Gastric cancer (GC) remains one of the most common malignant diseases in the world[1, 2]. Although the treatment has made some progress over the decades, the 5-year survival rate of patients with advanced GC remains low [3]. Exploration and analysis of tumor prognostic biomarkers are crucial for assessing tumor progression, predicting the effect of treatment, reducing recurrence and mortality, and prolonging survival.
Metabolic reprogramming is one of the characteristics of cancer, promotes tumor cell proliferation and survival[4, 5]. A lot of studies have shown that the metabolism of sugar, lipid and amino acid ultimately affects tumor growth through nucleotide metabolism[6-10]. Nucleotide metabolism is a multi-step process containing a variety of enzymes, including common catalytic enzymes and rate-limiting enzymes such as lyase, synthase, amidotransferase, dehydrogenase, etc. Studies have also proved that restraining the activity of some rate-limiting enzymes in pyrimidine metabolism can directly affect tumor growth[11, 12]. For example, high expression levels of rate–limiting enzymes carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase (CAD), deoxythymidylate kinase, 5'-nucleotidase, cytosolic II (NT5C2), NT5C3, ribonucleotide reductase catalytic subunit M1 (RRM1), RRM2, thymidine kinase 1 (TK1), TK2, dihydroorotate dehydrogenase (DHODH), thymidylate synthetase, uridine-cytidine kinase 2 (UCK2), UCKL1 in pyrimidine metabolism are described in liver cancer and lung cancer patients and related to poor clinical prognosis[11, 13]. Using pyrimidine metabolism rate-limiting enzymes CAD and DHODH as targets to inhibit pyrimidine synthesis enhances the molecular therapeutic response to glioblastoma[14]. 5'-nucleotidase ecto (NT5E) is related to poor clinical prognosis and regulates cell proliferation and migration in many cancers including GC[15, 16]. Historically, pyrimidine nucleotide synthesis has been the pathway of choice to target tumors, because pyrimidine nucleotides are the fundamental building block of DNA synthesis in cells and are increasingly needed by cancer cells due to its rapid growth[12]. Pyrimidine analogue 5- Fluorouracil (5FU) is one of the most extensively used drugs in cancer treatment. 5FU can inhibit thymidylate synthase and prevent the conversion of deoxyuridine acid to thymidylate, thus interfering with DNA synthesis[17]. It is commonly employed to treat breast, colorectal, pancreatic, gastric, liver, and ovarian cancer[18]. However, pyrimidine analogues like 5FU not only target the pyrimidine metabolism of tumor cells, but also partially affect the pyrimidine metabolism of normal cells, causing great side effects[19]. Thus, it is of great significance to search for genes differentially expressed in pyrimidine metabolism according to different cancer types for the treatment and prognosis of different cancer.
In this research, we employed The Cancer Genome Atlas (TCGA) cohort to explore the differentially expressed genes (DEGs) in pyrimidine metabolism in GC and verified them through in vitro experiments. A prognostic risk models was established based on these DEGs. Stratified survival analysis, univariate and multivariate Cox analysis of this model confirmed that it is a reliable and independent clinical factor. Therefore, we made nomograms to visually depict the survival rate of GC patients according to some important clinical factors including our risk model. These conclusions have been verified in the Gene Expression Omnibus (GEO) database. The detailed workflow chart of our article was shown in Figure 1A.