Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Jan 25:14:1097075.
doi: 10.3389/fimmu.2023.1097075. eCollection 2023.

Construction and systematic evaluation of a machine learning-based cuproptosis-related lncRNA score signature to predict the response to immunotherapy in hepatocellular carcinoma

Dingyu Lu  1 , Jian Liao  2 , Hao Cheng  1 , Qian Ma  1 , Fei Wu  1 , Fei Xie  1 , Yingying He  1
Affiliations

Construction and systematic evaluation of a machine learning-based cuproptosis-related lncRNA score signature to predict the response to immunotherapy in hepatocellular carcinoma

Dingyu Lu et al. Front Immunol. .

Abstract

Introduction: Hepatocellular carcinoma (HCC) is a common malignant cancer with a poor prognosis. Cuproptosis and associated lncRNAs are connected with cancer progression. However, the information on the prognostic value of cuproptosis-related lncRNAs is still limited in HCC.

Methods: We isolated the transcriptome and clinical information of HCC from TCGA and ICGC databases. Ten cuproptosis-related genes were obtained and related lncRNAs were correlated by Pearson's correlation. By performing lasso regression, we created a cuproptosis-related lncRNA prognostic model based on the cuproptosis-related lncRNA score (CLS). Comprehensive analyses were performed, including the fields of function, immunity, mutation and clinical application, by various R packages.

Results: Ten cuproptosis-related genes were selected, and 13 correlated prognostic lncRNAs were collected for model construction. CLS was positively or negatively correlated with cancer-related pathways. In addition, cell cycle and immune related pathways were enriched. By performing tumor microenvironment (TME) analysis, we determined that T-cells were activated. High CLS had more tumor characteristics and may lead to higher invasiveness and treatment resistance. Three genes (TP53, CSMD1 and RB1) were found in high CLS samples with more mutational frequency. More amplification and deletion were detected in high CLS samples. In clinical application, a CLS-based nomogram was constructed. 5-Fluorouracil, gemcitabine and doxorubicin had better sensitivity in patients with high CLS. However, patients with low CLS had better immunotherapeutic sensitivity.

Conclusion: We created a prognostic CLS signature by machine learning, and we comprehensively analyzed the signature in the fields of function, immunity, mutation and clinical application.

Keywords: cuproptosis-related lncRNA score; hepatocellular carcinoma; immunotherapy; machine learning; prognostic model.

PubMed Disclaimer

Conflict of interest statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Figures

Figure 1
Figure 1
Identification of cuproptosis-related genes and corresponding lncRNAs. (A) The fold change, mutational frequency, methylation level and Hazard ratio of the ten cuproptosis-related genes. (B) The correlated lncRNAs of the ten cuproptosis-related genes. *P-value < 0.05, **P-value < 0.01, ***P-value < 0.001.
Figure 2
Figure 2
Prognostic signature based on CLS was created. (A) Lasso regression of the cuproptosis-related lncRNAs. (B) Identification of the tuning parameter in Lasso model. (C) The coefficients in Lasso model. (D) The C-index of CLS, stage, age and gender in TCGA and ICGC databases. (E) The AUC of CLS, age, gender and stage in TCGA. (F) The survival status and the expression of the 13 cuproptosis-related lncRNAs of each sample ranked from high to low CLS. (G) The mRNA expression of 13 cuproptosis-related lncRNAs in HCC cell line Hep3B compared to the hepatic stellate cell line LX2. (H) Kaplan-Meier analysis of the high and low CLS patients. (I) The 1-, 3- and 5-year AUC of the prognostic signature. *P-value < 0.05, ***P-value < 0.001.
Figure 3
Figure 3
Construction of a CLS-based nomogram. (A) Univariate Cox regression in TCGA and ICGC cohorts. (B) Multivariate Cox regression in TCGA and ICGC cohorts. (C) Construction of a nomogram by various parameters. (D) Calibration curve of the CLS-based nomogram. (E) AUC analysis for the constructed nomogram. (F) One-year DCA for the nomogram. (G) Three-year DCA for the nomogram. (H) Five-year DCA for the nomogram.
Figure 4
Figure 4
Functional analyses of the CLS model. (A) The correlation between CLS and the Hallmark cancer-related pathways. (B) The enriched items in high CLS samples in Metascape. (C) The enriched items in low CLS samples in Metascape. (D) The top five enriched items in high CLS samples by GSEA. (E) The top five enriched items in low CLS samples by GSEA. (F) The expression of the interested pathways in each sample and the correlation between interested pathways and CLS.
Figure 5
Figure 5
Immune analyses of the CLS model. (A) The expression and correlation between the TICs and CLS. (B) The ESTIMATE score (including immune and stromal score) and tumor purity in high and low CLS samples. (C) The relative expression of six immune checkpoints in high and low CLS samples. (D) The correlation between CLS and immune checkpoints/ESTIMATE. (E) The correlation between the CTA score and the CLS, and the level of CTA score in high and low CLS samples. (F) The correlation between the HRD score and the CLS, and the level of HRD score in high and low CLS samples. (G) The correlation between the intratumor heterogeneity and the CLS, and the level of the intratumor heterogeneity in high and low CLS samples. *P-value < 0.05, **P-value < 0.01, ***P-value < 0.001, ****P-value < 0.0001. ns, not significant.
Figure 6
Figure 6
Mutational analyses of the CLS model. (A) The correlation between the all mutation counts and the CLS, and the number of all mutation counts in high and low CLS samples. (B) The correlation between the non-synonymous mutation counts and the CLS, and the number of all mutation counts in high and low CLS samples. (C) The waterfall plot of the top 20 altered mutation in high and low CLS samples. (D) The differentially mutated genes between high and low CLS samples. (E) The proportion and the types of the TP53 mutation in high and low CLS samples. (F) The number of mutations in five mutational signatures in high CLS samples. (G) The number of mutations in five mutational signatures in low CLS samples. (H) The amplification and deletion frequency in each arms between high and low CLS samples. (I) The total frequency of amplification in high and low CLS samples. (J) The total frequency of deletion in high and low CLS samples. TMB, Tumor mutational burden.
Figure 7
Figure 7
The clinical application of CLS model. (A) The expression and the correlation of the neoantigens in high and low CLS samples. (B) The expression and correlation of the proliferation score in high and low CLS samples. (C) The estimated IC50 of 5-fluorouracil, cisplatin, gemcitabine and doxorubicin in high and low CLS samples. (D) The IPS of each patients with high or low CLS. (E) TIDE analysis of the PD1 and CTLA4 response in patients with high and low CLS. (F) The proportion of the TIDE response in high and low CLS patients. (G) The AUC analysis of the CLS and biomarkers. IPS, Immunophenoscore.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Degasperi E, Colombo M. Distinctive features of hepatocellular carcinoma in non-alcoholic fatty liver disease. Lancet Gastroenterol Hepatol (2016) 1:156–64. doi: 10.1016/S2468-1253(16)30018-8 - DOI - PubMed
    1. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: Cancer J Clin (2018) 68:394–424. doi: 10.3322/caac.21492 - DOI - PubMed
    1. Marasco G, Colecchia A, Colli A, Ravaioli F, Casazza G, Bacchi Reggiani ML, et al. . Role of liver and spleen stiffness in predicting the recurrence of hepatocellular carcinoma after resection. J Hepatol (2019) 70:440–8. doi: 10.1016/j.jhep.2018.10.022 - DOI - PubMed
    1. Bruix J, Qin S, Merle P, Granito A, Huang YH, Bodoky G, et al. . Regorafenib for patients with hepatocellular carcinoma who progressed on sorafenib treatment (RESORCE): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet (2017) 389:56–66. doi: 10.1016/S0140-6736(16)32453-9 - DOI - PubMed
    1. Liu Z, Lin Y, Zhang J, Zhang Y, Li Y, Liu Z, et al. . Molecular targeted and immune checkpoint therapy for advanced hepatocellular carcinoma. J Exp Clin Cancer Res: CR (2019) 38:447. doi: 10.1186/s13046-019-1412-8 - DOI - PMC - PubMed