GCs, as well as their corresponding adjacent non-neoplastic tissues, intestinal metaplasia and dysplasia tissues were obtained from the tumor bank of the Beijing Cancer Hospital and Institute. Each gastric tissue was collected following patient consent and the approval of the ethics committee of the Beijing Cancer Hospital and Institute. The cancer tissues were excised from the central section of the GC. As a control for each GC patient, normal tissue was excised at least 5 cm from the border of the GC. The specimens were routinely diagnosed by senior pathologists according to pathological biopsy and Lauren's classification. The stage of GC was determined according to the tumor-node-metastasis (TNM) classification of the American Joint Committee on Cancer. Follow-up interviews were conducted with patients for at least 5 years or until the patient succumbed to the disease. For tissue microarray analysis, 40 pairs of GC samples, 40 metaplasia tissues and 30 dysplasia tissues were fixed in formalin and embedded in paraffin. For RT-PCR experiments, 12 pairs of samples were frozen in liquid nitrogen within 30 min after surgery and then stored at −70˚C until required.

Total RNA from gastric samples was extracted using Trizol reagent (Invitrogen, USA) following procedures described by the manufacturer. To remove any DNA contamination, total RNA was treated with RNase-free DNase I (Takara Bio Inc., Japan). The reverse transcription (RT) reaction was carried out with 2 μg total RNA in a 25-μl reaction containing M-MLV reverse transcriptase (Promega, USA) using oligo dT primers. An RT reaction was carried out without reverse transcriptase as a negative control. The PCR for NANOG/NANOGP8 was carried out using LATaq^® DNA polymerase with GC buffer (Takara Bio Inc.) in a 50-μl volume. The NANOG/NANOGP8 primers were: forward, 5′-CCTACCCCAGCCTCTACTCT-3′ and reverse, 5′-CGTCTTCAGGTTGCATGTTC-3′. This pair of primers were able to identify not only NANOG/NANOGP8, but also other pseudogenes of NANOG. Amplification of β-actin was used as a normalizing control. The β-actin primers were: forward, 5′-CGGGACCTGACTGACTACCTC-3′ and reverse, 5′-TCGTCATACTCCTGCTTGCTG-3′. PCR was performed with the following cycling profile: 5 min denaturation at 94˚C followed by 38 (NANOG/NANOGP8) or 30 cycles (β-actin) of 30 sec at 94˚C, 30 sec at 55˚C and 50 sec at 72˚C, with a final extension step at 72˚C for 10 min. Products were analyzed by electrophoresis on a 1.2% (w/v) agarose gel and the fragments extracted using a gel extraction kit (Omega, USA). Six clones of each specific PCR product were verified by sequencing analysis.

Tissue microarray blocks of gastric tissues were constructed for immunohistochemistry staining. For each case, five tissue cylinders (0.6 mm diameter, 1 mm high) were removed from individual GC or adjacent normal tissue. The tissue array blocks were arranged using a puncher (Beecher Instruments, Micro-Array Technologies, USA). Sections (4 μm) were obtained from the block and transferred to glass slides. The block contained a representative control of well-matched cancer and adjacent non-neoplastic tissue.

For immunohistochemical staining, the slides were baked overnight at 60˚C, deparaffinized with xylene and rehydrated in a graded ethanol series. Antigen retrieval was carried out by microwaving tissue in 1 mM EDTA solution (pH 8.0). The endogenous peroxidase activity was blocked by incubation in 3% H₂O₂ at room temperature for 10 min. After incubation in 3% milk to prevent non-specific binding, the slides were immunostained with mouse anti-human NANOG monoclonal antibody (1:100; Abcam, USA) at 4˚C overnight in a humidified chamber. Slides were washed in phosphate-buffered saline and incubated with a secondary antibody conjugated to peroxidase (Zhongshan Golden Bridge Biotechnology, China). Immunohistochemical staining was performed using a commercial DAB kit (Zhongshan Golden Bridge Biotechnology) which yielded a positive brown signal. Slides were then counterstained with hematoxylin, dehydrated in a graded ethanol series and mounted.

The data were analyzed using the Chi-square and Fisher's exact test. The association of NANOG/NANOGP8 expression with clinicopathological features was analyzed using SPSS software (version 14). The cumulative survival curve was compared by the log-rank test. For the analyses, P<0.05 was considered to be statistically significant.

To compare the expression of NANOG/NANOGP8 between tumor and normal tissues of GC patients, we initally detected mRNA expression using RT-PCR in 12 pairs of specimens. Due to the existence of many NANOG pseudogenes, the primers used in this study were likely able to identify these pseudogenes. Therefore, we examined the PCR products by sequence analysis. As shown in Fig. 1, the presence of NANOG and/or its pseudogene RNA was noted in 10 of the 12 GC cases as compared to normal tissues (6/12). The sequencing results obtained from the PCR products indicated that 4 of the 12 GC cases were specific for NANOGP8 and 2 cases contained NANOG, while only 1 of the 12 cases in normal tissues contained the NANOGP8 sequence in the PCR product (Table I). The NANOG pseudogenes, NANOGP2, NANOGP5, NANOGP7 and NANOGP4, were also detected in both GC (8/12) and normal tissues (6/12) (Table I).

Of note is that pseudogenes, including NANOG, cannot be translated into functional protein products (27). We attempted to detect NANOG/NANOGP8 protein in GC and normal tissues using immunohistochemical staining. Results indicated that there was an increased NANOG/NANOGP8 protein expression in GC tissues compared to normal matched tissues (Fig. 2, Table II). Positive NANOG/NANOGP8 staining was detected in 75% (30/40) of the GC tissues and only in 12.5% (5/40) of the corresponding adjacent normal tissues (P<0.001). The majority of NANOG/NANOGP8 cells were diffusely localized in both the nucleus and cytoplasm (Fig. 2A and B), whereas in other GC tissues, NANOG/NANOGP8 cells were clearly localized in the nucleus (Fig. 2C and D).

A correlation analysis of NANOG/NANOGP8 expression and the clinicopathological features of GC was conducted. NANOG/NANOGP8 protein expression did not correlate to the following clinicopathological features: GC histological grade (well vs. poor differentiation, P=0.589), lymph node metastasis (N₀ vs. N₁, P=0.473), distant metastasis (M₀ vs. M₁, P=1) or stage (TNM classification, stages I and II vs. stages III and IV, P=0.915) (Table III). Additionally, the cumulative survival curve indicated no significant difference in the survival time of patients with higher levels vs. those with low levels of NANOG/NANOGP8 expression (P=0.998; data not shown).

As NANOG/NANOGP8 protein expression in GC did not correlate to the afore-mentioned clinicopathological features, we hypothesized that NANOG/NANOGP8 plays a role in the early developmental stages of gastric carcinogenesis. We were able to detect the NANOG/NANOGP8 protein in the intestinal metaplasia and dysplasia tissues, representative of the early developmental stages of GC (28,29). As shown in Fig. 3, NANOG/NANOGP8 was highly expressed in the intestinal mucosa of the metaplasia tissues (60%, 24/40; Fig. 3A and B). The dysplasia tissues also demonstrated high levels of NANOG/NANOGP8 expression (66.7%, 20/30), whereas adjacent normal gastric tissues exhibited no expression (Fig. 3C and D).