Pancreatic cancer is the seventh leading cause of cancer-related death worldwide due to its poor prognosis with almost as many deaths (n=466,003) as cases (n=495,773) (1), which was estimated to become the second most common cause of cancer-related death by 2030 in the U.S (2). Pancreatic ductal adenocarcinoma (PDAC), the most common type of pancreatic cancer, is the deadliest malignancy with a 5-year survival rate of less than 8% (3). The worse clinical outcomes of PDAC are due to the lack of early typical symptoms and the highly aggressive biological characteristics, emphasizing the urgent need for searching promising prognostic markers for clinical practice.

Malignant proliferating cancer cells usually rewire energy metabolism manners to meet the requirements of rapid cell growth, division, invasion, and migration, which is an emerging hallmark of cancer known as “energy metabolism reprogramming” (4). It has been well known that tumor cells tend to increase their glucose uptake and lactate production even in the presence of ample oxygen, which is described as the “aerobic glycolysis” or “Warburg effect” (5). The enhancement of glycolysis has been demonstrated to play critical roles in pancreatic tumorigenesis and development (6, 7).

Previous studies have shown that multiple key glycolytic enzymes such as glucose transporter 1 (GLUT1), monocarboxylate transporter 4 (MCT4), hexokinase 2 (HK2), pyruvate kinase M2 (PKM2) and enolase 1 (ENO1) are usually overexpressed in tumors and play critical roles in glycolysis pathway (8). Transport of glucose across the plasma membrane is an important regulator of glucose metabolism and is mediated by glucose transporter (GLUT) family proteins. Increased rate of glucose uptake in malignant cells is associated with elevated expression of GLUTs, especially GLUT1 (9). Since enhanced glycolysis is associated with lactate production, tumor cells must eliminate excessive lactate out of the cell to prevent cellular acidification, which is achieved by the upregulation of MCT4, a proton-coupled lactate transporter (10). Besides cellular membrane transporters, several enzymes located in cytoplasmic fractions also contribute to glycolytic flux. As the most influential functional enzyme of glycolysis, overexpression of HK2 promotes glycolysis process by catalyzing phosphorylation of glucose to glucose-6-phosphate (11). PKM2 is the key rate-limiting enzyme involved in the final step of the glycolysis pathway and catalyzes the last irreversible step in glycolysis, the conversion of phosphoenolpyruvate to pyruvate, while ADP is phosphorylated to form ATP (12). Moreover, ENO1, also known as alpha-enolase, a critical glycolytic enzyme that dehydrates 2-phosphoglycerate to phosphoenolpyruvate, is present both in cell surface and cytoplasm (13). It has been well demonstrated that these glycolysis enzymes contribute to biological behaviors of cancer.

Although emerging evidence has focused on the correlation between various glycolysis markers and pancreatic cancer, the conclusions remain controversial. In terms of GLUT1, previous studies investigating the relationship between GLUT1 expression and clinical outcomes of patients with pancreatic cancer have yet yielded conflicting results. For example, some studies have demonstrated the independent prognostic role of GLUT1 in pancreatic cancer (14–19), while other investigators revealed that the expression of GLUT1 was not associated with clinicopathological characteristics and prognosis of pancreatic cancer (20, 21). Therefore, a comprehensive meta-analysis was performed to clarify the prognostic role of glycolysis markers GLUT1, MCT4, HK2, PKM2, and ENO1 in pancreatic cancer.

Energy metabolism reprogramming including aberrant glucose, amino acid, lipid metabolism and other bioenergetic metabolism pathways confers tumor cells the ability to acquire necessary nutrients and biomass in order to survive and proliferate under a frequently severe and nutrient-limited microenvironment (33, 34). Among various metabolic manners, aerobic glycolysis which is the central pathway of dysregulation of glucose metabolism has been well demonstrated in many tumors. Notably, upregulation of key enzymes including GLUT1, HK2, phosphofructokinase (PFK), TP53-induced glycolysis and apoptosis regulator (TIGAR), PKM2, ENO1, lactate dehydrogenase A (LDHA) and MCT4 in glycolytic pathway contributes to the increased glycolytic flux and promotion of malignant behaviors of various cancers including pancreatic cancer (7, 35). In addition, therapeutic strategies for targeting glycolytic enzymes are emerging as promising anti-cancer candidates. 3-bromopyruvate (3-BrPA), an inhibitor of HK2, has been well proved to block the growth and progression of pancreatic cancer both in vitro and in vivo (36, 37). In addition to HK2, the other glycolytic rate-limiting enzyme PFK1 could be allosterically activated by fructose-2,6-bisphosphate (F2,6BP) mediated by well-established glucose metabolism regulators 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatases (PFKFBs) (38). PFK15, a PFKFB3 inhibitor, has been demonstrated to reduce glycolysis and plasma membrane calcium ATPases (PMCAs) activity, resulting in calcium overload and cell apoptosis in PDAC (39). Treating pancreatic cancer xenografts with FX11 which is a small-molecule inhibitor of LDHA suppressed tumor progression (40). Additionally, the combination of LDHA inhibitors and gemcitabine displayed synergistic cytotoxic activity against pancreatic cancer cells under hypoxic conditions (41). In terms of clinical applications, in a phase I clinical trial in patients with advanced solid tumors, including pancreatic cancer, the glycolytic inhibitor 2-deoxy-D-glucose (2DG) in combination with chemotherapeutic agent docetaxel produced tolerable adverse effects and feasible clinical benefit (NCT00096707) (42). Given the potential benefits for targeting glycolytic enzymes, more clinical trials of specific glycolysis inhibitors in patients with pancreatic cancer should be considered.

In the present meta-analysis, we analyzed data from 28 studies consisting of 2010 patients to determine the prognostic value of five glycolysis markers in pancreatic cancer. The results showed that high expression levels of glycolysis markers significantly correlated with worse OS, DFS, RFS and DMFS. The conclusion is consistent with the results of most studies included in this meta-analysis. In terms of subgroup analysis for OS, analysis by ethnicity showed that higher expression of glycolysis markers was correlated with shorter OS in both Asians and Caucasians. This implies that malignant glycolysis phenotypes of pancreatic cancer are common phenomenon in humans. Moreover, subgroup analysis of the expression levels of specific glycolysis markers highlighted that higher expression levels of GLUT1, MCT4, and ENO1 correlated with poorer OS in pancreatic cancer patients, while the expression levels of HK2 and PKM2 did not. These results indicated that the glycolysis pathways of pancreatic cancer may vary in diverse glycolysis markers and GLUT1, MCT4, and ENO1 may be potential therapeutic targets in treating pancreatic cancer. Due to the limited number of included studies, subgroup analysis for DFS, RFS and DMFS in pancreatic cancer patients was not performed in the present meta-analysis. Hence, more studies examining the correlation between expression levels of glycolysis markers and DFS, RFS and DMFS of pancreatic cancer patients need to be conducted. Moreover, we also analyzed the association between GLUT1 and PKM2 and clinicopathologic features including gender, age, tumor location, tumor differentiation, TNM stage, lymph node metastasis, perineural invasion, vascular invasion, and resection margin in pancreatic cancer. However, no significant correlations were observed. The reason may be attributed to limited number of studies and samples and the high levels of heterogeneity between various studies. Therefore, further studies are warranted to evaluate the role of glycolysis markers in clinicopathologic characteristics of pancreatic cancer patients in the future.

In recent years, ¹⁸F-fluorodeoxyglucose positron emission tomography (FDG-PET) based on the increased glucose metabolism of malignant cells has been being used as a valuable functional imaging technology in clinical practice for various malignancies. Altered expression levels of GLUT has been demonstrated to be associated with enhanced FDG uptake quantified by standardized uptake values (SUVs) in tumor cells (43). Interestingly, previous studies investigating the association between the expression of GLUT1 and SUVs in pancreatic cancer have yielded conflicting results. Chikamoto et al. suggested that a significant relationship existed between high maximum standardized uptake value (SUV_max) and the expression of GLUT1 (18), whereas other studies indicated that GLUT1 expression was not significantly correlated with SUV_max (14, 20). Similarly, a recent meta-analysis suggested that only a moderate association was identified between the expression of GLUT1 and SUVs derived from FDG-PET in multiple tumor types (44). Therefore, it can be deduced that the correlations between glucose hypermetabolism in vivo displayed by FDG-PET and enhanced glycolytic phenotype of tumor cells are more complex than the single expression of GLUT1. Other key proteins of the glucose metabolism such as other members of GLUT family and HK2 might contribute to the alterations in SUVs, which needs to be confirmed in further studies.

The pancreatic tumor microenvironment (TME) consisting of cancer cells, stromal cells, and extracellular components is considered to play a pivotal role in multiple malignant biological behaviors including carcinogenesis, proliferation, invasion, metastasis, angiogenesis, immunosuppression and chemoresistance. It should be noted that stromal cells could be induced by cancer cells to present a glycolytic phenotype and support mitochondrial oxidative phosphorylation (OXPHOS) in cancer cells to fuel cell proliferation and growth, which was described as “the reverse Warburg effect” (45). As a key component in pancreatic TME, pancreatic stellate cells (PSCs) could be induced a shift to glycolysis in Caveolin-1-ROS positive feedback signaling, presenting upregulated expression of glycolytic enzymes HK2, 6-phosphofructokinase (PFKP), PKM2 and GLUT1 with downregulated expression of OXPHOS enzymes (46). Cancer-associated fibroblasts (CAFs), which develops mostly from activated PSCs, showed elevated expression of lactate dehydrogenase (LDHA), PKM2 and enhanced glucose uptake capability and lactate production (47, 48). The enhanced glycolysis of both PSCs and CAFs contribute to the proliferative and invasive capacities of pancreatic cancer cells. Moreover, pancreatic cancer is characterized by extensive infiltration of immunosuppressive cells such as regulatory T cells (Tregs), tumor-associated macrophages (TAMs) and myeloid-derived suppressor cells (MDSCs). It has been well demonstrated that glycolytic by-product lactic acid not only blunts tumor immunosurveillance by T and NK cells but also enhances immunosuppressive phenotype of Tregs, TAMs and MDSCs (49–52). A recent study further indicated that a terminally differentiated subpopulation of tumor-associated neutrophils exhibited hyperactivated glycolytic activity in PDAC TME promotes pro-tumor and immunosuppression functions of neutrophils (53). Therefore, targeting enhanced glycolysis of multiple components in the TME, not just tumor cells, may contribute to the development of novel therapeutic strategies against PDAC.

Several limitations should be addressed for this meta-analysis. First, given the current lack of a universal cut-off value for various glycolysis markers, different cut-off values were applied in different included studies. Second, some of the HRs and 95% CIs were estimated by extracting data from the survival curves, which might produce statistical deviations. Third, many studies included did not provide follow-up durations and they may have contributed to the bias. Fourth, the sample size of some studies is small. Considering the limitations of the present study, additional high-quality, large-scale, long-term studies need to be conducted.

This meta-analysis demonstrated that high expression of the five glycolysis markers correlated with poorer OS, DFS, RFS and DMFS in pancreatic cancer patients, suggesting that glycolytic pathway enzymes are potential prognostic biomarkers and promising therapeutic targets in patients with pancreatic cancer. However, due to the limitations of this study, more high-quality and well-designed studies need to be performed in the future.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

QX, CW, and RX conceived and designed the study. CW, RX, and JS drafted the manuscript, acquired the data, and performed the meta-analysis. YC, XY, and RR revised the manuscript for important intellectual contents. All authors read and approved the final manuscript.

This study was supported by the CAMS Innovation Fund for Medical Sciences (2021-1-I2M-002), National Natural Science Foundation of China (82102810 and 81970763) and fellowship of China Postdoctoral Science Foundation (2022T150067).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.