SLC1A5 is a novel biomarker associated with ferroptosis and the tumor microenvironment: a pancancer analysis

Basic information on SLC1A5

To first acquire basic information on SLC1A5 in cancer, we examined the distribution and expression level of SLC1A5 in tumors and normal tissues derived from different organs. The human organ map shows the pancancer distribution of SLC1A5 in the organs of male and female patients (Supplementary Figure 1A). Then, based on the expression data in TCGA and GTEx databases, the expression differences of SLC1A5 in different tumors and their corresponding normal tissues were analyzed. SLC1A5 was highly expressed in most tumors, including ACC, BRCA, CESC, CHOL, COAD, DLBC, ESCA, GBM, HNSC, KICH, KIRC, LAML, LGG, LIHC, LUAD, LUSC, OV, PAAD, PRAD, SKCM, STAD, TGCT, THCA, UCEC, and UCS (Figure 1A). Therefore, we performed immunohistochemical analysis, and as expected, SLC1A5 was significantly upregulated in GBM samples compared with LGG samples (Figure 1B and Supplementary Figure 1B). The analysis of genomic alterations of SLC1A5 showed that alterations of SLC1A5 are common across cancers; they are most UCS at a rate of 5%, and amplification is the most common alteration type (Figure 1C). In addition, immunofluorescence (IF) staining showed that SLC1A5 was mainly distributed and localized in the cell membrane of A-431 and U251 tumor cell lines (Figure 1D). Finally, a protein‒protein interaction network was constructed using protein interaction data obtained from the ComPPI online website, which described the subcellular localization of proteins closely related to SLC1A5 distributed in the cytoplasm, mitochondria, nucleus, extracellular, secretory pathway, and membrane (Figure 1E).

Figure 1. Basic information on SLC1A5 across cancers. (A) Expression levels of SLC1A5 in normal and cancerous tissues. (B) Immunohistochemistry and HE staining of SLC1A5 in normal brain and glioma tissues. (C) Alteration frequency of SLC1A5 across cancers based on the cBioPortal database. (D) Immunofluorescence images of SLC1A5 protein expression in the nucleus, endoplasmic reticulum (ER), and microtubules in A-431 and U-251 cells. (E) The protein‒protein interaction (PPI) network shows the proteins interacting with SLC1A5.

Single-cell analysis of SLC1A5 in cancers

To gain insight into the major cell types expressing SLC1A5 in the tumor microenvironment, we performed single-cell analysis of SLC1A5 expression in 79 tumor single-cell databases. Based on the data obtained in the TISCH database, a heatmap was drawn to show the expression level of SLC1A5 in each cell type (including immune cells, stromal cells, malignant cells, and functional cells) in 79 single-cell datasets. The results indicated that across cancers, SLC1A5 was mainly expressed in immune cells (especially monocytes/macrophages) and malignant cells (Figure 2A). In the GSE120575 dataset containing 16,291 cells from 32 SKCM patients treated with immune checkpoint inhibitors, SLC1A5 was widely expressed in immune cell types in the SKCM microenvironment, such as T cells, dendritic cells, NK cells, monocytes or macrophages (Figure 2B). In the Glioma_GSE131928_Smartseq2 dataset, which contains 7,930 cells from 28 glioma patients, SLC1A5 was highly expressed in malignant cells and monocytes/macrophages in the glioma microenvironment (Figure 2C).

Figure 2. Correlation of SLC1A5 with single-cell classes in the pancancer microenvironment. (A) Heatmap showing the expression levels of SLC1A5 in 33 single cell types. (B) Scatterplot showing the GSE120575 dataset and the distribution of 10 different types of single cells. (C) Scatter plot showing the GSE131928 dataset and the distribution of 10 different types of single cells.

Pancancer prognostic analysis of SLC1A5

For the survival landscape of SLC1A5 in pancancer, we showed the prognostic analysis of SLC1A5 with a heatmap, and the results showed that SLC1A5 is associated with the prognosis of most cancer types (BLCA, BRCA, CESC, CHOL, COAD, DLBC, GBM, HNSC, KICH, READ, SKCM, STAD, TGCT, THCA, UCEC, and UCS) (Figure 3A). The OS results showed that SLC1A5 was a protective factor for ACC, KIRC, LAML, LGG, LIHC, MESO, PAAD, SARC, and UVM. Since OS patient outcomes include noncancer mortality events, we used DSS to evaluate the association between SLC1A5 and patient cancer survival time. SLC1A5 was a protective factor in terms of DSS in LUSC and PCPG patients, but the trends for OS and DSS were consistent for all other cancer types. The DFI and PFI results were also examined to fully demonstrate that SLC1A5 is a risk factor for most cancer types and is significantly associated with the prognosis of cancer patients.

Figure 3. Correlation analysis between SLC1A5 and the prognosis of different cancer patients. (A) The summary of correlation analysis between SLC1A5 and patients’ overall survival (OS), disease-specific survival (DSS), disease-free interval (DFI) and progression-free interval (PFI) was analyzed by univariate Cox regression and Kaplan‒Meier. Red indicates that SLC1A5 is a risk factor for this type of cancer, and green indicates a protective factor. Only p values < 0.05 are shown. (B) Forest plot of the relationship between SLC1A5 and the prognosis of cancer patients analyzed by univariate Cox regression. Red markers indicate that SLC1A5 is a risk factor for this cancer type. (C–F) Kaplan‒Meier overall survival curves of SLC1A5 in LIHC (C), LGG (D), KIRC (E) and MESO (F).

To further understand the prognostic potential of SLC1A5, after obtaining the OS data of 32 cancers in TCGA, the prognostic value of SLC1A5 in each cancer was analyzed using univariate Cox regression. The forest plot shows that high expression of SLC1A5 predicted shorter OS time in UVM (HR = 2.635 (95% CI, 1.400 to 4.960), p = 0.002), KIRC (HR = 1.871 (95% CI, 1.494 to 2.343), p < 0.001), MESO (HR = 1.761 (95% CI, 1.302 to 2.381), p < 0.001), LGG (HR = 1.443 (95% CI, 1.210 to 1.720), p < 0.001), LIHC (HR = 1.365 (95% CI, 1.217 to 1.531), p < 0.001), and PAAD (HR = 1.310 (95% CI, 1.055 to 1.628), p < 0.001) (Figure 3B). Kaplan‒Meier analysis was used to analyze the relationship between SLC1A5 expression and patient survival, and the results showed that high expression of SLC1A5 was associated with a poor prognosis in patients with KIRC (p < 0.004), LGG (p < 0.001), LIHC (p < 0.001), MESO (p < 0.001), STAD (p < 0.032), and UVM (p < 0.009) (Figure 3C–3F).

GSEA of SLC1A5 in pancancer

GSEA was performed based on the differentially expressed genes between the low-SLC1A5 subgroup and the high-SLC1A5 subgroup to evaluate the cancer characteristics related to SLC1A5 in each cancer type. To present the results of the GSEA in an intuitive and explicit way, a bubble diagram was used. As shown in Figure 4, SLC1A5 expression was strongly associated with immune-related pathways, such as TNFA signaling via NFKB, IFN-α response, IFN-γ response, and inflammatory response, especially in ACC, GBM, LGG, OV, PCPG, SARC, THCA, and UCEC. In addition, the E2F target pathway was enriched in BRCA, HNSC, LUAD, MESO, OV, PAAD, STAD, and THYM. E2F transcription factors have been reported to play an important role in preventing tumorigenesis by strictly controlling the cell cycle, maintaining genome integrity, and responding to replication stress and DNA damage. In addition, epithelial-mesenchymal transition, oxidative phosphorylation, unfolded protein response, MYC targets, and MTORC1 signaling are closely related to the expression of SLC1A5 in cancer. In conclusion, the above results indicate that the expression of SLC1A5 is closely related to the activation of the tumor immune microenvironment and the malignant phenotype of various cancers, which provides new clues for us to further explore the role of SLC1A5 in tumor progression.

Figure 4. Gene set enrichment analysis (GSEA) of the SLC1A5 gene set across cancers. The size of the circle indicates the FDR value of the enriched element in each cancer, and the color indicates the normalized enrichment score (NES).

TIMER immune cell infiltration analysis

Based on the indication from the GSEA results that SLC1A5 is closely related to cancer immunity, we investigated the correlation between the expression of SLC1A5 and immune cell infiltration across cancers. Spearman correlation analysis was performed using pancancer immune infiltration data from the TIMER2 database. The results showed the infiltration levels of CD4+ T cells, CAFs, lymphoid progenitors, myeloid progenitors, monocyte progenitors, Endos, Eos, HSCs, Tfhs, γ/δ T cells, NK T cells, Tregs, B cells, neutrophils, monocytes, macrophages, dendritic cells, NK cells, mast cells, and CD8+ T cells in different cancer types (Figure 5). In most cancers, the expression of SLC1A5 is positively correlated with the infiltration of macrophages, CD4+ T cells, CD8+ T cells, dendritic cells, monocytes, MDSCs, and CAFs and negatively correlated with the infiltration of monocyte, HSC and Endo progenitors. In addition, SLC1A5 was closely associated with monocyte, dendritic cell, CAF and macrophage infiltration in LGG. Given the significant correlation between SLC1A5 expression and macrophage infiltration, we further investigated the relationship between SLC1A5 expression and macrophage subtype infiltration. The results revealed a positive correlation between SLC1A5 expression and M2 macrophage infiltration in LGG, but there was no significant correlation with M1 macrophages. Related reports have noted that the accumulation of tumor-associated M2 macrophages composed of brain-resident microglia and monocyte-derived macrophages is associated with a poor prognosis in patients with glioma [38]. Therefore, the important role of immune cells in tumor therapy cannot be ignored. Taken together, our results suggest that SLC1A5 affects the progression, treatment and prognosis of cancer patients by regulating immune cells.

Figure 5. Correlation of SLC1A5 expression with immune infiltration levels of B cells, CD4+ T cells, CD8+ T cells, dendritic cells, Endo, Eos, CAF, progenitor, mast cells, HSC, macrophages, monocytes, Tfh, γδT, NKT, regulatory T cells (Tregs), neutrophils, NK cells, and MDSCs in cancers. Red and blue indicate positive and negative correlations, respectively.

Relationships between SLC1A5 and immune regulators, TMB, and MSI

The correlations of the expression levels of 70 genes encoding immunomodulators, including 46 genes encoding immunostimulators and 24 genes encoding immunoinhibitors, and those of SLC1A5 across cancers is shown in Figure 6A. We found that SLC1A5 was positively correlated with most immunomodulators in ACC, LGG, and THYM but negatively correlated with LUAD, LUSC, PRAD, and READ. Recently, immune checkpoint inhibitors (ICIs) have been highlighted as potential clinical treatments, so we next analyzed the association between SLC1A5 expression and TMB and MIS. Our results showed that SLC1A5 expression had a strong positive association with TMB in BRCA, HNSC, KIRC, LUAD, PAAD, SARC, STAD, and THYM (Figure 6B). In addition, the correlation between SLC1A5 expression and MSI was positive in GBM, HNSC, LUAD, MESO, and SARC and negative in READ and UCEC (Figure 6C). Our results demonstrate that SLC1A5 can predict the efficacy of ICIs in the treatment of various cancers.

Figure 6. Relationship between SLC1A5 and immunotherapy-independent efficacy predictive biomarkers. (A) The Spearman correlation heatmap shows that the expression level of SLC1A5 is correlated with 46 immunostimulators and 24 immunoinhibitors. Red indicates a positive correlation and blue indicates a negative correlation. (B) The relationship between SLC1A5 expression level and tumor mutation burden (TMB) in cancer patients. (C) The relationship between the SLC1A5 expression level and microsatellite instability (MSI) in cancer patients. The labeled asterisk indicates the statistical p value (^*p < 0.05, ^**p < 0.01, ^***p < 0.001).

The predictive role of SLC1A5 in cancer immunotherapy

By targeting and inhibiting the actions of immune checkpoint molecules, immune checkpoint blockade (ICB) therapy can activate the immune system to mount an attack against the tumor. PD-1/PD-L1 and CTLA-4 are the most commonly targeted molecules. Clinical trials have demonstrated the remarkable efficacy of immune checkpoint inhibitors across various tumor types, particularly melanoma, non-small cell lung cancer, renal cell carcinoma (RCC), and bladder cancer, among others [39]. Building upon the aforementioned findings, we further investigated the predictive role of SLC1A5 in the ICB therapy cohort. In the GBM-PRJNA482620 cohort, there was no statistically significant difference in the expression level of SLC1A5 between responders and nonresponders (Supplementary Figure 2A). However, high expression of SLC1A5 in patients treated with anti-PD1 was associated with a poorer prognosis (Supplementary Figure 3A). In melanoma patients, it was observed that low expression levels of SLC1A5 in the GSE78220_anti-PD1 and GSE106128_DCs cohorts were significantly associated with a favorable response to ICB therapy (Supplementary Figure 2B and 2C). Furthermore, patients with low SLC1A5 expression exhibited improved prognosis (Supplementary Figure 3B and 3C). In contrast, in melanoma patients from the Nathanson_2017 cohort, it was observed that patients with low expression of SLC1A5 who received ai_CTLA-4 treatment had a worse prognosis (Supplementary Figures 2D and 3D). In RCC patients from the RCC_2020 cohort, contrasting ICB treatment responses and prognoses were observed in the high SLC1A5 expression group after ai-PD-1 and EVEROLIM treatment (Supplementary Figures 2E, 2F and 3E, 3F). These data confirmed that SLC1A5 expression levels in different cohorts of cancer patients were associated with varied immune checkpoint inhibitor treatment responses and prognoses, highlighting its potential as a predictive biomarker.

Connectivity map (CMap) analysis of SLC1A5 in pancancer

The potential drugs and components targeting SLC1A5 in pancancer are shown as a heatmap in Figure 7. The heatmap shows the SLC1A5-related drug components that appear in 3 or more cancer types, and the enrichment parameters of each drug in pancancer are shown in Supplementary Table 1. As shown, ingenol-related genes were significantly enriched in 16 cancers, while prostratin- and parthenolide-related genes were enriched in 12 cancers. These drugs are generally related to the prevention and treatment of cancer; for example, ingenol plays an anticancer role by inhibiting the migration and invasion of colon cancer cells [40], and prostratin inhibits the proliferation of breast cancer cells [41]. The association of these drugs with SLC1A5 increased our interest in them. Our results showed that these drugs have only a small part in the treatment of cancer, and their roles and potential mechanisms in the occurrence and development of various cancers need to be further explored.

Figure 7. The heatmap represents the enrichment score (positive in blue, negative in red) for each drug in each cancer in the CMap database. Components or drugs are sorted from right to left with a decreasing number of enriched cancers.

SLC1A5 regulates the proliferation and migration of glioma cells

To determine the expression levels of SLC1A5 in tumor and normal cell lines, Western blot and RT‒qPCR analyses were performed. The results showed that the expression levels of SLC1A5 were higher in tumor cell lines than in normal cells (Figure 8A, 8B). Based on these results, the functions of SLC1A5 in the U251 and U118 cell lines were further investigated. The efficiency of SLC1A5 knockdown by siRNA was verified at the RNA and protein levels (Figure 8C–8E). Therefore, a loss-of-function experiment was conducted in which SLC1A5 was knocked down. The colony formation assay indicated that the suppression of SLC1A5 expression inhibited the proliferation of U251 and U118 cells (Figure 8F–8H). Moreover, the CCK-8 assay showed that cell growth was inhibited in the SLC1A5 knockdown group compared to the control group (Figure 8I, 8J). Furthermore, the transwell migration assay showed that SLC1A5 knockdown inhibited the migration of U251 and U118 cells (Figure 8K and Supplementary Figure 1C, 1D). Therefore, these results indicated that SLC1A5 promotes the proliferation and migration of glioma cells.

Figure 8. In vitro evidence that SLC1A5 is involved in the proliferation and migration of glioma cells. (A) RNA and (B) protein expression levels of SLC1A5 in astrocytes and glioma cell lines. (C–E) RT‒qPCR and Western blotting verified the knockdown efficiency of SLC1A5 in U251 and U118 cell lines, respectively. (F) Colony formation experiments were performed to analyze the effect of SLC1A5 downregulation on the colony formation ability of U251 and U118 cell lines, and (G, H) statistical analysis was performed. (I, J) Cell proliferation ability was evaluated using the CCK-8 assay, and proliferation curves were plotted. (K) Cell migration ability was determined by transwell migration assay. The labeled asterisk indicates the statistical p value (^*p < 0.05, ^**p < 0.01, ^***p < 0.001).

SLC1A5 participates in glioma progression by inhibiting ferroptosis

To study the effect of SLC1A5 knockdown on ferroptosis in glioma cells, we evaluated the level of ferroptosis in cells by measuring ferroptosis indexes with kits. GSH is an important antioxidant in cells that can scavenge lipid peroxides via glutathione peroxides and inhibit the occurrence of ferroptosis. As expected, knockdown of SLC1A5 resulted in a downregulation of GSH levels in U251 and U118 cells (Figure 9A, 9B). Since MDA is an end product of ferroptosis, we detected MDA and found that it was upregulated after SLC1A5 knockdown (Figure 9C, 9D). The RT‒qPCR and WB results showed that the expression of GPX4 was decreased and the expression of ACSL4 was increased in the siRNA group compared with the NC group (Figure 9E–9I). Thus, these data suggest that SLC1A5 promotes glioma progression by inhibiting ferroptosis in glioma cells.

Figure 9. SLC1A5 inhibits ferroptosis in glioma cells. (A, B) The level of GSH was measured in U251 and U118 cells after SLC1A5 knockdown. (C, D) The level of MDA was measured in U251 and U118 cells after SLC1A5 knockdown. (E, F) The RNA expression levels of the ferroptosis gene GPX4 were detected in U251 and U118 cell lines after SLC1A5 knockdown. (G, H) The RNA expression levels of the ferroptosis gene ACSL4 in U251 and U118 cell lines were detected after SLC1A5 knockdown. (I) Western blotting was used to detect the protein expression levels of GPX4 and ACSL4 after SLC1A5 knockdown. The labeled asterisk indicates the statistical p value (^*p < 0.05, ^**p < 0.01, ^***p < 0.001).

Data source

The mRNA expression profile and clinical characteristics of patients from the TCGA pancancer and GTEx cohorts were downloaded from the UCSC Xena database (https://xenabrowser.net/datapages/). The cBioPortal for Cancer Genomics (http://cbioportal.org) web tool was used to analyze the genomic alteration frequency of SLC1A5 in the 33 cancer types. Human Protein Atlas (THPA, https://www.proteinatlas.org/) was used to detect SLC1A5 protein expression levels in pathological tissues and to confirm protein distribution at the subcellular level. The compartmentalized protein–protein interaction database (ComPPI) (http://comppi.linkgroup.hu) was used to detect the protein‒protein interaction information. A list of the cancers and their abbreviations are presented in Supplementary Table 2.

Single-cell analysis of SLC1A5

The Tumor Immune Single-cell Hub (TISCH) web tool was used to analyze gene expression profiles in tumor microenvironment cells at the single-cell level. Analysis parameters included genes, major lineages, and all cancers selected in the online tool. SLC1A5 expression levels in each cell type were quantified and visualized by a heatmap. Documentation of the data collection, processing and analysis procedures is available online (http://tisch.comp-genomics.org/documentation/).

Pancancer prognosis analysis of SLC1A5

Four key outcome measures were obtained from the UCSC Xena database (https://xenabrowser.net/datapages/), including overall survival (OS), disease-specific survival (DSS), disease-free interval (DFI), and progression-free interval (PFS). Log-rank and univariate Cox regression tests were used to evaluate the association between the expression of SLC1A5 and the prognosis of the patients in each type of cancer. In univariate Cox regression models, SLC1A5 expression was treated as a continuous variable, and Wald’s test was used to assess significant differences in survival. Kaplan‒Meier (KM) curves were used to assess survival and compared using log-rank tests. Risk ratios (HRs) with 95% confidence intervals (95% CIs) were calculated for log-rank and univariate Cox regression, with HR > 1 indicating a risk factor and HR < 1 indicating a protective factor, and the results are presented as a heatmap.

Genes with differential expression between the low- and high-SLC1A5 subgroups

Patients were ranked by SLC1A5 expression level, with the top 30% defined as the high SLC1A5 group and the bottom 30% defined as the low SLC1A5 group. The log2 (fold change) and adjusted p value of each gene in each tumor were calculated using the limma package to determine the differentially expressed genes (DEGs) between the high and low SLC1A5 groups (Supplementary Table 3). Genes with p-adjusted values < 0.05 were considered DEGs.

Gene set enrichment analysis

For enrichment analysis, the normalized enrichment score (NES) and false discovery rate (FDR) of each biological process in each cancer type were calculated by downloading the “gmt” file of the hallmark gene set (h.all.v7.4.symbols.gmt) from the GSEA website (https://www.gsea-msigdb.org/gsea/index.jsp). GSEA was implemented using the clusterProfiler R package, and the R package ggplot2 was used to generate the bubble plot.

Immune cell infiltration analysis in TIMER2

Tumor Immune Estimation Resource (TIMER) is a comprehensive resource for the systematic analysis of immune infiltrates across different types of cancer. For all cancers in TCGA, immune cell infiltration data were downloaded from TIMER2.0 (http://timer.cistrome.org). To investigate the immune infiltration data, we identified the correlations between the expression of SLC1A5 mRNA and 21 immune cells, including B cells, CD4+ T cells, CD8+ T cells, dendritic cells, endothelial cells (Endo), eosinophils (Eos), cancer-associated fibroblasts (CAFs), lymphoid progenitors, myeloid progenitors, monocyte progenitors, mast cells, macrophages, monocytes, neutrophils, hematopoietic stem cells (HSCs), NK cells, T-cell follicular helper cells, γ/δ T cells, NK T cells and regulatory T cells (Tregs), in pancancer by Spearman correlation analysis.

Immunotherapy prediction analysis

A Spearman correlation test was used in this study to determine the correlation between SLC1A5 expression levels and levels of immunomodulators, including 46 immunostimulators and 24 immunoinhibitors with information downloaded from the TISIDB database (http://cis.hku.hk/TISIDB/download.php). Furthermore, we investigated the relationship between SLC1A5 expression and tumor mutation burden (TMB) and microsatellite instability (MSI) across cancers. We performed visual analysis of the immunotherapy response and survival of SLC1A5 in different cancers using the Immunotherapy Response module in the TIGER (Tumor Immunotherapy Gene Expression Resource) online database (http://tiger.canceromics.org/).

Compounds correlating with SLC1A5 in pancancer

The Connectivity Map (CMap) is a gene expression profiling database based on drug-related gene expression developed by the Broad Institute. It is mainly used to reveal the functional relationship between small molecule compounds, genes and disease states. We analyzed the correlation of SLC1A5 expression levels and the levels of genes related to small molecule inhibitors in each cancer by CMap based on DEGs between the low-SLC1A5 and high-SLC1A5 subgroups. In a previous report, the steps for applying the web tool and processing the data for heatmap visualization were described in detail [67].

Cell culture, small interfering RNA transfection, RT‒qPCR, and Western blotting

NHA, T98, U118, LN229, U251, and U87 cell lines were obtained from the Cancer Cell Bank of the Chinese Academy of Medical Sciences (Beijing, China). The cells were cultured as described in our previous studies [68, 69]. Small interfering RNA (siRNA) targeting SLC1A5 was synthesized by General Biologicals (AnHui, China). The RNA Isolater Total RNA Extraction Reagent (BioFlux, HangZhou, China) was used for RNA extraction. The reverse transcription and real-time PCR procedures were performed as previously described [68, 69]. The following antibodies were used for the Western blot (WB) assay: anti-Actin antibody (Abcam, 1:1000), anti-SLC1A5 antibody, anti-GPX4 antibody, anti-ACSL4 antibody, and anti-TFRC antibody. Western blot assays were performed as described in our previous study [68, 69].

Cell proliferation analyses, colony formation assay and migration assay

For the in vitro cell proliferation assay, the CCK-8 assay was performed using the CCK-8 kit (Bioss, Beijing, China) according to the manufacturer’s protocol. An enzyme immunoassay instrument (Thermo Fisher Scientific, USA) was used to detect the OD value at 450 nm. An in vitro colony formation assay was performed to determine whether colony formation was possible in vitro, and the procedure was performed as previously described [68, 69]. Finally, the migration ability of the cells was determined using the Transwell chamber system (BD Biosciences, BD Biosciences, Franklin Lakes, NJ, USA).

Assay for GSH/GSSG and MDA

The cells were harvested, and the intracellular GSH/GSSG ratio was measured spectrophotometrically using a GSH/GSSG assay kit (Beyotime Biotechnology, Shanghai, China) according to the kit instructions. The absorbance at 593 nm was measured using a microplate reader (Bio-Rad, Hercules, CA, USA) for colorimetric analysis. A lipid peroxidation (MDA) assay kit (Beyotime Biotechnology, Shanghai, China) was used to measure the lipid peroxidation product malondialdehyde (MDA). The operational steps of the MDA assay were performed strictly according to the instructions of the kit. The absorbance of the supernatant was measured at 532 nm using a microplate reader.

Statistical analysis

To determine the significance of differences in expression levels between normal and tumor tissues, the Wilcoxon rank sum test was performed. The prognostic effect of SLC1A5 in each cancer was evaluated using the Kaplan-Meier method and univariate Cox proportional hazards regression analysis. For correlation analyses of SLC1A5 and other factors, a Spearman correlation analysis was performed. All the R packages were run through R Studio version 1.3.959, while all the statistical analyses were conducted using R version 4.2.2 (https://www.r-project.org/) and Prism 8 (GraphPad Software Inc., La Jolla, CA, USA).