This article is mentioned in:

Introduction

Inhibins and activins, members of the transforming growth factor-β (TGF-β) superfamily, are polypeptides that were originally isolated from ovarian fluid, based on their effect on pituitary follicle-stimulating hormone (FSH) production and secretion. Inhibins are heterodimers that are composed of a common α subunit and one of two homologous β subunits (βA and βB). Activins are either heterodimers or homodimers of the inhibin β subunits (βAβA, βBβB, and βAβB) (1–3). Activin βC, βD, βE chains, and partially characterized activin AC(βAβC) and activin BC(βBβC) proteins have also been reported (4–6).

Besides their classical endocrine function in suppressing of FSH production and secretion, inhibins are thought to act in an autocrine or paracrine manner within reproductive tissues. In females, inhibins appear to influence folliculogenesis by regulating granulose cell maturation and proliferation, steroid hormone production and oocyte maturation within the ovary (7–10). Reduced production of inhibins or production of mutant forms of inhibins is linked to several ovarian diseases, including premature ovarian failure (11) and polycystic syndrome (12), whereas increased inhibin is found in pre-eclampsia (13) and in certain ovarian cancers (14,15). Overexpression of inhibin-α in mice results in abnormal reproductive function including decreased litter size (16) and decreased embryo size (17). In males, inhibins are potential autocrine or paracrine regulators of Leydig and Sertoli cell proliferation, differentiation and steroidogenesis (18).

More recently, inhibins and activins have both been implicated in endocrine-related cancers (19). The inhibin-α gene was identified as a tumor suppressor gene in the gonads and adrenals by functional studies using knockout mice (20–22). This has raised the question of whether it plays a broader role as a tumor suppressor outside the reproductive axis. A second interesting model has been prostate carcinoma, where it was observed that hypermethylation of the inhibin-α gene promoter and LOH at 2q32-36, the chromosome region harboring the inhibin-α gene, occurred in 42% of prostate carcinomas (23). Moreover, there was a positive correlation between loss of inhibin expression and malignancy of these human prostate carcinomas cells (24). Recently, seemingly conflicting functions of the inhibin-α gene in the prostate were reported. Inhibin-α expression was reduced in early-stage tumors but increased in late-stage, metastatic prostate cancers, suggesting that inhibins act as tumor suppressors during early tumorigenesis but might act as tumor promoters during late-stage disease (24,25). Moreover, it was reported that inhibin negatively regulated matrix metalloproteinase (MMP) levels and MMPs have an integral role in the increased metastatic and invasive potential of cancer cells (26–28). Thus considerable evidence supports roles of inhibin-α beyond the gonad and adrenal gland and indicates its importance in regulating cell growth. Whether or not inhibin-α has a general function in cancerous cells and tissues other than prostate is not yet clear. Having the long-term goal of addressing this question, we studied inhibin-α gene expression changes in human gastric cancer carcinoma cells.

Materials and methods

Cell cultures

Human gastric cancer cell lines AGS and KATO III were purchased from the American Tissue Culture Collection (Manassas, VA, USA). SNU-1, SNU-5, SNU-16, SNU-484, SNU-601, SNU-638, SNU-668, and SNU-719 cell lines were supplied by the Korean Cell Line Bank (Cancer Research Center, Seoul, Korea). Cells were cultured in RPMI-1640 medium (Gibco-BRL, Carlsbad, CA, USA) containing 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 μg/ml streptomycin at 37°C in a humidified atmosphere of 5% CO2 in 95% air.

Bisulfite modification

The methylation status of the promoter CpG islands of the inhibin-α gene in all sample DNAs was analyzed by PCR on the sodium-bisulfite converted DNA (29). Genomic DNA was extracted by the Wizard Genomic DNA purification kit (Promega Corporation, Madison, WI, USA). DNA (2 μg) in a volume of a 50 μl was denatured with NaOH (final concentration, 0.2 M) and incubated at 37°C for 15 min, then 30 μl of 10 mM hydroquinone and 520 μl of 3 M sodium bisulfite (Sigma-Aldrich, St. Louis, MO, USA) at pH 5.0 were added into the tube. After mixing, samples were incubated under mineral oil at 55°C for 16 h. Then, DNA was desalted with Wizard DNA Clean-Up system (Promega), desulfonated by addition of NaOH (final concentration, 0.3 M), and incubated at 37°C for 15 min. The solution was neutralized by addition of ammonium acetate (final concentration, 3.0 M), and the DNA was ethanol-precipitated, dried, and re-suspended in 20 μl of water and used immediately or stored at −20°C.

Detection of methylation

Methylation was assessed by PCR and sequence analysis of bisulfite-treated DNA. The bisulfite reaction converted unmethylated cytosines to uracil, whereas methylated cytosines were unchanged. The 5′-UTR region of inhibin-α was amplified by nested PCR using primers designed to the bisulfite-converted sequence (23). Primer sequence 1 (5′-GATAAGAGTTTAGATTGGTTTTATTGGTT-3′) and 2 (5′-ACACCATAACTCACCTAACCCTACTAATAA-3′) were used for the first round of PCR and primer sequences 3 (5′-ACCCCTTCTACCAAAATCTACCCAAAA-3′) and 4 (5′-GAAGGTGTTGTATGTTTGTATGTGTGAGTT-3′) were used for the second round of PCR. The first round of PCR was performed in 25 μl reactions with 2 μl of bisulfite-converted DNA, 1X PCR buffer (10 mM Tris, pH 8.3, 50 mM KCl, 1.5 mM MgCl2), 200 μM of each dNTPs, 10 pmol of each primer 1 and 2, and 1 unit of AmpliTaq Gold DNA polymerase (Applied Biosystems, Carlsbad, CA, USA). PCR cycles consisted of 95°C for 15 min followed by 5 cycles of 95°C for 1 min, 50°C for 2 min, and 72°C for 3 min and followed by 30 cycles of 95°C for 1 min, 55°C for 2 min, and 72°C for 2 min with a final incubation step of 72°C for 10 min. A sample of 2 μl from the first PCR was amplified in a 25 μl reaction as above except that primer 3 and 4 were used. PCR cycling conditions were as for the first reaction, with the exception that the annealing temperature was increased to 60°C. PCR products were gel purified, ligated into the PCR 2.1 cloning vector, and cloned using the TOPO® TA Cloning® Kit according to the manufacturer’s instructions (Invitrogen, Carlsbad, CA, USA). For each PCR, 10 clones were sequenced and the methylation at each of the seven CpGs in the inhibin-α proximal promoter was determined.

DNA analysis

DNA was isolated from cultured cells by standard methods. Two regions of inhibin-α gene were amplified from genomic DNA by PCR with specific oligonucleotide primers (30). The first region of 240 bp (fragment A), which includes 140 bp of 5′-UTR and 100 bp of exon 1, was amplified by primers AF (5′-GACTGGGGAAGACTGGATGA-3′) and AR (5′-TCACCTTGGCCAGAACAAGT-3′). The second region of 396 bp (fragment B), which comprises part of exon 2, was amplified by primers BF (5′-AGCAGCCTCCAATAGCTCTG-3′) and BR (5′-AGCTCCTGGAAGGAGATGTTC-3′). Genomic DNA (200 ng) was amplified in a 50 μl volume reaction containing 1X PCR buffer, 2 mM MgCl2, 2.5% DMSO, 0.2 mM of each dNTP, 20 pmol of each specific primer and 1.5 units of AmpliTaq Gold DNA polymerase. The condition for amplification was as follows: first after denaturation at 95°C for 14 min, then denaturation at 95°C for 40 sec, annealing at 57°C for 30 sec, and extension at 72°C for 1 min for 35 cycles and final extension at 72°C for 7 min. Polymorphism −16C>T in the 5′-UTR was screened in the samples by restriction enzyme analysis using SpeI (New England Biolabs, Ipswich, MA, USA). Briefly, fragment A was amplified by PCR and 5 μl of purified PCR product was digested overnight at 37°C with 5 units of SpeI, electrophoresed on 8% polyacrylamide gels, stained with ethidium bromide and visualized by using a Gel Doc 1,000 Gel Documentation System (Bio-Rad, Hercules, CA, USA). Presence of the 240 bp fragment indicated a homozygous variant for wild-type, whereas presence of two fragments of 120 bp corresponded to homozygous variant T. Substitution 769G>A of exon 2 was analyzed by digestion of fragment B with different restriction enzymes. Five microliters of purified PCR product was digested overnight at 37°C with 5 units of BsrFI (New England Biolabs) and analyzed as described above. The restriction site that renders two fragments of 340 and 56 bp is abolished in the variant allele. In addition, 5 μl of purified PCR product were digested overnight at 37°C with 5 units of Fnu4HI (New England Biolabs), electrophoresed on 15% polyacrylamide gels, stained with ethidium bromide and visualized by image analysis. The 396 bp fragment renders four fragments of 153, 107, 51, and 25 bp, among others of lower molecular weight, in the wild-type allele, whereas the allele with substitution 769G>A renders four fragments of 153, 107, 76 and 51 bp, among others of lower molecular weight.

Loss of heterozygosity (LOH) analysis

LOH was determined using microsatellite markers on 2q32-q33 (D2S389) and 2q33-q36 (D2S128) which were previously described (23,31). Oligonucleotide primer sequences were D2S389 (5′-TAAAGCCTAGTGGAAGATCATC-3′, 5′-GCTGAGTTAACAGTTATCAACAATT-3′) and D2S128 (5′-AAACTGAGATTTGTCTAAGGGG-3′, 5′-AGCCAGGAATTTTTGCTATT-3′). PCR was performed in 20 μl reactions consisting of 200 ng of DNA, 1X PCR buffer, 0.2 mM dNTPs, 10 pmol of each primer, and 1 unit of AmpliTaq Gold DNA polymerase. The condition for amplification was as follows: after denaturation at 95°C for 14 min, then denaturation at 95°C for 1 min, annealing at 55°C for 1 min, and extension at 72°C for 1 min for 35 cycles and final extension at 72°C for 10 min. Ten microliters of PCR products was mixed with 10 μl of stop solution containing 95% formamide, 10 mM NaOH, 0.25% bromophenol blue and 0.25% xylene cyanol FF. The mixture was denatured at 95°C for 5 min, put onto ice for 5 min, electrophoresed on 12% polyacrylamide gels containing 10% glycerol with 1X TBE buffer and stained with ethidium bromide. LOH was defined as reduction of the intensity of the signal of a single allele by >50% in the tumor DNA by direct visualization when compared with the corresponding DNA of peripheral blood lymphocytes.

RNA extraction and RT-PCR procedures

Total-RNA was extracted from cultured cells using the RNA-Bee solution kit following the manufacturer’s protocol (Tel-Test, Friendswood, TX, USA). Total-RNA was treated with RQ1 RNase-free DNase (1 μl/μg) (Promega) at 37°C for 30 min and added 1 μl of RQ1 DNase stop solution to terminate the reaction, followed by heat inactivation at 65°C for 10 min. The RNA was purified with a phenol/chloroform extraction and precipitated with ethanol. First-strand cDNA synthesis was performed with 2 μg of DNase-treated mRNA. The cDNA was made with random hexamers using a reverse transcription system (Promega) according to the manufacturer’s protocol. PCR was performed with 2 μl cDNA in a 25 μl reaction mixture of 1X PCR buffer, 0.2 mM each dNTP, 10 pmol of each primer inhibin-α (5′-AGGAAGAGGAGGATGTCTCC-3′, 5′-GAGTAACCTCCATCCGAGGT-3′; 823 bp), betaglycan (5′-ACATGGATAAGAAGCGATTCAGC-3′, 5′-AACGCAATGCCCATCACGGTTAG-3′; 331 bp), and β-actin (5′-CTTCTACAATGAGCTGCGTG-3′, 5′-TCATGAGGTAGTCAGTCAGG-3′; 305 bp) and 1 unit of AmpliTaq Gold DNA polymerase. The reactions were carried out in a thermal cycler with an initial denaturation step at 95°C for 14 min followed by 35 cycles (30 cycles for betaglycan and 22 cycles for β-actin) of denaturation at 95°C for 1 min, primer annealing at 50°C (inhibin-α), 64°C (betaglycan), and 55°C (β-actin) for 1 min, and extension at 72°C for 1 min. The reaction was terminated at 72°C for 10 min and samples were stored at 4°C. Ten microliters of PCR products was separated by electrophoresis on a 2% agarose gel containing ethidium bromide (0.5 μg/ml) and visualized by image analysis.

Immunohistochemical staining for the inhibin-α protein

Monolayer cell lines were seeded in eight-well Lab-Tek II chamber slides with covers (Nalgene, Rochester, NY, USA) and the cells were treated with 5-AzaC. At the end of the treatment period, suspension cell lines were attached on collagen coated-slides by cytospin. After a PBS wash, the cells were fixed on the slides using −10°C methanol for 5 min and air dried. The chambers were taken off before the immunostaining procedure. After incubation with 0.3% H2O2 for 10 min to block the endogenous peroxidase, the cells were incubated for 1 h with human inhibin-α mouse monoclonal antibody diluted to 5 μg/ml (Serotec, Kidlington, England). Subsequently, the biotin-conjugated secondary antibody was incubated for 30 min using mouse ABC staining systems according to the manufacturer’s protocol (Santa Cruz Biothechnology, Santa Cruz, CA, USA), and incubated with horseradish peroxidase (HRP)-streptavidin complex for 30 min and visualized by reaction with diaminobenzidine (DAB) for 5 min. Slides were stained with Mayer’s haematoxylin and dehydrated through ethanol and xylene. Concentration-matched mouse IgG2a was used as a negative control.

Flow cytometric analysis

The cultured cells were detached with 0.05% trypsin-EDTA solution. After washing with cold PBS, cells were then incubated with a 1:50 dilution of anti-inhibin-α goat polyclonal antibody (Santa Cruz Biotechnology) or normal goat serum as a negative control for 30 min at 4°C. After being washed three times with cold PBS, cells were stained with fluorescein isothiocyanate (FITC)-labeled donkey antibody to 1:50 diluted rabbit immunoglobulin for 30 min at 4°C. Washing was repeated in the same manner and cell surface immunofluorescence was analyzed using a FACSCalibur with CellQuest software (Becton Dickinson, Franklin Lakes, NJ, USA).

5-aza-2′-deoxycytidine (5-AzaC) treatment

Cells were seeded at a density of 5×105 cells and then allowed to attach during a 24-h period, and treated with 5-AzaC (Sigma, St. Louis, MO, USA) at 10 μM for 5 days, which was consistent with another report (32). For the doubling time experiment, 5-AzaC was used at 5 μM. The medium and the drug were replaced every 2 day. At the end of the treatment period, the medium was removed and the cell pellets were used for analysis.

Determination of cell doubling time

Cells were treated with 5-AzaC and washed with PBS. Cells were seeded at 2×104 cells/ml in 12-well plates with culture medium, and cell number/dish was counted with a trypan blue assay each day for 5 consecutive days. Untreated cells were analyzed under similar conditions as a control. The average cell number from two plates was determined, and the mean cell numbers were plotted to define the cell population doubling times. The cell population doubling time was calibrated by a formula of Kuchler (33).

Cell cycle analysis

Cells (5×105/100-mm dish) were treated with 5-AzaC. At the end of the treatment period, cells were harvested and washed with PBS. Cells were fixed with 70% ethanol for 1 h, treated with RNasin (20 μg/ml) at 37°C for 1 h, stained with PI (50 μg/ml) (Sigma). DNA content at each cell cycle stage was analyzed using a FACSCalibur with CellQuest software (Becton Dickinson).

Results

Methylation status of the inhibin-α gene in human gastric cancer cell lines

Transcriptional silencing of tumor suppressor genes mediated by hypermethylation is a common feature of human cancer (23,34). As well, change in promoter methylation has been reported as a potential regulatory mechanism for the inhibin-α gene in prostate tumors (23). Thus, the methylation state of inhibin-α gene was investigated in human gastric cancer cells. Methylation was determined at the seven CpG sites which are located −149 to −284 bp from the ATG site of the inhibin-α gene, using bisulfite DNA sequencing (23) (Fig. 1A). We observed that inhibin-α promoter was heavily methylated in the SNU-1, SNU-5, and SNU-484 cell lines. The promoter was moderately methylated in SNU-719, AGS, and KATO III cell lines. In contrast, the promoters were largely unmethylated in the SNU-16, SNU-601, SNU-638 and SNU-668 cell lines. The promoter of inhibin-α in normal epithelium (N. Epithelium) was relatively unmethylated whereas that in normal peripheral blood leukocytes (N. PBL) was heavily methylated (Fig. 1B). To determine how methylation status might affect the growth characteristics of these cells (reported in a later section), the lines were treated with 5-AzaC, a demethylating agent. The result was that many CpG sites were unmethylated after treatment with 5-AzaC in SNU-1, SNU-5, SNU-484 and SNU-719 although the degree of methylation loss was different for each line. Of interest was that specifically CpG 5 site was strongly influenced by 5-AzaC (Fig. 1C).

Figure 1 Methylation analysis of the inhibin-α promoter region. (A) Map of the seven CpG sites in the inhibin-α promoter within a 135 bp regions from −149 to −284 of the ATG start. (B) Methylation profile of the inhibin-α promoter region in human gastric cancer cell lines. (C) Methylation profile of the inhibin-α promoter region in human gastric cancer cell lines after 5-AzaC treatment. Cells were exposed to 5-AzaC for 5 days. Methylation levels were determined by sequencing of 10 independent clones derived from amplified bisulfite-treated DNA isolated from cancer cell lines. Methylated and unmethylated CpGs are represented by closed and open circles, respectively. Demethylated CpGs after 5-AzaC treatment are represented by half-closed circles. AP1, 2, and 3, activator proteins 1, 2, and 3; CRE, cAMP response element; SP1, specific protein 1, SBE; Smad binding element; N. Epithelium, normal epithelium; N. PBL, normal peripheral blood leukocytes.

LOH and mutation of the inhibin-α gene in human gastric cancer cell lines

Transcriptional silencing of tumor suppressor genes can also occur through LOH of the gene (23,35). Thus, we investigated the LOH of inhibin-α gene in human gastric cancer cells. Microsatellite markers on 2q32-q33 (D2S389) and 2q33-q36 (D2S128) was amplified by PCR. The result was that LOH occurred with at least one microsatellite marker at 2q32-33 in SNU-5 and SNU-484 whereas LOH at 2q33-36 was observed in AGS, KATO III and SNU-484. Interestingly, LOH both at 2q32-33 and 2q33-36 was observed in SNU-484. In addition, allelic imbalances were observed both at 2q32-33 locus in SNU-16, SNU-601, and SNU-719 and at 2q33-36 locus in SNU-1, SNU-5, SNU-16, SNU-601, SNU-638, and SNU-719 (Fig. 2A).

Figure 2 (A) LOH analysis of chromosome 2q and (B–D) mutation analysis at the sites of −16 bp of 5′-UTR and +769 bp of exon 2. (A) Genomic DNAs obtained from human gastric cancer cell lines were amplified using PCR with primers as described in Materials and methods. Primers covered the microsatellite markers on 2q32-q33 and 2q33-q36. PCR product was separated on 12% polyacrylamide gels, stained with ethidium bromide, and visualized by image analysis. LOH was expressed when band intensity was decreased below 50% compared with normal band. AI was expressed when band size was different from normal allele. NI, not informative; LOH, loss of heterozygosity; AI, allelic imbalance. (B) Fragment A (240 bp) which contained region of −16 bp of 5′-UTR was digested with SpeI after PCR as described in Materials and methods. Fragment of 240 bp indicated a wild-type allele, whereas two fragments of 120 bp corresponded to an allele which contains T at −16 bp of 5′-UTR. (C) Fragment B (396 bp) which contained region of 769G>A of exon 2 digested with BsrFI after PCR as described in Materials and methods. PCR product rendered two fragments of 340 and 56 bp in wild-type allele G and remained intact in the mutated allele A which was not observed. (D) Fragment B was digested with Fnu4HI. PCR product rendered four fragments of 153, 107, 51 and 25 bp in the wild-type allele, whereas the allele with substituted 769G>A rendered four fragments of 153, 107, 76 and 51 bp. The result of (C) and (D) revealed that substituted 769G>A of exon 2 was not occurred in the human gastric cancer cell lines. PCR product incubated with restriction enzyme overnight and separated on 8 and 15% polyacrylamide gels, stained with ethidium bromide, and visualized by image analysis. +E, with restriction enzyme; −E, without restriction enzyme.

Transcriptional silencing of tumor suppressor genes can also occur through mutation of the gene. Two polymorphic sites were indentified in the inhibin-α gene: −16C>T in the 5′-UTR (30) and 769G>A in exon 2 in previous report (11,30,36). Thus, we investigated these variants of the inhibin-α gene in human gastric cancer cells. The first investigated region was located at the 5′-UTR region (Fig. 2B). The second investigated region was located at 769 bp within exon 2 region (Fig. 2C and D). Thus, the single base change at 769G>A of exon 2 was not found in all human gastric cancer cell lines.

The low levels of mRNA and protein of inhibin-α in human gastric cancer cell lines

To determine how the observed differences in promoter DNA methylation or LOH might relate to differences in inhibin-α gene expression among cell lines, we measured inhibin-α mRNA levels by PCR, and observed low or undetectable levels of mRNA in all gastric cancer cell lines except in SNU-668. In most lines, the low level of mRNA was increased following 5-AzaC treatment (Fig. 3). Recently, inhibin-α was shown to act through a betaglycan signaling system (28). When we measured the level of mRNA of beta-glycan, it was not closely associated with the mRNA level of inhibin-α. Moreover, mRNA level of betaglycan was not changed with 5-AzaC treatment in contrast to inhibin-α (Fig. 3).

Figure 3 mRNA levels of inhibin-α and betaglycan after 5-AzaC treatment. Cells were cultured with 5-AzaC. Then mRNA was extracted and RT-PCR was performed. PCR product was separated on a 2% agarose gel, stained with ethidium bromide and visualized by image analysis. The mRNA of β-actin was amplified as a control.

The low level of mRNA of inhibin-α in most cell lines is expected to result in low levels of inhibin-α protein. When we measured the inhibin-α protein by IHC using a monoclonal inhibin-α antibody, faint signals which meant very low level of inhibin-α was observed in all human gastric cell lines. The low level of inhibin-α could be increased with 5-AzaC treatment at detectable levels in AGS, SNU-601, SNU-638, SNU-668 and SNU-719 (monolayer cell lines) and KATO III (suspension cell line). Other suspension cell line such as SNU-1, SNU-5 and SNU-16 were not immunostained both on collagen-coated slide and poly-D-lysine coated slide, resulting in no signals (Fig. 4A). The protein profile was reinvestigated using flow cytometry. When we measured inhibin-α at 5 days after 5-AzaC treatment, the levels of inhibin-α protein were increased between 1.25 and 28.67-fold compared with control among the various cell lines (Table II). Fluorescence intensities after 5-AzaC treatment were particularly higher in AGS, KATO III, SNU-1, SNU-16, SNU-668 and SNU-719 compared to control (Fig. 4B).

Figure 4 Inhibin-α protein detected by (A) immunohistochemistry and (B) flow cytometry. (A) For monolayer cells, cells were culture in chambered slide with 5-AzaC. For suspension cells, cells were cultured with 5-AzaC, washed with PBS and attached on collagen coated-slides by cytospin. The imunoreactivity of inhibin-α protein was detected using inhibin-α monoclonal antibody. Immunostained cells were examined under ×200 magnifications using a light microscope. (B) Cells were cultured with 5-AzaC. These cells were incubated with anti-inhibin-α polyclonal antibody and FITC conjugated antibody. Diluted normal goat serum was used as a negative control for the primary antibody. Then inhibin-α protein was detected by flow cytometry. 5-AzaC-treated cells, thick line; control cells, thin line.

Table II Summary.

Table II

Summary.

			Chromosome		mRNA expression		Protein expression (flow cytometry)		Apoptosis		G0/G1
Cell lines	Methylation status	Mutation in 5′ UTR	2q32-33	2q33-36	Control	5-AzaC (fold)	Control	5-AzaC (fold)	Control	5-AzaC (fold)	Control (%)	5-AzaC (%) (fold)
AGS		Wild	NI	LOH	0.10	0.69 (6.90)	2.7	46.9 (17.37)	7.1	17.3 (2.44)	33.9	11.8 (0.34)
KATO III		Wild	NI	LOH	0.30	0.74 (2.47)	1.3	23.1 (17.77)	2.6	10.3 (3.96)	40.5	18.4 (0.45)
SNU-1	Hyper	M (hetero)	NI	AI	0.45	0.90 (2.00)	1.5	43.0 (28.67)	9.6	45.3 (4.72)	57.4	39.0 (0.68)
SNU-5	Hyper	M (hetero)	LOH	AI	0.76	0.81 (1.07)	2.9	14.8 (5.10)	2.2	34.6 (15.73)	59.2	30.3 (0.51)
SNU-16	Un	Wild	AI	AI	0.03	0.36 (12.00)	2.6	22.3 (8.58)	3.1	26.8 (8.65)	41.3	20.6 (0.50)
SNU 484	Hyper	M (homo)	LOH	LOH	0.56	0.66 (1.18)	0.4	0.50 (1.25)	18.4	28.0 (1.52)	43.1	32.5 (0.75)
SNU-601	Un	Wild	AI	AI	0.06	0.61 (10.17)	1.1	6.90 (6.27)	4.6	33.7 (7.33)	45.4	22.0 (0.48)
SNU-638	Un	Wild	NI	AI	0.11	0.24 (2.18)	1.4	9.40 (6.71)	4.2	32.3 (7.69)	51.1	33.8 (0.66)
SNU-668	Un	Wild	NI	LOH	0.99	1.03 (1.04)	3.4	22.2 (6.56)	4.0	14.8 (3.70)	56.8	36.0 (0.63)
SNU-719		Wild	AI	AI	0.04	0.14 (3.50)	2.8	35.6 (12.71)	5.0	22.2 (4.44)	62.5	41.1 (0.66)

[i] Level of mRNA and protein were expressed after treatment with 5-AzaC as indicated compared to the control. AI, allelic imbalance; NI, not informative; Hyper, hypermethylated; Un, unmethylated; M(hetero), mutant(hetero); M(homo), mutant(homo).

Cell growth suppression and apoptosis induced by 5-AzaC

Low levels of inhibin-α were expected in cells undergoing rapid division since inhibin-α is known as a tumor suppressor. Reactivation of the inhibin-α gene was observed after 5-AzaC treatment, so we sought to determine if this altered the growth characteristics of the cells. When we measured cell viability after 5-AzaC treatment (5 μM) it was decreased between 47.6% in SNU-5 and 35.6% in SNU-638 (Table I). When we measured the doubling time after 5-AzaC treatment it was increased 1.24 (SNU-484) to 2.56-fold (SNU-16) (Table I). Related to the low cell viability, apoptosis was also investigated. When we analyzed cancer cells using flow cytometry, apoptosis was increased 1.52 to 15.73-fold in human gastric cancer cells after 5-AzaC treatment. In addition, the G0/G1 phase was decreased from 0.34 (AGS) to 0.75-fold (SNU-484) (Fig. 5, Table II).

Figure 5 Apoptosis and cell cycle after 5-AzaC treatment. Cells were cultured with 5-AzaC. Cells were fixed in 70% ethanol after media were removed. Cells were then washed with PBS, treated with RNasin, and stained PI. DNA content at each cell cycle stage was analyzed by flow cytometry. The percentage of cells in apoptosis and G0/G1 phase were calculated.

Table I Effects of 5-AzaC on growth suppression.

Table I

Effects of 5-AzaC on growth suppression.

	Viability (%) 5-AzaC			Doubling time (h) 5-AzaC		Growth suppression (fold)
Gastric cancer cell lines	0 μM	5 μM	10 μM	0 μM	5 μM
AGS	100	34.4	33.9	20	50	2.50
KATO III	100	39.4	38.9	36	68	1.89
SNU-1	100	41.0	30.7	26	57	2.19
SNU-5	100	47.6	22.2	34	55	1.62
SNU-16	100	45.5	25.4	27	69	2.56
SNU-484	100	47.4	27.4	67	83	1.24
SNU-601	100	37.0	26.0	47	84	1.79
SNU-638	100	35.6	28.8	58	82	1.41
SNU-668	100	46.1	32.7	74	135	1.82
SNU-719	100	46.0	23.3	184	412	2.24

[i] Cells (5×10⁵/100-mm dish) were treated with 5-AzaC for 5 days and then changed with fresh medium and drug every 2 days. Viable cells were counted by the trypan blue assay. At the end of the treatment period, the cells were seeded in 12-well plate at 2×10⁴ cells/ml. Cell number/plates was counted by trypan blue assay each day for 5 consecutive days. Untreated cells were analyzed under similar condition as a control. The average cell number from two plates was determined, and the mean cell numbers were plotted to define the cell population doubling times.

When the cells undergoing apoptosis and cell cycle changes were investigated by microscopy, morphological changes were observed in epithelial-like cells (AGS, SNU-484, SNU-601, SNU-638, SNU-668 and SNU-719) and spherical cells (KATO III) after 5-AzaC treatment. Minor morphological changes were observed in lymphoblast-like cells (SNU-1, SNU-5 and SNU-16) (Fig. 6).

Figure 6 Morphology of cells after 5-AzaC treatment. Cells were cultured with 5-AzaC and examined under ×200 magnifications using light microscopy.

Discussion

Numerous reports link the inhibin-α gene with cancer since it was initially identified as a tumor suppressor gene in the gonads and adrenals by functional studies using inhibin-α deficient mice (20–22). Roles for inhibin-α in cell growth have been proposed in prostate, endometrium, and breast, with loss of inhibin expression or sensitivity linked to tumor initiation and progression and to poor patient survival (37). Although it was suggested that inhibin-α acts as a tumor suppressor during early tumorigenesis, it might also act as a tumor promoter during late-stage disease in prostate cancer (24,25).

Hypermethylation, LOH, and polymorphisms of inhibin-α were investigated in this study. Hypermethylation of the inhibin-α gene promoter was observed in several human gastric cancer cell lines although the degree of methylation varied substantially among cell lines. Heavy methylation was found in SNU-1, SNU-5 and SNU-484 whereas moderate methylation was observed in SNU-719, AGS and KATO III. Little or no methylation was observed in SNU-16, SNU-601, SNU-638 and SNU-668. Demethylation colud be induced by 5-AzaC in heavily or moderately methylated CpG sites in SNU-1, SNU-5, SNU-484 and SNU-719. In these cells, all CpG sites were influenced by 5-AzaC, but CpG5 was more strongly influenced by 5-AzaC than other sites (Fig. 1). Previously, it was reported that the inhibin-α gene promoter was hypermethylated in prostate cancer (23). In this report, four CpG sites (CpG1, CpG2, CpG5 and CpG6) were unique in human whereas three CpG sites (CpG3, CpG4 and CpG7) were conserved between species. CpG5 differed from the bovine, rat, and mouse sequence (23). Of interest is that this site that varies among species corresponds to the site that is strongly influenced by 5-AzaC.

LOH of the inhibin-α gene locus was also investigated. Initially, LOH was observed in inactivation of tumor suppressor genes (23). In this study, LOH at 2q32-33 was observed in SNU-5 and SNU-484 whereas LOH at 2q33-36 was observed in AGS, KATO III and SNU-484 (Fig. 2A). LOH at 2q32-36 contains the chromosome region harboring the inhibin-α gene (38). LOH at the inhibin-α locus on chromosome 2q was also reported in 6% granulosa cell tumors (35) and in 42% of prostate carcinomas (23).

Mutation of the inhibin-α gene was investigated. Mutation at −16 bp site of the 5′-UTR was observed in SNU-1 (one allele), SNU-5 (one allele) and SNU-484 (both alleles). However, mutation at 769 bp site within exon 2 region was not observed (Fig. 2B–D). Although controversy exists, mutation at −16, and 769 of the inhibin-α gene has been related to premature ovarian failure (POF) (11,30,36). The observed DNA changes directly influences the inhibin-α mRNA levels. In our study, it was observed that all human gastric cancer cell lines had at least one of the discussed DNA alterations including methylation, allelic imbalance including LOH, and nucleotide mutation, and these lines all exhibit low inhibin-α expression.

It was reported that levels of mRNA and protein of tumor suppressor gene were very low in cancer cells (24,39). In this study, low level of inhibin-α gene transcript existed in human gastric cancer cell lines except SNU-668 (Fig. 3). Parallel with this, low level of inhibin-α protein was observed in most of cell lines (Fig. 4A and B). Consistent with this, it was reported that malignant tissues lacked inhibin-α gene transcripts and protein whereas non-malignant regions of human primary prostate carcinomas expressed inhibin-α (24). These low levels of mRNA and protein inhibin-α could be increased to some degree with 5-AzaC treatment, consistent with the view that hypermethylation contributes to suppression of inhibin-α expression. These low levels of mRNA and protein of inhibin-α seemed to be correlated with the cancer state, whereas there was no such correlation to betaglycan mRNA levels. The levels of betaglycan mRNA were not changed after treatment of 5-AzaC (Fig. 3).

5-AzaC treatment enhanced apoptosis 1.52 to 15.73-fold compared to the controls (Fig. 5, Table II) and 5-AzaC suppressed cell growth by increasing the doubling time 1.24 to 2.56-fold (Table I). Morphology of the cells was changed after 5-AzaC treatment as presented in Fig. 6. The cell can be influenced by inhibin-α induced by 5-AzaC or 5-AzaC itself. It was reported that inhibin-α influences cell proliferation, cell growth, and metastasis (25,40,41). In normal cells, it was reported that inhibin A increased apoptosis in early ovarian antral follicles (40). It was also reported that overexpressed inhibin-α (1–32 amino acid) fragment inhibited bovine granulose cell (GC) proliferation and induced apoptosis in GC (42). Inhibin-α decreased the proliferation of ovarian cancer cell lines such as SKOV3, OCC1, OVCAR3 and A2780-s (41). Inhibin-α reduced tumor growth in LNCaP which is androgen-dependent prostate cancer (PCa) whereas inhibin-α increased tumor growth and metastasis in PC3 which is androgen-independent PCa (25). Importantly, targeted disruption of the inhibin-α gene resulted in an ovarian phenotype of granulosa cell tumors (20,43). Combined with our results and previous reports (20,43), it is evident that inhibin-α must be maintained at low levels in many cancerous cell types. These results suggest that inhibin-α has a critical function at the cellular level. 5-AzaC itself is known to have a cytotoxic effect. However, it was reported that this occurs through DNA methyltransferase (Dnmt) 3a, Dnmt3b (44) and caspase-8 (45). Moreover, it was reported that 5-AzaC altered the expression of several proteins involved in cell cycle regulation, apoptosis, and survival (32,46). In our study, it seemed that 5-AzaC influenced the cell through inhibin-α induction as well as other known gene induction.

As to the mechanism of action of inhibin-α in the process of carcinogenesis, there is little data and much to be learned. Carcinogenesis is a complex process whereby malignant transformation occurs through a sequence of events. The hallmarks of cancer include sustaining proliferation, evading growth suppressors, resisting cell death, enabling replicative immortality, inducing angiogenesis, and activating invasion and metastasis. Underlying these hallmarks are genome instability and inflammation (47). Our result suggested that inhibin-α acts at steps including sustaining proliferation, evading growth suppressors, and altering cell death. Among these steps, apoptosis has been relatively well studied. In cancer cells, inhibin-α is maintained at low level and there is little apoptosis to maintain cancer state. If we limit the function of inhibin-α within apoptosis, two routes are possible. One possibility is that inhibin-α itself reduce apoptosis. Another possible is that inhibin-α reduce apoptosis through inhibin A. In both cases, however, inhibin-α or inhibin A seem to act without changeing betaglycan levels since there was no change in betaglycan mRNA level after 5-AzaC treatment (Fig. 3).

Acknowledgements

We thank Professor Kelly E. Mayo (Northwestern University) for helpful suggestion and discussion. This study was supported by the Research Fund, 2010 of The Catholic University of Korea.

References