Purpose:

Metallothionein 2A (MT2A) suppresses the progression of human gastric cancer potentially through an “MT2A–NF-κB pathway” with unclear mechanisms. This study explored the role of a transcription factor, myeloid zinc-finger 1 (MZF1), in MT2A-NF-κB pathway and its clinical significance in gastric cancer.

Experimental Design:

MZF1 expression and function in gastric cancer were investigated in vitro and in vivo. The relationship between MZF1 and MT2A was determined by gain-of-function and loss-of-function assays in gastric cancer cells and an immortalized gastric cell line GES-1. The prognostic value of MZF1 expression in association with MT2A was evaluated using IHC in two cohorts.

Results:

MZF1 was epigenetically silenced in human gastric cancer cell lines and primary tumors. Overexpression of MZF1 in gastric cancer cells suppressed cell proliferation and migration, as well as the growth of xenograft tumors in nude mice. Knocking-down of MZF1 transformed GES-1 cells into a malignant phenotype characterized by increased cell growth and migration. Mechanistically, MZF1 was upregulated in both GC and GES-1 cells by MT2A ectopically expressed or induced upon treatment with a garlic-derived compound, diallyl trisulfide (DATS). MZF1 associated with MT2A was colocalized in the nuclei of GES-1 cells to target the promoter of NF-κB inhibitor alpha (NFKBIA). Clinically, MT2A and MZF1 were progressively downregulated in clinical specimens undergoing gastric malignant transformation. Downregulation of MT2A and MZF1 was significantly correlated with poorer patient prognosis.

Conclusions:

MT2A exerts its anti-gastric cancer effects by complexing with MZF1 to target NFKBIA. MT2A/MZF1 may serve as a valuable prognostic marker and a novel therapeutic target for human gastric cancer.

This article is featured in Highlights of This Issue, p. 897

Translational Relevance

Both metallothionein 2A (MT2A) and NF-κB are involved in the carcinogenesis and chemosensitivity of gastric cancer. Our previous studies suggest that diallyl trisulfide (DATS), a garlic-derived compound, suppresses gastric cancer growth and enhances gastric cancer chemosensitivity via an “MT2A–NF-κB” pathway. However, the precise link between MT2A and NF-κB remains unclear. Here, we found epigenetic silencing of transcription factor myeloid zinc-finger 1 (MZF1) is associated with MT2A in human gastric cancer cell lines and primary tumors. MZF1 suppresses gastric carcinogenesis by interacting with MT2A to bind to NFKBIA promoter. Downregulation of MZF1/MT2A was significantly correlated with the malignant potential of gastric cancer and poor survival of patients. Our results thus reveal an important role of MT2A/MZF1–NF-κB signaling cascade elicited by DATS in regulating gastric carcinogenesis and chemosensitivity. Thus, MT2A/MZF1 appears to be a diagnostic and prognostic biomarker for gastric cancer and MT2A/MZF1–NF-κB pathway as a potential therapeutic target.

Gastric cancer is one of the most common malignant cancers worldwide with a high mortality rate (1–3). Gastric carcinogenesis is a complicated process due to its histologic and etiologic heterogeneity (3, 4). Gastric cancer becomes symptomatic in the advanced stage, resulting in a 5-year overall survival rate below 20% (1, 2). The highest incidence, despite progressive decrease, occurs in China, Japan, Latin America, and Eastern Europe. However, higher survival rate can be achieved in Japan by early diagnosis and consecutive early tumor resection (2, 5). To reduce further the incidence of gastric cancer and improve therapeutic efficacy, it is crucial to delineate the molecular mechanisms of gastric cancer progression for discovery of unique biomarkers (3).

Metallothionein 2A (MT2A) and NF-κB are involved in the carcinogenesis and chemosensitivity of gastric cancer (6, 7–11). However, the primary role of MT2A in NF-κB activation in tumorigenesis and chemoresistance differs depending on tumor cell types and it remains unclear in gastric cancer (6, 12, 13). We recently showed that decreased expression of MT2A and NFKB inhibitor alpha (NFKBIA, also called as IκB-α) in human gastric cancer is associated with poor prognosis of patients (14). We also provided the first evidence for epigenetic upregulation of MT2A in gastric cancer by diallyl trisulfide (DATS), a garlic-derived compound that suppresses gastric cancer progression (15). Our studies suggested ability of DATS and its ability to sensitize gastric cancer cells to the chemotherapeutic reagent docetaxel (DOC) through an “MT2A–NF-κB” pathway, in which MT2A induced by DATS is likely to suppress NF-κB activation by directly enhancing NFKBIA (IκB-α) transcription in gastric cancer (15). However, MT2A is a low molecular-weight, heavy metal-binding protein that is unlikely to regulate NFKBIA activity as a transcription factor by direct binding to its promoter (6, 13, 16). Therefore, the precise link between MT2A and NF-κB remains to be elucidated.

Analysis of TRANSFAC database for putative transcription factors binding to NFKBIA promoter revealed multiple potential transcription factor-binding sites (TFBS), among which a consensus sequence of 5′ AGTGGGGA 3′ located in the −498/−493 region of NFKBIA promoter represents a typical binding motif for the transcription factor myeloid zinc-finger 1 (MZF1; refs. 17–19). MZF1 belongs to the mammalian Krüppel-like family of transcription factors and is involved in cellular proliferation and differentiation (20–22). MZF1 has shown to be crucial for regulation of proliferation, migration, and invasion of malignant tumor cells by acting both as a transcription activator and repressor (17–19, 23–31). However, few studies have revealed controversial effects of MZF1 on cancer development (32–34), raising the need for clarification on the biofunctional roles of MZF1 and its regulation in gastric tumorigenesis (21).

In this study, we investigated the function and clinical significance of MZF1 in the MT2A–NF-κB pathway in gastric cancer cells. We found epigenetic silencing of MZF1 in association with MT2A in human gastric cancer cell lines and primary tumors. MZF1 suppresses gastric carcinogenesis by interacting with MT2A to target directly the promoter region of NFKBIA. Downregulation of both MZF1 and MT2A was correlated with the malignant potential of gastric cancer and poor survival of patients. Our results thus delineate a mechanistic basis of MT2A/MZF1–NF-κB signaling cascade elicited by DATS to suppress tumor growth and increase the chemosensitivity of human gastric cancer.

Cell culture and treatment

Human GC cell lines and GES-1, an immortalized human gastric mucosal epithelial cell line, were cultured and treated as previously described (14, 15). Details are provided in Supplementary Materials and Methods.

Cell line construction

BGC823 cells stably expressing ectopic MT2A (BGC823-MT2A), and BGC823-MT2A with short hairpin RNA against MT2A (BGC823-MT2A/shMT2A) or nonspecific control (BGC823-MT2A/shNC) were constructed as described (14, 15). GES-1 cells with shMT2A stable transfection (GES-1-shMT2A) or nonspecific control (GES-1-shNC) was generated as previously procedure. The cDNA coding human MZF1 (sequence identification number NM_003422.2) was amplified by PCR and cloned into pEZ-M02 vector (GeneCopoeia, Inc.) to construct recombinant plasmid. In vitro transfection was performed using Lipofectamine 2000 (Invitrogen) following the manufacturer's instruction. Short hairpin RNA against MZF1 (shMZF1-1 and -2) was purchased from Santa Cruz (sc-45714-SH) and transfected into BGC823 or GES-1 cells for transient expression as previously described (14, 15). Transfection efficiency was evaluated by Western blotting or qPCR. All oligonucleotide sequences are listed in Supplementary Table S7.

Patients and specimens

Two independent cohorts were used in this study. Written informed consents were obtained from the patients or their legal guardians. Ethical clearance was approved for this study in accordance with the ethical standards of the respective institutional ethics committees on human experimentation and with the revised Helsinki Declaration before starting the study. For details, see Supplementary Materials and Methods.

RNA isolation, reverse transcription, and PCR

See details in Supplementary Materials and Methods.

DNA extraction, methylation-specific PCR/qPCR (MSP-PCR/qPCR)

Total DNA was extracted from cell lines and tissues using the genomic DNA Rapid Extraction Kit (Aurora Biomed, Inc.). Bisulfite modification of DNA was performed using Zymo DNA Methylation Kit (Zymo Research). PCR or qPCR was performed to amplify bisulfite modified DNA by using primer pairs that specifically amplify either methylated or unmethylated sequences of the MZF1 gene (Fig. 1C; Supplementary Table S7). The in vitro methylation DNA (IVD) that served as the positive control was the A & D Human Methylated DNA Standard (A & D Technology), and the negative control was the genomic DNA from normal human peripheral lymphocytes as described (35). MSP products were analyzed using a 2% agarose gel electrophoresis.

Figure 1.

Epigenetic silencing of MZF1 in human gastric cancer. A, Gastric cancer cell lines and immortalized human gastric cell GES-1 were treated with 5-AZA and/or TSA (5 μmol/L, 96 hours; TSA: 5 μmol/L, 24 hours). MZF1 mRNA expression, relative to GAPDH as internal control, was measured by quantitative RT-PCR using the 2−ΔΔCt method. Results represent the mean ± SD. *, P < 0.05; **, P < 0.01 (unpaired t test) compared with untreated control. B, qRT-PCR measurement of the mRNA expression levels of MZF1 in tumor and adjacent tissue samples from patients with gastric cancer (n = 24). GAPDH served as the internal control. **, P < 0.01. C, Schematic diagram of MZF1 gene. CpG islands and the primer pairs for MSP and ChIP within the MZF1 promoter were shown as indicated. D, CpG methylation status in the promoter region of MZF1 was analyzed by qPCR using MSP primer 2 on the bisulfite-treated genomic DNA isolated from gastric cancer cell lines and GES-1 cells. The results were presented with percentage of methylation (M) and unmethylation (UM), respectively, related to the methylated or unmethylated status in GAPDH promoter. E, The methylation frequency of MZF1 promoter determined using MSP or qMSP in tumor and adjacent tissue samples from patients with gastric cancer (n = 18). The frequency of methylation (M) and unmethylation (UM) in the samples were presented as percentage of cases, respectively. **, P < 0.01 (unpaired t test). F, MSP analysis of methylation frequency of MZF1 promoter in FFPE tumor tissues from patients with gastric cancer (n = 34). **, P < 0.01 (unpaired t test). G, ChIP assay was performed on BGC823 cells after TSA (5 μmol/L, 24 hours) treatment using antibody against histone 3 acetylation at lysine 9 (H3K9ac) or control IgG. ChIP precipitated DNA fractions were analyzed by qPCR for the enrichment of H3K9ac or IgG in in the promoter region of MZF1. The position of ChIP-qPCR primer is shown in Fig. 1C. Results are expressed as the percentage of input.

Figure 1.

Epigenetic silencing of MZF1 in human gastric cancer. A, Gastric cancer cell lines and immortalized human gastric cell GES-1 were treated with 5-AZA and/or TSA (5 μmol/L, 96 hours; TSA: 5 μmol/L, 24 hours). MZF1 mRNA expression, relative to GAPDH as internal control, was measured by quantitative RT-PCR using the 2−ΔΔCt method. Results represent the mean ± SD. *, P < 0.05; **, P < 0.01 (unpaired t test) compared with untreated control. B, qRT-PCR measurement of the mRNA expression levels of MZF1 in tumor and adjacent tissue samples from patients with gastric cancer (n = 24). GAPDH served as the internal control. **, P < 0.01. C, Schematic diagram of MZF1 gene. CpG islands and the primer pairs for MSP and ChIP within the MZF1 promoter were shown as indicated. D, CpG methylation status in the promoter region of MZF1 was analyzed by qPCR using MSP primer 2 on the bisulfite-treated genomic DNA isolated from gastric cancer cell lines and GES-1 cells. The results were presented with percentage of methylation (M) and unmethylation (UM), respectively, related to the methylated or unmethylated status in GAPDH promoter. E, The methylation frequency of MZF1 promoter determined using MSP or qMSP in tumor and adjacent tissue samples from patients with gastric cancer (n = 18). The frequency of methylation (M) and unmethylation (UM) in the samples were presented as percentage of cases, respectively. **, P < 0.01 (unpaired t test). F, MSP analysis of methylation frequency of MZF1 promoter in FFPE tumor tissues from patients with gastric cancer (n = 34). **, P < 0.01 (unpaired t test). G, ChIP assay was performed on BGC823 cells after TSA (5 μmol/L, 24 hours) treatment using antibody against histone 3 acetylation at lysine 9 (H3K9ac) or control IgG. ChIP precipitated DNA fractions were analyzed by qPCR for the enrichment of H3K9ac or IgG in in the promoter region of MZF1. The position of ChIP-qPCR primer is shown in Fig. 1C. Results are expressed as the percentage of input.

Close modal

Western blot analysis

Protein was measured using a BCA protein Assay Kit (CWBIO). Proteins were resolved by SDS-PAGE and transferred onto PVDF membranes using a Bio-Rad Mini PROTEAN 3 system. The membranes were blocked with PBS containing 5% milk and 0.1% Tween-20 at room temperature for 1 hour. The membranes were then immunoblotted with desired primary antibodies and incubated with the responding horseradish peroxidase conjugated anti-mouse or anti-rabbit secondary antibodies. The primary antibodies were as follows: antibodies against MT2A (mouse monoclonal, 1:1,000; ab12228), H3K9ac (rabbit polyclonal, 1:1,000; ab4441), H4K5ac (rabbit polyclonal, 1:1,000; ab114146) were purchased from Abcam, anti-HDAC1 (rabbit polyclonal, 1:800; AH379) and anti-HDAC2 (rabbit polyclonal, 1:800; AH382) were from Beyotime Ltd. Anti-β-actin (mouse monoclonal, 1:10,000; A4551) was from Sigma. Anti-MZF1 (rabbit polyclonal, 1:800, BS5810), horseradish peroxidase-conjugated anti-mouse (1:2,500 dilution), or anti-rabbit (1:2,500 dilution) secondary antibodies were purchased from Bioworld Technology, Inc. Immunoreactive bands were visualized by using the Amersham ECL Western Blotting Detection Kit according to the manufacturer's instructions as described (15, 35). β-Actin served as a loading control.

Chromatin immunoprecipitation

Chromatin immunoprecipitation (ChIP) assay was performed following the EpiTech ChIPOneDay Kit protocol (QIAGEN) as described (15, 35). GES-1 cells with shMT2A stable transfection (GES-1-shMT2A), nonspecific control (GES-1-shNC) or BGC823 cells were fixed with 1% formaldehyde. Chromatin was prepared by sonication of cell lysate and preclearing with protein A beads. Aliquots of precleared chromatin solution, named as IP fractions, were incubated with 2 μg of specific anti-MZF1(BS5810), anti-MT2A (ab12228), or preimmune rabbit IgG on a rotation platform at 4°C overnight. 1% of IP fraction was served as the ChIP input control. The antibody-enriched protein–DNA complexes were precipitated with protein A beads from the IP fractions. DNA fragments were released by reversal crosslinking and purified by using a QIA Quick Purification Kit (Qiagen). Immunoprecipitated DNA fractions were analyzed by qPCR.

Immunofluorescence

GES-1 cells, BGC823 cells transfected or not with MT2A, or treated with DATS (40 μmol/L, 12 hours) were fixed with 4% paraformaldehyde at room temperature for 10 minutes followed by washing and preblocking. The cells were incubated overnight at 4°C with antibodies against MZF1 (1:50, BS5810) or MT2A (1:50, ab12228), respectively, followed by incubation with FITC-conjugated secondary antibody (1:50; Santa Cruz) for 1 hour. DAPI was used for nuclear staining (10 μg/mL in PBS, Invitrogen, Life Technologies). Images were then recorded by laser confocal microscopy (Leica Sp5 Laser Scanning Confocal Microscope, GE).

Immunohistochemistry

IHC was performed on gastric mucosal biopsy specimens, or tumor sections from GC patients using anti-MZF1 or MT2A antibody. The expression of MZF1 or MT2A was defined as MT2ALow, MT2AHigh, MZF1Low, or MZF1High, respectively, in accordance with the procedure described previously (14, 15). For details, see Supplementary Materials and Methods.

Cell viability

BGC823 or GES-1 cells were seeded into 96-well plates at 2 × 103 cells/well. Cell viability was measured every day by using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay Kit (KeyGEN Biotech) as described (14, 15). MTT readout was presented as the mean value of the absorbance at 490 nm subtracted by the reference absorbance at 570 nm.

Colony formation

BGC823 or GES-1 cells were seeded in soft agar in six-well tissue culture plates (0.1–0.2 × 103 cells/well) in triplicate. Colonies with more than 50 cells were counted after 2 weeks by following the previous protocol (15, 35).

Cell monolayer scratching, migration

The cell monolayer-scratching assay was performed with serum-starved cells and images were captured using IncuCyte Zoom (Essen Bioscience) at every 12 hours after cell monolayer scratching. Transwell apparatus was used with 8-μm pore inserts (Corning Inc.). For migration assay, the upper chambers were seeded with 200 μL of serum-free medium containing 1 × 104 of serum-starved cells in GES-1, 2.5 × 104 in BGC823, respectively. The lower chambers were filled with 500 μL of 10% FBS-DMEM. After 18 hours for BGC823, 24 hours for GES-1 cells, the cells that migrated to the lower chamber were fixed and stained with 0.2% crystal violet (Beyotime; ref. 35).

Tumorigenicity

The animal handling and all in vivo experimental procedures were approved by the Institutional Animal Ethics Committee of Peking University Cancer Hospital. A total of 1 × 106 BGC823 cells transfected with MZF1 (MZF1), or vector control (Vector) in 0.1 mL PBS were subcutaneously implanted in 4-week-old Balb/c female athymic mice (Vital River Laboratories, Beijing, China) with both flank injection. Animal handling and all in vivo experiments were performed as described (15). Tumor diameters and body weight in gastric cancer xenograft tumor-bearing nude mice were measured and documented every 5 days until the animals were sacrificed at day 20. Gastric cancer tumor xenografts were isolated and weighted. Tumor volume was calculated by measuring the longest (a) and shortest (b) diameters of the tumor and calculated by the formula: ab2/2, where a is for the longest diameter, b is for the shortest diameter.

Statistical analysis

All statistical analyses were performed using SPSS version 23.0 (SPSS Inc.). The data are expressed as means ± SD of at least three independent experiments. Statistical significance was considered when P < 0.05. DNA methylation in human gastric cancer was analyzed by the Student t test. The relationship between MZF1/MT2A and clinicopathologic characteristics was measured by χ2 or Fisher exact tests. Cancer-related survival was analyzed using Kaplan–Meier method and compared using log-rank tests. Spearman rank test and Fisher exact test were used to analyze clinicopathologic correlation. Kaplan–Meier analysis with log-rank test was used to calculate the prognostic value. Univariate and multivariate survival analyses were performed using the Cox proportional hazard model with a forward stepwise procedure (the entry and removal probabilities were 0.05 and 0.10, respectively).

MZF1 is epigenetically silenced in gastric cancer cells

To determine whether MZF1 participates in MT2A–NF-κB pathway in gastric cancer, we first assessed the expression of MZF1 in human gastric cancer cell lines, BGC823, MGC803, AGS, SGC7901, and NCI-N87, as well as in an immortalized gastric mucosal epithelial cell line, GES-1. MZF1 mRNA expression was barely detectable in BGC823, MGC803, NCI-N87, and AGS cell lines, weakly expressed in SGC7901 cell lines in contrast to GES-1 cells, in which MZF1 was highly expressed (Fig. 1A; Supplementary Fig. S1A). Western blotting showed that the MZF1 protein was completely absent in BGC823, MGC803, and NCI-N87 cells, slightly expressed in SGC7901 and AGS cells compared with GES-1 cells (Supplementary Fig. S1B). Reverse transcription (RT)-qPCR analysis of MZF1 expression in 24 pairs of gastric tumor samples and matched adjacent normal tissues showed that the mRNA level of MZF1 was significantly lower in gastric tumors compared to matched adjacent tissues (Fig. 1B; Supplementary Fig. S1C).

To elucidate the involvement of epigenetic regulation in the downregulation of MZF1 expression in gastric cancer, gastric cancer cell lines were treated with 5-aza-2′-deoxycytidine (5-AZA) in combination with trichostatin A (TSA). As shown in Fig. 1A and Supplementary Fig. S2A, the expression of MZF1 mRNA was restored in N87 and AGS, and was weakly upregulated in BGC823 and MGC803 cells by 5-AZA treatment, whereas MZF1 mRNA expression significantly increased by TSA, and further enhanced by combination of TSA with 5-AZA in BGC823, MGC803, NCI-N87, and AGS cells. However, MZF1 expression was not significantly affected in SGC7901 and GES-1 cells upon the treatments.

We then examined epigenetic status of MZF1 promoter in gastric cancer (Fig. 1C). Both MSP-PCR and MSP-qPCR showed complete or partial methylation of the promoter region of MZF1 in gastric cancer cells in contrast to an unmethylated status in either GES-1, or gastric embryonic cells (Fig. 1D; Supplementary Fig. S2B). The methylation level of MZF1 promoter was further determined in 18 pairs of gastric tumor samples and matched adjacent normal tissues, in which MZF1 promoter was methylated in 89% (16/18) of gastric cancer tumors and 27% (5/18) of adjacent normal tissues, respectively (Fig. 1E; Supplementary Fig. S2C and S2D). The results were confirmed in 31 cases of formalin-fixed paraffin-embedded (FFPE) tissue samples from patients with gastric cancer (Fig. 1F; Supplementary Fig. S2E).

We further investigated histone acetylation status associated with DNA methylation in MZF1 promoter region in gastric cancer cells. Western blotting analysis showed that TSA treatment of BGC823 cells increased histone 3 acetylation at lysine 9 (H3K9ac) and histone 4 acetylation at lysine 5 (H4K5ac), accompanied with reduction of histone deacetylase 1 and 2 (HDAC1/2; Supplementary Fig. S2F). To analyze the association of histone hyperacetylation with transcriptional activity of MZF1 gene, ChIP was performed with TSA-treated BGC823 cells. Real-time qPCR analysis of genomic DNA immunoprecipitated with anti-H3K9ac antibody showed that H3K9ac was significantly enriched in the MZF1 promoter region with a methylated status (Fig. 1C and G). These findings suggest that epigenetic silencing of MZF1 in gastric cancer is attributable to DNA hypermethylation and to histone hypoacetylation in the promoter region.

MZF1 suppresses malignant behavior of gastric cancer

The ectopic MZF1 was stably overexpressed in BGC823 cells (BGC823-MZF1) to evaluate its potential roles (Fig. 2A). Cell viability was significantly reduced in BGC823-MZF1 as compared with vector transfected control cells (BGC823-Vector; Fig. 2B). The result was confirmed by colony formation assay, in which BGC823-MZF1 cells formed significantly fewer colonies than vector control cells (Fig. 2C). We next examined the role of MZF1 in the metastatic ability of gastric cancer cells. In a cell monolayer scratching assay BGC823-MZF1 cells showed a remarked inhibition of cell migration ability (Fig. 2D). Consistent with this, overexpression of MZF1 caused a significant reduction in the migration ability of gastric cancer cells in matrigel (Fig. 2E). To evaluate the role of MZF1 in gastric cancer cell tumorigenesis in vivo, BGC823-MZF1 and BGC823-Vector cells were injected subcutaneously into nude mice. MZF1 overexpression caused a significant reduction in xenograft tumor volume formed by BGC823 cells in nude mice, without affecting body weights of the mice (Fig. 2F–H; Supplementary Fig. S3). These results suggest that MZF1 suppresses the tumorigenesis of human gastric cancer cells.

Figure 2.

The suppressive effects of ectopic MZF1 expression on the proliferation, migration, and tumorigenicity of human gastric cancer cells. The ectopic MZF1 was transfected into BGC823 GC cells for establishment of the stable cell line with overexpressed MZF1 (BGC823-MZF1), or vector control (BGC823-Vector) to evaluate the potential roles of MZF1 in gastric tumorigenesis in vitro and in vivo. A, Western blotting analysis of the protein level of MZF1 transfected with MZF1 or vector control. ß-Actin was sample loading control. B, MTT assays showed the viability of BGC823 cells with MZF1 overexpression (BGC823-MZF1), or vector control (BGC823-Vector). The results were presented as the mean ± SD. **, P < 0.01 vs. control. C, Colony formation in soft agar for 2 weeks by BGC823-MZF1 or BGC823-Vector cells. Top: Representative image. Bottom: The colony numbers counted shown as the mean ± SD. **, P < 0.01 vs. control. D, Cell monolayer scratching assay showing the different time points of cell migration ability of BGC823 cells with or without ectopic PRSS3 expression. E, Transwell migration assay assessing cell movement ability of BGC823 cells with or without MZF1 expression. Representative images presented on the top, quantification of the data shown on the bottom of the images as the mean ± SD. **, P < 0.01 vs. control. F–H, The effect of MZF1 on tumorigenicity of human gastric cancer cells was evaluated by subcutaneously injecting BGC823-MZF1 or BGC823-Vector cells into the left and right flanks of nude mice (n = 2 flanks × 5 mice in each group). F, Photographs of dissected BGC823 xenograft tumors at day 20 from BGC823-MZF1 or BGC823-Vector cells. G and H, The measurement of tumor volume (G) and body weight (H) of tumor-bearing mice at the times indicated. Results shown represent the mean ± SD; **, P < 0.01 (unpaired t test) vs. control.

Figure 2.

The suppressive effects of ectopic MZF1 expression on the proliferation, migration, and tumorigenicity of human gastric cancer cells. The ectopic MZF1 was transfected into BGC823 GC cells for establishment of the stable cell line with overexpressed MZF1 (BGC823-MZF1), or vector control (BGC823-Vector) to evaluate the potential roles of MZF1 in gastric tumorigenesis in vitro and in vivo. A, Western blotting analysis of the protein level of MZF1 transfected with MZF1 or vector control. ß-Actin was sample loading control. B, MTT assays showed the viability of BGC823 cells with MZF1 overexpression (BGC823-MZF1), or vector control (BGC823-Vector). The results were presented as the mean ± SD. **, P < 0.01 vs. control. C, Colony formation in soft agar for 2 weeks by BGC823-MZF1 or BGC823-Vector cells. Top: Representative image. Bottom: The colony numbers counted shown as the mean ± SD. **, P < 0.01 vs. control. D, Cell monolayer scratching assay showing the different time points of cell migration ability of BGC823 cells with or without ectopic PRSS3 expression. E, Transwell migration assay assessing cell movement ability of BGC823 cells with or without MZF1 expression. Representative images presented on the top, quantification of the data shown on the bottom of the images as the mean ± SD. **, P < 0.01 vs. control. F–H, The effect of MZF1 on tumorigenicity of human gastric cancer cells was evaluated by subcutaneously injecting BGC823-MZF1 or BGC823-Vector cells into the left and right flanks of nude mice (n = 2 flanks × 5 mice in each group). F, Photographs of dissected BGC823 xenograft tumors at day 20 from BGC823-MZF1 or BGC823-Vector cells. G and H, The measurement of tumor volume (G) and body weight (H) of tumor-bearing mice at the times indicated. Results shown represent the mean ± SD; **, P < 0.01 (unpaired t test) vs. control.

Close modal

To verify further the role of MZF1 in suppressing gastric malignancy, an immortalized human gastric mucosal epithelial cell line GES-1 with constitutive expression of MZF1 was stalely transfected with MZF1-specific short hairpin RNA (shMZF1) to generate an MZF1 knockdown cell line (GES-1-shMZF1; Fig. 3A). shMZF1 significantly increased the proliferation rate and the number of colonies formed in the soft agar by GES-1 cells as compared with control cells (Fig. 3B and C). Moreover, the migration ability of GES-1 cells was significantly augmented by specific shMZF1 transfection (Fig. 3D and E). Therefore, the presence of MZF1 reduces the malignancy of gastric cancer cells.

Figure 3.

The effect of MZF1 on gastric malignancy. GES-1 cells were stably transfected with MZF1-specific shRNA (GES-1-shMZF1), or scrambled shRNA as negative control (GES-1-shNC) to evaluate the potential roles of MZF1 in gastric cancer development. A, Western blotting analysis of MZF1 in GES-1-shMZF1, or GES-1-shNC cells. B, MTT assays showing the viability of GES1 cells transfected with shMZF1 or shNC. C, Colony formation of GES-1-shMZF, or GES-1-shNC cells in soft agar for 2 weeks. Top: Representative image. Bottom: Quantitative analysis shown as the mean ± SD. *, P < 0.05 vs. control. D, Cell monolayer scratching assay for monitoring the cell migration ability of GES-1 cells transfected with shMZF1 or shNC. E, Cell movement ability of GES-1 upon transfection with shMZF1 or shNC was assessed using Transwell migration assay. The results were shown as representative image (top) and quantitative analysis (bottom: mean ± SD; *, P < 0.05 vs. shNC), respectively.

Figure 3.

The effect of MZF1 on gastric malignancy. GES-1 cells were stably transfected with MZF1-specific shRNA (GES-1-shMZF1), or scrambled shRNA as negative control (GES-1-shNC) to evaluate the potential roles of MZF1 in gastric cancer development. A, Western blotting analysis of MZF1 in GES-1-shMZF1, or GES-1-shNC cells. B, MTT assays showing the viability of GES1 cells transfected with shMZF1 or shNC. C, Colony formation of GES-1-shMZF, or GES-1-shNC cells in soft agar for 2 weeks. Top: Representative image. Bottom: Quantitative analysis shown as the mean ± SD. *, P < 0.05 vs. control. D, Cell monolayer scratching assay for monitoring the cell migration ability of GES-1 cells transfected with shMZF1 or shNC. E, Cell movement ability of GES-1 upon transfection with shMZF1 or shNC was assessed using Transwell migration assay. The results were shown as representative image (top) and quantitative analysis (bottom: mean ± SD; *, P < 0.05 vs. shNC), respectively.

Close modal

MT2A regulates MZF1 expression

To explore the position of MZF1 in MT2A-NF-κB pathway in gastric cancer, we first examined the expression of MZF1 upon knockdown of ectopic MT2A in BGC823-MT2A cells (BGC823-MT2A/shMT2A). As shown in Fig. 4A and B, MZF1 expression was restored in both protein and mRNA levels in BGC823-MT2A, but diminished in BGC823-MT2A/shMT2A cells, suggesting a regulatory effect of MT2A on MZF1 expression. We then investigated whether MZF1 is associated with MT2A in GES-1 cells, in which both MT2A and MZF1 are constitutively expressed (Fig. 1A; Supplementary Fig. S1A and S1B; Fig. 4C and D). Knockdown of MT2A with transfection of specific shMT2A in GES-1 cells significantly reduced the expression of MZF1 (Fig. 4C and D), confirming the transcriptional regulation of MZF1 by MT2A.

Figure 4.

The transcriptional regulation of MT2A on MZF1 expression. A and B, MT2A overexpressed BGC 823 cells (BGC823-MT2A) were transfected with MT2A-specific shRNA (BGC823-MT2A/shMT2A), or scrambled shRNA as negative control (BGC823-MT2A/shNC). The expression of MT2A and MZF1 was determined by Western blotting (A) or RT-qPCR (B), respectively. C and D, GES-1 cells were transfected with MT2A-specific shRNA (GES-1-shMT2A), or negative control (GES-1-shNC). The expression of MT2A and MZF1 was determined by Western blotting (C), or RT-qPCR (D). E and F, BGC823 and GES-1 cells with or without the transfection of shMT2A were treated by DATS (40 μmol/L, 12 hours). The cells were subjected to RT-qPCR analysis (E), or Western blotting (F). G and H, qRT-PCR analysis of the expression of MT2A and MZF1 in BGC823-MT2A cells (G), or GES1 cells (H) upon shMZF1 transfection. The results presented as the bar charts show the mean ± SD of the fold changes in mRNA levels relative to GAPDH mRNA levels. *, P < 0.05; **, P < 0.01, vs. control, respectively.

Figure 4.

The transcriptional regulation of MT2A on MZF1 expression. A and B, MT2A overexpressed BGC 823 cells (BGC823-MT2A) were transfected with MT2A-specific shRNA (BGC823-MT2A/shMT2A), or scrambled shRNA as negative control (BGC823-MT2A/shNC). The expression of MT2A and MZF1 was determined by Western blotting (A) or RT-qPCR (B), respectively. C and D, GES-1 cells were transfected with MT2A-specific shRNA (GES-1-shMT2A), or negative control (GES-1-shNC). The expression of MT2A and MZF1 was determined by Western blotting (C), or RT-qPCR (D). E and F, BGC823 and GES-1 cells with or without the transfection of shMT2A were treated by DATS (40 μmol/L, 12 hours). The cells were subjected to RT-qPCR analysis (E), or Western blotting (F). G and H, qRT-PCR analysis of the expression of MT2A and MZF1 in BGC823-MT2A cells (G), or GES1 cells (H) upon shMZF1 transfection. The results presented as the bar charts show the mean ± SD of the fold changes in mRNA levels relative to GAPDH mRNA levels. *, P < 0.05; **, P < 0.01, vs. control, respectively.

Close modal

We previously showed the capacity of a garlic component DATS to induce MT2A expression in gastric cancer cells (15). We therefore investigated the interaction between MZF1 and MT2A in gastric cancer and GES-1 cells upon DATS treatment. As shown in Fig. 4E and F, DATS treatment induced the expression of MZF1 at both mRNA and protein levels in association with MT2A expression in BGC823 cells. MT2A knockdown suppressed DATS-induced expression of MZF1 in BGC823 cells. In GES-1 cells, introduction of shMT2A also attenuated the expression of MZF1 and MT2A upon DATS treatment, although the expression of MZF1 and MT2A was not affected by DATS. Interestingly, knockdown of MZF1 with shMZF1 decreased MT2A expression in BGC823-MT2A (Fig. 4G) and GES-1 cells (Fig. 4H), indicating a positive feedback interaction between MZF1 and MT2A at the transcriptional level.

MZF1 associated with MT2A binds to NFKBIA promoter

To investigate the interaction between MZF1 and MT2A, co-immunoprecipitation (co-IP) performed in BGC823-MT2A showed that MT2A or MZF1 was detected in cell lysate pulled down by antibody against MZF1 or MT2A, respectively (Fig. 5A), suggesting a physical interaction between MZF1 and MT2A. Our recent study demonstrated that DATS increased the sensitivity of gastric cancer cells to the DOC by enhancing MT2A expression to inhibit NF-κB activation (15). Fig. 5B showed an association of MZF1 with MT2A in gastric cancer cells treated with DATS, or DOC, or DATS/DOC, suggesting the involvement of MZF1 in MT2A–NF-κB pathway elicited by DATS/DOC in gastric cancer. Immunofluorescence (IF) imaging demonstrated MZF1 colocalization with MT2A in the nucleus of GES-1 cells, or BGC823 with MT2A expression, or with DATS treatment (Fig. 5C). To address whether MZF1 directly binds to MT2A, we performed a GST pull down assay with GST-MT2A and Flag-MZF1 purified from BGC823 cells. The result showed that GST-MT2A pulled down Flag-MZF1, whereas GST alone failed to associate with Flag-MZF1, indicating that MT2A associates directly with MZF1(Fig. 5D). These data suggest a regulatory function of MZF1 in cooperation with MT2A.

Figure 5.

MZF1 in cooperation with MT2A to target NFKBIA promoter. A, BGC 823 cells overexpressing MT2A (MT2A), or control (Vector) were subject to IP with MZF1 or MT2A antibodies followed by immunoblotting (IB) using MT2A or MZF1 antibodies, respectively. IP with IgG was used as a negative control. IB of the input with ß-actin antibody served as sample loading control. B, IP with MZF1 or MT2A antibodies was performed on BGC823 cells treated with DATS (40 μmol/L, 24 hours), or DOC (10 μmol/L, 24 hours), or DATS/DOC followed by IB with MT2A or MZF1 antibodies, respectively. IgG and ß-actin antibody used as a negative or loading control as described. C, Confocal IF staining of MT2A and MZF1 in GES-1 cells or BGC823 cells transfected or not with MT2A, or treated with DATS (40 μmol/L, 12 hours). Nuclei were shown by DAPI staining (blue). Shown are results from one of two comparable experiments. Scale bars, 10 μm. D, BGC823 cells were transfected with Flag-MZF1 or GST-MT2A expression plasmids. The purified proteins with Flag-tagged MZF1 were pulled down with purified GST or GST-MT2A. A 1/10 fraction of the sample input and the precipitates were subjected to immunoblotting with anti-MZF1 or anti-MT2A, respectively. E, A schematic illustration of the promoter region of NFKBIA showing the positions of the consensus sequence of 5′-AGTGGGGA-3′ (498/−493 region of the NFKBIA promoter) as well as the ChIP-PCR primer set P4 in the previous experiments. F, ChIP was performed on GES-1 and BGC823 cells with anti-MZF1 or MT2A antibody, or IgG as control. Precipitated ChIP DNA fractions were analyzed by qPCR for the enrichment of MZF1 or MT2A in the promoter region of the NFKBIA. Results are expressed as the percentage of input. G, ChIP on GES-1 cells after being treated or not with DATS (40 μmol/L, 12 hours). ChIP on GES-1 and BGC823 cells with, or IgG as control. Precipitated DNA by anti-MZF1 or MT2A antibody was analyzed by qPCR. Results are expressed as the percentage of input.

Figure 5.

MZF1 in cooperation with MT2A to target NFKBIA promoter. A, BGC 823 cells overexpressing MT2A (MT2A), or control (Vector) were subject to IP with MZF1 or MT2A antibodies followed by immunoblotting (IB) using MT2A or MZF1 antibodies, respectively. IP with IgG was used as a negative control. IB of the input with ß-actin antibody served as sample loading control. B, IP with MZF1 or MT2A antibodies was performed on BGC823 cells treated with DATS (40 μmol/L, 24 hours), or DOC (10 μmol/L, 24 hours), or DATS/DOC followed by IB with MT2A or MZF1 antibodies, respectively. IgG and ß-actin antibody used as a negative or loading control as described. C, Confocal IF staining of MT2A and MZF1 in GES-1 cells or BGC823 cells transfected or not with MT2A, or treated with DATS (40 μmol/L, 12 hours). Nuclei were shown by DAPI staining (blue). Shown are results from one of two comparable experiments. Scale bars, 10 μm. D, BGC823 cells were transfected with Flag-MZF1 or GST-MT2A expression plasmids. The purified proteins with Flag-tagged MZF1 were pulled down with purified GST or GST-MT2A. A 1/10 fraction of the sample input and the precipitates were subjected to immunoblotting with anti-MZF1 or anti-MT2A, respectively. E, A schematic illustration of the promoter region of NFKBIA showing the positions of the consensus sequence of 5′-AGTGGGGA-3′ (498/−493 region of the NFKBIA promoter) as well as the ChIP-PCR primer set P4 in the previous experiments. F, ChIP was performed on GES-1 and BGC823 cells with anti-MZF1 or MT2A antibody, or IgG as control. Precipitated ChIP DNA fractions were analyzed by qPCR for the enrichment of MZF1 or MT2A in the promoter region of the NFKBIA. Results are expressed as the percentage of input. G, ChIP on GES-1 cells after being treated or not with DATS (40 μmol/L, 12 hours). ChIP on GES-1 and BGC823 cells with, or IgG as control. Precipitated DNA by anti-MZF1 or MT2A antibody was analyzed by qPCR. Results are expressed as the percentage of input.

Close modal

MT2A induced by DATS in gastric cancer cells is likely to suppress NF-κB activation by directly enhancing NFKBIA transcription (14, 15). Transcription factors binding to NFKBIA promoter searched in the TRANSFAC database revealed multiple potential TFBS (data not shown), among which a consensus sequence of 5′-AGTGGGGA-3′ located in the −498/−493 region of NFKBIA promoter represents a typical binding motif for MZF1 (refs. 17–19; Fig. 5E). The sequences flanking this region were previously identified as a targeting region for MT2A to transcriptionally regulate NFKBIA (15). To verify whether MZF1 directly binds NFKBIA promoter in association with MT2A, ChIP assay was performed in BGC823 cells and GES-1 cells with antibody against MT2A or MZF1. Immunoprecipitated genomic DNA fragments were amplified using ChIP primer set (14, 15), and the antibody against MT2A or MZF1 enriched more DNA fragments of NFKBIA promoter region in GES-1 cells than in BGC823 cells (Fig. 5F). Disruption of MZF1 expression in GES-1 cells by transfection with shMZF1 reduced the enrichment of DNA fragments in the region of NFKBIA promoter either by anti-MT2A or anti-MZF1 antibody (Fig. 5G), confirming our recent studies and further suggesting that MT2A recruits the transcription factor MZF1 to regulate NFKBIA transcription. Therefore, the interaction between MZF1 and MT2A regulates NF-κB pathway, in which MT2A attenuates NF-κB activity through recruitment of MZF1 to target NFKBIA promoter.

Expression of both MT2A and MZF1 decreased progressively through gastric malignant transformation

To explore the clinical relevance of MZF1 association with MT2A in gastric carcinogenesis, we analyzed their expression in a cohort of 293 tissue samples including 36 chronic superficial gastritis (CSG), 85 chronic atrophic gastritis (CAG), 69 intestinal metaplasia (IM), 47 dysplasia (DYS), and 54 gastric cancer, as well as 4 normal mucosal tissue samples as control (Supplementary Table S1). Immunohistologic staining of clinical specimens showed that both the expression of MZF1 and MT2A was progressively reduced with the degree of gastric malignant transformation (Fig. 6A and B), and significantly with gastric cancer occurrence (P < 0.001; Supplementary Table S1). Lower expression of MT2A (MT2ALow) in the cohort samples was correlated with lower expression of MZF1 (MZF1Low; P = 0.005) and the Helicobacter pylori infection at diagnosis (P = 0.02), but not with the other parameters (Supplementary Table S2). The decreased association between MT2A and MZF1 was further confirmed by group cross tabulation analysis of the expression of MT2A and MZF1 in the cohort samples (Supplementary Table S3). These clinical data suggest downregulated expression of MT2A/MZF1 during the process of gastric malignant transformation.

Figure 6.

Co-expression of MT2A and MZF1 in human gastric specimens and their prognostic significance in patients with gastric cancer. A and B, IHC staining was performed to determine the expression of MT2A and MZF1 in a cohort of gastric mucosal biopsy specimens with premalignant and malignant gastric lesions. The results were presented as MT2ALow, MT2AHigh, MZF1Low, or MZF1High, respectively, in accordance with the procedure described. A, Representative immunohistochemical staining of MT2A and MZF1. Magnifications: ×200. B, Analysis of the expression of MT2A and MZF1 associated with premalignant and malignant gastric lesions in the cohort. C, The expression of MZF1 in six paired gastric cancer specimens identified with MT2A expression previously. Top: qPCR; bottom: Western blotting. D and E, IHC staining of the expression of MT2A and MZF1 in a gastric cancer cohort (n = 255). D, Analysis of the association of coexpression of MT2A and MZF1 in the tumor of patients with gastric cancer. The significances among different groups are shown analyzed by χ2 tests. E, Kaplan–Meier analysis of survival of patients with GC in association with the expression of MT2A, or MZF1, and their combination. *, P < 0.05; **, P < 0.01, vs. MT2AhighMZF1high, respectively.

Figure 6.

Co-expression of MT2A and MZF1 in human gastric specimens and their prognostic significance in patients with gastric cancer. A and B, IHC staining was performed to determine the expression of MT2A and MZF1 in a cohort of gastric mucosal biopsy specimens with premalignant and malignant gastric lesions. The results were presented as MT2ALow, MT2AHigh, MZF1Low, or MZF1High, respectively, in accordance with the procedure described. A, Representative immunohistochemical staining of MT2A and MZF1. Magnifications: ×200. B, Analysis of the expression of MT2A and MZF1 associated with premalignant and malignant gastric lesions in the cohort. C, The expression of MZF1 in six paired gastric cancer specimens identified with MT2A expression previously. Top: qPCR; bottom: Western blotting. D and E, IHC staining of the expression of MT2A and MZF1 in a gastric cancer cohort (n = 255). D, Analysis of the association of coexpression of MT2A and MZF1 in the tumor of patients with gastric cancer. The significances among different groups are shown analyzed by χ2 tests. E, Kaplan–Meier analysis of survival of patients with GC in association with the expression of MT2A, or MZF1, and their combination. *, P < 0.05; **, P < 0.01, vs. MT2AhighMZF1high, respectively.

Close modal

Higher levels of MT2A/MZF1 are associated with favorable outcome in patients with gastric cancer

To examine the association of MZF1 with MT2A in gastric cancer, we determined the expression of MZF1 in six-paired gastric cancer tissues, in which decreased expression of MT2A was demonstrated previously (15). The results showed that MZF1 expression was decreased in all gastric cancer tissues compared with the paired normal tissues (Fig. 6C). To analyze the relationship between the coexpression of MT2A/MZF1 and gastric cancer prognosis, immunostaining with anti-MZF1 and anti-MT2A antibodies was performed on tumor samples from a gastric cancer cohort (n = 255). As shown in Supplementary Table S4, 85.10% (217 of 255) patients with gastric cancer were with MZF1Low in tumor tissues, significantly correlated with poor clinicopathological features including TNM stage (P = 0.011), invasion depth (P = 0.002), lymph node metastasis (P = 0.028), and more significantly with the patients expressing MT2ALow in tumors (81.57%, 208 of 255). There was no correlation with other parameters. Fig. 6D showed that the numbers of the patients with MT2ALow MZF1Low, MT2ALow MZF1High, MT2AHigh MZF1Low, and MT2AHigh MZF1High were 202, 6, 15, and 32, respectively. Among 208 MT2ALow patients with gastric cancer, 202 (97.07%) simultaneously expressed low MZF1 (MT2ALow MZF1Low) in contrast to six (2.93%) patients in the high MZF1 expression (MT2ALow MZF1High; P < 0.001), suggesting a close association of MZF1 with MT2A in gastric cancer. Logistic regression analysis of clinicopathologic features and prognosis in patients with gastric cancer revealed that both the MZF1Low and MT2ALow patients had significantly shorter survival time (P < 0.001), and MZF1Low MT2ALow defined patients with the shortest median time among the different groups (Supplementary Table S5). Kaplan–Meier survival analysis of patients showed a significantly shorter overall 5-year survival time in the cases with MZF1Low MT2ALow tumors as compared with patients with MZF1Low or MT2ALow alone (P = 0.001; Fig. 6E). Interestingly, although multivariate cox regression analysis of clinicopathologic features and molecular signatures indicated that MZF1 expression was an independent positive prognostic factor in patients with GC (95% CI, 0.112–0.453; P < 0.001; Supplementary Table S6), the patients with MT2Ahigh MZF1Low showed a comparable overall survival time to the MT2Ahigh MZF1high patients (P = 0.57; Fig. 6E; Supplementary Table S5). Thus, these results reveal that GC patients with MT2ALow MZF1Low had a poor clinical outcome, suggesting that MT2A/MZF1 provides a more reliable diagnosis and prognosis prediction of gastric cancer.

MZF1 is a member of the SCAN-Zinc finger (SCAN-ZF) transcription factor superfamily and plays an important role in the regulation of gene transcription, cell differentiation, and embryogenesis (21, 22, 26). Because initial studies of the role of MZF1 in myeloid differentiation and leukemia (18, 23), it has been demonstrated that the MZF1-dependent transcriptional changes occur in malignant cellular processes (24, 27, 28, 30, 36–43), and aberrant expression of MZF1 correlates with poor prognosis of several types of cancers (29, 32, 34, 44, 45). However, the contribution of MZF1 to tumorigenesis has not been fully elucidated (20). In this study, we found that epigenetic silencing of MZF1 is contributed to gastric carcinogenesis and provided evidence for MZF1 as a transcription factor in MT2A–NF-κB pathway in response to DATS treatment. In line with our previous observation that epigenetic upregulation of MT2A by DATS attenuates NF-κB activation to enhance chemosensitivity of gastric cancer (15), this study revealed a molecular basis for MT2A–NF-κB pathway in DATS-mediated anti-GC effects through MZF1 targeting NFKBIA to regulate gastric carcinogenesis.

Although several studies showed the impact of miRNAs on the capacity of MZF1 to regulate gene expression by targeting its 3′ untranslated regions (34, 39), or DNA methylation involved in genes regulated by MZF1 (41, 46), the epigenetic regulation of MZF1 was not clear. Here we showed that the promotor region of MZF1 was methylated associated with hypoacetylation of histone in gastric cancer cells and primary tumor tissues, and epigenetic silencing of MZF1 expression in gastric cancer cells was reversible by treatment with epigenetic agents, or DATS. Gain- and loss-of-function analysis of MZF1 in gastric cancer cells demonstrated the capacity of MZF1 to reduce gastric cancer cell malignancy, in which we showed that disruption of MZF1 expression rendered growth and migration advantages to an immortalized human gastric mucosal epithelial cell line. Oncogenic cell transformation by ectopic expression of MZF1 was indeed demonstrated using NIH3T3 transformation model (21). The results were confirmed in clinical specimens showing progressive downregulation of MT2A/MZF1 through gastric malignant transformation from CAG, IM, DYS to gastric cancer and association of epigenetic silencing of MZF1 with poor clinicopathological features and shortened overall survival of patients with gastric cancer. These data, in support of our previous findings of DATS-mediated epigenetic upregulation of MT2A (15), demonstrate that gastric tumorigenesis was attributable to epigenetic silencing of MZF1 in association with MT2A. Because epigenetic alterations, rather than genetic variations (35), are progressive and reversible through cancer development and progression, downregulation of MT2A/MZF1 should have significant implications for early detection and targeted treatment of gastric cancer.

It has been demonstrated that constitutive hyperactivation of NF-κB contributes to tumorigenesis and generates chemotherapeutic resistance in gastric cancer (8–11). Blocking NF-κB activation through producing nondegradable form of NFKBIA in cancer cells has shown promising anticancer effects. NFKBIA is thought to be a potential target for small molecules such as antioxidants to suppress NF-κB activation (11, 47, 48). Although antioxidants were suggested as possible NF-κB inhibitors many years ago, the mechanisms for antioxidants to block NF-κB activation remain unclear (11). MT2A is a nonenzymatic antioxidant and detoxicant involved in metal ion homeostasis (6, 12). We previously showed anti-gastric cancer effects of a garlic antioxidant, DATS via MT2A-NF-κB pathway (15). The gain- and loss-of-function assays in this study further demonstrated the regulatory capacity of MT2A in the control of MZF1 expression. Through in vitro pull-down assays, we showed a direct interaction between MT2A and MZF1 in gastric cancer cells at exogenous condition by ectopic expression of MT2A, or endogenously treated with DATS and/or DOC. Moreover, we demonstrated that MT2A recruited MZF1 to form MT2A/MZF1 complex attaching to the −498/−493 region of NFKBIA promoter, and to mediate the antitumor effects of DATS on gastric cancer. Our findings were confirmed by downregulation of MT2A/MZF1 in gastric tumor tissues associated with significant value for predicting the outcome of patients with gastric cancer. We also observed a positive feedback regulation at the transcriptional level between MZF1 and MT2A to enhance DATS-mediated anti-gastric cancer effects. In agreement with the capacity of zinc to inhibit NF-κB activation in cancer cells (7, 13, 49, 50), zinc chelation is likely involved in the anti-gastric cancer activity of MT2A/MZF1–NF-κB pathway mediated by DATS, in which MT2A on one hand regulates zinc-binding proteins by supplying or removing zinc, but on the other hand is transcriptionally inducible by zinc-binding proteins to target its promoter region containing multiple regulatory elements, such as metal responsive element (MRE). As the three-dimensional structural models of both MT2A and MZF1 are available (6), future work to determined three-dimensional structure of MT2A/MZF1 complex by NMR for atomic solution should provide precise visualization.

Studies so far presented a complicated and heterogeneous interaction network for the expression of MZF1 and its biofunctional role in cancer. For instance, Rafan and colleagues recently identified a complex signaling network for ErbB2-driven breast cancer invasion through upregulation of MZF1 activity leading to increased cathepsin B (CTSB) transcription and cancer cell invasiveness, in which MZF1 may also amplify ErbB2 signaling network via protein kinase C alpha (PKCα)-ERK2-mediated recycling of ErbB2 (37). Lee and colleagues further demonstrated that in breast cancer, MZF interacting with Ets-like protein-1 (Elk-1) to form MZF1/Elk-1 complex promoted tumor progression by upregulation of PKCα expression through directly targeting its promoter region (42). MZF1/Elk-1 in cooperation with PKCα promoting tumorigenesis and metastasis were also observed in human hepatocellular carcinoma (HCC) to be associated with poor outcomes of patients (44), even though MZF1 gene amplification or mutation was relatively low in HCC in database (21). However, high frequency of MZF1 alterations profiled in bladder cancers (21) was not supported by insignificant changes of MZF1 expression in various differentiated urinary bladder transitional cell carcinoma (TCC) cell lines (36), or in clinical specimens (44). In contrast, Gaboli and colleagues demonstrated that Mzf1 acts as a growth and tumor suppressor to control cell proliferation and tumorigenesis through targeted disruption of Mzf1 in mouse germ line (26). In human prostate cancer, Chen and colleagues found that MZF1 expression was decreased in tumor cells and tissues associated with downregulation of iron exporter Ferroportin (FPN). A suppressive effect of MZF1 on prostate tumor growth may be based on elevating FPN-conducted iron egress out of tumor cells because loss of FPN resulting in excess tumor iron stimulates tumor progression (39). Our findings clearly demonstrated a tumor suppressive role of MT2A/MZF1 in gastric malignancy. Moreover, our study unveils the important interaction between MT2A and MZF1 via targeting NFKBIA promoter, through which MT2A/MZF1 cooperatively suppresses NF-κB pathway in response to chemotherapeutic agent treatment of gastric cancer with DATS, DOC, as well as epigenetic agents. In agreement with the repressive role of MZF1 in gastric cancer, Lee reported that MZF1 expression was decreased in gastric cancer cell lines with close correlation with SMAD4, a key regulator of TGF-β signaling and suggested a tumor suppressor function of MZF1 in gastric tumorigenesis via targeting SMAD4 (33). Li and colleagues recently provided further evidence showing reduced MZF1 expression in gastric cancer tissues with unfavorable outcome of patients (32). Controversial reports recently showed that MZF1 may exert oncogenic function in gastric cancer by facilitating the expression of matrix metalloproteinase 14 (MMP14), in which miR-337-3p suppresses the transcription of MMP14 via epigenetically suppressing the binding of MZF1 to its promoter (34), despite the previous study showing the inhibitory capability of MZF1 in cervical cancer cells through reducing MMP-2 expression (30). Therefore, our findings support the hypothesis that the bi-functional roles of MZF1 in tumorigenesis may be dependent on cancer types and on association with co-activators or co-repressors.

Overall, our study reveals the molecular basis of DATS-mediated anti-gastric cancer effects via cooperation of MT2A/MZF1 binding to NFKBIA promoter and proposes an important role of MT2A/MZF1–NF-κB pathway in suppressing gastric carcinogenesis and enhancing chemosensitivity (Supplementary Fig. S4; refs. 14, 15). MT2A/MZF1 represents a diagnostic and prognostic biomarker for gastric cancer and MT2A/MZF1–NF-κB pathway represents a therapeutic target.

No potential conflicts of interest were disclosed.

Conception and design: Y. Pan, Z. Chang, J. Sheng, Y. Lu, J. Huang

Development of methodology: Y. Pan, R. Tian, K. Chen, L. Zhang, Y. Lu, J. Huang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Lin, X. Wang, Y. Pan, R. Tian, B. Lin, Y. He, L. Zhang, P. Jin, L. Yang, G. Li, Y. Wu, W. Gong

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Lin, X. Wang, Y. Pan, B. Lin, G. Jiang, Y. He, P. Jin, L. Yang, G. Li, Y. Wu, Z. Chang, J. Sheng, J. Huang

Writing, review, and/or revision of the manuscript: S. Lin, X. Wang, Y. Pan, J.M. Wang, J. Huang

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Pan, J. Hu, J. Sheng, J. Huang

Study supervision: J. Hu, Y. Lu, J. Huang

Other (financial support): G. Jiang, J. Huang

Other (the GST pull-down assay for the direct interaction between GST-MT2A and Flag-MZF1): W. Zhai

This work was supported by grants from the National Science Foundation of China (Grant no. 81872021) to S. Lin, X. Wang, R. Tian, B. Lin, Y. Lu, and J. Huang. S. Lin, X. Wang, K. Chen, J. Huang, and J.M. Wang were funded in part by Federal funds from the National Cancer Institute, National Institutes of Health, under Contract No. HHSN261200800001E and were also supported in part by the Intramural Research Program of the NCI, NIH. G. Jiang was supported by grants from the National Science Foundation of China (Grant no. 81573467), the key project of the transformation of independent achievements in Shangdong (2015ZDJS04003) and the project for laureate of Taishan Scholar (no. ts 201511075). Y. Pan, Y. He, P. Jin, L. Yang, and J. Sheng were supported by the Project of Army Special Care (no. 12BJZ04), the Beijing Municipal Natural Science Foundation (no. 7172213), the China Postdoctoral Science Foundation (no. 2017M613421). Y. Pan, G. Li, Y. Wu, and J. Hu were supported by the Inner Mongolia Autonomous Region Natural Science Foundation Project (no. 2016MS0893) and the Inner Mongolia Province Natural Science Foundation of China (no. 2016MS0890).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Global Burden of Disease Cancer C
,
Fitzmaurice
C
,
Allen
C
,
Barber
RM
,
Barregard
L
,
Bhutta
ZA
, et al
Global, Regional, and National Cancer Incidence, Mortality, Years of Life Lost, Years Lived With Disability, and Disability-Adjusted Life-years for 32 Cancer Groups, 1990 to 2015: a systematic analysis for the global burden of disease study
.
JAMA Oncol
2017
;
3
:
524
48
.
2.
Sitarz
R
,
Skierucha
M
,
Mielko
J
,
Offerhaus
GJA
,
Maciejewski
R
,
Polkowski
WP
. 
Gastric cancer: epidemiology, prevention, classification, and treatment
.
Cancer Manag Res
2018
;
10
:
239
48
.
3.
Liu
X
,
Meltzer
SJ
. 
Gastric cancer in the era of precision medicine
.
Cell Mol Gastroenterol Hepatol
2017
;
3
:
348
58
.
4.
Cancer Genome Atlas Research N
. 
Comprehensive molecular characterization of gastric adenocarcinoma
.
Nature
2014
;
513
:
202
9
.
5.
Lee
KJ
,
Inoue
M
,
Otani
T
,
Iwasaki
M
,
Sasazuki
S
,
Tsugane
S
, et al
Gastric cancer screening and subsequent risk of gastric cancer: a large-scale population-based cohort study, with a 13-year follow-up in Japan
.
Int J Cancer
2006
;
118
:
2315
21
.
6.
Ruttkay-Nedecky
B
,
Nejdl
L
,
Gumulec
J
,
Zitka
O
,
Masarik
M
,
Eckschlager
T
, et al
The role of metallothionein in oxidative stress
.
Int J Mol Sci
2013
;
14
:
6044
66
.
7.
Habel
N
,
Hamidouche
Z
,
Girault
I
,
Patino-Garcia
A
,
Lecanda
F
,
Marie
PJ
, et al
Zinc chelation: a metallothionein 2A's mechanism of action involved in osteosarcoma cell death and chemotherapy resistance
.
Cell Death Dis
2013
;
4
:
e874
.
8.
Kwon
HC
,
Kim
SH
,
Oh
SY
,
Lee
S
,
Lee
JH
,
Jang
JS
, et al
Clinicopathologic significance of expression of nuclear factor-kappaB RelA and its target gene products in gastric cancer patients
.
World J Gastroenterol
2012
;
18
:
4744
50
.
9.
Pires
BRB
,
Silva
R
,
Ferreira
GM
,
Abdelhay
E
. 
NF-kappaB: Two Sides of the Same Coin
.
Genes (Basel)
2018
;
9
:
24
47
.
10.
Karin
M
. 
Nuclear factor-kappaB in cancer development and progression
.
Nature
2006
;
441
:
431
6
.
11.
Gupta
SC
,
Sundaram
C
,
Reuter
S
,
Aggarwal
BB
. 
Inhibiting NF-kappaB activation by small molecules as a therapeutic strategy
.
Biochim Biophys Acta
2010
;
1799
:
775
87
.
12.
Krizkova
S
,
Ryvolova
M
,
Hrabeta
J
,
Adam
V
,
Stiborova
M
,
Eckschlager
T
, et al
Metallothioneins and zinc in cancer diagnosis and therapy
.
Drug Metab Rev
2012
;
44
:
287
301
.
13.
Zalewska
M
,
Trefon
J
,
Milnerowicz
H
. 
The role of metallothionein interactions with other proteins
.
Proteomics
2014
;
14
:
1343
56
.
14.
Pan
Y
,
Huang
J
,
Xing
R
,
Yin
X
,
Cui
J
,
Li
W
, et al
Metallothionein 2A inhibits NF-kappaB pathway activation and predicts clinical outcome segregated with TNM stage in gastric cancer patients following radical resection
.
J Translat Med
2013
;
11
:
173
84
.
15.
Pan
Y
,
Lin
S
,
Xing
R
,
Zhu
M
,
Lin
B
,
Cui
J
, et al
Epigenetic Upregulation of Metallothionein 2A by diallyl trisulfide enhances chemosensitivity of human gastric cancer cells to docetaxel through attenuating NF-kappaB Activation
.
Antioxidants Redox Signal
2016
;
24
:
839
54
.
16.
Babula
P
,
Masarik
M
,
Adam
V
,
Eckschlager
T
,
Stiborova
M
,
Trnkova
L
, et al
Mammalian metallothioneins: properties and functions
.
Metallomics
2012
;
4
:
739
50
.
17.
Morris
JF
,
Hromas
R
,
Rauscher
FJ
 3rd
. 
Characterization of the DNA-binding properties of the myeloid zinc finger protein MZF1: two independent DNA-binding domains recognize two DNA consensus sequences with a common G-rich core
.
Mol Cell Biol
1994
;
14
:
1786
95
.
18.
Perrotti
D
,
Melotti
P
,
Skorski
T
,
Casella
I
,
Peschle
C
,
Calabretta
B
. 
Overexpression of the zinc finger protein MZF1 inhibits hematopoietic development from embryonic stem cells: correlation with negative regulation of CD34 and c-myb promoter activity
.
Mol Cell Biol
1995
;
15
:
6075
87
.
19.
Morris
JF
,
Rauscher
FJ
 3rd
,
Davis
B
,
Klemsz
M
,
Xu
D
,
Tenen
D
, et al
The myeloid zinc finger gene, MZF-1, regulates the CD34 promoter in vitro
.
Blood
1995
;
86
:
3640
7
.
20.
Nygaard
M
,
Terkelsen
T
,
Vidas Olsen
A
,
Sora
V
,
Salamanca Viloria
J
,
Rizza
F
, et al
The mutational landscape of the oncogenic MZF1 SCAN Domain in Cancer
.
Front Mol Biosci
2016
;
3
:
78
.
21.
Eguchi
T
,
Prince
T
,
Wegiel
B
,
Calderwood
SK
. 
Role and regulation of myeloid zinc finger protein 1 in cancer
.
J Cell Biochem
2015
;
116
:
2146
54
.
22.
Hromas
R
,
Davis
B
,
Rauscher
FJ
 3rd
,
Klemsz
M
,
Tenen
D
,
Hoffman
S
, et al
Hematopoietic transcriptional regulation by the myeloid zinc finger gene, MZF-1
.
Curr Top Microbiol Immunol
1996
;
211
:
159
64
.
23.
Hromas
R
,
Boswell
S
,
Shen
RN
,
Burgess
G
,
Davidson
A
,
Cornetta
K
, et al
Forced over-expression of the myeloid zinc finger gene MZF-1 inhibits apoptosis and promotes oncogenesis in interleukin-3-dependent FDCP.1 cells
.
Leukemia
1996
;
10
:
1049
50
.
24.
Hsieh
YH
,
Wu
TT
,
Tsai
JH
,
Huang
CY
,
Hsieh
YS
,
Liu
JY
. 
PKCalpha expression regulated by Elk-1 and MZF-1 in human HCC cells
.
Biochem Biophys Res Commun
2006
;
339
:
217
25
.
25.
Deng
Y
,
Wang
J
,
Wang
G
,
Jin
Y
,
Luo
X
,
Xia
X
, et al
p55PIK transcriptionally activated by MZF1 promotes colorectal cancer cell proliferation
.
Biomed Res Int
2013
;
2013
:
868131
.
26.
Gaboli
M
,
Kotsi
PA
,
Gurrieri
C
,
Cattoretti
G
,
Ronchetti
S
,
Cordon-Cardo
C
, et al
Mzf1 controls cell proliferation and tumorigenesis
.
Genes Dev
2001
;
15
:
1625
30
.
27.
Inoue
M
,
Takahashi
K
,
Niide
O
,
Shibata
M
,
Fukuzawa
M
,
Ra
C
. 
LDOC1, a novel MZF-1-interacting protein, induces apoptosis
.
FEBS Lett
2005
;
579
:
604
8
.
28.
Yan
QW
,
Reed
E
,
Zhong
XS
,
Thornton
K
,
Guo
Y
,
Yu
JJ
. 
MZF1 possesses a repressively regulatory function in ERCC1 expression
.
Biochem Pharmacol
2006
;
71
:
761
71
.
29.
Mudduluru
G
,
Vajkoczy
P
,
Allgayer
H
. 
Myeloid zinc finger 1 induces migration, invasion, and in vivo metastasis through Axl gene expression in solid cancer
.
Mol Cancer Res
2010
;
8
:
159
69
.
30.
Tsai
SJ
,
Hwang
JM
,
Hsieh
SC
,
Ying
TH
,
Hsieh
YH
. 
Overexpression of myeloid zinc finger 1 suppresses matrix metalloproteinase-2 expression and reduces invasiveness of SiHa human cervical cancer cells
.
Biochem Biophys Res Commun
2012
;
425
:
462
7
.
31.
Vishwamitra
D
,
Curry
CV
,
Alkan
S
,
Song
YH
,
Gallick
GE
,
Kaseb
AO
, et al
The transcription factors Ik-1 and MZF1 downregulate IGF-IR expression in NPM-ALK(+) T-cell lymphoma
.
Mol Cancer
2015
;
14
:
53
70
.
32.
Li
GQ
,
He
Q
,
Yang
L
,
Wang
SB
,
Yu
DD
,
He
YQ
, et al
Clinical significance of myeloid zinc finger 1 expression in the progression of gastric tumourigenesis
.
Cell Physiol Biochem
2017
;
44
:
1242
50
.
33.
Lee
JH
,
Kim
SS
,
Lee
HS
,
Hong
S
,
Rajasekaran
N
,
Wang
LH
, et al
Upregulation of SMAD4 by MZF1 inhibits migration of human gastric cancer cells
.
Int J Oncol
2017
;
50
:
272
82
.
34.
Zheng
L
,
Jiao
W
,
Mei
H
,
Song
H
,
Li
D
,
Xiang
X
, et al
miRNA-337-3p inhibits gastric cancer progression through repressing myeloid zinc finger 1-facilitated expression of matrix metalloproteinase 14
.
Oncotarget
2016
;
7
:
40314
28
.
35.
Lin
B
,
Zhou
X
,
Lin
S
,
Wang
X
,
Zhang
M
,
Cao
B
, et al
Epigenetic silencing of PRSS3 provides growth and metastasis advantage for human hepatocellular carcinoma
.
J Mol Med (Berl)
2017
;
95
:
1237
49
.
36.
Jou
YC
,
Chiu
YW
,
Chen
YH
,
Hwang
JM
,
Chao
PY
,
Shiu
JJ
, et al
Expression of protein kinase Calpha and the MZF-1 and elk-1 transcription factors in human bladder transitional cell carcinoma cells
.
Chin J Physiol
2012
;
55
:
75
81
.
37.
Rafn
B
,
Nielsen
CF
,
Andersen
SH
,
Szyniarowski
P
,
Corcelle-Termeau
E
,
Valo
E
, et al
ErbB2-driven breast cancer cell invasion depends on a complex signaling network activating myeloid zinc finger-1-dependent cathepsin B expression
.
Mol Cell
2012
;
45
:
764
76
.
38.
Chen
PM
,
Cheng
YW
,
Wang
YC
,
Wu
TC
,
Chen
CY
,
Lee
H
. 
Up-regulation of FOXM1 by E6 oncoprotein through the MZF1/NKX2-1 axis is required for human papillomavirus-associated tumorigenesis
.
Neoplasia
2014
;
16
:
961
71
.
39.
Chen
Y
,
Zhang
Z
,
Yang
K
,
Du
J
,
Xu
Y
,
Liu
S
. 
Myeloid zinc-finger 1 (MZF-1) suppresses prostate tumor growth through enforcing ferroportin-conducted iron egress
.
Oncogene
2015
;
34
:
3839
47
.
40.
Yue
CH
,
Huang
CY
,
Tsai
JH
,
Hsu
CW
,
Hsieh
YH
,
Lin
H
, et al
MZF-1/Elk-1 complex binds to protein kinase calpha promoter and is involved in hepatocellular carcinoma
.
PLoS One
2015
;
10
:
e0127420
.
41.
Jia
N
,
Wang
J
,
Li
Q
,
Tao
X
,
Chang
K
,
Hua
K
, et al
DNA methylation promotes paired box 2 expression via myeloid zinc finger 1 in endometrial cancer
.
Oncotarget
2016
;
7
:
84785
97
.
42.
Lee
CJ
,
Hsu
LS
,
Yue
CH
,
Lin
H
,
Chiu
YW
,
Lin
YY
, et al
MZF-1/Elk-1 interaction domain as therapeutic target for protein kinase Calpha-based triple-negative breast cancer cells
.
Oncotarget
2016
;
7
:
59845
59
.
43.
Verma
NK
,
Gadi
A
,
Maurizi
G
,
Roy
UB
,
Mansukhani
A
,
Basilico
C
. 
Myeloid Zinc Finger 1 and GA Binding Protein Co-Operate with Sox2 in regulating the expression of yes-associated protein 1 in cancer cells
.
Stem Cells
2017
;
35
:
2340
50
.
44.
Ye
JC
,
Hsu
LS
,
Tsai
JH
,
Yang
HL
,
Hsiao
MW
,
Hwang
JM
, et al
MZF-1/Elk-1/PKCalpha is associated with poor prognosis in patients with hepatocellular carcinoma
.
J Cancer
2017
;
8
:
3028
36
.
45.
Lee
EH
,
Ji
KY
,
Kim
EM
,
Kim
SM
,
Song
HW
,
Choi
HR
, et al
Blockade of Axl signaling ameliorates HPV16E6-mediated tumorigenecity of cervical cancer
.
Sci Rep
2017
;
7
:
5759
.
46.
Lee
YK
,
Park
UH
,
Kim
EJ
,
Hwang
JT
,
Jeong
JC
,
Um
SJ
. 
Tumor antigen PRAME is up-regulated by MZF1 in cooperation with DNA hypomethylation in melanoma cells
.
Cancer Lett
2017
;
403
:
144
51
.
47.
Li
Y
,
Kong
D
,
Ahmad
A
,
Bao
B
,
Sarkar
FH
. 
Antioxidant function of isoflavone and 3,3′-diindolylmethane: are they important for cancer prevention and therapy? Antioxid Redox Signal
2013
;
19
:
139
50
.
48.
Ahmad
A
,
Ginnebaugh
KR
,
Li
Y
,
Padhye
SB
,
Sarkar
FH
. 
Molecular targets of naturopathy in cancer research: bridge to modern medicine
.
Nutrients
2015
;
7
:
321
34
.
49.
Prasad
AS
,
Beck
FW
,
Snell
DC
,
Kucuk
O
. 
Zinc in cancer prevention
.
Nutr Cancer
2009
;
61
:
879
87
.
50.
Kolenko
V
,
Teper
E
,
Kutikov
A
,
Uzzo
R
. 
Zinc and zinc transporters in prostate carcinogenesis
.
Nat Rev Urol
2013
;
10
:
219
26
.