All mice were housed in an air-conditioned room under adequate temperature (20.7°C-22.8°C) and humidity (41.6%-59.9%) control with a 12-hour dark/light cycle and free access to water and food (Lab diet 5002, LabDiet, Richmond, IN). All animal experiments were performed after receiving approval from the Institutional Animal Care and Use Committee (IACUC) of Clinical Research Institute at Seoul National University Hospital (SNU-070509-4), and the National Research Council (NRC) guidelines for the care and use of laboratory animals were followed (revised in 1996).

To produce miR-222/221 transgenic mice, miR-222/221 was first amplified using C57/BL6J mice. It was cloned into pGEM-T easy vector and subsequently transferred into the pcDNA3 vector using the BamHI and XhoI sites. The nucleotide sequence of the plasmid was confirmed by sequencing. The plasmid was then injected into fertilized eggs. Quantitative polymerase chain reaction screening was performed to verify that the transgene was transferred to the mice. Genomic DNA was extracted from the tail of each mouse. The presence of the miRNA-222/221 transgene was assessed by using LaboPass Tissue Mini DNA Purification Kit (Cosmo Genetech, Seoul, Korea) according to the manufacturer's protocol. The primer sequences are listed in Supplementary Table 1.

MNU (Sigma Chemical Co., St. Louis, MO) was dissolved in distilled water at a concentration of 240 ppm and added to drinking water in light-shielded bottles. It was freshly prepared thrice per week. At 6 weeks of age, mice from each group were given drinking water containing MNU on alternating weeks for a total of 10 weeks of exposure as previously described [16]. At the end of the treatment, the drinking water was replaced by autoclaved distilled water.

A total of 53 wild-type mice and 65 transgenic mice were used. At 6 weeks of age, mice from each group were treated with MNU five times during alternate weeks. The animals were carefully autopsied at 36 weeks of age, under anesthesia. At the end of the experiment, all mice were fasted for 24 hours, but had free access to water. Prior to euthanasia, blood samples were collected from the vena cava for blood chemistry tests. The stomachs and other organs were carefully examined macroscopically. After pathologic observation of the stomach and intestine, immunohistochemistry and validation studies were performed (Fig. 1).

All mice underwent a thorough postmortem examination at the time of euthanasia. The stomach was tied with silk thread at the esophagus and duodenum and then removed. Thereafter, phosphate buffered saline was injected into the stomach and 10 minutes later, the stomach was excised along the greater curvature with scissors and fixed with a pin. Half of the excised stomach was frozen in liquid nitrogen for molecular biology experiments and the remaining stomach was fixed for 24 hours, processed by standard methods, embedded in paraffin, sectioned at 5 μm, and stained with hematoxylin and eosin (H&E). The intestine was removed from the abdomen and tied with silk thread at the duodenum and the edge of the rectum. It was then injected with 10% neutral buffered formaldehyde solution. Twenty-four hours later, the knots were untied, the intestine was excised longitudinally with scissors and the contents were carefully removed. Next, the segment was rolled up longitudinally, with the mucosa outwards, using forceps [13]. Finally, the intestine was placed in a fixative for paraffin-wax embedding and sectioned. The 5-μm-thick sections were stained with H&E. Formalin-fixed, paraffin-embedded, 4-μm sections were dewaxed in xylene, rehydrated through graded alcohol, and placed in an endogenous peroxide block for 15 minutes. Sections were washed in water, antigen retrieved, and then placed in citrate buffer. Nonreactive staining was blocked by treating sections with 1% horse serum in Tris-buffered saline (pH 6.0) for 3 minutes.

Anti-CD31 (1:80, goat polyclonal, Santa Cruz Biotechnology, Santa Cruz, CA), anti–Ki-67 (1:200, rat monoclonal, DAKO, Glostrup, Denmark), anti-cleaved caspase-3 (1:100, rabbit monoclonal, Cell Signaling, Danvers, MA), anti-p27 (1:500, rabbit polyclonal, Santa Cruz Biotechnology), and anti-p57 (1:300, rabbit monoclonal, Epitomics, Burlingame, CA) were then applied. Antibody binding was detected using an avidin-biotin-peroxidase complex (Universal Elite ABC Kit, Vectastain, Burlingame, CA) for 10 minutes. Diaminobenzidine tetrahydrochloride solution (Kit HK153-5K, Biogenex, San Ramon, CA) was then used as the chromogen.

Total RNA was extracted from mouse tissues by using TRI-zol reagent (Invitrogen, Carlsbad, CA) and 30 μg was loaded on a 12.50% urea-polyacrylamide gel. The resolved RNA was transferred to a Zeta-Probe GT blotting membrane (Bio-Rad Laboratories, Hercules, CA) by capillary method and fixed by ultraviolet crosslinking. Oligonucleotides complementary to each miRNA were end-labeled at the 5′-end with 32P-γ-ATP using T4 polynucleotide kinase (Takara, Tokyo, Japan) and used as probes. Prehybridization and hybridization were carried out using ExpressHyb Hybridization Solution (Takara) according to the manufacturer's instructions. The sequences of the antisense oligos are complimentary to mature miR-222 and miR-221. The sequences of the probes are shown in Supplementary Table 1. Membranes were prehybridized for 30 minutes at 37°C using hybridization solution in hybridization tubes. Subsequently, the probes were added into the tubes and incubated for 1 hour at 37°C. After hybridization, the membranes were washed twice for 30 minutes using buffer 1 (0.05% sodium dodecyl sulphate [SDS], 2× sodium saline citrate [SSC]) at room temperature and then washed twice for 15 minutes using buffer 2 (10% SDS, 20× SSC) at room temperature. BAS 2500 Image Analysis System and Multi Gauge V3.0 software (Fujifilm, Tokyo, Japan) were used for quantitation. Ethidium bromide staining of rRNA was used for RNA loading control.

Ten human GC cell lines, SNU-638, and AGS were obtained from the Korean cell line bank. RPMI 1640 (Welgene, Daegu, Korea) supplemented with 10% fetal bovine serum (Thermo Scientific, Waltham, MA) and 1% penicillin and streptomycin (Life Technologies, Carlsbad, CA) was used for cell culture. The cells were maintained at 37°C under an atmosphere of 5% CO₂-95% air.

Transfection to AGS cells was performed one day after seeding using Lipofectamine 2000 (Invitrogen).

One microgram of total RNA was used for single-stranded cDNA using the SuperScript First Strand cDNA synthesis (Life Technologies). RNA quantification was conducted by real-time polymerase chain reaction (PCR) using the comparative Ct method with the Bio-Rad iCycler Real-Time PCR system. Real-time PCR was performed in a 20 μL final volume containing 10 μL of SYBR 2× master mix (Thermo Fisher Scientific), 10 pmol of each primer in a volume of 1 μL, and 1 μL of cDNA. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) primer sets were used for normalization. Each sample was run in triplicate (Bio-Rad Laboratories). Primers for the amplification of mature miR-221 and miR-222 are presented in Supplementary Table 1.

Statistical analyses were performed using SPSS ver. 13.0 (SPSS Inc., Chicago, IL) and graphs were generated using GraphPad Prism 5.0 (GraphPad Software Inc., San Diego, CA). Chi-square was used to measure the carcinogenesis association between the miR-222/221 transgenic and wild-type mice. p-values less than 0.05 are considered statistically significant.

The expression level of pre-miRNA-222/221 and mature miR-222/221 was increased in the stable cell line transfected with miR-222/221 as compared to the cell line transfected with pcDNA3 null vector (Fig. 2A). To examine whether off-springs contained the injected foreign DNA, germ-line transmission was tested by PCR using the CMV promoter primer and miR-222 primer at 3 weeks of age, all mice were screened by PCR (Fig. 2B). At 10 weeks of age, the expression levels of pre-miRNA-222/221 and mature miR-222/221 were compared between wild-type and transgenic mice in nine organs (brain, thymus, heart, lung, stomach, liver, intestine, spleen, and kidney) by northern blotting (Fig. 2C). miR-222/221 was expressed in transgenic mice, but not in wild-type mice. Compared to wild-type mice, transgenic mice strongly overexpressed both pre-miRNA-222/221 and mature miR-222/221 in the stomach, brain, lung, and kidney. The organ weight and blood chemistry were investigated in 56-week-old mice to compare the phenotype of transgenic and wild-type mice. There was no difference in organ weight between the two groups. When we compared body weights between male mice, transgenic mice were about 20% heavier than wild-type mice. The white blood cell (WBC) count was 1.8-fold lowerin transgenic mice in blood chemistry tests (Table 1).

MNU treated mice were euthanized at 36 weeks of age. Transgenic and wild-type mice were randomly selected and the expression levels of miR-222 and miR-221 were confirmed by northern blot and real-time PCR (Supplementary Fig. S1). Approximately 85% of transgenic mice survived until 36 weeks of age (Table 2). At the end of the experiment, all surviving mice were euthanized, and their stomachs and other organs were carefully examined. From the pathological findings, after treatment with the carcinogen MNU, carcinoma and other lesions, including hyperplasia, microcarcinoma in adenoma, angiosarcoma, microscopic dysplasia, adenoma in pylorus region, and squamous papilloma in the fundus region, were detected in the mice stomachs. However, only flat serrated adenoma was observed from the intestines of three mice. From these results, adenoma and hyperplasia, the two most frequently observed pre-cancerous lesions, were selected and included for MNU-induced multigastric carcinogenesis steps. Representative histological images of normal, pre-cancerous lesions and carcinoma are shown in Fig. 3. Hyperplasia was observed in 3.77% of the wild-type mice and in 18.46% of the transgenic mice (p=0.020). Adenoma was observed in 20.75% of the wild-type mice and 26.15% of the transgenic mice (p=0.522). Carcinoma was observed in 32.08% of the wild-type mice and 41.54% of the transgenic mice (p=0.341). The frequency of hyperplasia, adenoma, and carcinoma was higher in transgenic mice, but the difference was statistically significant only in hyperplasia (Table 3). When hyperplasia and adenoma were combined as a pre-cancerous lesion, it was observed in 44.62% of transgenic mice and 24.53% of wild-type mice; this difference was statistically significant (p=0.033).

To test the correlation between cancer and miR-222/221 in our in vivo model, cancer related markers were selected for immunohistochemistry. CD31 was used as a marker of angiogenesis, Ki-67 was used to assess cell proliferation, and cleaved caspase 3 was used as an apoptosis marker. Moreover, two antibodies against proteins encoded by two miR-222/221 target genes, p27 and p57, were used to compare p27 and p57 protein expression in the stomachs of transgenic and wild-type mice by immunohistochemistry. Based on our hypothesis, in transgenic mice overexpressing miR-222/221, the expression of p27 and p57 should be repressed as compared to that in wild-type mice. However, there was no significant difference in term of p27 and p57 expression between the transgenic and wild-type mice. Compared to wild-type mice, in transgenic mice, CD31 (angiogenesis marker) and Ki-67 (proliferation marker) were highly expressed, suggesting that cancer-related biological processes, angiogenesis and proliferation, were activated. However, the expression of cleaved caspase 3 (apoptosis marker), p27, and p57 was not significantly different between the two groups (Fig. 4A). Additionally, p27 expression was similar in the stomach of untreated wild-type and transgenic mice (Fig. 4B).