The Oncomine database (http://www.oncomine.com) was queried to assess the expression of NAMPT. Set filter conditions included: i. Cancer Type: Glioblastoma; ii. Gene: NAMPT; iii. Data Type: mRNA; iv. Sample Type: Clinical Specimen; v. Analysis Type: Cancer vs. Normal Analysis; vi. Critical setting condition: P-value <1×10⁻⁴, fold change >2, gene rank=top 10%.

The Kaplan-Meier survival analysis was performed using the ‘Tumor glioblatoma-TCGA-540 MAS5.0-u133a’ dataset in the Genomics Analysis and Visualization Platform (R2 platform) (https://hgserver1.amc.nl/cgi-bin/r2/main.cgi). The median was used as the threshold to distinguish the high and low expression of NAMPT.

A glioblastoma tissue microarray (US Biomax GL805b) was obtained from GeneChem (Shanghai, China) containing two spots each of tissues from 35 cases of glioblastoma, paracarcinoma tissues from 2 cases (1.5-cm distance from the paracarcinoma tissue to the cancer tissue) and normal neural tissues from 3 cases. A two-step immunohistochemical protocol was used. The primary antibody was an anti-NAMPT rabbit monoclonal antibody (dilution 1:500, cat. no. ab58640; Abcam), and the secondary antibody was from an UltraSensitive SP KIT-9720 (Fujian MXB). Three random high-magnification fields were observed under an optical microscope (Caikon XDS-100, original magnification, ×200).

The immunohistochemical microarray staining was scored on a semiquantitative scale by assessing the intensity and proportion of staining. The staining intensity was scored using the following scale: 0 (negative), 1 (weak), 2 (moderate), and 3 (strong). According to the positive rate of staining, the proportion of staining was scored as 0 (0–4%), 1 (5–25%), 2 (26–50%), 3 (51–75%), or 4 (76–100%). The total score was calculated by multiplying the proportion score and the intensity score. Specimens with a total score of ≤6 were assigned to the low-expression group. Specimens with a total score of >6 were assigned to the high-expression group.

GBM gene expression profile was downloaded from The Cancer Genome Atlas (TCGA) database (https://www.cancer.gov/tcga). The queue were divided into NAMPT high and low expression groups according to the median expression of NAMPT. Gene set enrichment analysis was performed using the GSEA software, version 2.0.1, obtained from the Broad Institute (http://www.broad.mit.edu/gsea). Gene set permutations were performed 1,000 times for each analysis. Family error rate (FWER) and false discovery rate (FDR) were used to screen the enriched pathways in each group.

The human glioblastoma cell lines, U87 and U251, were obtained from the American Type Culture Collection (ATCC). STR profiling indicated that the U87 was a probable glioblastoma of unknown origin. These cell lines were cultured with Corning^® DMEM medium (Corning Inc.) containing 10% FBS. Cell lines were cultured at 37°C in a humidified atmosphere of 5% CO₂.

A small hairpin RNA (shRNA) targeting NAMPT with high specificity was designed and synthesized by GeneChem (Shanghai, China) and cloned into a lentiviral vector (GeneChem), which also encoded green fluorescent protein (GFP). Then, the lentivirus expressing NAMPT shRNA (LV-NAMPT-RNAi) was prepared and collected. The NAMPT shRNA sequence was as follows: 5′-AACTTAGATGGTCTGGAAT-3′. A scrambled shRNA sequence was used as the negative control: 5′-TTCTCCGAACGTGTCACGT-3′.

The human glioblastoma cell lines U87 and U251 were used for lentivirus infection. Briefly, cells in the logarithmic phase were digested by trypsinase. A 4×10⁴/ml cell suspension was prepared, and 1 ml of the cell suspension was inoculated into a 12-well plate. When the cell confluence was 20%, the appropriate lentiviral particle solution containing the lentivirus expressing either NAMPT or the scrambled shRNA was added to the target plate (MOI=5). The infection efficiency was determined by the GFP expression status as observed with a fluorescence microscope (Olympus Corp., Tokyo, Japan) 72 h after infection.

Total RNA was extracted and reverse transcribed. The resulting cDNA was used as the template for RT-PCR using a standard SYBR Green PCR kit (Takara) in a LightCycler 480 Real-Time PCR machine (Roche). Independent experiments were conducted in triplicate. GAPDH served as the internal control. Gene expression was calculated using the ΔΔCq method (20).

Cell lysates were prepared in ice-cold RIPA lysis buffer. Total protein concentration was determined using an enhanced BCA protein assay kit. An amount of 30 µg of total protein per lane was separated by 10% SDS-PAGE, transblotted onto PVDF membranes, which were then blocked with 5% milk dissolved in TBST for 1 h at room temperature and incubated overnight with primary antibodies at 4°C. The primary antibodies used were as follows: Rabbit anti-NAMPT (cat. no. ab58640; dilution 1:500; Abcam), mouse anti-GAPDH (cat. no. ab8245; dilution 1:3,000; Abcam). Then the blots were incubated with anti-rabbit (cat. no. ab205718; dilution 1:5,000; Abcam) or anti-mouse (cat. no. ab205719; dilution 1:5,000; Abcam) HRP-labeled secondary antibody at room temperature for 1.5 h. The immunoactivity was detected using an ECL-Plus kit (Thermo Fisher Scientific. Inc.). The immunoreactive bands were quantified by densitometry with ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Cells at a density of 2,000 cells/well in logarithmic phase were cultured in 96-well culture plates. A 10-µl volume of CCK-8 reagent (Sigma-Aldrich; Merck KGaA) was added at various time points (24, 48, 72, 96 and 120 h). After cells were cultured for 4 h, the absorbance of the plates was read at a wavelength of 450 nm using a microplate reader (M2009PR; Tecan Infinite). The experiment was repeated five times.

Three days after infection with the lentivirus, cells were seeded at a density of 800 cells per well in 6-well plates with triplicate wells for each group. Cell colonies were grown in a humidified 5% CO₂ incubator at 37°C for 16 days. Images were acquired with an inverted microscope at ×100 magnification. Then, cells were fixed, stained with Giemsa solution and imaged. Colonies measuring 0.3 mm or more were counted.

Cells at a density of 50,000 cells/well were seeded in 96-well wounding replicator and grown to over 90% confluence. Then the culture medium was replaced with 0.5% FBS. Wounds were generated using the tip with 200 µm width of a wounding replicator, VP scientific, VP408FH. Cells were incubated for up to 24 h. Wound closure was observed at 0, 8 and 24 h post scratching. These assays were performed with three independent replicates. Cell images were acquired with an inverted microscope (Caikon XDS-100, original magnification, ×10). The migration rate was calculated as follows: Migration rate=[(width of wound at 0 h-width of wound at the time point tested)/width of wound at 0 h] ×100%.

A Corning invasion kit was used to perform this assay according to the manufacturer's instructions. Cell suspensions (10⁵ cells per well in a 24-well culture plate) in FBS-free DMEM were added to the top chamber, and DMEM containing 30% FBS was added to the bottom chamber. Cells that migrated through the filter were stained with Giemsa. Images of the migrated cells were captured with a digital camera connected to an inverted microscope (Caikon XDS-100). Five random fields of vision at ×100 magnification and 9 random fields of vision at ×200 magnification were observed per well. The migrated cells were counted in the ×200 images. The invasion assay was carried out with three independent replicates.

The first step was to prepare tumor-forming cells according to the method described above. U87 cells infected with a lentiviral vector expressing NAMPT-shRNA and GFP sequences were generated as the knockdown (KD) group; cells infected with lentiviral vector expressing scrambled shRNA and GFP sequences were generated as the negative control (NC) group. Mice were randomized into the two groups. BALB/c nude mice (male, 4 weeks old, ~20 g) were purchased from the HUST Animal Center, Wuhan, China. They were maintained in a constant temperature, humidity, sterile environment and routinely fed with food and water according to the national regulations. A total of 1×10⁷ logarithmically growing U87 cells in 0.1 ml of PBS were subcutaneously injected into the right flank of the mice, 10 mice in each group. According to the Institutional Animal Care and Use Committees (IUCUC) protocol (AN-5803), the maximal allowable size of the tumours could not exceed 10% of mouse body weight. Tumor volumes were measured every four days after injection with a caliper and calculated using the formula V=π/6 (lxwxw), where l is the length and w is the width. Forty days after injection, animals were anesthetized using an intraperitoneal injection of 150 mg/kg pentobarbital sodium. Tumor tissues were excised and weighed. Immunohistochemical staining of Ki-67 and CD31 was performed on the transplanted tumors, and the Ki-67 index and microvascular density (MVD) were calculated. All animal experiments were approved by the Institutional Animal Care and Use Committee of Huazhong University of Science and Technology.

Data are expressed as the means ± standard errors of the mean. Student's t-test was used to compare the data between two groups. The data between three groups were compared by ANOVA, and then Tukey test was used to compare the data between the two groups if P<0.05. P<0.05 was considered to indicate a statistically significant difference.

Six studies in the Oncomine database, comprising a total of 942 samples, met the criteria (21–25). Meta-analysis of the 6 studies showed that NAMPT was statistically overexpressed in glioblastoma relative to its expression in normal neural tissue (P=1.68E-7) (Fig. 1A). Immunohistochemical microarray analysis showed that NAMPT was mainly localized in the cytoplasm (Fig. 1C-E). The average total staining scores of the glioblastoma and nontumor tissues (paracarcinoma and normal neural tissues) were 7.54 and 4.600, respectively (P=0.000637). These results suggest that the expression of NAMPT in human glioblastoma is higher than that in normal tissues.

Data from the R2 platform (540 glioblastoma samples) showed that patients with higher NAMPT expression had a lower overall survival rate than patients with lower NAMPT expression. The 2-year survival rates were 4 and 1% for the high- and low-expression groups, respectively (P=0.0063) (Fig. 1B).

Gene set enrichment analysis (GSEA) was performed using data from the TCGA GBM cohort comprising 169 samples. The expression level of NAMPT was used as the phenotype label. A pathway was classified as enriched by setting the FWER P-value to <0.01 and the FDR to <0.05. The ‘GO acute inflammatory response’, ‘positive regulation of I-κB kinase NF-κB signaling’, ‘I-κB kinase NF-κB signaling’, and ‘acute phase response’ pathways were enriched in the NAMPT high-expression phenotype (Fig. 1F). Therefore, NAMPT may play a role via these pathways.

A lentiviral vector expressing both shRNA targeting human NAMPT and the GFP sequence was used to establish knockdown (KD) cells, while a lentiviral vector expressing both scrambled shRNA and the GFP sequence was used to establish negative control (NC) cells in the U87 and U251 cell lines. Cells without lentivirus infection were considered blank control (CON) cells. GFP expression was observed in more than 90% of cells infected with the lentivirus 72 h after infection (Fig. 2A and B).

RT-PCR and western blot analysis were performed to assess the knockdown efficiency. RT-PCR results showed that NAMPT-shRNA was effective in both U251 and U87 GBM cells, with a knockdown efficiency of approximately 74.9% in U87 GBM cells (Fig. 2C) and 87.7% in U251 GBM cells (Fig. 2D). The location (Fig. 2E) and quantification (Fig. 2F) of the western blot band further confirmed the above results.

The CCK-8 assay showed that the proliferation of cells in the KD group was significantly inhibited at the 4 and 5 days in U87 and U251 GBM cells (P<0.001 KD vs. NC for both cell lines) (Fig. 3A and B).

The soft agar colony formation assay showed that in the KD, NC and CON groups of U87 and U251 GBM cells the average colony numbers were 19, 27 and 40 and 6, 51 and 88, respectively (P<0.05 KD vs. NC for both cell lines) (Fig. 3C and D).

As shown by the wound healing assay in U87 cells, the migration rates of KD cells at 8 and 24 h were 0.14±0.08 and 0.4±0.06, respectively, lower than those of the CON group (0.21±0.08 and 0.64±0.05) and NC (0.19±0.07 and 0.54±0.07) cells (Fig. 4A and B).

The Transwell assay showed that the average numbers of migrated cells in the KD, NC and CON groups were 56±5.55, 101±3.88 and 106±3.74 in the U87 cells and 16±0.44, 107±5.00 and 116±2.92 in the U251 cells, respectively (P<0.001, KD vs. NC and CON) (Fig. 4C and D).

The apoptosis assay showed that the percentage of apoptotic U87 GBM cells in the KD group was 7.68±0.3386%, which was higher than that in the NC group (3.74±0.3103%) (P<0.001). Similarly, the percentage of apoptotic U251 GBM cells in the KD group was 7.43±0.4921%, which was also higher than that in the NC group (3.88±0.3553%, P<0.001) (Fig. 4E-G).

The growth curve of subcutaneously transplanted tumors in nude mice showed that tumors derived from the KD cells grew much more slowly than those derived from the NC tumors. Forty days after inoculation, the maximum diameter of the KD and NC transplanted tumors were 11.40 and 18.50 mm, the maximum weights of the KD and NC tumors were 0.643 and 1.049 g, respectively. The average weights of the KD and NC transplanted tumors were 0.269±0.244 and 0.788±0.186 g, respectively (P<0.0001) (Fig. 5A-C).

The Ki-67 indices of the KD and NC tumors was 17.7±3.1 and 55.3±8.2%, and the MVD in the KD and NC tumors was 15.9±2.8 and 28.2±5.4, respectively. These results suggest that NAMPT knockdown inhibits the growth and microvessel formation of U87 glioblastoma xenografts in nude mice (Fig. 5D-F).