Gene and protein expression of glucose transporter 1 and glucose transporter 3 in human laryngeal cancer—the relationship with regulatory hypoxia-inducible factor-1α expression, tumor invasiveness, and patient prognosis

The study material constituted 106 fresh biopsy tissue samples obtained from squamous cell laryngeal carcinoma (SCC) cases, and the control constituted 73 tissue samples from a morphologically estimated noncancerous laryngeal mucosa (NCM) from individuals who were qualified for total laryngectomy. According to ethical rules, the adjacent normal epithelium of the larynx in patients qualified for partial laryngectomy was not sampled. The patients (100 men and 6 women, mean age 62.4 ± 9.1 years) were recruited between January 2003 and December 2011 and were under treatment at the Department of Otolaryngology and Laryngological Oncology, Medical University of Łódź, Poland. All individuals had a confirmed diagnosis of laryngeal carcinoma based on histopathological evaluation and had undergone partial or total laryngectomy, depending on the extent of neoplastic lesions. All fresh samples were stored at −80 °C before the analyses. The tissue specimens collected in the operation room were prepared and evaluated by an experienced pathologist. Normal laryngeal tissues were collected from the sites as far as possible from the margins of the tumor by individually harvesting samples from presumptive noncancerous regions. Hematoxylin and eosin (H&E)-stained sections provided histological confirmation of noncancerous and cancerous tissues. The criteria for patient participation in this study were as follows: (1) a pathologically confirmed diagnosis of squamous cell planoepitheliale carcinoma; (2) primary surgical resection without receiving prior immunotheraphy, radiotheraphy, or chemotherapy; (3) absence of distant metastases and second primary neoplasms; (4) a negative history of previously diagnosed with other types of primary cancers; and (5) a negative history of recurrences of laryngeal cancer. Informed consent was obtained from each subject. The investigations were performed with the approval of the Bioethical Commission of the Medical University of Łódź and the National Science Council, Poland (approval No RNN/60/13/KE). In all cases, surveys were performed to complete the cancer registry database. The database catalog was queried every 2 months and identified all histopathologically confirmed incident primary squamous cell laryngeal cancer cases reported within 4 months of diagnosis preceding the recruitment. The sociodemographic features of the study subjects are shown in Table 1.

Archival formalin-fixed paraffin-embedded tissue samples were used for the histological classification of tumors. All specimens were assessed according to the criteria conducted in accordance with the AJCC TNM classification of 2010 for laryngeal cancers [34]. Morphological estimation was performed on H&E-stained sections in the most invasive, peripheral zones of the tumor, according to tumor front grading (TFG), which is one of the most reliable pathological methods for the analysis of neoplastic progress and determination of the dynamics of tumor growth, as well as a reasonably precise prognostic factor in laryngeal carcinoma [35]. The histological evaluation considers the mode and depth of invasion as well as total TFG score. The factors were assessed in at least five different regions of the peripheral part of the tumor (magnification ×200, number of mitoses magnification ×400). Each factor was graded according to a scale ranging from 1 to 4. The numeric morphological TFG score was computed as the sum of tumor-related features (cytoplasmic differentiation, nuclear polymorphism, number of mitoses) and adjacent stroma-related characteristics of the peripheral edge of tumor infiltration (mode of invasion, depth of invasion, and plasmalymphocytic infiltration), with a maximum score of 24 points. According to the TFG total score, tumors were divided into four groups: 6–9, 10–13, 14–17, 18–21, and >22 TFG points. The clinicopathological characteristics of the laryngeal cancers are shown in Table 2.

The tissue specimens collected in the operation room were prepared and evaluated by an experienced pathologist. Samples were stored at −80 °C until RNA preparation. Total RNA was isolated using Trizol® Reagent (Sigma-Aldrich, USA) according to manufacturer’s protocol and quantified spectrophotometrically. RNA was eluted in 20 μl RNase-free water, quantified by spectrophotometry at 260 nm and stored at −20 °C. RNA with a 260/280 nm ratio in the range 1.8–2.0 was considered high quality. First-strand complementary DNAs (cDNAs) were obtained by reverse transcription of 1 μg of total RNA using RevertAid™ first-strand cDNA synthesis kit (Fermentas International, Lithuania) following the manufacturer’s protocol.

Real-time gene expression analysis of target genes (SLC2A1, SLC2A3, and HIF-1α) was performed using TaqMan® Gene Expression Assays (Applied Biosystems, USA) according to manufacturer’s instructions. The hypoxanthine phosphoribosyltransferase 1 (HPRT1) gene was used as internal control. The assay numbers for these genes were as follows: Hs00892681_m1, Hs00359840_m1, Hs00936368_m1, and Hs02800695_m1. Each PCR reaction was performed in a 10-μl volume that included 5 μl of 2× TaqMan Universal PCR MasterMix (Applied Biosystems, USA), 4.5 μl of water diluted cDNA template (50 ng), and 0.5 μl of TaqMan® Gene Expression Assay consisted of a pair of unlabeled PCR primers and TaqMan probe with a FAM™. The RT-qPCR reaction was carried out using the Mastercycler ep realplex (Eppendorf) under the following conditions: denaturation for 10 min at 95 °C followed by 50 cycles of 15 s at 95 °C, 1 min annealing and extension at 60 °C. Relative RNA quantification was performed using the ΔCt method. ΔCt (Ct_gene − Ct _HPRT1 ) values were recalculated into relative copy number values (number of SLC2A1, SLC2A3, and HIF-1α mRNA copies per 1000 copies of HPRT1 mRNA).

To determine the SLC2A1 and SLC2A3 gene amplification, copy number quantification was carried out using quantitative real-time PCR Mastercycler ep realplex (Eppendorf) with the glucokinase (GCK) gene used as the reference gene. The real-time PCR primers are as follows: SLC2A1 (f) 5′-TGTGCAACCCATGAGCTAA-3′, SLC2A1 (r) 5′-CCTGGTCTCATCTGGATTCT-3′; SLC2A3 (f) 5′-TTCGTCTCTAGCCTGCACTG-3′, SLC2A3 (r) 5′-ACACAACTTCTCCGGGTGAC-3′; GCK(f) 5′-CGGATGCAGAAGGAGATGGA-3′, and GCK(r) 5′-CATCTTCACACTGGCCTCTTCA-3′. Real-time PCR was performed in 50-μl reaction volumes that contained 2× Power SYBR Green PCR Master Mix (Applied Biosystems, USA) and 0.9 mM forward and reverse primers. PCR conditions were as follows: 5 s at 95 °C followed by 40 cycles consisting of 15 s at 95 °C and 30 s at 60 °C. ΔCt was calculated by a Ct value of GCK taking away that of SLC2A1 and SLC2A3 and three or more ΔCt was defined as amplified.

The protein content of the tissue homogenate fraction was estimated by means of the modified Lowry procedure [36] using bovine serum albumin (BSA) as standard. The samples (50 μg protein/lane) of homogenates were resolved by 8 % SDS-PAGE and electroblotted onto Immobilon-P transfer membranes (Millipore, Bedford, MA, USA). The blots were incubated 1 h with rabbit polyclonal anti-GLUT1 (Abcam, UK), mouse monoclonal anti-GLUT3 (Santa Cruz Biotechnology, Inc., USA) or rabbit polyclonal anti-HIF-1α antibodies (Santa Cruz Biotechnology, Inc., USA) in a 1:1000, 1:400, and 1:1000 dilution, respectively. After being washed three times with Tris-buffered saline with Tween-20 (TBST), the membranes were incubated 1 h with goat anti-rabbit or anti-mouse antibodies conjugated with horseradish peroxidase (1:5000 dilution). The membranes were again washed three times with TBST and incubated with peroxidase substrate solution (3,3′-diaminobenzidine (DAB)). Gel-Pro® Analyzer software (Media Cybernetics Inc., USA) was used for densitometry analysis of protein bands. The integrated optical density (IOD) of the bands, in a digitized picture, was measured. For the immunoblot analysis, an IOD less than 5 was taken as negative.

All RT-PCR and Western blot reactions were repeated three times for each sample.

The statistical analyses were performed using STATISTICA version 9.0 (StatSoft, Poland). Since levels of expression in endometrial and breast cancer specimens did not follow a normal distribution (Kolmogorov-Smirnov test), nonparametrical statistical tests were applied (Mann-Whitney U test, Spearman’s rank analysis). Kruskal-Wallis test with post hoc multiple comparisons was used according to clinical data. Kaplan-Meier survival analysis was performed to determine the association of SLC2A1, SLC2A3, and HIF-1α mRNA expression with overall survival. The cutoff value was established to be the median of SLC2A1, SLC2A3, and HIF-1α mRNA level. The survival curves were compared between two groups: high (≥ median value) and low (< median value) expression using log-rank tests. Distribution of quantitative variables was described using means and standard deviations. A p value <0.05 was considered as statistically significant.