Distribution of Hypoxia-Inducible Factor-1α and Glucose Transporter-1 in Human Tongue Cancers

Distribution of Hypoxia-Inducible Factor-1

a

and Glucose Transporter-1 in Human

Tongue Cancers

Marcelo Gadelha Vasconcelos, DDS, MSc, PhD,*

Rodrigo Gadelha Vasconcelos, DDS, MSc, PhD,

y

Denise H
elen Imaculada Pereira de Oliveira, DDS, MSc,z

Edilmar de Moura Santos, MSc,

x

Le~ao Pereira Pinto, DDS, MSc, PhD,k

Ericka Janine Dantas da Silveira, DDS, MSc, PhD,{

and L
elia Maria Guedes Queiroz, DDS, MSc, PhD#

Purpose: Oral squamous cell carcinomas have the potential for rapid and unlimited growth. Therefore, hypoxic tissue areas are common in these malignant tumors and contribute to cancer progression, therapy resistance, and poor outcomes. The aim of the present study was to analyze the gene product distribution of hypoxia-inducible factor-1a (HIF-1a) and glucose transporter-1 (GLUT-1) in cases of tongue squamous cell carcinoma (TSCC) and to identify a preliminary correlation between these proteins and clinical staging and Brynes’s histologic grading system (HGS).

Materials and Methods: The sample included 57 cases of TSCC. Histologic sections of 3mm were sub-mitted to the immunoperoxidase method and semiquantitative analysis. The association between HIF-1a and GLUT-1 expression in TSCC and the clinical stage and the HGS of Bryne (1998) was evaluated using the c2_{test, with the significance level set at 0.05 (a = 0.05).}

Results: HIF-1a was mainly expressed in the nucleus/cytoplasm of neoplastic cells, most specimens exhibited diffuse staining in neoplastic cells (84.2%), and focal staining was only observed in perinecrotic areas (15.8%). GLUT-1 was expressed in the cytoplasm and membrane of malignant cells, and diffuse stain-ing was observed in all cases. The intensity of HIF-1a expression correlated significantly with clinical stage (P = .011) and HGS (P = .002). A significant association was observed between the distribution of HIF-1a expression and metastasis (P = .040). Immunoexpression of GLUT-1 correlated significantly with clinical stage (P = .002) and HGS (P = .000). GLUT-1 expression in the peripheral island was predominant in most low-grade tumors (78.6%); in the high-grade cases, the expression prevailed in the location center/periphery (55.8%). Comparison of the location of the tumor island in the different histologic grades showed a statistically significant difference (P = .025).

*Professor, State University of Para
ıba, Campina Grande, Para
ıba, Brazil.

yProfessor, State University of Para
ıba, Campina Grande, Para
ıba, Brazil.

zPhD Student, Postgraduate Program, Oral Pathology, Federal University of Rio Grande do Norte, Natal, Rio Grande do Norte, Brazil.

xPhD Student, Postgraduate Program, Oral Pathology, Federal University of Rio Grande do Norte, Natal, Rio Grande do Norte, Brazil.

kProfessor, Postgraduate Program, Oral Pathology, Federal University of Rio Grande do Norte, Natal, Rio Grande do Norte, Brazil.

{Professor, Postgraduate Program, Oral Pathology, Federal University of Rio Grande do Norte, Natal, Rio Grande do Norte, Brazil.

#Professor, Postgraduate Program, Oral Pathology, Federal University of Rio Grande do Norte, Natal, Rio Grande do Norte, Brazil.

This work was supported by the Brazilian Research Council for Science and Technology Development (CNPq; Edital MCT/CNPq no. 14/2009).

Address correspondence and reprint requests to Dr Pereira de Oliveira: Departamento de Odontologia, Universidade Federal do Rio Grande do Norte, Av Senador Salgado Filho, 1787, Lagoa Nova CEP 59056-000, Natal, RN, Brazil; e-mail: denisehelen2011@ hotmail.com

Received August 7 2014 Accepted March 2 2015

Ó 2015 American Association of Oral and Maxillofacial Surgeons 0278-2391/15/00262-1

http://dx.doi.org/10.1016/j.joms.2015.03.013

Conclusion: The expression of HIF and GLUT proteins within TSCC appears to be associated with dis-ease stage, grade, and the presence of metastases. Additional studies are needed to evaluate the diagnostic and prognostic uses of these proteins in the treatment of TSCC.

Ó 2015 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg 73:1753-1760, 2015

Oral squamous cell carcinoma (OSCC) is the most common head and neck cancer, accounting for more than 90% of all tumors in this region. This lesion can involve areas not easily visualized, such as the lateral and ventral tongue and posterior oropharynx.1It has a relatively high tumor recurrence rate and poor prog-nosis, probably because it is often detected late, frequently progresses to lymph node metastasis and local invasion, and requires aggressive surgical inter-vention and chemo- or radiotherapy to establish local control.1,2 Despite recent advances in the treatment of this lesion, the overall survival rate of patients has remained less than 50% for many decades.2

The aggressiveness of these tumors is related to several factors, including the histologic degree, tumor size, involvement level of the adjacent tissues, the presence of metastases at diagnosis, and the anatomic location of the tumor.3Many studies have attempted to establish a correlation between the clinical and histo-logic characteristics of oral cancer and its prognosis, but the results have been unsatisfactory.4The biologic behavior of OSCC is uncertain; many of these tumors have aggressive biologic behavior at the initial stages with early regional metastases and death. However, advanced tumors can metastasize slowly, and these patients can remain disease free for long periods after surgery.4,5 Therefore, additional independent prognostic markers need to be identified to improve the diagnosis and therapeutic regimens.6,7

The presence of neck metastases is the most impor-tant prognostic factor for OSCC; if present, it results in a 50% reduction in survival rates.5 The growth and metastasis of solid tumors require the induction of angiogenesis (ie, the formation and remodeling of new blood vessels from a pre-existing vascular network to ensure the delivery of oxygen, nutrients, and growth factors to the rapidly dividing transformed cells).8,9One method for tumor cells to acquire the ability to induce angiogenesis, and hence to grow and metastasize, is through activity of the hypoxia-inducible factor (HIF) family of transcription regula-tors, the most important effectors of the adaptive response to hypoxia in multicellular organisms.10

In OSCC, high levels of the HIF-1a subunit correlate with a poor prognosis, reduced disease-specific sur-vival, and tumor progression, including an increase in the size of the primary lesions and lymph node metastasis. HIF-1a is also overexpressed in most pa-tients with SCC of the oropharynx, and this has

corre-lated with a low rate of remission, a high incidence of lymph node metastasis, and poor disease-free and overall survival.11,12

The 2 main issues of cancer progression are insuffi-cient glucose uptake and oxygen support if a tumor outstrips its vasculature.13Malignant cells can present with an increased rate of glucose consumption, and glucose is generally metabolized through glycolysis. It is well-known that glucose uptake is triggered by HIF-1a. One of the most relevant downstream proteins activated by HIF-1a is glucose transporter-1 (GLUT-1), a member of glucose transport molecules.14,15 The rate of glucose transport by GLUT-1 is altered in situa-tions in which the metabolic rate needs to be adjusted, such as cell division (mitosis and meiosis), differentia-tion, transformation, and nutrient starvation.8,16 Increased GLUT-1 expression correlated with poor sur-vival and poor therapeutic outcomes in OSCCs on uni-variate analysis in previous studies.17,18

Therefore, we analyzed the expression of HIF-1 and GLUT-1 using immunohistochemistry and examined the correlation between these proteins and the clinical stage and Brynes’s histologic grading system (HGS).

Materials and Methods

The present retrospective study analyzed 57 speci-mens of TSCC, embedded in paraffin, and obtained from surgical biopsies of patients who had undergone tumor resection without previous chemotherapy and/ or radiotherapy. The Federal University of Rio Grande do Norte research ethics committee approved the pre-sent study (protocol no. 029/10). From the registry files analyzed, information was obtained regarding pa-tient gender, age, race, habits (smoker and alcohol), clinical TNM stage of the lesion, presence or absence of metastasis, and disease outcome (remission or death). Metastases were demonstrated by imaging examinations such as computed tomography or magnetic resonance.

For histologic analysis, 5-mm-thick sections were cut from paraffin-embedded tumor material and stained with hematoxylin-eosin. Malignancy at the invasion front was evaluated using the HGS proposed by Bryne.3For each tumor, the degree of keratinization, nuclear pleomorphism, invasion pattern, and inflam-matory response were analyzed semiquantitatively, attributing scores of 1 to 4. Cases with a total score of 4 to 8 were classified as low-grade malignancy and

cases with a total score higher than 8 were classified as high-grade malignancy.3,4

For immunohistochemistry, 3-mm-thick sections were cut from paraffin-embedded tumor material. The specifications of the primary antibodies (clone, manufacturer, dilution, antigen retrieval, and incuba-tion) are listed inTable 1. Antibodies were detected by immunoperoxidase staining using a dextran polymer-based signal enhancement technique (ADVANCE, Dako, Carpinteria, CA). The reaction was developed with diaminobenzidine as the chro-mogen. The sections were counterstained with Mayer’s hematoxylin and mounted in Permount (Fisher Scientific, Fair Lawn, NJ). Negative controls consisted of replacement of the primary antibodies with bovine serum albumin. A colon carcinoma spec-imen previously shown to be strongly positive for anti–HIF-1a and erythrocyte for anti–GLUT-1 were used as positive controls.

The specimens stained with the anti–HIF-1a and anti–GLUT-1 antibodies were examined under a light microscope (Olympus CH30, Olympus, Tokyo, Japan). The slides were assessed independently by 2 investiga-tors, who were unaware of the clinical data; any disagreement was resolved by discussion. The staining intensity and distribution were evaluated according to the method of Choi et al,19Tanaka et al,20and Ander-sen et al,21 to determine the immunoexpression in stroma or parenchymal cells, the cellular location of this staining (membrane, cytoplasm, nucleus), and location at the periphery or in the center of the tumor islands. Staining intensity was evaluated using scores of 0 to 2, with 0 indicating no staining, 1, weak stain-ing, and 2, strong staining. Cases in which more than 10% of cells were stained were defined as positive.21,22 The staining distribution was analyzed quantitatively using the following criteria adapted from Lee et al23: focal (<50% of immunostained cells) and diffuse (>50% of immunostained cells).

The results were submitted to statistical analysis. The data were originally presented in Excel format (Office XP Professional, Microsoft, Redmond, WA) were exported to database format and analyzed using

the Statistical Package for Social Sciences for Windows XP, version 10.0 (IBM, Armonk, NY). The association between HIF-1a and GLUT-1 expression patterns in TSCC, and the clinical and morphologic variables was evaluated using thec2test, with the level of signif-icance set at 0.05 (a = 0.05).

Results

PROFILE OF THE SAMPLE

Most patients with TSCC were men (n = 40, 70.2%) and ranged in age from 61 to 70 years (n = 20, 35.1%). White patients predominated in the sample (n = 32, 56.2%). Smoking and alcohol consumption were the most frequent habits (n = 24, 42.1%), followed by pa-tients who only smoked (n = 18, 31.6%). TNM stage III was the most frequent (n = 22, 38.6%), followed by stage I (n = 14, 24.6%), IV (n = 11, 19.3%), and II (n = 10, 17.5%). With respect to outcome, tumor remission was observed in 77.2% (n = 44) of the cases, and 57.9% (n = 57.9%) developed metastases (Table 2).

HIF-1a IMMUNOSTAINING

Staining for HIF-1a was positive in the stroma and parenchyma in 100% of the TSCC cases. This protein was expressed mainly in the nucleus/cytoplasm of neoplastic cells (n = 44, 77.2%), followed by expres-sion in the nucleus only (n = 12, 21.1%) and cytoplasm (n = 1, 1.8%). In addition, most specimens exhibited diffuse staining in the neoplastic cells (n = 48, 84.2%). Focal staining was only observed in perine-crotic regions (n = 9, 15.8%). Evaluation of the staining intensity revealed strong staining in 68.4% (n = 39) and weak staining in 31.6% (n = 18). Analysis of the inva-sion front showed expresinva-sion of HIF-1a at the periph-ery and in the center of the tumor islands and/or sheets (P = .040;Table 3).

The intensity of HIF-1a staining correlated signifi-cantly with the clinical stage (P = .011), with 13 (22.8%) of the 18 (31.6%) cases exhibiting weak stain-ing bestain-ing in the early disease stages (stages I and II), and with the histologic grade of malignancy (P = .002). In the latter case, strong staining was

Table 1. SPECIFICITY, DILUTION, MANUFACTURER, ANTIGEN RETRIEVAL, AND INCUBATION OF ANTIBODIES USED

Specificity Clone Manufacturer Dilution Antigen Retrieval Incubation

HIF-1a H1a67 Santa Cruz Biotechnology* 1:200 Tris/EDTA, pH 9.0, Pascal, 3 min

Overnight (18 h)

GLUT-1 SLC2A1 GeneTexy 1:400 Citrate, pH 6.0, Pascal, 3 min 60 min

Abbreviations: GLUT-1, glucose transporter-1; HIF-1a, hypoxia-inducible factor-1a. * Santa Cruz, CA.

y Irvine, CA.

mainly observed in the high-grade cases (n = 34, 79.1%) and weak staining in the low-grade cases (n = 9, 64.3%;Figs 1, 2).

GLUT-1 IMMUNOSTAINING

GLUT-1 was expressed in the cytoplasm and mem-brane of the malignant cells. Diffuse staining was observed in all cases. In the stroma, red blood cells and some blood vessels were positive for this protein. Most cases exhibited strong immunoexpression (n = 44, 77.2%). GLUT-1 was expressed more at the periph-ery of the tumor islands (n = 30, 52.6%) than in the center or periphery (n = 27, 47.4%).

The intensity of GLUT-1 staining correlated signifi-cantly with the clinical stage (P = .002). In this respect, 11 (19.3%) of the 13 cases (22.8%) exhibiting weak staining were in the early disease stages (stages I and II;Fig 3). In contrast, a predominance of strong stain-ing was observed in the more advanced disease stages (stage III followed by stage IV;Table 4,Fig 4). A signif-icant association was also observed between the inten-sity of GLUT-1 staining and the histologic grade of malignancy (P = .000), with 90.7% (n = 39) of high-grade TSCC cases exhibiting strong staining.

With respect to the location of staining in the tumor islands, peripheral staining was observed in most

Table 3. ASSOCIATION OF DISTRIBUTION AND INTENSITY OF HIF-1a STAINING WITH METASTASIS, CLINICAL STAGE, AND HISTOLOGIC GRADE OF MALIGNANCY

HIF-1a Variable Total c2 P Value Metastasis Absent Present Staining distribution 4.212 .040 Focal 1 (4.2) 8 (24.2) 9 (15.78) Diffuse 23 (95.8) 25 (75.8) 48 (84.2) Clinical Stage Total c2 P Value I II III IV Staining intensity 11.049 .011 Weak 8 (57.1) 5 (50) 2 (9.1) 3 (27.3) 18 (31.6) Strong 6 (42.9) 5 (50) 20 (90.9) 8 (72.7) 39 (68.4) Histologic Grade Total c2 P Value Low High Staining intensity 9.188 .002 Weak 9 (64.3) 9 (20.9) 18 (31.6) Strong 5 (35.7) 34 (79.1) 39 (68.4) Data presented as n (%).

Abbreviation: HIF-1a, hypoxia-inducible factor-1a.

Vasconcelos et al. HIF-1a and GLUT-1 in Human Tongue Cancer. J Oral Maxillofac Surg 2015. Table 2. CLINICAL AND PATHOLOGIC

CHARACTERISTICS OF PATIENTS WITH SQUAMOUS CELL CARCINOMA OF THE TONGUE

Variable Patients (n) Gender Male 40 (70.2) Female 17 (29.8) Race Black 14 (24.5) Mulatto 11 (19.3) White 32 (56.2)

Age group (yr)

31-40 3 (5.3) 41-50 6 (10.5) 51-60 18 (31.6) 61-70 20 (35.1) 71-80 7 (12.3) 81-90 3 (5.2) Habits Smoker 18 (31.6)

Smoker and alcohol 24 (42.1)

Nonsmoker and alcohol 1 (1.8)

Unknown 14 (24.5)

Data in parentheses are percentages.

Vasconcelos et al. HIF-1a and GLUT-1 in Human Tongue Cancer. J Oral Maxillofac Surg 2015.

low-grade specimens (n = 11, 78.6%; Fig 4). In contrast, central/peripheral staining predominated in high-grade cases (n = 24, 55.8%;Table 4).

Discussion

Great variations exist in the treatment response and prognosis (ie, patients with tumors at a similar site and stage are treated similarly, with some patients experi-encing prolonged survival and others dying within a short period of regional or distant metastasis). Major difficulties in determining the most appropriate treat-ment of OSCC is that it is still based mainly on the

site and clinical tumor stage.24Because of these diffi-culties, additional independent prognostic markers need to be identified to improve the diagnosis and therapeutic regimens.6,7 It has been suggested that hypoxia is an independent adverse prognostic factor in patients with head and neck cancer.25 The main transcription factor activated by hypoxia is HIF-1a, which regulates cell metabolism as part of the cell’s response to hypoxia. The largest functional group of genes consistently regulated by HIF-1a are the genes associated with glucose metabolism.6HIF-1a increases the rate of glucose uptake through the transcriptional activation of the facultative glucose transporters GLUT-1 and GLUT-3. Elevated expression of either HIF1-a or GLUT-1 has been detected in many different cancers, including head and neck tumor.14Therefore, we analyzed the expression of HIF-1a and GLUT-1 and examined the correlation between these proteins and the clinical stage and Brynes’s HGS. We were able to show that the intensity of HIF-1a expression corre-lated significantly with the clinical stage (P = .011) and HGS (P = .002). A significant association was observed between the distribution of HIF-1a staining and metastasis (P = .040). Immunoexpression of GLUT-1 correlated significantly with clinical stage (P = .002) and HGS (P = .000).

Zhu et al26and Liang et al27reported an association between the overexpression of HIF-1a in more advanced OSCC and poor prognosis, lower survival rate, and resistance to antineoplastic treatment. Roh et al,28 analyzing the role of 5 hypoxia markers (HIF-1a, HIF-2a, carbonic anhydrase IX, GLUT-1, and erythropoietin receptor), found that HIF-1a is signifi-cantly associated with disease-specific survival in stage

FIGURE 2. Strong immunohistochemical expression of hypoxia-inducible factor-1a in a low-grade squamous cell carcinoma of the tongue (ADVANCE horseradish peroxidase method, original magnification200).

Vasconcelos et al. HIF-1a and GLUT-1 in Human Tongue Cancer. J Oral Maxillofac Surg 2015.

FIGURE 3. Weak immunohistochemical expression of glucose transporter-1 in a low-grade squamous cell carcinoma of the tongue (ADVANCE horseradish peroxidase method, original magnification 200).

Vasconcelos et al. HIF-1a and GLUT-1 in Human Tongue Cancer. J Oral Maxillofac Surg 2015.

FIGURE 1. Weak immunohistochemical expression of hypoxia-inducible factor-1a in a high-grade squamous cell carcinoma of the tongue (ADVANCE horseradish peroxidase method, original magnification200).

Vasconcelos et al. HIF-1a and GLUT-1 in Human Tongue Cancer. J Oral Maxillofac Surg 2015.

T2 tongue carcinoma, with a relative risk of 0.352, but not with other clinical or pathologic parameters such as tumor thickness, lymph node involvement, or resec-tion margin status. Liang et al27reported an association between HIF-1a overexpression in these tumors and shorter disease-free survival.

The present study demonstrated a significant corre-lation between the HIF-1a staining intensity and the clinical stage and histologic grade of malignancy, with high- and low-grade cases exhibiting strong and weak staining, respectively. Liu et al29observed over-expression of HIF-1a in 69.64% of OSCC cases analyzed, with a positive correlation with the rate of tumor progression (ie, tumor size, lymph node metas-tasis, histologic differentiation, clinical stage, and pres-ence of tumor cell-lined vessels) and poor survival. In agreement with the findings of Fillies et al30and Lin et al,31 most specimens studied had diffuse HIF-1a staining, with focal staining observed in the surround-ing necrotic areas. In addition, a significant correlation was found with metastasis, with the presence of me-tastases in 43.8% of the 84.2% of cases with diffuse staining. According to Fillies et al30 and Semenza,32 the predominantly diffuse expression of HIF-1a in tumors is related, not only to the hypoxic environ-ment, but also to alterations in oncogenes and tumor suppressor genes and the lack of GLUT-1 expression. Because most of the diffuse distribution of this mole-cule in normal epithelium is found in deep layers when associated with keratinized areas, HIF-1a prob-ably has a more physiologic function within this context, participating in cell differentiation instead of exerting a carcinogenic effect.

Table 4. ASSOCIATION OF DISTRIBUTION AND INTENSITY OF GLUT-1 STAINING WITH METASTASIS, CLINICAL STAGE, AND HISTOLOGIC GRADE OF MALIGNANCY

GLUT-1 Variable Total c2 P Value Metastasis Absent Present Staining intensity 0.952 .329 Weak 7 (29.2) 6 (18.2) 13 (22.8) Strong 17 (70.8) 27 (81.8) 44 (77.2) Clinical Stage Total c2 P Value I II III IV Staining intensity 15.270 .002 Weak 8 (57.1) 3 (30) 2 (9.1) 0 (0) 13 (22.8) Strong 6 (42.9) 7 (70) 20 (90.9) 11 (100) 44 (77.2) Histologic Grade Total c2 P Value Low High Staining intensity 18.136 .000 Weak 9 (64.3) 4 (9.3) 13 (22.8) Strong 5 (35.7) 39 (90.7) 44 (77.2) Data presented as n (%).

Abbreviation: GLUT-1, glucose transporter-1.

Vasconcelos et al. HIF-1a and GLUT-1 in Human Tongue Cancer. J Oral Maxillofac Surg 2015.

FIGURE 4. Strong immunohistochemical expression of glucose transporter-1 in a high-grade squamous cell carcinoma of the tongue (ADVANCE horseradish peroxidase method, original magnification 200).

Vasconcelos et al. HIF-1a and GLUT-1 in Human Tongue Cancer. J Oral Maxillofac Surg 2015.

A significant association was observed between the expression of GLUT-1 and the clinical stage. Similar re-sults have been reported by Kato et al33for esophageal carcinoma, Choi et al19for lip SCC, and Ayala et al34for OSCC, highlighting the importance of the association between this immunomarker and tumor stage and its use as a diagnostic and prognostic marker. However, Schrijvers et al35 found no association between the parameters of clinical evolution and GLUT-1 staining intensity in cases of SCC of the larynx.

Several mechanisms have been proposed to explain the upregulation of anaerobic glycolysis in cancer, in particular, those involving the activation of oncogenes and/or inactivation of tumor suppressor genes that affect glycolytic activity. For example, the c-Myc gene is known to activate most genes that encode glycolytic enzymes, such as lactate dehydrogenase A and GLUT-1, which increase lactate production and improve glucose uptake,36respectively. According to Vander Heiden et al,37 mutations that result in the loss of function of succinate dehydrogenase and fuma-rate hydrase, enzymes participating in the Krebs cycle, cause the accumulation of succinate and fumarate in the cytoplasm, respectively. These metabolites, in turn, are responsible for inhibition of prolyl hydro-lyase, an enzyme that catalyzes the hydroxylation of HIF-1a, with a consequent increase in the concentra-tion of this protein and signaling of a Warburg effect emergency, thus explaining the greater expression of GLUT-1 in malignant cells. Similarly, mutations in isoci-trate dehydrogenase types 1 (cytoplasmic IDH1) and 2 (mitochondrial IDH2) result in the reduction of a-keto-glutarate into 2-oxoa-keto-glutarate, also contributing to the constitutive activation of HIF-1a. These facts indicate that GLUT-1 is involved in the complex interactions that occur between the glucose metabolism pathways and the mechanisms of tumor hypoxia, which are essential for cell growth and proliferation.

Analysis of the association between the histologic grade of malignancy and intensity of GLUT-1 staining showed a greater frequency of high-grade and low-grade cases exhibiting strong and weak staining, respectively. Similarly, Kato et al33in esophageal carci-noma, Li et al38in head and neck SCC, and Ohba et al12 in OSCC observed significantly weaker immunohisto-chemical expression of this molecule in primary and well-differentiated tumors compared with recurrent and undifferentiated tumors. This finding suggests that GLUT-1 can be used as a molecular target for the diagnosis, prognosis, and adequate therapeutic choice of head and neck SCCs.

GLUT-1 staining was detected at the periphery of islands of neoplastic epithelial cells, in agreement with the findings from Ohba et al.12In addition, a sig-nificant correlation was found between the positive staining in the tumor islands and the histologic grade

of malignancy, with central/peripheral staining pre-dominating in high-grade tumors. However, no corre-lation was observed with metastasis or clinical stage. Studies have shown that GLUT-1 is expressed mainly at the periphery of the tumor and is absent in the cen-ter because of a greacen-ter concentration of glycogen and well-differentiated cells.12These findings suggest that the expression of GLUT-1 is inversely proportional to the quantity of glycogen stored in the cells. Cells located at the periphery of the tumor invasion front require high glycolytic activity for differentiation, inva-sion, and metastasis. In the center of the tumor islands, the glycogen reserves inside differentiated cells and minimum oxygen tension will be sufficient to maintain satisfactory aerobic respiration despite the hypoxic environment. It is believed that GLUT-1 is one of the adaptive strategies that permit tumor growth and pro-gression in an environment characterized by marked oxygenation instability.12,39

GLUT-1 was expressed in the cytoplasm and mem-brane of neoplastic cells in 100% of the TSCC cases studied, in agreement with the findings of Kunkel et al18and Choi et al,19who observed greater mem-brane and/or cytoplasmic expression in OSCCs. In contrast, Li et al38and Ogane et al40reported a predom-inance of membrane expression in head and neck SCC. Demasi et al39 and Ohba et al12 observed greater expression in the cytoplasm of neoplastic cells of mucoepidermoid carcinoma and OSCC, respectively. These findings suggest greater transcription of GLUT-1 during cell growth and proliferation, which is stored in intracytoplasmic endosomes and expressed on the plasma membrane surface, depending on the need for higher energy uptake by neoplastic cells.16

The significance of induction of GLUT-1 expression for neoplastic transformation of the epithelium is not well established. Schrijvers et al35 and Zhu et al26 demonstrated the coexpression of GLUT-1 and HIF-1a in head and neck SCC cases, supporting the hypoth-esis that GLUT-1 is expressed in response to tissue hypoxia. However, Pedersen et al41 suggested that GLUT-1 expression is a constitutive characteristic of the malignant phenotype and is induced by Ras and Src oncogenes. In the present study, low-grade cases or cases in the early stages of disease (stage I and II) ex-hibited strong staining for these markers of hypoxia. In contrast, high-grade cases or cases in the late disease stages (stage III and IV) showed weak expression.

In conclusion, our study has provided additional ev-idence that hypoxia-regulated proteins, such as HIF-1a and GLUT-1, can be used in the clinic to infer the pres-ence of hypoxia in tumors. However, the type of scoring system used to determine the predictive qual-ity of these hypoxia markers requires future studies of their combined use in a variety of clinical situations. The results of the present study suggest that the

expression of HIF and GLUT proteins within TSCC ap-pears to be associated with disease stage, grade, and the presence of metastases. Furthermore, we empha-size the difficulty in standardizing parameters that can be universally used as prognostic markers. There-fore, additional studies are required to validate the util-ity of these proteins as independent prognostic markers for patients with OSCC.

References

1. Alves PM, Godoy GP, Gomes DQ, et al: Significance of galectins-1, -3, -4 and -7 in the progression of squamous cell carcinoma of the tongue. Pathol Res Pract 207:236, 2011

2. Zhou H, Yang YH, Binmadi NO, et al: The hypoxia-inducible fac-tor-responsive proteins semaphorin 4D and vascular endothelial growth factor promote tumor growth and angiogenesis in oral squamous cell carcinoma. Exp Cell Res 318:1685, 2012

3. Bryne M: Is the invasive front of an oral carcinoma the most important area for prognostication? Oral Dis 4:70, 1998

4. Silveira EJ, Godoy GP, Lins RD, et al: Correlation of clinical, his-tological, and cytokeratin profiles of squamous cell carcinoma of the oral tongue with prognosis. Int J Surg Pathol 15:376, 2007

5. Montoro JR, Ricz HA, Souza LD, et al: Prognostic factors in squa-mous cell carcinoma of the oral cavity. Braz J Otorhinolaryngol 74:861, 2008

6. Eckert AW, Lautner MH, Sch€utze A, et al: Co-expression of Hif1al-pha and CAIX is associated with poor prognosis in oral squa-mous cell carcinoma patients. J Oral Pathol Med 39:313, 2010

7. Luo XM, Zhou SH, Fan J: Glucose transporter-1 as a new therapeu-tic target in laryngeal carcinoma. J Int Med Res 38:1885, 2010

8. Burrows N, Resch J, Cowen RL, et al: Expression of hypoxia-inducible factor 1 alpha in thyroid carcinomas. Endocr Relat Cancer 17:61, 2010

9. Hosokawa Y, Okumura K, Terashima S, et al: Radiation protec-tive effect of hypoxia-inducible factor-1a (HIF-1a) on human oral squamous cell carcinoma cell lines. Radiat Prot Dosimetry 152:159, 2012

10. P
erez-Say
ans M, Su
arez-Pe~naranda JM, Pilar GD, et al: Hypoxia-inducible factors in OSCC. Cancer Lett 313:1, 2011

11. Uehara M, Sano K, Ikeda H, et al: Hypoxia-inducible factor 1 alpha in oral squamous cell carcinoma and its relation to prog-nosis. Oral Oncol 45:241, 2009

12. Ohba S, Fujii H, Ito S, et al: Overexpression of GLUT-1 in the invasion front is associated with depth of oral squamous cell car-cinoma and prognosis. J Oral Pathol Med 39:74, 2010

13. Eckert AW, Kappler M, Schubert J, et al: Correlation of expres-sion of hypoxia-related proteins with prognosis in oral squa-mous cell carcinoma patients. Oral Maxillofac Surg 16:189, 2012

14. Denko NC: Hypoxia, HIF 1and glucose metabolism in the solid tumor. Nat Rev Cancer 8:705, 2008

15. Rankin EB, Giaccia AJ: The role hypoxia-inducible factors in tumorigenesis. Cell Death Differ 15:678, 2008

16. Mendes AI, Matos P, Moniz S, et al: Protein kinase WNK1 pro-motes cell surface expression of glucose transporter GLUT1 by regulating a Tre-2/USP6 BUB2-Cdc16 domain family member 4 (TBC1D4)-Rab8A complex. J Biol Chem 285:39117, 2010

17. Jonathan RA, Wijffels KI, Peeters W, et al: The prognostic value of endogenous hypoxia-related markers for head and neck squa-mous cell carcinomas treated with ARCON. Radiother Oncol 79: 288, 2006

18. Kunkel M, Moergel M, Stockinger M, et al: Overexpression of GLUT-1 is associated with resistance to radiotherapy and adverse prognosis in squamous cell carcinoma of the oral cavity. Oral Oncol 43:796, 2007

19. Choi YS, Kim SJ, Kim DS, et al: Glucose transporter-1 expression in squamous cell carcinoma of the tongue. Cancer Res Treat 39: 109, 2007

20. Tanaka N, Kato H, Inose T, et al: Expression of carbonic anhy-drase 9, a potential intrinsic marker of hypoxia, is associated

with poor prognosis in oesophageal squamous cell carcinoma. Br J Cancer 99:1468, 2008

21.Andersen S, Eilertsen M, Donnem T, et al: Diverging prognostic impacts of hypoxic markers according to NSCLC histology. Lung Cancer 72:294, 2011

22.Kim SJ, Shin HJ, Jung KY, et al: Prognostic value of carbonic an-hydrase IX and Ki-67 expression in squamous cell carcinoma of the tongue. J Clin Oncol 37:812, 2007

23.Lee SS, Tsai CH, Yang SF, et al: Hypoxia inducible factor-1a expression in areca quid chewing-associated oral squamous cell carcinomas. Oral Dis 16:696, 2010

24.Rodrigues PC, Miguel MC, Bagordakis E, et al: Clinicopatholog-ical prognostic factors of oral tongue squamous cell carcinoma: A retrospective study of 202 cases. Int J Oral Maxillofac Surg 43: 795, 2014

25.Le QT, Kong C, Lavori PW, et al: Expression and prognostic signif-icance of a panel of tissue hypoxia markers in head-and neck squa-mous cell carcinomas. Int J Radiat Oncol Biol Phys 69:167, 2007

26.Zhu GQ, Tang YL, Li L, et al: Hypoxia inducible factor 1a and hypoxia inducible factor 2a play distinct and functionally over-lapping roles in oral squamous cell carcinoma. Clin Cancer Res 16:4732, 2010

27.Liang X, Zheng M, Jiang J, et al: Hypoxia-inducible factor-1 alpha, in association with TWIST2 and SNIP1, is a critical prognostic factor in patients with tongue squamous cell carcinoma. Oral Oncol 47:92, 2011

28.Roh JL, Cho KJ, Kwon GY, et al: The prognostic value of hypoxia markers in T2-staged oral tongue cancer. Oral Oncol 45:63, 2009

29.Liu TR, Chen FJ, Yang AK, et al: Elective neck dissection in clin-ical stage I squamous cell carcinoma of the tongue: Does it improve regional control or survival time? Oral Oncol 47:136, 2011

30.Fillies T, Werkmeister R, Van Diest PJ, et al: HIF1-alpha overex-pression indicates a good prognosis in early stage squamous cell carcinoma of the oral floor. BMC Cancer 5:84, 2005

31.Lin PY, Yu CH, Wang JT, et al: Expression of hypoxia-inducible factor-1 alpha is significantly associated with the progression and prognosis of oral squamous cell carcinomas in Taiwan. J Oral Pathol Med 37:18, 2008

32.Semenza GL: Development of novel therapeutic strategies that target HIF-1. Expert Opin Ther Targets 10:267, 2006

33.Kato H, Takita J, Miyazaki T, et al: Glut-1 glucose transporter expression in esophageal squamous cell carcinoma is associated with tumor aggressiveness. Anticancer Res 22:2635, 2002

34.Ayala FR, Rocha RM, Carvalho KC, et al: GLUT1 and GLUT3 as potential prognostic markers for oral squamous cell carcinoma. Molecules 15:2374, 2010

35.Schrijvers ML, van der Laan BF, de Bock GH, et al: Overexpression of intrinsic hypoxia markers HIF1alpha and CA-IX predict for local recurrence in stage T1-T2 glottic laryngeal carcinoma treated with radiotherapy. Int J Radiat Oncol Biol Phys 72:161, 2008

36.Macheda ML, Rogers S, Best JD: Molecular and cellular regula-tion of glucose transporter (GLUT) proteins in cancer. J Cell Physiol 202:654, 2005

37.Vander Heiden MG, Cantley LC, Thompson CB: Understanding the Warburg effect: The metabolic requirements of cell prolifer-ation. Science 324:1029, 2009

38.Li SJ, Guo W, Ren GX, et al: Expression of Glut-1 in primary and recurrent head and neck squamous cell carcinomas, and compared with 2-[18F]fluoro-2-deoxy-D-glucose accumulation in positron emission tomography. Br J Oral Maxillofac Surg 46: 180, 2008

39.Demasi AP, Costa AF, Altemani A, et al: Glucose transporter protein 1 expression in mucoepidermoid carcinoma of salivary gland: Cor-relation with grade of malignancy. Int J Exp Pathol 91:107, 2010

40.Ogane N, Yasuda M, Shimizu M, et al: Clinicopathological impli-cations of expressions of hypoxia related molecules in esopha-geal superficial squamous cell carcinoma. Ann Diagn Pathol 14:23, 2010

41.Pedersen MW, Holm S, Lund EL, et al: Coregulation of glucose uptake and vascular endothelial growth factor (VEGF) in two small-cell lung cancer (SCLC) sublines in vivo and in vitro. Neoplasia 3:80, 2001