database, and compared the effect of \u003cem\u003ePD-1\u003c/em\u003e and \u003cem\u003eCTLA-4\u003c/em\u003e treatment on patients with HCC in the high-low-risk group, we found that the high-low-risk group in the GSE10141 dataset had a significant difference in immunotherapy (\u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/em\u003e). Compared to the high-risk group, the low-risk group fared better when treated with \u003cem\u003eCTLA4\u003c/em\u003e alone, \u003cem\u003ePD-1\u003c/em\u003e alone, or both in combination (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC-F). However, the response of HCC patients in the GSE14520 and GSE27150 datasets to immunotherapy is different from that in the GSE10141 dataset. Only the high-risk and low-risk groups of HCC patients in the double negative group in the GSE14520 dataset are statistically different. There was no difference in immunotherapy efficacy between high-risk and low-risk HCC patients in the GSE27150 dataset (Supplementary Fig.\u0026nbsp;7). In conclusion, our model can predict the differences in immune status and immunotherapy efficacy between high-risk and low-risk groups of HCC patients for most of the datasets investigated here.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eTumor immunity-related biological features are associated with the prediction results\u003c/h2\u003e \u003cp\u003eSimilarly, our model is advantageous for assessing the association between risk differences and immune infiltration in HCC patients. Compared to the low-risk group, the high-risk group had more immune cells, such as NK cells, monocytic lineage, endothelial cell, T cells, myeloid dendritic cells, but the low-risk group had more neutrophils (\u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/em\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA, Supplementary Fig.\u0026nbsp;8). Furthermore, neutrophils were negatively correlated with the risk of HCC patients, and other immune cells were positively correlated with the risk of HCC patients (\u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/em\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWhile tumor mutation burden (TMB) has been considered as a biomarker for predicting cancer progression and treatment outcome [22], physiological mismatch repair (MMR) proteins aim to detect and correct errors in mismatched nucleotides, the defects of which can lead to microsatellite instability (MSI) in DNA [23]. Therefore, we also analyzed the role of our predictive model in TMB. The TMB of HCC patients was divided into high-TMB and low-TMB groups, and we found that the survival rate of the high-TMB group was significantly lower than that of the low-TMB group (\u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/em\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC, Supplementary Table\u0026nbsp;6). When the patient risk score was combined TMB, we found that HCC patients with high-TMB in the high-risk group had a lower survival rate, while HCC patients with low-TMB in the high-risk group had a lower survival rate than those in the low-risk group (\u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/em\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD), suggesting that the survival risk of HCC patients and TMB are independent of each other. We analyzed the correlation between HCC risk score, TMB, MSI and immune cells in HCC patients. According to Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eE, MSI was positively correlated with TMB and patient risk score, except for neutrophils, HCC patients risk scores were positively correlated with most immune cells. In fact, the survival risk of HCC patients and TMB are indeed independent of each other. These results indicate that NMIRGs can be used to predict the prognostic risk of HCC patients, and the risk is positively correlated with TMB and MSI, and the high risk is closely correlated with immune-related processes.\u003c/p\u003e \u003cp\u003eWe further analyzed the functions or pathways of enrichment among the high-risk and low-risk groups by GSEA enrichment method. As shown from in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eF, five biological pathways including cell cycle, cytokine-cytokine receptor interaction, ECM receptor interaction, hematopoietic cell lineage and neuroactive ligand-receptor interaction were enriched in the high-risk group Five biological processes including cytochrome p450, fatty acid metabolism, glycine, serine, and threonine metabolism, limonene and degradation, and primary bile acid biosynthesis were enriched in the low-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eG). This suggests that immune-related biological processes are mainly enriched in the high-risk group of HCC patients, whereas metabolism-related processes are mainly enriched in the low-risk group of HCC patients.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eComparison with previous HCC prognostic models\u003c/h2\u003e \u003cp\u003eFinally, we compared the prediction score of the external validation set GSE14520 with the models already reported in the references, including Zhang_signature [24](Chemokines and chemokine receptors related three-gene model), Cui_signature [25] (Immune index related ten-gene model), Su_signature [26] (Immunogenic cell death related three-gene model), Lu_signature[27](inflammatory response and immunity status related seven-gene model). Using the same calculation method, HCC samples were divided into high-risk and low-risk groups, and the survival analysis showed that our model is better than Zhang_signature and Su_signature, and is at least comparable to Cui_signature and Lu_signature (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA-E). Furthermore, the ROC curve analysis showed that our NMIRGs_signature indeed showed better performance, where the area under ROC curve for 1-year, 3-year and 5-year survival rates in our model were 0.780, 0.695 and 0.710, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eF-J). In comparison with other models, the C-index of our model was 0.697, which is also higher than Zhang_signature (C-index\u0026thinsp;=\u0026thinsp;0.595), Cui_signature (C-index\u0026thinsp;=\u0026thinsp;0.587), Su_signature (C-index\u0026thinsp;=\u0026thinsp;0.621), Lu_signature (C-index\u0026thinsp;=\u0026thinsp;0.622) (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eK). These results indicate that our model has a high accuracy in predicting the prognosis of HCC patients.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eNMRIGs signature were differentially expressed in HCC patients\u0026rsquo; tissues and cells\u003c/h2\u003e \u003cp\u003eThe expression of these nine NMRIGs signature genes used in the construction of the model was searched in the HPA online database. We found that the protein of GHR, HDAC1, MAPT, HSP90AA1, CHGA and STC1 were expressed in the tissues of HCC patients (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA-E), while RAC3 was expressed in cholangiocarcinoma, but was not detected in HCC. And GAL was not detected in HCC patients (Supplementary Table\u0026nbsp;7). What\u0026rsquo;s more, the protein of GHR, HDAC1, MAPT, HSP90AA1, CHGA were significantly correlated with the prognosis of HCC patients (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eF-J). We analyzed liver cancer cell lines in the GDSC database and found that HuH7 cells were more resistant to HCC therapeutics than other HCC cell lines (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eK, Supplementary Table\u0026nbsp;8). The relative expression levels of NMRIGs signature genes mRNA were expressed in different degrees in Huh7 cell lines, among which the relative expression level of GHR gene was the highest (\u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/em\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eL), which was consistent with the results of immunohistochemistry.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eHCC remains a challenge to global health. Although immunotherapy has made a breakthrough, the treatment outcome is not the same due to the different of tumor microenvironment, and there is still a significant proportion of HCC patients still with poor prognosis [12]. Therefore, it is urgent to investigate its effective prognostic indicators. Given the role of nucleotide metabolism in HCC progression and the immune microenvironment [3,5,6,9]. The combination of both predictors has a higher predictability[28]. Therefore, this study integrates nucleotide metabolism and immune-related differential genes and HCC patients clinical features, established the prognosis prediction model of HCC patients, in order to help clinicians more accurately evaluate the prognosis of HCC patients and provide personalized treatment and find new prognostic markers.\u003c/p\u003e \u003cp\u003eWe first classified HCC patients into C1 and C2 subtypes by NMF clustering, and found that C1 cluster had high expression of nucleotide metabolism and immune-related genes, low tumor purity, more immune cell infiltration, and overall survival (OS) and progression-free survival were significantly lower than C2. This suggests that high expression of NMIRGs is associated with poor prognosis and immune infiltration in HCC patients, and has predictive potential. Therefore, next, we screened out 9 nucleotide metabolism and immune signature genes to construct a prognostic model of HCC. \u003cem\u003eHSP90AA1\u003c/em\u003e, \u003cem\u003eHDAC1\u003c/em\u003e, \u003cem\u003eRAC3\u003c/em\u003e, \u003cem\u003eSTC1\u003c/em\u003e, \u003cem\u003eMAPT\u003c/em\u003e and \u003cem\u003eBTC\u003c/em\u003e genes are high-risk genes, and \u003cem\u003eGHR\u003c/em\u003e gene is a low-risk gene. Receptor-associated coactivator 3 (\u003cem\u003eRAC3\u003c/em\u003e) is an oncogene that is highly expressed in many malignant tumors and regulates cell cycle progression and cell proliferation, invasion, and migration [29]. In fact, RAC3 only showed strong staining in cholangiocarcinoma. Betacellulin (\u003cem\u003eBTC\u003c/em\u003e) proteins are ligands for EGF receptors, and binding to EGFR, ERBB 4, and other members of the EGF receptor family members can effectively promote angiogenesis [30]. RT-PCR showed that BTC gene was expressed in HuH7 cell line. Chromogranin A (\u003cem\u003eCHGA\u003c/em\u003e) may play a role in early tumor progression [31], it is mentioned in most liver cancer prediction models, and we found that it is indeed differentially expressed in HCC patients and is associated with the prognosis of patients with liver cancer. Human Heat Shock Protein 90 Alpha Family Class A Member 1 (\u003cem\u003eHSP90AA1)\u003c/em\u003e [32], microtubule-associated protein tau (\u003cem\u003eMAPT\u003c/em\u003e) [33], and histone deacetylase 1 (\u003cem\u003eHDAC1\u003c/em\u003e) [34] are associated with a poor prognosis in HCC. Knockdown of growth hormone receptor (\u003cem\u003eGHR\u003c/em\u003e) increases the sensitivity of HCC cells to sorafenib [35]. We also detected differential expression of these genes in HCC cells and tissues. In addition, the AUC of the nomogram reached 0.803, indicating that our model has good predictive performance, and these genes are expected to become potential targets of HCC. According to the expression of NMRIGs signature genes, personalized guidance for HCC patients can be targeted to reduce the burden of HCC patients.\u003c/p\u003e \u003cp\u003eWe also evaluated the predictive performance of the model by immune invasion and the effect of immunotherapy. Because of studies have shown that tumor cells evade immune killing of the immune system by altering the local immune microenvironment [36]. Therefore, changes in the immune microenvironment are critical for assessing tumor progression. We found that our model could well evaluate the correlation between the difference in prognostic risk of HCC patients and immune checkpoint genes and immune infiltration, and the changes in neutrophils were also consistent with the negative correlation between the number of tumor- infiltrating neutrophils and the number of T cells reported in the reference [37]. In addition, when evaluating the effect of \u003cem\u003ePD-1\u003c/em\u003e and \u003cem\u003eCTLA-4\u003c/em\u003e immunotherapy in high-risk and low-risk HCC patients, the results of the external validation dataset were also in line with expectations. There are differences in the effectiveness of immunotherapy in different HCC patients, so the efficacy of immunotherapy is not the same in different datasets. For example, the GSE10141 dataset included more HCC patients with high alcohol consumption than the GSE14520 dataset, and our model showed that they had better immunotherapy outcomes for \u003cem\u003ePD-1\u003c/em\u003e and \u003cem\u003eCTLA-4\u003c/em\u003e in the GSE10141 dataset.\u003c/p\u003e \u003cp\u003eIn order to explore the potential regulatory mechanism of NMRIGs signature on HCC, we performed GSEA enrichment analysis in the high-risk and low-risk groups, and found that the enrichment of immune-related processes was mainly in the high-risk group of HCC patients, while the enrichment of metabolism-related processes was mainly in the low-risk group. This is an interesting result, which shows that the analysis of HCC progression cannot only consider a single feature, and the combined analysis of multiple features can enable the more comprehensive interpretation of the HCC prognosis. TMB and MSI, as indicators of somatic mutation [23], have been used as biomarkers to predict cancer progression and treatment outcome. High TMB indicates increased production of neoantigens, which can affect the efficacy of immunotherapy. Our analysis showed that the risk score of HCC patients was positively correlated not only with MSI and TMB, but also with most immune cell abundances (except neutrophils). Nonetheless, the NMIRGs risk score and TMB are largely independent to each other in predicting HCC prognosis.\u003c/p\u003e \u003cp\u003eIn summary, we found that NMIRGs can classify HCC and thus established a predictive model with advantages in predicting survival, assessing clinical relevance, prognostic risk, and immune microenvironment in HCC patients. Potential targets associated with HCC prognosis were also discovered. Most importantly, our predictive model has higher accuracy in predicting the prognosis of HCC patients than some single-feature predictive models.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn conclusion, we have constructed a clinical model of 9 nucleotide metabolic and immune signature genes that predicting HCC prognosis, which has advantages in predicting the risk and immune microenvironment of HCC patients. It is expected that this model will help clinicians to more accurately assess the prognosis of HCC patients and provide personalized treatment options. In addition, the differential expression of NMIRGs signature genes suggests their potential as biomarkers, providing a new perspective for the diagnosis and treatment of HCC. However, it should be noted that our study is based on re-analysis of public data, and therefore further analysis and validation through experimental studies and prospective clinical investigations are needed.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eHCC: hepatocellular carcinoma\u003c/p\u003e\n\u003cp\u003eLIHC: Liver hepatocellular carcinoma\u003c/p\u003e\n\u003cp\u003eNMIRGs: nucleotide metabolism and immune-related genes\u003c/p\u003e\n\u003cp\u003eC1: cluster 1\u003c/p\u003e\n\u003cp\u003eDEGs: Differential Expressed Genes\u003c/p\u003e\n\u003cp\u003eTMB: Tumor mutation load\u003c/p\u003e\n\u003cp\u003eMSI: mitochondrial microsatellite instability\u003c/p\u003e\n\u003cp\u003eTCGA: The Cancer Genome Atlas\u003c/p\u003e\n\u003cp\u003eGEO: Gene Expression Omnibus\u003c/p\u003e\n\u003cp\u003eDAVID: The Database for Annotation, Visualization and Integrated Discovery\u003c/p\u003e\n\u003cp\u003eLASSO: Least Absolute Shrinkage and Selection Operator\u003c/p\u003e\n\u003cp\u003eMCPcounter: Microenvironment Cell Populations counter\u003c/p\u003e\n\u003cp\u003eESTIMATE: Estimation of STromal and Immune cells in MAlignant Tumour tissues using Expression\u003c/p\u003e\n\u003cp\u003eGSEA: Gene Set EnrichmentAnalysis\u003c/p\u003e\n\u003cp\u003eNMF: Non-Negative Matrix Factorization\u003c/p\u003e\n\u003cp\u003ePFS: progression-free survival\u003c/p\u003e\n\u003cp\u003eDCA: Decision Curve Analysis\u003c/p\u003e\n\u003cp\u003eROC: Receiver Operating Characteristic\u003c/p\u003e\n\u003cp\u003eRAC3: Receptor-associated coactivator 3\u003c/p\u003e\n\u003cp\u003eHSP90AA1: Heat Shock Protein 90 Alpha Family Class A Member 1\u003c/p\u003e\n\u003cp\u003eBTC: Betacellulin\u003c/p\u003e\n\u003cp\u003eMAPT: microtubule-associated protein tau\u003c/p\u003e\n\u003cp\u003eHDAC1: histone deacetylase 1\u003c/p\u003e\n\u003cp\u003eSTC1: stanniocalcin-1\u003c/p\u003e\n\u003cp\u003eCHGA: Chromogranin A\u003c/p\u003e\n\u003cp\u003eGAL: galactoside\u003c/p\u003e\n\u003cp\u003eGHR: growth hormone receptor\u003c/p\u003e\n\u003cp\u003ePD-1: programmed cell death protein 1\u003c/p\u003e\n\u003cp\u003eCTLA-4: Cytotoxic T lymphocyte associate protein-4\u003c/p\u003e\n\u003cp\u003eADORA2A: Adenosine Receptor A2a\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthical Statement:\u003c/strong\u003e Not applicable, due to it is a re-analysis of public data and no new human and animal subjects are involved.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions:\u0026nbsp;\u003c/strong\u003eConceptualization, X.W.; Methodology, X.W., Y.Z.; Formal Analysis, X.W.; Data Curation, X.W.; Writing \u0026ndash; Original Draft Preparation, X.W.; Writing \u0026ndash; Review \u0026amp; Editing, Y.Z., Q.C.; Supervision, Y.Z., Q.C.; Funding Acquisition, Y.Z.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e: This work was supported by the National Natural Science Foundation of China (32222020 to Yuan Zhou).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInformed Consent Statement:\u003c/strong\u003e Not applicable, due to it is a re-analysis of public data and no new human and animal subjects are involved.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement\u003c/strong\u003e: The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of interest:\u0026nbsp;\u003c/strong\u003eThere is no conflicts of interest in this study.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eVogel, A.; Meyer, T.; Sapisochin, G.; Salem, R.; Saborowski, A. Hepatocellular carcinoma. \u003cem\u003eLancet \u003c/em\u003e\u003cstrong\u003e2022\u003c/strong\u003e, \u003cem\u003e400\u003c/em\u003e, 1345-1362, doi:10.1016/S0140-6736(22)01200-4.\u003c/li\u003e\n\u003cli\u003eAnwanwan, D.; Singh, S.K.; Singh, S.; Saikam, V.; Singh, R. Challenges in liver cancer and possible treatment approaches. \u003cem\u003eBiochim Biophys Acta Rev Cancer \u003c/em\u003e\u003cstrong\u003e2020\u003c/strong\u003e, \u003cem\u003e1873\u003c/em\u003e, 188314, doi:10.1016/j.bbcan.2019.188314.\u003c/li\u003e\n\u003cli\u003eMa, J.; Zhong, M.; Xiong, Y.; Gao, Z.; Wu, Z.; Liu, Y.; Hong, X. Emerging roles of nucleotide metabolism in cancer development: progress and prospect. \u003cem\u003eAging (Albany NY) \u003c/em\u003e\u003cstrong\u003e2021\u003c/strong\u003e, \u003cem\u003e13\u003c/em\u003e, 13349-13358, doi:10.18632/aging.202962.\u003c/li\u003e\n\u003cli\u003eWu, H.L.; Gong, Y.; Ji, P.; Xie, Y.F.; Jiang, Y.Z.; Liu, G.Y. Targeting nucleotide metabolism: a promising approach to enhance cancer immunotherapy. \u003cem\u003eJ Hematol Oncol \u003c/em\u003e\u003cstrong\u003e2022\u003c/strong\u003e, \u003cem\u003e15\u003c/em\u003e, 45, doi:10.1186/s13045-022-01263-x.\u003c/li\u003e\n\u003cli\u003eJain, S.; Jacobson, K.A. Purinergic Signaling in Liver Pathophysiology. \u003cem\u003eFront Endocrinol (Lausanne) \u003c/em\u003e\u003cstrong\u003e2021\u003c/strong\u003e, \u003cem\u003e12\u003c/em\u003e, 718429, doi:10.3389/fendo.2021.718429.\u003c/li\u003e\n\u003cli\u003eMullen, N.J.; Singh, P.K. Nucleotide metabolism: a pan-cancer metabolic dependency. \u003cem\u003eNat Rev Cancer \u003c/em\u003e\u003cstrong\u003e2023\u003c/strong\u003e, \u003cem\u003e23\u003c/em\u003e, 275-294, doi:10.1038/s41568-023-00557-7.\u003c/li\u003e\n\u003cli\u003eXu, D.; Tian, Y.; Xia, Q.; Ke, B. The cGAS-STING Pathway: Novel Perspectives in Liver Diseases. \u003cem\u003eFront Immunol \u003c/em\u003e\u003cstrong\u003e2021\u003c/strong\u003e, \u003cem\u003e12\u003c/em\u003e, 682736, doi:10.3389/fimmu.2021.682736.\u003c/li\u003e\n\u003cli\u003eDonne, R.; Lujambio, A. The liver cancer immune microenvironment: Therapeutic implications for hepatocellular carcinoma. \u003cem\u003eHepatology \u003c/em\u003e\u003cstrong\u003e2023\u003c/strong\u003e, \u003cem\u003e77\u003c/em\u003e, 1773-1796, doi:10.1002/hep.32740.\u003c/li\u003e\n\u003cli\u003eHan, T.; Kang, D.; Ji, D.; Wang, X.; Zhan, W.; Fu, M.; Xin, H.B.; Wang, J.B. How does cancer cell metabolism affect tumor migration and invasion? \u003cem\u003eCell Adh Migr \u003c/em\u003e\u003cstrong\u003e2013\u003c/strong\u003e, \u003cem\u003e7\u003c/em\u003e, 395-403, doi:10.4161/cam.26345.\u003c/li\u003e\n\u003cli\u003eShen, B.; Zhang, G.; Liu, Y.; Wang, J.; Jiang, J. Identification and Analysis of Immune-Related Gene Signature in Hepatocellular Carcinoma. \u003cem\u003eGenes (Basel) \u003c/em\u003e\u003cstrong\u003e2022\u003c/strong\u003e, \u003cem\u003e13\u003c/em\u003e, 1834, doi:10.3390/genes13101834.\u003c/li\u003e\n\u003cli\u003eChen, Y.; Shen, C.; Wu, J.; Yan, X.; Huang, Q. Role of immune related genes in predicting prognosis and immune response in patients with hepatocellular carcinoma. \u003cem\u003eJ Biochem Mol Toxicol \u003c/em\u003e\u003cstrong\u003e2023\u003c/strong\u003e, e23519, doi:10.1002/jbt.23519.\u003c/li\u003e\n\u003cli\u003eXu, Y.J.; He, M.K.; Liu, S.; Huang, L.C.; Bu, X.Y.; Kan, A.; Shi, M. Construction of a single nucleotide variant score-related gene-based prognostic model in hepatocellular carcinoma: analysis of multi-independent databases and validation in vitro. \u003cem\u003eCancer Cell Int \u003c/em\u003e\u003cstrong\u003e2021\u003c/strong\u003e, \u003cem\u003e21\u003c/em\u003e, 610, doi:10.1186/s12935-021-02321-z.\u003c/li\u003e\n\u003cli\u003eWei, T.; Liu, J.; Ma, S.; Wang, M.; Yuan, Q.; Huang, A.; Wu, Z.; Shang, D.; Yin, P. A Nucleotide Metabolism-Related Gene Signature for Risk Stratification and Prognosis Prediction in Hepatocellular Carcinoma Based on an Integrated Transcriptomics and Metabolomics Approach. \u003cem\u003eMetabolites \u003c/em\u003e\u003cstrong\u003e2023\u003c/strong\u003e, \u003cem\u003e13\u003c/em\u003e, 1116, doi:10.3390/metabo13111116.\u003c/li\u003e\n\u003cli\u003eSherman, B.T.; Hao, M.; Qiu, J.; Jiao, X.; Baseler, M.W.; Lane, H.C.; Imamichi, T.; Chang, W. DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update). \u003cem\u003eNucleic Acids Res \u003c/em\u003e\u003cstrong\u003e2022\u003c/strong\u003e, \u003cem\u003e50\u003c/em\u003e, W216-W221, doi:10.1093/nar/gkac194.\u003c/li\u003e\n\u003cli\u003eBhattacharya, S.; Dunn, P.; Thomas, C.G.; Smith, B.; Schaefer, H.; Chen, J.; Hu, Z.; Zalocusky, K.A.; Shankar, R.D.; Shen-Orr, S.S.; et al. ImmPort, toward repurposing of open access immunological assay data for translational and clinical research. \u003cem\u003eSci Data \u003c/em\u003e\u003cstrong\u003e2018\u003c/strong\u003e, \u003cem\u003e5\u003c/em\u003e, 180015, doi:10.1038/sdata.2018.15.\u003c/li\u003e\n\u003cli\u003eYoshihara, K.; Shahmoradgoli, M.; Martinez, E.; Vegesna, R.; Kim, H.; Torres-Garcia, W.; Trevino, V.; Shen, H.; Laird, P.W.; Levine, D.A.; et al. Inferring tumour purity and stromal and immune cell admixture from expression data. \u003cem\u003eNat Commun \u003c/em\u003e\u003cstrong\u003e2013\u003c/strong\u003e, \u003cem\u003e4\u003c/em\u003e, 2612, doi:10.1038/ncomms3612.\u003c/li\u003e\n\u003cli\u003eBecht, E.; Giraldo, N.A.; Lacroix, L.; Buttard, B.; Elarouci, N.; Petitprez, F.; Selves, J.; Laurent-Puig, P.; Sautes-Fridman, C.; Fridman, W.H.; et al. Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression. \u003cem\u003eGenome Biol \u003c/em\u003e\u003cstrong\u003e2016\u003c/strong\u003e, \u003cem\u003e17\u003c/em\u003e, 218, doi:10.1186/s13059-016-1070-5.\u003c/li\u003e\n\u003cli\u003eMoons, K.G.M.; Altman, D.G.; Reitsma, J.B.; Ioannidis, J.P.A.; Macaskill, P.; Steyerberg, E.W.; Vickers, A.J.; Ransohoff, D.F.; Collins, G.S. Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD): Explanation and Elaboration. \u003cem\u003eAnnals of Internal Medicine \u003c/em\u003e\u003cstrong\u003e2015\u003c/strong\u003e, \u003cem\u003e162\u003c/em\u003e, W1-W73, doi:10.7326/m14-0698.\u003c/li\u003e\n\u003cli\u003eHu, F.F.; Liu, C.J.; Liu, L.L.; Zhang, Q.; Guo, A.Y. Expression profile of immune checkpoint genes and their roles in predicting immunotherapy response. \u003cem\u003eBrief Bioinform \u003c/em\u003e\u003cstrong\u003e2021\u003c/strong\u003e, \u003cem\u003e22\u003c/em\u003e, bbaa176, doi:10.1093/bib/bbaa176.\u003c/li\u003e\n\u003cli\u003eLong, M.; Zhou, Z.; Wei, X.; Lin, Q.; Qiu, M.; Zhou, Y.; Chen, P.; Jiang, Y.; Wen, Q.; Liu, Y.; et al. A novel risk score based on immune-related genes for hepatocellular carcinoma as a reliable prognostic biomarker and correlated with immune infiltration. \u003cem\u003eFront Immunol \u003c/em\u003e\u003cstrong\u003e2022\u003c/strong\u003e, \u003cem\u003e13\u003c/em\u003e, 1023349, doi:10.3389/fimmu.2022.1023349.\u003c/li\u003e\n\u003cli\u003eCharoentong, P.; Finotello, F.; Angelova, M.; Mayer, C.; Efremova, M.; Rieder, D.; Hackl, H.; Trajanoski, Z. Pan-cancer Immunogenomic Analyses Reveal Genotype-Immunophenotype Relationships and Predictors of Response to Checkpoint Blockade. \u003cem\u003eCell Rep \u003c/em\u003e\u003cstrong\u003e2017\u003c/strong\u003e, \u003cem\u003e18\u003c/em\u003e, 248-262, doi:10.1016/j.celrep.2016.12.019.\u003c/li\u003e\n\u003cli\u003eChoucair, K.; Morand, S.; Stanbery, L.; Edelman, G.; Dworkin, L.; Nemunaitis, J. TMB: a promising immune-response biomarker, and potential spearhead in advancing targeted therapy trials. \u003cem\u003eCancer Gene Ther \u003c/em\u003e\u003cstrong\u003e2020\u003c/strong\u003e, \u003cem\u003e27\u003c/em\u003e, 841-853, doi:10.1038/s41417-020-0174-y.\u003c/li\u003e\n\u003cli\u003eGabbia, D.; De Martin, S. Tumor Mutational Burden for Predicting Prognosis and Therapy Outcome of Hepatocellular Carcinoma. \u003cem\u003eInt J Mol Sci \u003c/em\u003e\u003cstrong\u003e2023\u003c/strong\u003e, \u003cem\u003e24\u003c/em\u003e, 3441, doi:10.3390/ijms24043441.\u003c/li\u003e\n\u003cli\u003eZhang, Z.; Mao, M.; Wang, F.; Zhang, Y.; Shi, J.; Chang, L.; Wu, X.; Zhang, Z.; Xu, P.; Lu, S. Comprehensive analysis and immune landscape of chemokines- and chemokine receptors-based signature in hepatocellular carcinoma. \u003cem\u003eFront Immunol \u003c/em\u003e\u003cstrong\u003e2023\u003c/strong\u003e, \u003cem\u003e14\u003c/em\u003e, 1164669, doi:10.3389/fimmu.2023.1164669.\u003c/li\u003e\n\u003cli\u003eCui, X.; Han, L.; Cui, L.; Fu, G.; Liu, E.; Wang, D.; Song, B.; Zhang, Y.; Zhou, W.; Wang, H.; et al. Immune index: A gene and cell prognostic signature for immunotherapy response prediction in hepatocellular carcinoma. \u003cem\u003ePharmacol Res \u003c/em\u003e\u003cstrong\u003e2023\u003c/strong\u003e, \u003cem\u003e187\u003c/em\u003e, 106583, doi:10.1016/j.phrs.2022.106583.\u003c/li\u003e\n\u003cli\u003eSu, D.; Zhang, Z.; Xia, F.; Liang, Q.; Liu, Y.; Liu, W.; Xu, Z. ICD-related risk model predicts the prognosis and immunotherapy response of patients with liver cancer. \u003cem\u003eFront Pharmacol \u003c/em\u003e\u003cstrong\u003e2023\u003c/strong\u003e, \u003cem\u003e14\u003c/em\u003e, 1202823, doi:10.3389/fphar.2023.1202823.\u003c/li\u003e\n\u003cli\u003eLu, G.; Du, R.; Feng, B.; Wang, J.; Zhang, F.; Pei, J.; Wang, Y.; Shang, Y. A Novel Gene Signature Associated with Inflammatory Responses and Immune Status Assists in Prognosis and Intervention for Patients with HCC. \u003cem\u003eJ Inflamm Res \u003c/em\u003e\u003cstrong\u003e2022\u003c/strong\u003e, \u003cem\u003e15\u003c/em\u003e, 6729-6743, doi:10.2147/JIR.S390113.\u003c/li\u003e\n\u003cli\u003eXu, Y.Y.; Qin, Z.; Gao, J.M.; Yang, Y.; Lu, Y.Z.; Yu, F.Y.; Lv, Y.; Sun, Z.W.; Zhang, J.Y.; Tang, J.N. Combination of Neutrophil Count and Gensini Score as a Prognostic Marker in Patients with ACS and Uncontrolled T2DM Undergoing PCI. \u003cem\u003eCardiovasc Innov App \u003c/em\u003e\u003cstrong\u003e2023\u003c/strong\u003e, \u003cem\u003e8\u003c/em\u003e, doi:10.15212/Cvia.2023.0051.\u003c/li\u003e\n\u003cli\u003ede Curtis, I. The Rac3 GTPase in Neuronal Development, Neurodevelopmental Disorders, and Cancer. \u003cem\u003eCells \u003c/em\u003e\u003cstrong\u003e2019\u003c/strong\u003e, \u003cem\u003e8\u003c/em\u003e, 1063, doi:10.3390/cells8091063.\u003c/li\u003e\n\u003cli\u003eDahlhoff, M.; Wolf, E.; Schneider, M.R. The ABC of BTC: structural properties and biological roles of betacellulin. \u003cem\u003eSemin Cell Dev Biol \u003c/em\u003e\u003cstrong\u003e2014\u003c/strong\u003e, \u003cem\u003e28\u003c/em\u003e, 42-48, doi:10.1016/j.semcdb.2014.01.002.\u003c/li\u003e\n\u003cli\u003eMahata, S.K.; Corti, A. Chromogranin A and its fragments in cardiovascular, immunometabolic, and cancer regulation. \u003cem\u003eAnn N Y Acad Sci \u003c/em\u003e\u003cstrong\u003e2019\u003c/strong\u003e, \u003cem\u003e1455\u003c/em\u003e, 34-58, doi:10.1111/nyas.14249.\u003c/li\u003e\n\u003cli\u003eZuehlke, A.D.; Beebe, K.; Neckers, L.; Prince, T. Regulation and function of the human HSP90AA1 gene. \u003cem\u003eGene \u003c/em\u003e\u003cstrong\u003e2015\u003c/strong\u003e, \u003cem\u003e570\u003c/em\u003e, 8-16, doi:10.1016/j.gene.2015.06.018.\u003c/li\u003e\n\u003cli\u003eChang, C.W.; Shao, E.; Mucke, L. Tau: Enabler of diverse brain disorders and target of rapidly evolving therapeutic strategies. \u003cem\u003eScience \u003c/em\u003e\u003cstrong\u003e2021\u003c/strong\u003e, \u003cem\u003e371\u003c/em\u003e, eabb8255, doi:10.1126/science.abb8255.\u003c/li\u003e\n\u003cli\u003eCai, D.; Yuan, X.; Cai, D.Q.; Li, A.; Yang, S.; Yang, W.; Duan, J.; Zhuo, W.; Min, J.; Peng, L.; et al. Integrative analysis of lactylation-related genes and establishment of a novel prognostic signature for hepatocellular carcinoma. \u003cem\u003eJ Cancer Res Clin Oncol \u003c/em\u003e\u003cstrong\u003e2023\u003c/strong\u003e, \u003cem\u003e149\u003c/em\u003e, 11517-11530, doi:10.1007/s00432-023-04947-0.\u003c/li\u003e\n\u003cli\u003eGao, S.; Ni, Q.; Wu, X.; Cao, T. GHR knockdown enhances the sensitivity of HCC cells to sorafenib. \u003cem\u003eAging (Albany NY) \u003c/em\u003e\u003cstrong\u003e2020\u003c/strong\u003e, \u003cem\u003e12\u003c/em\u003e, 18127-18136, doi:10.18632/aging.103625.\u003c/li\u003e\n\u003cli\u003eGiannini, E.G.; Aglitti, A.; Borzio, M.; Gambato, M.; Guarino, M.; Iavarone, M.; Lai, Q.; Levi Sandri, G.B.; Melandro, F.; Morisco, F.; et al. Overview of Immune Checkpoint Inhibitors Therapy for Hepatocellular Carcinoma, and The ITA.LI.CA Cohort Derived Estimate of Amenability Rate to Immune Checkpoint Inhibitors in Clinical Practice. \u003cem\u003eCancers (Basel) \u003c/em\u003e\u003cstrong\u003e2019\u003c/strong\u003e, \u003cem\u003e11\u003c/em\u003e, 1689, doi:10.3390/cancers11111689.\u003c/li\u003e\n\u003cli\u003eHe, G.; Zhang, H.; Zhou, J.; Wang, B.; Chen, Y.; Kong, Y.; Xie, X.; Wang, X.; Fei, R.; Wei, L.; et al. Peritumoural neutrophils negatively regulate adaptive immunity via the PD-L1/PD-1 signalling pathway in hepatocellular carcinoma. \u003cem\u003eJ Exp Clin Cancer Res \u003c/em\u003e\u003cstrong\u003e2015\u003c/strong\u003e, \u003cem\u003e34\u003c/em\u003e, 141, doi:10.1186/s13046-015-0256-0.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"hepatocellular carcinoma, nucleotide metabolism, immune microenvironment, prognostic risk, biomarker","lastPublishedDoi":"10.21203/rs.3.rs-6048868/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6048868/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003eGiven the persistent poor prognosis of hepatocellular carcinoma (HCC), it is imperative to establish a multi-feature prognostic prediction model.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eLIHC patient data were obtained from TCGA, and LIHC patients were clustered by NMF and a nucleotide metabolism and immune-related genes (NMIRGs) prognostic risk model was built byby cox regression and Lasso regression, the gene expression was detected by immunohistochemistry and qPCR. The GSE14520, GSE10141 and GSE27150 datasets were used as external validation sets. The prognostic value of the model was evaluated by immune checkpoint genes, immune infiltration, functional enrichment analysis (GSEA) and tumor mutation load, and its potential mechanism was explored. The accuracy of our model was evaluated by the C-index value, compared to the reported prediction model.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003eThe HCC patients were grouped into two subtypes based on nucleotide metabolism and immune differentially expressed genes (DEGs), with cluster 1 (C1) having more DEGs and worse prognosis than cluster 2 (C2). \u0026nbsp;A NMIRGs signature gene prognostic model was established, with HSP90AA1, HDAC1, STC1, MAPT and CHGA genes as high-risk genes and GHR genes as low-risk genes. We found that the stage and risk score of HCC patients were independent prognostic factors. The model also assess the prognostic risk of patients in relation to immune checkpoint genes and immune infiltration. Further analysis showed that our model's risk score from our model and tumor mutation burden (TMB) were largely independent in predicting the efficacy of immunotherapy and were closely related to immune-related processes. Finally, the model was found to have a higher C-index and accuracy in predicting the prognosis of HCC patients.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions: \u003c/strong\u003eOur study provides a helpful predictive model for the prognosis of HCC and its immune microenvironment. In addition, it identifies potential new biomarkers associated with HCC prognosis.\u003c/p\u003e","manuscriptTitle":"Risk model of nucleotide metabolism and immunity signature genes in predicting of hepatocellular carcinoma prognosis and immune microenvironment","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-25 09:43:51","doi":"10.21203/rs.3.rs-6048868/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"549bc0ab-a2bf-4365-b6bf-39e522298f2f","owner":[],"postedDate":"March 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-05-09T20:52:47+00:00","versionOfRecord":[],"versionCreatedAt":"2025-03-25 09:43:51","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6048868","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6048868","identity":"rs-6048868","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}