Comprehensive Analysis of RAPGEF2 for Predicting Prognosis and Immunotherapy Response in Patients with Hepatocellular Carcinoma

Abstract

Background: Hepatocellular carcinoma (HCC) is the sixth most common tumor worldwide. Additionally, deletion of RAPGEF2 plays a critical role in CNV and related to tumor immune microenvironment, whereas the prognostic potential of RAPGEF2 in HCC patient needs to be explored.

Methods: We looked for prognostic potential genes in HCC using a variety of R programs. Then, using the LASSO Cox regression, we thoroughly evaluated and integrated the RAPGEF2-related genes from TCGA database. Meanwhile, utilizing TCGA and ICGA databases, the link between RAPGEF2 and immunotherapy response in HCC was studied. In vivo, the effect of RAPGEF2 on tumor development and the capacity of natural killer (NK) cells to recruit were confirmed. To ascertain the connection between RAPGEF2-related genes and the prognosis of HCC, a prognostic model was created and validated.

Result: We demonstrated RAPGEF2 has a differential expression, and patients with deletion of RAPGEF2 gene get shorter survival in HCC. Additionally, the tissues without RAPGEF2 have a weaker ability to recruit the NK cells and response to immunotherapy. After that, we scoured the database for eight RAPGEF2-related genes linked with a better prognosis in HCC patients. Additionally, silencing RAPGEF2 accelerated tumor development in the HCC mouse model and decreased CD56+ NK cell recruitment in HCC tissues. TCGA database was used to classify patients into low- and high-risk categories based on the expression of related genes. Patients in the low-risk group had a significantly greater overall survival than those in the high-risk group (P < 0.001). Meanwhile, the low-risk group demonstrated connections with the NK cell and immunotherapy response. Finally, the prognostic nomogram showed a high sensitivity and specificity for predicting the survival of HCC patients at 1, 2, and 3 years.

Conclusion: The prognostic model based on RAPGEF2 and RAPGEF2-related genes showed an excellent predictive performance in terms of prognosis and immunotherapy response in HCC, therefore establishing a unique prognostic model for clinical assessment of HCC patients.

Figure 1

Expression variation and survival curves of RAPGEF2 genes in hepatocellular carcinoma (HCC). (a) Kaplan-Meier curves for the OS of patients in the different RAPGEF2 expression from TCGA database. (b) The location of copy number variation (CNV) alteration of RAPGEF2 on chromosomes using TCGA datasets. (c) The Kaplan-Meier curves for the OS of patients between the deletion and normal of RAPGEF2 expression from TCGA database. (d) The expression of RAPGEF2 in different tissues.

Figure 2

Relationship among RAPGEF2, immune infiltration tumor immune dysfunction exclusion, and immunotherapy response to HCC patients. (a) The scores of 16 immune cells. (b) The scores of 13 immune-related functions. (c) The value of the tumor immune dysfunction and exclusion (TIDE). (d) The response to immunotherapy (^∗P < 0.05; ^∗∗P < 0.01; ^∗∗∗P < 0.001; ns: no significant).

Figure 3

RAPGEF2 knockdown promotes tumor growth in HCC cancer xenografts. (a) Western blotting examination of MHCC97 cells transfected with either control (Ctrl) or RAPGEF2 shRNA. (b, c) The tumor growth curve and weights of the tumors were determined. (d, e) Fluorescence-activated cell sorting (FACS) study of CD56+CD3-natural killer cells in MHCC97 xenograft tumors with either control (Ctrl) or RAPGEF2 shRNA. Data represents the mean ± SD, n = 10 per group. ^∗P < 0.05, ^∗∗P < 0.01, and ^∗∗∗P < 0.001.

Figure 4

Risk model and the Kaplan-Meier curves for the OS in HCC patients based on RAPGEF2 and its related genes. (a) The univariate Cox regression analysis demonstrated a substantial correlation between the identified genes and clinical prognosis. (b) To cross-validate the error curve, the tuning parameters (log λ) of OS-related proteins were selected. Perpendicular imaginary lines were drawn at the ideal value using the minimum and 1-se criterion. (c) The LASSO coefficient profile of 13 OS-associated genes was drawn together with the perpendicular imaginary line at the value determined by 9-fold cross-validation. (d) Kaplan-Meier curves for the OS of patients in TCGA database who were classified as high-risk or low-risk. (e) ROC analysis using TCGA database. (f) ROC analysis using the ICGC database.

Figure 5

Relationship among risk model, immune infiltration tumor immune dysfunction exclusion, and immunotherapy response to HCC patients. (a) The scores of 16 immune cells. (b) The scores of 13 immune-related functions. (c) The value of the TIDE from TCGA database. (d) The response to immunotherapy from TCGA database. (e) The value of the TIDE from ICGC database. (f) The response to immunotherapy from ICGC database. (^∗P < 0.05; ^∗∗P < 0.01; ^∗∗∗P < 0.001; ns: no significant).

Figure 6

Predictive nomogram construction and validation. (a) The univariate Cox regression analysis's results. (b) The multivariate Cox regression analysis's results. (c) A nomogram for predicting the 1-, 2-, and 3-year OS of patients with HCC. (d) Nomogram calibration curves for OS prediction at 1, 2, and 3 years. (e) ROC analysis using TCGA database. (f) ROC analysis using the ICGC database.