Bioinformatics analysis of the role of aldolase A in tumor prognosis and immunity

Aldolase A (ALDOA) is an enzyme that plays an important role in glycolysis and gluconeogenesis, which is closely related to tumor metabolism. In this study, the overall roles of ALDOA in pan-cancer have been investigated from several aspects using databases and online analysis tools. Using the ONCOMINE database, the expression of ALDOA in various cancers was analyzed. The prognostic role of ALDOA was explored by PrognoScan, GEPIA, and Kaplan–Meier Plotter. The immune-related role of ALDOA and its downstream substrates was decided by TIMER, cBioPortal and String. Our data indicate that ALDOA expression level in lung adenocarcinoma, liver hepatocellular carcinoma, head and neck squamous cell carcinoma is higher than that in normal tissues. Increased expression of ALDOA often indicates a poor prognosis for patients. The correlation between ALDOA and immune infiltration among different tumors is very different. We also investigate the relationship between ALDOA and its upstream/downstream proteins. Our results showed that ALDOA could be used as a biomarker for the tumor prognosis, and could be correlated with the infiltrating levels of macrophages, CD4+ T cells and CD8+ T cells.

In addition, the RNA expression level of ALDOA of different tumors in TCGA was analyzed by GEPIA2. We found that ALDOA mainly affected the OS in the pan-cancer rather than DFS (OS: HR = 1.3, logrank P = 7.5e−15; DFS: HR = 1.1, logrank P = 0.079) ( Supplementary Fig. 3A). High expression of ALDOA often indicated poor overall survival in SKCM, THCA, THYM and PAAD. However, ALDOA is associated with DFS in PRAD (Supplementary Fig. 3). All the above results indicate that there is a certain relationship between the expression of ALDOA and the prognosis of tumor.
Analysis of the correlation between ALDOA and tumor immunity. Next, we explore the relationship of ALDOA with stromal cells and infiltrated immune cells in the TME. We have observed that the expression of ALDOA in DLBC (lymphoid neoplasm diffuse large B-cell lymphoma), GBM, LIHC and PRAD is positively correlated with the infiltration of CAFs and negatively correlated with BRCA-lumA, THCA and THYM (thymoma) (Fig. 4A,B). We further evaluated the correlation between the expression of ALDOA and tumor immune cell infiltration in pan-cancer (Fig. 4C). In BRCA, LUSC and SKCM (skin cutaneous melanoma), this relationship is obviously negatively correlated, while in LIHC it is positively correlated. www.nature.com/scientificreports/ In recent years, research on immune checkpoints and tumor immunotherapy has continued to progress 24 . Although PD1 (Programmed Cell Death Protein 1) and CTLA4 (Cytotoxic T-Lymphocyte Associated Antigen 4) are among the best, there are still other immune checkpoints under constant research 25 . Therefore, we selected LIHC and BRCA tumors and expanded 30 immune checkpoints to predict their prognostic relationship using Kaplan-Meier Plotter based on the TCGA database ( Supplementary Fig. 4). We found that TMIGD2, CD27 and CD40LG have the effect of OS and RFS on LIHC and BRCA (Table 1). In addition, CD274, HHLA2, ICOS,  www.nature.com/scientificreports/ BTLA, TNFRSF18, TNFSF4, HAVCR2 and NT5E all have an impact on the OS of LIHC and BRCA. Next, we studied the co-expression relationship between ALDOA and immune checkpoints in BRCA, LUAD, and SKCM (Supplementary Table 2). We found that in these three tumors, ALDOA and TNFSF4 are co-expressed (Table 2).

ALDOA related gene and protein analysis. Protein as the expresser of genetic information is closely
related to the life activity characteristics of cells. Figure 5A shows the interaction network of 50 ALDOA binding proteins that have been experimentally confirmed based on the STRING tool. We collected 100 genes related to ALDOA expression in pan-cancer using GEPIA2, and obtained the first 6 most closely related genes. As shown in Fig. 5B, the expression level of ALDOA is positively correlated with PKM (pyruvate kinase M1/2) (R = 0.54), BCKDK (branched chain keto acid dehydrogenase kinase) (R = 0.51), ENO1 (enolase 1) (R = 0.48), GAPDH (glyceraldehyde-3-phosphate dehydrogenase) (R = 0.48), GPI (glucose-6-phosphate isomerase) (R = 0.51) and PGK1 (phosphoglycerate kinase 1) (R = 0.48). Corresponding to this, the heatmap also showed that the above six genes are positively correlated with ALDOA in most tumors (Fig. 5C).

Discussion
The energy metabolism of tumors has its unique characteristics, such as increased glucose metabolism and increased lactate production 26 . In order to explore the role of ALDOA-regulated glycolysis in this change, we conducted a comprehensive analysis of ALDOA genes in 33 different tumors based on data from TCGA and GEO databases, prognostic effects, immune correlation, and related gene and protein analysis.
Most of the research on ALDOA focuses on liver cancer and lung cancer. For liver diseases, the high expression of ALDOA in patients with liver cirrhosis is closely related to the risk of liver cancer 27 . ALDOA has been proven to be an important regulator of the growth and progression of liver cancer cells under hypoxic conditions 8,28 . And ALDOA is also related to the prognosis of liver cancer patients 29 . This is consistent with our www.nature.com/scientificreports/ results. At the same time, we propose that this poor prognosis may be caused by the influence of ALDOA on immune infiltration. For lung cancer, our results indicate that ALDOA affects the OS of LUAD rather than LUSC, which is consistent with the results of Wang Zhihao et al. 15 . However, the opposite result was shown in the immune-related screening, so we suspect that this result may be related to the level of infiltration of CAFs (Fig. 4A). But some studies have shown that ALDOA can promote the progression of lung squamous cell carcinoma and affect the prognosis 30 , which contradicts our results. This difference in results may be related to the sample size. At the same time, ALDOA has been proven to regulate the progression and metastasis of lung cancer through a variety of ways 12,31-33 . Therefore, the regulatory role of ALDOA in lung cancer and the prediction of prognosis are obvious.
For other tumors, ALDOA can also be used as an important biomarker for monitoring progress and predicting prognosis 10,[34][35][36][37] . ALDOA plays the role as an oncogene in a variety of tumors, but its regulation methods are diverse. In cervical adenocarcinoma and bladder cancer, ALDOA affects tumor progression by regulating EMT process 9,36 . The interaction between ALDOA and ncRNA regulates tumor progression, a major research direction 10,11,32 . The discovery of the role of ALDOA in exosomes is a new direction in the field of ALDOA research 38 . At present, there are still few studies on ALDOA in tumors, and different ways of action are still being explored. Therefore, we hope to provide a general direction for future research through the pan-cancer analysis of ALDOA.
TNFSF4, also known as OX-40L, is a member of the TNF superfamily, which provides signals for CD4 T cell responses and plays an important role in tumor immunotherapy 24,39,40 . At present, the research of TNFSF4 mainly focuses on the immunomodulatory function in tumors and immune-related diseases 41 . Our results showed that ALDOA regulates the co-expression of TNFSF4 in a variety of tumors, which may have a certain correlation with the tumor's immune microenvironment and prognosis. Other immune checkpoints in different tumors, and their relationship with ALDOA are also worth in-depth investigation. Perhaps there is a certain connection between glycolysis and tumor immunity, which needs to be further explored by other researchers.
Here, we describe the prognostic value of ALDOA in pan-cancer and report its immunological correlation with different cancers. As mentioned earlier, ALDOA is closely related to tumor metabolism, which plays vital roles in the tumor immune microenvironment 42 , but the relationship between ALDOA and tumor immune infiltration has not been clearly reported. We are surprised to find that there is a clear correlation between the two in LIHC, LUSC, BRCA and SKCM. However, it is far from enough to draw such a conclusion based on data analysis, and it needs to be verified by experiments in the future research. This study may provide new ideas for subsequent research and build a new relationship between metabolism and immunity. www.nature.com/scientificreports/

Conclusions
Based on our study, the expression level of ALDOA in most tumors is higher than that in normal tissues. Increased expression of ALDOA often indicates a poor prognosis for patients. Differences in the correlation between ALDOA and immune infiltration among different tumors was observed. For example, there is a positive correlation between the expression of ALDOA and immune infiltration in LIHC. However, there is a negative correlation in LUSC, BRCA and SKCM. At the same time, we found that ALDOA may be co-expressed with TNFSF4 to regulate tumor immune infiltration. Finally, we made a statistical analysis of the expression of ALDOA-related genes in pan-cancer. As far as we know, this is the first report on the correlation between ALDOA and tumor immune infiltration. Based on this comprehensive analysis, we believed that ALDOA can be used as a prognosis biomarker in pan-cancer and is related to immune infiltration. Further studies can be performed to elucidate functions and detailed molecular mechanisms of ALDOA in tumor metabolism and immune microenvironment.

Materials and methods
Gene expression analysis. We used the ONCOMINE database (www. oncom ine. org) to analyze the mRNA expression of ALDOA in different types of cancer. The significant expression of ALDOA between tumors and normal was recorded with a P < 0.001 and the fold change to 1.5.
Survival prognosis analysis. We first analyzed the relationship between ALDOA and patient prognosis through PrognoScan (http:// dna00. bio. kyute ch. ac. jp/ Progn oScan/ index. html) 43 . We collected information about overall survival (OS), disease-free survival (DFS), relapse-free survival (RFS), distant metastasis-free survival (DMFS) and disease-specific survival (DSS). We followed the methods of Qingchen Yuan et al. 2020. and set the threshold to Cox P-value < 0.05 44 . Then we input tumor information from TCGA database and normal sample information from TCGA and GTEX project in GEPIA2 (http:// gepia2. cancer-pku. cn/) 45 to investigate the effect of ALDOA expression on OS and DFS in various tumors (n = 33). We also used Kaplan-Meier Plotter (https:// kmplot. com/ analy sis/) 46 to complement with above analysis. We calculate the hazard ratios (HRs) and log-rank P values with 95% confidence intervals (CI).

Immune infiltration analysis.
We searched ALDOA on the TIMER2.0 database (http:// timer. cistr ome. org/) to compare its relationship with immune infiltration in different tumors or specific tumor subtypes. We also selected cancer-associated fibroblasts for further analysis. At the same time, in order to perform reliable immune correlation assessment, we used immuneeconv, which is an R software package that integrates six latest algorithms, including TIMER, MCP-counter, xCell, EPIC, CIBERSORT and quanTIseq 47 . Perform a rank sum test on the data, and consider that P value of < 0.05 is statistically significant. Plot these data as heatmaps and scatter plots.
Mutual-exclusivity analysis between ALDOA and multiple-immune checkpoints. We searched ALDOA and 30 multiple-immune checkpoints in the cBioPortal (http:// www. cbiop ortal. org) to obtain coexpression or mutual exclusion information between them. Then we selected the genes that are significantly related to ALDOA and organized them into a table.
ALDOA-related gene or protein analysis. We followed the methods of Cui et al. 47 . We searched ALDOA in the String (https:// www. string-db. org/) database and obtained the ALDOA binding protein that has been verified by experiments. Next, we searched ALDOA in the "Similar Gene Detection" module of GEPIA2, and obtained the top 100 targeted genes related to ALDOA. We also applied the "correlation analysis" module of GEPIA2 to perform a pairwise gene Pearson correlation analysis of ALDOA and selected genes. Moreover, we used the data obtained by the "Gene_Corr" module of TIMER2 to draw a heatmap of the correlation between the target gene and ALDOA expression in pan-cancer.
Statistical analysis. T-test, fold change, and gene grade determine the P value generated in Oncomine.
Kaplan-Meier plotter and GEPIA2 curve analyses were generated by calculating HR and logrank P. HR and Cox P values in PrognoScan are calculated using univariate Cox regression model. The correlation of gene expression is assessed using Spearman's correlation 44 . In general, we determined P < 0.05 represents a statistically significant.

Data availability
The data and materials can be obtained by contacting the corresponding author. www.nature.com/scientificreports/