The current study was approved by the Research Ethics Committee of The First Affiliated Hospital of Medical College, Shantou University (Shantou, China). All patients enrolled in the study provided written informed consent. All specimens were handled and made anonymous according to ethical and legal standards of The First Affiliated Hospital of Medical College, Shantou University, and were obtained under sterile conditions during surgery.

For western blotting and qPCR, 30 pairs of freshly prepared glioma and non-neoplastic brain tissues were obtained from the Department of Neurosurgery, The First Affiliated Hospital of Medical College, Shantou University (Shantou, China) between January 2010 and December 2014. The mean patient age was 50.6 years (range, 12–88 years), and 20 (66.67%) of them were male, while 10 (33.33%) were female. According to the clinicopathological criteria provided by the World Health Organization (WHO) (14): 3 patients were WHO grade I as pilocytic astrocytomas; 10 patients were grade II, including 5 patients with fibrillary astrocytoma, 3 patients with protoplasmic astrocytoma and 2 patients with oligodendroglioma; 12 patients were WHO grade III, including 8 patients with anaplastic astrocytoma and 4 patients with anaplastic oligodendroglioma; and 5 patients were WHO grade IV, all GBM.

For immunohistochemistry, 180 patients-derived paraffin-embedded glioma tissues were obtained from the Department of Neurosurgery, The First Affiliated Hospital of Medical College, Shantou University between January 2005 and December 2014. The mean patient age was 50.8 years (range, 10–86 years). Of these patients, 110 (61.11%) were male and 70 (38.89%) were female. According to the clinicopathological criteria provided by WHO (14), 15 patients were WHO grade I as pilocytic astrocytomas; 55 patients were WHO grade II, including 25 patients with fibrillary astrocytoma, 20 patients with protoplasmic astrocytoma and 10 patients with oligodendroglioma; 85 patients were WHO garde III, including 50 patients with anaplastic astrocytoma and 35 patients with anaplastic oligodendroglioma; and 25 patients were WHO grade IV, all GBM. Follow-up data were completed for all 180 patients, with a median follow-up time of 32 months (range, 2–118 months). The clinicopathological characteristics are summarized in Table I. All patients did not undergo any other treatments prior to surgery.

qPCR was performed to detect the expression of LKB1 mRNA in 30 pairs of freshly prepared glioma and non-neoplastic brain tissues. Total RNA was extracted from 30 pairs of freshly prepared glioma and non-neoplastic brain tissues using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) according to the manufacturer's protocol. The RNA was pretreated with DNase, and single-stranded complementary DNA was synthesized using the SuperScript First-Strand Synthesis System or RT-PCR (Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Relative expression of LKB1 mRNA was determined using a SYBR Green PCR Master Mix (Applied Biosystems; Thermo Fisher Scientific, Inc.) and an ABI 7500 Real-Time PCR System (Applied Biosystems; Thermo Fisher Scientific, Inc.), and normalized using GAPDH. The primer sequences used in the present study were as follows: LKB1 forward, 5′-AGGGATGCTTGAGTACGAACC-3′ and reverse, 5′-GTCCTCCAAGTACGGCACC-3′; GAPDH forward, 5′-TGAACGGGAAGCTCACTGG-3′ and reverse, 5′-TCCACCACCCTGTTGCTGTA-3′. PCR amplifications for each gene were repeated three times. The fold-change of each gene was calculated using the 2^−ΔΔCq method (15,16).

Western blot analysis was performed to detect the expression of LKB1 protein in 30 pairs of freshly prepared glioma and non-neoplastic brain tissues. Tissue samples were ground with liquid nitrogen and lysed at 48°C for 30 min in lysis buffer (50 mM Tris, pH 7.4, 100 mM NaCl₂, 1 mM MgCl₂, 2.5 mM Na₃VO₄, 1 mM phenylmethylsulfonyl fluoride, 2.5 mM EDTA, 0.5 % Triton X-100, 0.5% NP-40, and 5 mg/ml each of aprotinin, pepstatin A and leupeptin), and centrifuged at 10,000 × g for 30 min to collect the supernatant. Protein concentration was determined using the Bradford method. An equivalent amount of protein from each sample was separated by 10% SDS-PAGE and then transferred to a polyvinylidene fluoride membrane (EMD Millipore, Billerica, MA, USA), followed by incubation in blocking buffer (PBS containing 5% nonfat milk) for 2 h at room temperature. Subsequently, the membrane was incubated with a goat polyclonal antibody against LKB1 (dilution, 1:200; catalogue no., sc-5638; Santa Cruz Biotechnology, Inc., Dallas, TX, USA) overnight at 4°C. Next, the membrane was washed twice with PBS for 5 min and incubated with secondary horseradish peroxidase-conjugated rabbit anti-goat immunoglobulin G (IgG) antibody (dilution, 1:1,000; catalogue no., sc-2949; Santa Cruz Biotechnology, Inc.) for 2 h at room temperature. GAPDH was detected using a specific goat polyclonal antibody (dilution, 1:500; catalogue no., sc-20358; Santa Cruz Biotechnology, Inc.) as a loading control. Proteins were visualized using an enhanced chemiluminescence reagent (Santa Cruz Biotechnology, Inc.). Relative expression of LKB1 protein was normalized to GAPDH. The experiments were repeated three times.

Immunohistochemistry was performed to detect the expression pattern and subcellular localization of LKB1 protein in 180 pairs of paraffin-embedded glioma and non-neoplastic brain tissues. Tissue sections were dewaxed in xylene, rehydrated using graded alcohol and soaked in 0.3% hydrogen peroxide to block endogenous peroxidase activity. The sections were then rinsed in PBS, pH 7.2, and 10% goat serum (Gibco; Thermo Fisher Scientific, Inc.) was applied for 1.5 h at room temperature to block nonspecific reactions. The aforementioned goat polyclonal antibody against LKB1 was incubated with the tissue sections overnight at 4°C. Negative control slides were processed in parallel using a nonspecific IgG (dilution, 1:1,000; catalogue no., sc-2949; Santa Cruz Biotechnology, Inc.) at the same concentration as the primary antibody. Subsequent to being rinsed in PBS, the peroxidase reaction was visualized by incubating the sections with a solution containing 0.1% phosphate buffer, 0.02% 3,3′-diaminobenzidine tetrahydrochloride and 3% H₂O₂. Finally, the sections were counterstained with hematoxylin, dehydrated and mounted in mounting resin.

The immunohistochemical results were evaluated by two independent pathologists blinded to the patients' clinicopathological characteristics. The immunoreactive scoring (IRS) of LKB1 expression was calculated by combining the intensity and the percentage of positive cells. The percentage of positive cells was as follows: 0–5% scored 0; 6–35% scored 1; 36–70% scored 2; and >70% scored 3. The staining intensity was as follows: No staining scored 0; weakly staining scored 1; moderately staining scored 2; and strongly staining scored 3. The final IRS was designated as low or high expression using the percentage of positive cell score × staining intensity score.

SPSS software version 11.0 for Windows (SPSS, Inc., Chicago, IL, USA) was used for the statistical analyses. All data obtained from experiments were expressed as the mean ± standard deviation. The differences of LKB1 mRNA and protein expression between glioma and non-neoplastic brain tissues were evaluated by Student's paired t-test. Spearman's rank correlation analysis was used to analyze the correlation between the level of LKB1 mRNA and protein expression. The associations between LKB1 expression and various clinicopathological features were analyzed by the χ² test. Kaplan-Meier survival plots were constructed to analyze survival data. Univariate and multivariate analyses were performed using Cox's proportional hazards model. P<0.05 was considered to indicate a statistically significant difference.

LKB1 expression was markedly decreased at the mRNA (tumor vs. normal, 1.96±0.59 vs. 3.66±0.92; P<0.001; Fig. 1A) and protein (tumor vs. normal, 1.89±0.54 vs. 3.71±0.87; P<0.001; Fig. 1B) levels in 30 freshly prepared glioma tissues, compared with non-neoplastic brain tissues. Spearman's rank correlation analysis revealed that the expression levels of LKB1 mRNA in glioma tissues were closely correlated with those of LKB1 protein in gliomas (r=0.32; P=0.02; Fig. 1C). Notably, the present results indicated a relatively lower expression level of LKB1 mRNA and protein in freshly prepared high-grade glioma tissue samples (WHO grades III–IV) compared with those in the low-grade glioma samples (WHO grades I–II; Fig. 1D).

The positive expression of LKB1 protein was examined in 36/180 (20.00%) gliomas in the present cohort (Fig. 2A), and markedly strong LKB1 protein expression was observed in 132/180 (73.33%) non-neoplastic brain tissue samples (Fig. 2B). Statistically, the IRS of LKB1 protein expression in gliomas was significantly decreased compared with that in the corresponding non-neoplastic brain tissues (tumor vs. normal, 2.38±1.00 vs. 4.38±1.19; P<0.001; Fig. 2C). Notably, the expression of LKB1 in high-grade glioma was decreased compared with that in low-grade glioma (P<0.01; Fig. 2D).

To explore the clinical significance of LKB1 downregulation in human gliomas, the associations between LKB1 and various clinicopathological parameters of glioma patients was statistically analyzed. The median value (2.31) of the IRS of LKB1 protein expression in glioma tissues was used as a cut-off point to divide all 180 patients with glioma into low LKB1 expression group (LKB1-low; n=98) and high LKB1 expression group (LKB1-high; n=82). Table I summarizes the associations between LKB1 expression and various clinicopathological parameters of glioma patients. As a result, low LKB1 expression was significantly associated with large tumor size (P=0.02; Table I), advanced WHO grade (P=0.006; Table I) and low Karnofsky performance scale (P=0.01; Table I). No significant associations of LKB1 with patients' age, gender, tumor location or recurrence were observed (Table I).

To additionally evaluate the prognostic implication of LKB1 downregulation in human gliomas, univariate and multivariate analyses using the Cox's proportional hazards model were constructed, including patients' age, gender, tumor size, tumor location, WHO grade, Karnofsky performance scale, tumor recurrence and LKB1 expression. Kaplan-Meier survival plots in Fig. 3A revealed that the overall survival of glioma patients with low LKB1 expression was clearly shorter compared with that of patients with high LKB1 expression (P<0.001). Notably, the subgroup analysis based on WHO classification demonstrated that low LKB1 expression indicated a poorer survival in high-grade glioma patients (P<0.001; Fig. 3B), but not in low-grade glioma patients (Fig. 3C). Additionally, the results of univariate and multivariate analyses revealed that LKB1 expression (P<0.001), WHO grade (P<0.001) and Karnofsky performance scale (P=0.01) were independent prognostic factors for overall survival of glioma patients (Table II).