Expression of Glucose Metabolism-Related Proteins in Adrenal Neoplasms

Adrenal gland tumors include adrenal cortical neoplasms (ACN) originating from the adrenal cortex and pheochromocytomas (PCC) originating from the adrenal medulla. ACN consist of adrenal cortical adenoma (ACA) and adrenal cortical carcinoma (ACC). ACA is often challenging to histologically differentiate from ACC [1], a rare and highly aggressive tumor for which effective targeted therapies are lacking [2]. Currently, all PCC cases are thought to have metastatic potential; previous categories of benign and malignant PCC have been eliminated [3]. Although the cells of origin of ACN and PCC are different, ACN and PCC may show histologically similar appearance; therefore, differential diagnosis is often required. In addition, since it is difficult to predict the biologic behavior of the tumors based on histological findings alone, combinations of various clinical, biochemical, and/or histological features are needed to predict the biologic behavior of adrenal neoplasms.

Tumor cells have different metabolic features from those of normal cells. Unlike normal cells, which produce energy by mitochondrial oxidative phosphorylation, cancer cells acquire energy by aerobic glycolysis, called the Warburg effect [4, 5]. This accelerated glycolysis is characterized by high rates of glucose consumption and lactate secretion in tumor tissues and is stimulated by various oncogenes. Deregulated cell metabolism is a crucial consequence of oncogenic mutations and a hallmark of cancer [6]. Glucose transporter 1 (GLUT1), hexokinase II, and carbonic anhydrase IX (CAIX) play important roles in this glycolytic process. GLUT1 is a glucose transporter, acting as a channel for glucose to enter the cell [7]. Hexokinase II is an enzyme that phosphorylates glucose in the cell to produce glucose-6-phosphate [8]. CAIX reversibly hydrates carbon dioxide and neutralizes acidification by lactate formed during glycolysis, allowing glycolysis to continue [9, 10]. Two major glucose catabolic pathways are associated with glycolysis. The pentose phosphate pathway links the synthesis of the nucleotide precursor ribose and nicotinamide adenine dinucleotide phosphate. In the pentose phosphate pathway, the oxidative branch produces nicotinamide adenine dinucleotide phosphate and ribonucleotides, and glucose-6-phosphate dehydrogenase (G6PDH) acts as an important enzyme [11]. The other is the glycine and serine metabolic pathway, which is a glycolytic intermediate metabolic pathway [12-15]. In the serine biosynthesis pathway, 3-phosphoglycerate generated during glycolysis is oxidized to 3-phosphohydroxypyruvate by phosphoglycerate dehydrogenase (PHGDH). Serine hydroxymethyltransferase (SHMT) causes the reversible conversion of serine and glycine.

We have previously identified differences in metabolic phenotypes among thyroid tumor subtypes [16-18]. Furthermore, several studies have shown that metabolism-related proteins may be predictors of cancer prognosis [19-21]. We hypothesized that adrenal gland tumors (ACA, ACC, and PCC) have different metabolic characteristics and that these properties could serve as prognostic factors in ACC and PCC, for which it is currently difficult to predict prognosis. In this study, we investigated the expression levels of glucose metabolism-related proteins in adrenal gland tumors and their clinicopathologic implications.

Formalin-fixed paraffin-embedded tissue samples from patients diagnosed with ACN and PCC in surgically resected adrenal glands at Severance Hospital from January 2000 to December 2012 were used. The study cohort consisted of 132 ACN (115 adrenal cortical adenoma [ACA] and 17 adrenal cortical carcinoma [ACC]) and 189 PCC cases. This study was approved by the Institutional Review Board of Severance Hospital. All cases were retrospectively reviewed by pathologists (K.J.S. and K.H.M.). For ACN, parameters corresponding to Weiss criteria were examined [22]. For PCC, parameters corresponding to the grading system for adrenal pheochromocytoma and paraganglioma (GAPP) were investigated [23]. Disease-free survival was calculated from the date of the first curative surgery to the date of the first locoregional or systemic relapse or death without any type of relapse. Overall survival (OS) was estimated from the date of the first curative operation to the date of the last follow-up or death from any cause. Clinical information was obtained from electronic medical records.

Representative areas were selected on hematoxylin-eosin-stained slides. Three-millimeter core biopsies were taken from selected areas and placed in a 6 × 5 recipient block. More than 2 tissue cores were extracted from each case to minimize extraction bias.

Antibodies used for immunohistochemistry are listed in online suppl. Table 1 (see www.karger.com/doi/10.1159/000518208 for all online suppl. material). All immunohistochemistry was performed using an automatic immunohistochemistry staining device (Benchmark XT; Ventana Medical System, Tucson, AZ, USA). In brief, 5-µm-thick formalin-fixed paraffin-embedded tissue sections were transferred onto adhesive slides and dried at 62°C for 30 min. Standard heat epitope retrieval was performed for 30 min in ethylene diamine tetra-acetic acid, pH 8.0, using an autostainer.

Immunohistochemical markers were semiquantitatively evaluated according to previously reported methods [24]. Tumor cell and stromal cell staining was assessed as follows: 0, negative or weak immunostaining in <1% of the tumor/stroma; 1, focal expression in 1–10% of the tumor/stroma; 2, positive staining in 11–50% of the tumor/stroma; and 3, positive staining in 51–100% of the tumor/stroma. Scores of 0 and 1 were defined as negative, while 2 and 3 were defined as positive.

Data were analyzed using SPSS for Windows, Version 22.0 (SPSS Inc., IBM Corporation, Chicago, IL, USA). For the determination of statistical significance, Student’s t tests and Fisher’s exact tests were used for continuous and categorical variables, respectively. For multiple comparisons, a corrected p value with the Bonferroni multiple comparison procedure was used. The threshold for statistical significance was set to p < 0.05. Kaplan-Meier survival curves and log-rank statistics were employed to evaluate the time to tumor recurrence and OS. A multivariate regression analysis was performed using the Cox proportional hazards model.

ACC was characterized by a significantly larger tumor size (p < 0.001) and younger age (41.0 ± 25.1 years; p = 0.048) than those of ACA. Various factors, including Weiss criteria, also showed significant differences (online suppl. Table 2). The basal characteristics of PCC are presented in online suppl. Table 3.

GLUT1 in tumor cells ([T]; p < 0.001), GLUT1 in stromal cells ([S]; p < 0.001), G6PDH (p < 0.001), and SHMT1 (p = 0.002) showed significant differences between ACN and PCC. ACN showed higher overall expression levels than those in PCC (Table 1; Fig. 1). In ACN, GLUT1 (T; p = 0.045) and PHGDH (p = 0.043) levels were significantly different between ACC and ACA. Expression was generally higher in ACC than in ACA (Table 2; Fig. 1).

In ACN, the expression of GLUT1 (T) was associated with increased mitosis (p = 0.007), and CAIX (S) expression was associated with venous invasion (p < 0.001; Fig. 2a, b). PHGDH expression was also related to increased mitosis (p = 0.001), a decreased clear cell proportion (p < 0.001), and venous invasion (p = 0.001; Fig. 2c–e). In PCC, CAIX (T) expression was associated with norepinephrine type (p = 0.002), and CAIX (S) expression was related to the nonzellballen pattern (p = 0.002; Fig. 2f, g).

A univariate analysis showed that GLUT1 (T; p = 0.017), CAIX (p = 0.003), and PHGDH (p = 0.009) in ACN were related to a shorter OS (online suppl. Table 4; Fig. 3a–c). In PCC, GLUT1 (T; p = 0.001) and PHGDH (p < 0.001) were associated with a shorter OS (online -suppl. Table 5; Fig. 3d, e).

In a multivariate Cox analysis of OS, GLUT1 positivity was an independent predictor of poor prognosis in ACN (HR: 404.8, 95% CI: 1.204–136,128, p = 0.043; Table 3). In PCC, a high GAPP score (>3, HR: 35.190, 95% CI: 1.528–810.4, p = 0.026), GLUT1 positivity (HR: 10.875, 95% CI: 2.330–50.76, p = 0.002), and PHGDH positivity (HR: 52.229, 95% CI: 7.672–355.5, p < 0.001) were independent factors associated with shorter OS (Table 4).

We examined the expression of glucose metabolism-related proteins in adrenal gland neoplasms. Previous studies have reported that glycolysis-related proteins are expressed in PCC [25] and ACN [26], but direct comparisons are difficult owing to the lack of investigations of expression differences between these tumors. We initially found that the expression levels of glucose metabolism-related proteins (GLUT1, G6PDH, and SHMT1) were higher in ACN than in PCC, suggesting that the glycolytic process is generally more active in ACN. This result can be explained by the genetic background of PCC. At least 30% of PCC and paraganglioma (PGL) cases are now known to be hereditary, and mutations in genes encoding succinate dehydrogenase enzymes (SDH gene family) have been identified as the most common cause (14–20%), followed by mutations in VHL (9%) and RET (5%) [3]. In loss-of-function SDH mutants, mitochondrial function is impaired, and metabolic reprogramming occurs from mitochondrial oxidative phosphorylation to glycolysis [27], which is also demonstrated by high radiolabeled deoxy-2-[¹⁸F]fluoro-D-glucose (¹⁸F-FDG) uptake in SDH-mutated PCC in positron emission tomography (PET) studies [25, 28]. In a recent comprehensive genomic study, PCC/PGL was classified into 4 molecular subtypes, including a pseudohypoxia subtype. Many tumors of this subtype have germline mutations in Krebs cycle genes, SDH and IDH, and had disruptions in the hypoxia signaling pathway [29]. Sporadic tumors, accounting for the remaining 70% of PCC/PGL, mainly show alterations, such as NF1, HRAS, and ATRX mutations, corresponding to kinase signaling and Wnt-altered subtypes. In addition, the mechanism underlying increased FDG uptake in SDH-mutated PCC is explained by accelerated glucose phosphorylation, rather than by increased glucose transporter expression [25].

GLUT1 (T) and PHGDH levels were higher in ACC than in ACA, which can be explained by an increase in glycolysis by glucose uptake, one of the metabolic characteristics of malignant tumors [4, 5]. The levels of GLUT1, CAIX, and MCT4, key metabolism-related proteins, have been reported to be higher in ACC than in ACA [26]. FDG-PET studies have also shown higher FDG uptake in ACC than in ACA, suggesting that glucose metabolism is increased in ACC [30, 31]. CAIX and GLUT1 are part of the hypoxia inducible factor-1 (HIF-1) pathway and have been suggested as endogenous markers for hypoxia in solid tumors [32]. HIF accumulates in response to decreased cellular oxygen levels and regulates the adaptation of tumor cells to hypoxic stress. HIF-1α raises the activity of several glycolytic protein isoforms, including GLUT-1, CAIX, GAPDH, PGK1, and others, and directly upregulates the genes coding for glucose transporters such as GLUT-1 and the enzymes of the glycolytic pathway [33, 34].

We examined the expression levels of glucose metabolism-related proteins in stromal cells as well as in tumor cells. GLUT1 was more highly expressed in stromal cells in ACN compared to PCC. Metabolic interactions between cancer cells and cells in the tumor microenvironment allow metabolites to migrate from stromal cells to meet metabolic demands and maintain ATP production in cancer cells. The expression of these proteins in stromal cells might be explained by the reverse Warburg effect, which describes a 2-compartment model in which stromal cells are induced by cancer cells to undergo aerobic glycolysis and then transfer the products back to the cancer cells for utilization for mitochondrial oxidative phosphorylation [35]. Metabolism occurs by oxidative phosphorylation in functional mitochondria in tumor cells, whereas metabolism in stromal cells by glycolysis results in dysfunctional mitochondria resulting from increased autophagy activity in breast cancers [36, 37]. Similar metabolic coupling between stromal cells and cancer cells has been observed in various cancers, including ovarian, prostate, liver, colon, and pancreatic cancers [38-42]. Therefore, further research is needed to determine whether the reverse Warburg effect is one of the metabolic features in adrenal cortical neoplasms.

The expression levels of GLUT1 (T), CAIX (S), and PHGDH in ACN, as well as GLUT1 (T) and PHGDH in PCC, were associated with poor prognosis in univariate analysis. In multivariate analysis, GLUT1 (T) in ACN and high GAPP score, GLUT1 (T), and PHGDH in PCC were independent prognostic factors for OS. GLUT1 is a hyp-oxia-responsive glucose transporter, and its overexpression has been identified in various cancers, including ACC, stomach, urinary bladder, ovary, oral cavity, esophagus, pancreas, colorectum, lung, and gallbladder cancer, resulting in increased glucose uptake into the cytoplasm of tumor cells under hypoxic conditions; it is considered a poor prognostic factor [26, 43, 44]. PHGDH is overexpressed in a substantial fraction of human cancers, mainly as a result of genomic amplification of the PHGDH gene on chromosome 1p12; it catalyzes a growth-promoting metabolic pathway. PHGDH is associated with poor prognosis in pancreatic cancer [45], lung cancer [46, 47], and gastric cancer [48]. Previous studies have suggested that PHGDH is a potential therapeutic target in tumors in which metabolism is twisted toward de novo serine biosynthesis [49]. Furthermore, inhibiting GLUT1 [50-52], CAIX [53, 54], G6PDH [55], PHGDH [56-59], and SHMT [60, 61] limits tumor growth. Therefore, glucose metabolism-related proteins, such as GLUT1 and PHGDH, may be prognostic markers and are expected to be effective treatment targets in adrenal gland tumors in the near future.

A limitation of this study is that assessing the expression of enzymes or proteins via immunohistochemistry may not necessarily equate to their activity. Additional factors that may affect the enzymatic activity include posttranslational modification (e.g., addition of functional groups or proteins/peptides, chemical conversion, and structural changes), various inhibitors (e.g., competitive, uncompetitive, noncompetitive, suicide, and mixed), activators (e.g., allosteric, co-factors like vitamins, and nonvitamins/minerals), and environmental factors (e.g., changing pH). Therefore, it is necessary to conduct more in-depth research while considering these various factors as much as possible.

In summary, the expression levels of GLUT1, G6PDH, and SHMT1, which are glucose metabolism-related proteins, differ between ACN and PCC and are higher in ACN. In addition, GLUT1 and PHGDH are associated with poor prognosis in adrenal gland tumors.

This study was approved by the Institutional Review Board of Severance Hospital (4-2019-0902). Informed consent from patients was exempted by IRB. This study was exempt from obtaining written informed consent due to the retrospective nature of the study.

The authors declare no conflicts of interest.

No external funding source was used in this study.

E.K.K. participated in the design of the study and performed the statistical analysis and carried out the immunoassays. H.M.K. conducted the data collection and analysis. J.S.K. conceived the study and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.

All data generated or analyzed during this study are included in this article and its online supplementary material. Further enquiries can be directed to the corresponding author.