Subtype-dependent difference of glucose transporter 1 and hexokinase II expression in craniopharyngioma: an immunohistochemical study

Papillary craniopharyngiomas are characterized by the BRAF V600E mutation. Enhancement of glucose metabolism may be involved in the downstream of the BRAF V600E mutation in many types of tumors. Glucose metabolism was investigated in craniopharyngioma using immunohistochemical analysis. The study included 29 cases of craniopharyngioma (18 adamantinomatous type [ACP], 11 papillary type [PCP]). Immunohistochemical analysis was performed with anti-glucose transporter-1 (GLUT-1), anti-hexokinase-II (HK-II), anti-BRAF V600E, and anti-beta-catenin antibodies. Expressions of GLUT-1 and HK-II were evaluated using a semiquantitative 4-tiered scale as 0, 1+, 2+, 3+, and divided into negative (0 or 1+) or positive (2+ or 3+) group. GLUT-1 expression level was significantly higher in PCPs than ACPs (0, 1+, 2+, 3+ = 2, 12, 4, 0 cases in ACP, respectively, 0, 1+, 2+, 3+ = 0, 2, 5, 4 in PCP, p = 0.001), and most PCPs were classified into positive group (positive rate, 22.2% [4/18] in ACP, 81.8% [9/11] in PCP; p = 0.003). HK-II expression was also conspicuous in PCPs (0, 1+, 2+, 3+ = 7, 9, 2, 0 cases in ACP, 0, 3, 3, 5 in PCP; p = 0.001), and most of them divided into positive group (positive rate, 11.1% [2/18] in ACP, 72.7% [8/11] in PCP; p = 0.001). Expression patterns of BRAF V600E and beta-catenin reflected the clinicopathological subtypes. Both GLUT-1 and HK-II expressions were prominent in PCP. Glucose metabolism might be more enhanced in PCP than ACP. PCP may use the glucose metabolic system downstream of the BRAF V600E mutant protein.

www.nature.com/scientificreports/ cancers with BRAF mutations have higher glucose uptake on [ 18 F]fluorodeoxyglucose positron emission tomography (FDG-PET) than those without BRAF mutations 5 . Glucose metabolism is reported to be enhanced in many tumors with BRAF mutations (Fig. 1). This immunohistochemical study examined the expression of glycometabolism-related enzymes in the two clinicopathological variants of craniopharyngioma.

Patients. A total of 41 craniopharyngioma cases removed at Gunma University Hospital from August 2000
to March 2019 were collected. We excluded 7 cases due to a few tumor tissues with marked xanthogranulomatous change. Five of the other 34 cases were recurrent after gamma knife treatment and were excluded because of irregular histopathological findings. Therefore, 29 cases including 15 men and 14 women (median age, 44 years, interquartile range [IQR], 38-58) were utilized in this study. This study was reviewed and approved by the institutional review board of Gunma university graduate school of medicine.

FDG-PET.
FDG-PET imaging before surgery was carried out in 7 of 29 cases (24.1%). Image findings and maximum standardized uptake value (SUVmax) were evaluated in 7 cases. FDG-PET imaging was performed with a dedicated scanner as described previously 10 . Image acquisition was initiated 50 min after injection of 5-6 MBq/kg of FDG after more than 6 h fasting. Three-dimensional data acquisition was performed for 3 min per bed position, followed by imaging reconstruction with the three-dimensional ordered subset expectation maximization method. Correction of segmented attenuation was performed by computed tomography (CT) radiography (140 kV, 120-240 mA) to produce 128 × 128 matrix images. SUV was calculated as follows: radioactive concentration in the region of interest (MBq/g)/injected dose (MBq)/patient body weight. Region of interest analysis was conducted by a nuclear physician with the aid of the corresponding CT scans. Only FDG-PET data partially overlaps with the previous report 10 .

Statistical analysis.
All values were reported as median (IQR). Between group comparisons were performed using the Mann-Whitney test. The Fisher exact probability test was used in 2 × 2 tables. p values of < 0.05 were considered to indicate significant difference. www.nature.com/scientificreports/ Ethical approval. This study was conducted under the institutional review board approval. The institutional review boards approved an opt-out method of informed consent. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Case 2: A 38-year-old woman developed visual field disturbance. MR imaging indicated a solid lesion with cysts in the suprasellar region (Fig. 3A). FDG-PET did not show uptake in the solid part of the tumor with SUVmax of 4.1 (Fig. 3B). The tumor was removed through the endoscopic extended transsphenoidal approach. The resected tumor showed characteristic features of ACP such as stellate reticulum and wet keratin formed by well-differentiated epithelium (Fig. 3C). The results of BRAF V600E and beta-catenin immunostaining were corresponded to the diagnosis (Fig. 3B). Immunopositive area was very focal in GLUT-1 (1+) and no positivity was detected in HK-II (0) (Fig. 3E,F). The tumor was safely removed by more than 90%, and recurred after 6 months. She received second surgery with the same approach. The tumor was completely removed on postoperative MR imaging. However, very small recurrence was observed, and was treated by gamma knife radiosurgery. She remains in the same job without recurrence for 3 years.

Discussion
This immunohistochemical study showed that expression of GLUT-1 and HK-II was common and strong in PCP, and rare and weak in ACP. Immunohistochemical staining for the BRAF V600E, and beta-catenin almost perfectly distinguished the clinicopathological variants. FDG-PET imaging showed very high uptakes of glucose with SUVmax over 9 only in several PCP cases.
A relationship between BRAF mutation and activation of MAPK downstream targets such as HIF1α and c-Myc and increased glucose metabolism were shown in melanoma 15 . HIF1α has been known to up-regulate Table 2. Immunostaining score for anti-GLUT-1 and anti-HK-II. www.nature.com/scientificreports/ www.nature.com/scientificreports/ expression of GLUT1 in hypoxic conditions. However, colorectal cancer cell lines with KRAS/BRAF mutations exhibited increased GLUT1 expression, independent of HIF1α status in normoxic conditions 4 . The mechanism between from BRAF mutation to up-regulates cellular glucose metabolism has not completely understood, but ERK regulates multiple transcription factors that control the expression of GLUT1 17,18 . Clinically, evaluation of glucose metabolism depends on FDG-PET. Clinical studies have evaluated BRAF activity with FDG-PET. FDG-PET showed higher SUVmax in mutant BRAF than in the wild type of colorectal cancer 5 , and differentiated thyroid carcinoma 7 . The therapeutic effect of a BRAF inhibitor (vemurafenib) was evaluated by FDG-PET SUVmax in BRAF mutant melanoma 17 . Almost all PCP have the BRAF V600E mutation. On the other hand, ACP has the CTNNB1 mutation, and these two genetic alteration are mutually exclusive 2,3,19 . Previous FDG-PET investigations of a series of sellar and parasellar tumors showed FDG uptake was high in PCP, but relatively low in ACP 10 .
Enhanced glucose metabolism in neoplastic lesions is associated with two independent systems, GLUT-1 and the glycolytic enzyme HK-II 5,9,[20][21][22] . Tumor cells require excess glucose to maintain growth and proliferation. High GLUT-1 expression has been reported in various cancers, including colorectal cancer, lung cancer, breast cancer, ovarian cancer, melanomas, head and neck squamous cell carcinoma, and others 20 . The present study observed common and strong expression (2+, 3+) of GLUT-1 in 82% of PCP cases. In addition, GLUT-1 expression was found in 90% of BRAF V600E-positive craniopharyngioma cases (all PCP). In contrast, ACP showed no expression of BRAF V600E, and GLUT-1 expression was observed in only 22% of ACPs. Activation of the BRAF/MEK/ ERK pathway in PCP may upregulate expression of the GLUT-1, and consequently affect glucose metabolism.
Hexokinase is an enzyme that catalyzes the first step of the glycolytic reaction. The type II isozyme is highly expressed in tumor cells 23 , and is thought to be the key enzyme in glucose metabolism and the Warburg effect 24 , Two steps are required to accumulate FDG in cancer cells: (1) facilitated diffusion through a glucose transport protein (GLUT-1); and (2) subsequent phosphorylation by one of the hexokinase isoforms (HK-II) to form FDG-6-phosphate. FDG-6-phosphate is not transported out of cells nor undergoes glycolytic breakdown 9,20-24 . HK-II expression is increased in many malignant tumors, including nasopharyngeal cancer, ovarian cancer, renal cell carcinoma, hepatocellular carcinoma, colon cancer, and glioma 20 . HK-II is related to the clinical stage, differentiation, metastasis, and poor prognosis of malignant tumors 9,21-24 . Many studies have focused on the correlation between GLUT-1 and HK-II activities and the FDG uptake in various types of tumors. Overexpression of GLUT-1 and HK-II is associated with enhanced tumor aggressiveness and poor survival [20][21][22][25][26][27] . However, HK-II expression is not completely consistent with GLUT-1 expression 4,5 . The present study showed common and strong expression (2+, 3+) of HK-II in 73% of PCP cases, and in 80% of BRAF V600E-positive craniopharyngioma cases (all PCP). In contrast, HK-II positivity was detected in only 11% of ACP cases. The signaling system from BRAF to HK-II remains unclear 15 . However, the present results suggest that HK-II acts to drive glucose metabolism in PCP, or BRAF mutated craniopharyngiomas.
BRAF inhibitors are highly effective against melanoma, non-small cell lung cancer, differentiated thyroid tumor, colorectal cancer, cholangiocarcinoma, and other tumors with the BRAF V600E mutation 28 . A case of PCP also showed remarkable response to a BRAF inhibitor 29,30 . So, identification of the clinicopathological variants of the tumor or the presence or absence of the BRAF mutation are necessary to optimize the chemotherapy regimen 28 . Craniopharyngioma is located in the deepest skull base region, and sample tissues are not easy to obtain (in contrast to melanoma). Therefore, imaging diagnosis of these clinicopathological variants is of great clinical importance 31,32 . Recently, a method was reported for diagnosing molecular variants (BRAF mutated or nonmutated craniopharyngioma) by preoperative MR imaging and CT 31,32 . However, this method depends on a combination of clinical and morphological phenotypes of tumors, which are essentially different from molecular imaging. Previous investigations of craniopharyngioma using FDG-PET showed strong uptake in PCP, but relatively low uptake in ACP 10 . The present immunohistological study revealed that BRAF V600E mutation-positive cases and GLUT-1-or HX-II-positive cases almost overlapped. FDG-PET can predict the effectiveness of BRAF inhibitor treatment and evaluate the subsequent effect. Recently, endoscopic extended transsphenoidal surgery without craniotomy has been widely introduced. In this series, 66% of cases were treated by endoscopic endonasal transsphenoidal surgery (Table 1). In cases of large intracranial tumor treated through a narrow nasal corridor, the extent of endoscopic transnasal tumor removal depends greatly on the size of the tumor 33 . Preoperative FDG-PET could be important in considering neoadjuvant BRAF inhibitor treatment strategy for craniopharyngioma in the future. Craniopharyngioma is a pathologically benign tumor classed as WHO grade I, and does not have specific carcinomatous properties such as metastasis. However, the presence of glucose-metabolizing activity with expression of GLUT-1 and HX-II as well as BRAF mutation may indicate potential malignant character.
The limitation of this study is the small number of cases collected over a long period and evaluated retrospectively. In particular, FDG-PET was performed for various reasons based on the clinical judgment of various neurosurgeons and physicians. Therefore, selection bias is present. Statistical analysis was not applicable to the small number of imaging results. Further prospective study is needed to establish the true diagnostic value of FDG-PET for subtype diagnosis of craniopharyngiomas.

Conclusion
Both GLUT-1 and HK-II expressions were prominent in PCP. Glucose metabolism might be more enhanced in PCP than ACP. PCP may use the glucose metabolic system downstream of the BRAF V600E mutant protein.
These new concepts may be useful in biological, pathological, and clinical considerations of craniopharyngiomas.