Page 79 - TD-3-3

Tumor Discovery PTMAP5–hsa-miR-22-3p–KIF2C axis in HCC development

significant miRNA-gene/pseudogene pairs. This threshold enables the screening of numerous genes involved in T-cell-
was set at an R-value of <0.1 and P < 0.05, ensuring robust mediated cytotoxicity and immunotherapy, integrating
statistical significance and reliability in our analysis. genomic mapping data for accurate and thorough analysis.
Through StarBase, we predicted upstream pseudogenes In this study, we investigated the correlation of KIF2C
potentially interacting with hsa-miR-22-3p, identifying with tumor immunology and its expression across various
271 candidates. We then assessed ncRNAs for their binding human cancer immunoisoforms and molecular subtypes.
affinity to miRNAs, predicting lncRNAs that may interact For the TISIDB database operation, the search criteria
with hsa-miR-22-3p. This analysis identified 116 upstream were as follows: (i) Gene Symbol: KIF2C; (ii) Search
lncRNAs in StarBase, with an additional 52 identified Result: KIF2C; and (iii) Categories: Immunomodulator,
through MiRNet. Chemokine, and Subtype.

2.5. The GEPIA database 2.8. Statistical analysis
GEPIA is an online platform that provides RNA sequencing The R language, a free, open-source software, was used
expression data from 9736 cancerous and 8587 healthy tissue for statistical analysis and plotting. One key method
specimens, sourced from the cancer genome atlas (TCGA) was the receiver operating characteristic (ROC) curve, a
and the genotype-tissue expression (GTEx) project. The fundamental tool for assessing binary classification models
platform supports comparative expression analysis, patient by plotting the true-positive rate (sensitivity) against the
survival analysis, and correlation analysis. In pursuit of false-positive rate (1-specificity). The ROC curve aids
a deeper understanding of the role of KIF2C in cancer in determining the optimal diagnostic threshold value,
immunity, we leveraged the comprehensive resources of with the area under the ROC curve (AUC) increasing as
the GEPIA database to scrutinize its intricate associations the curve approaches the upper left corner, indicating a
with immune cell biomarkers, specifically in the context of more accurate model. In addition, we employed LASSO
HCC. This meticulous investigation allowed us to delineate regression analysis, a least-squares estimation method for
potential correlations between KIF2C expression and linear models. LASSO compresses insignificant coefficients,
immune cell signatures, shedding light on its putative role facilitating variable selection and accurate estimation,
in modulating the immune microenvironment in HCC. which helps in screening variables and optimizing models.
2.6. The connection between KIF2C and immune 3. Results
checkpoints in liver cancer
3.1. Screening of significant DEGs in HCC
Immune checkpoints, including PD1/PD-L1 and CTLA-
4, are critical regulators of immune evasion by tumors. We screened two HCC-related datasets, GSE87630 and
To elucidate the potential association between KIF2C GSE45267, from the GEO database, which comprised
and HCC, we analyzed the relationship between KIF2C 181 samples (71 from liver cancer patients and 110 from
expression and the immune checkpoints PD1, PD-L1, and healthy individuals). By analyzing both datasets using the
CTLA-4 utilizing the GEPIA and TIMER databases. Our criteria (adjusted P < 0.05, |logFc| >1), we identified 1,995
findings revealed a notable positive link between KIF2C DEGs. A Venn diagram was employed to analyze common
and these immune checkpoints in HCC, suggesting that DEGs across the datasets, resulting in 346 DEGs. Of
KIF2C may contribute to immune evasion mechanisms in these, 69 were upregulated, and 277 were downregulated
HCC development. (Figure 1A and Table S1).
The following steps outline the GEPIA database 3.2. Construction and analysis of the PPI network
operation: To further explore the interactions among the DEGs,
i. Select “Correlation Analysis” we utilized STRING to construct a PPI network. The
ii. Enter “gene: PDCD1” and “gene: KIF2C” Cytoscape MCODE plugin was then applied to identify key
iii. Select “Used Expression Datasets: LIHC Tumor” clusters within the network, resulting in the identification
iv. Click “plot” of 29 key nodes (Figure 1B).
Repeat these steps for “gene: CD274” and “gene: CTLA-4.”
3.3. Confirmation of expression and evaluation of
2.7. Relevance of tumor immunology, survival in key genes
immunoisoforms, and molecular subtypes Utilizing the STRING platform, we mapped out a PPI
TISIDB (http://cis.hku.hk/TISIDB/) is a comprehensive network for the identified DEGs and highlighted the most
16
repository of tumor-immune system-related data. It significant nodes. By applying the Cytoscape MCODE

Volume 3 Issue 3 (2024) 3 doi: 10.36922/td.2846