Genomic alterations of tumor suppressorsoften encompass collateral protein-coding genes that create therapeutic vulnerability to further inhibition of their paralogs. Here, we report that malic enzyme 2 (ME2) is frequently hemizygously codeleted with SMAD4 in gastric cancer. Its isoenzyme ME1 was upregulated to replenish the intracellular reducing equivalent NADPH and to maintain redox homeostasis. Knockdown of ME1 significantly depleted NADPH, induced high levels of reactive oxygen species (ROS), and ultimately cell apoptosis under oxidative stress conditions, such as glucose starvation and anoikis, in ME2-underexpressed cells. Moreover, ME1 promoted tumor growth, lung metastasis, and peritoneal dissemination of gastric cancer in vivo. Intratumoral injection of ME1 siRNA significantly suppressed tumor growth in cell lines and patient-derived xenograft–based models. Mechanistically, ME1 was transcriptionally upregulated by ROS in an ETV4-dependent manner. Overexpression of ME1 was associated with shorter overall and disease-free survival in gastric cancer. Altogether, our results shed light on crucial roles of ME1-mediated production of NADPH in gastric cancer growth and metastasis.

Significance: These findings reveal the role of malic enzyme in growth and metastasis.

Graphical Abstract:http://cancerres.aacrjournals.org/content/canres/78/8/1972/F1.large.jpg. Cancer Res; 78(8); 1972–85. ©2018 AACR.

Multifaceted sequencing has revealed an unprecedentedly detailed blueprint for gene amplification or deletion in human genomes as well as in other mammals (1). However, vast targeted therapies often focused on the amplified, overexpressed, or mutant-driving oncoproteins (2, 3), whereas the deleted, underexpressed, or mutant-inactivated tumor suppressors received less attention (4). Recently, strategies such as synthetic lethality have been proposed to exploit genomic loss of suppressor genes, as these events often occur at large regions that may encompass critical fundamental housekeeping genes that are essential for cell growth and survival (4–8). Cancer cells may sometimes tolerate these stresses by rewiring the information flow into functionally redundant paralogs that maintain these essential cellular reactions (6, 8, 9). When several other homologous genes serving overlapping functions were shut down by inhibitors, the cells would experience lethal strike (4, 6, 10). Genetic and pharmacologic studies have evidenced the therapeutic exploit of collateral deletion in the tumor-suppressive loci (4, 10–12). One notable example is that hemizygous deletion of TP53 in colorectal cancer necessarily led to high vulnerability to inhibition of the neighboring gene POLR2A (5).

Malic enzymes are responsible for oxidative decarboxylation of malate to pyruvate, which is the primary substrate supporting the tricarboxylic acid (TCA) cycle and the major source of intracellular reducing equivalents (13, 14). In human cells, malic enzymes are encoded by three homologous genes (15, 16), including ME1, which is located in the cytoplasm and enzymic activities require NADP; malic enzyme 2 (ME2), which is located in the mitochondrion and enzymic activities require NAD; and ME3, which is located in the mitochondrion and enzymic activities require NADP. These enzymes are widely distributed in nature and have highly conserved sequences and similar structural topologies across different species, suggesting that they have important biological functions (17).

For chemical work, there is an equally important role for ATP and NADPH, which powers the redox defense and reductive biosynthesis (18). Tumor cells reprogram their metabolic patterns to satisfy the needs of rapid cell proliferation at the expense of overproduced reactive oxygen species (ROS), which requires plenty of NADPH supplementation (19). By recycling the TCA intermediate malate into the common carbon source pyruvate, malic enzymes may have a regulatory role in satisfying cellular demand for reducing equivalents, energy, and biosynthetic precursors (16). Malic enzymes are essential for NADPH production from both the oxidative pentose phosphate pathway (16) and glutamine metabolism (20) and thus have been evaluated as therapeutic targets. It has been reported that ME1 produces NADPH at levels as high as those produced by G6PD in the pentose phosphate pathway shunt (18). Repression of ME1 or ME2 results in altered metabolism (13, 14), reduced cell growth, migration, and elevated ROS level in nasopharyngeal carcinoma (13) and non–small cell lung cancer (21).

To mine the therapeutic targets with collateral lethality in gastric cancer, we analyzed The Cancer Genome Atlas (TCGA) and Cancer Cell Line Encyclopedia (CCLE) databases and found that the housekeeping gene ME2 was frequently codeleted or counderexpressed with the tumor suppressor gene SMAD4 in gastric cancer. During energy stress such as glucose deprivation, anchorage-independent growth, and solid tumor formation in vivo, ME1 played essential roles in supplying NADPH for elimination of intracellular ROS when its paralog ME2 was suppressed due to coalteration with SMAD4.

Cell culture

GES1, AGS, SGC7901 (originally purchased from ATCC on July 2014) and SNU216, BGC823, HGC27, MGC803, NUGC4, MKN45, MKN74 (originally purchased from the Institute of Basic Medical Sciences of the Chinese Academy of Medical Sciences on June 2014) were cultured in RPMI1640 or DMEM medium (Invitrogen) supplemented with 10% FBS (HyClone) at 37°C with 5% CO2 according to the suppliers' instructions. Glucose-free RPMI1640 medium (GIBCO/Thermo Fisher Scientific, cat. no. 11879020) supplemented with 10% dialyzed FBS (GIBCO/Thermo Fisher Scientific, cat. no. 26400-044) was used for glucose deprivation assays. All cells were tested negative for mycoplasma and authenticated by short tandem repeat DNA fingerprinting at the Medicine Lab of the Forensic Medicine Department of Sun Yat-sen University (Guangzhou, China). All cell lines have not been passaged for more than 6 months in our study after resuscitation.

Bioinformatic analysis

Gene copy number and corresponding gene expression was analyzed using data obtained from CCLE (http://www.broadinstitute.org/ccle) and TCGA (http://www.cbioportal.org/portal/) according to previously described methods (11). We searched potential target gene in the proximity of SMAD4 gene and analyzed its codeletion with SMAD4 in gastric cancer as in a recent report (6).

Protein extraction, immunoblotting, and antibodies

Proteins were extracted with the RIPA lysis buffer (Cell Signaling Technology, cat. no. 9806). Briefly, scrapped cells were collected after centrifugation at 2,000 rpm for 3 minutes. Then, the pelleted cells were lysed in RIPA buffer containing proteinase and phosphatase inhibitors for 15 minutes on ice. The supernatant was transferred to a new tube, and the protein concentrations were measured using BCA Protein Assay Kit (Thermo Fisher Scientific, cat. no. 23225). SDS-PAGE and immunoblotting was performed as described previously (22). The following antibodies were used: ME1 (Abcam, cat. no. ab97445); ME2 (Abcam, cat. no. ab139686); ME3 (Abcam, cat. no. ab172972); vinculin (Abcam, cat. no. ab129002); G6PD (Abcam, cat. no. ab993); PHGDH (Abcam, cat. no. ab211365); Flag (Abcam, cat. no. ab49763); E-cadherin (Cell Signaling Technology, cat. no. 3195); N-cadherin (Cell Signaling Technology, cat. no. 13116); cleaved PARP (Cell Signaling Technology, cat. no. 5625); cleaved caspase-3 (Cell Signaling Technology, cat. no. 9664); ETV4 (Aviva Systems Biology, cat. no. ARP32263_P050); ETV4 (Lifespan Biosciences, cat. no. LS-B1527); and β-actin (Cell Signaling Technology, cat. no. 4970).

RNA extraction and qPCR analysis

Total RNA was isolated from cells or tissues by TRIzol Reagent (Invitrogen, cat. no. 15596018). One microgram of RNA for each sample was reversed to cDNA by a Prime Script RT Master Mix Kit (Takara, cat. no. RR036Q), and 1 μL cDNA was used as a template to perform qPCR with GoTaq qPCR Master Mix (Promega, cat. no. A6002) according to the manufacturer's instructions. Primers used in our study were listed in Supplementary Table S1.

Tissue specimens and clinicopathologic characteristics

The total 207 paraffin-embedded, archived gastric samples used in this study were histopathologic and clinically diagnosed at the Sun Yat-sen University Cancer Center between 2007 and 2009. Written informed consent was obtained from all patients, and no patient received any chemo- or radiotherapy prior to surgery. The use of clinical specimens for research purposes was conducted in accordance with the Declaration of Helsinki and approved by the ethical committee of Sun Yat-sen University Cancer Center. The clinicopathologic characteristics of the samples are summarized in Supplementary Table S2. All patients were followed up regularly after the operation at 3-month intervals. The median follow-up time was 49 months (range, 3–102 months). Fifty freshly collected gastric cancer tissues and matched adjacent nontumoral gastric tissues from the same patient were frozen and stored in liquid nitrogen until required for RNA or protein extraction.

IHC and TUNEL analysis

IHC assays were conducted as reported previously (5). Briefly, the sections were deparaffinized and rehydrated before they were heated at a subboiling temperature in sodium citrate buffer (pH 6.0) for 10 minutes with a microwave oven for antigen retrieval. Samples were then incubated with 3% hydrogen peroxide for 10 minutes to block endogenous peroxidase activity and then with antibody against ME1 (Abcam, ab97445, 1:500), ME2 (Abcam, cat. no. ab139686), Ki-67 (Cell Signaling Technology, 9129, 1:500) or cleaved caspase-3 (Cell Signaling Technology, 9664, 1:1,000) at 37°C for 1 hour. The IHC Kit (Dako, cat. no. K5007) including second antibody and DAB substrate was used to detect protein expression according to the manufacturer's protocol. Counterstaining color was carried out using hematoxylin. Assessments of the staining were scored by two experienced pathologists blinded to the patients’ identity and clinical status. In discrepant cases, a pathologist reviewed the cases and reached the consensus. Expression level was determined according to our previous report (23). The terminal deoxynucleotidyl transferase–mediated nick end labeling (TUNEL) assays were performed with the In Situ Cell Death Detection Kit (Roche, cat. no. 11684795910) according to the manufacturer's instructions.

ROS, cell apoptosis detection, and measurement of NADPH

ROS levels were determined as described previously (24). Briefly, cells cultured in glucose-deprived medium or in matrix detachment conditions were incubated with 10 mmol/L 2′,7′-dichlorodihydrofluorescein diacetate (H2-DCFDA, Thermo Fisher Scientific, cat. no. D399) at 37°C for 30 minutes. Afterward, the cells were collected, washed twice in ice-cold PBS, and resuspended in PBS. Fluorescence was immediately measured using a FACScan Flow Cytometer (Beckman-Coulter). For apoptosis analysis, cells were collected and stained with Annexin V-FITC and PI (4A Biotech Co. cat. no. FXP018) before measurement with flow cytometer. The intracellular levels of NADPH, total NADP, and GSH were measured with the NADP/NADPH-Glo Kit (Promega, cat. no. G9081) or GSH/GSSG-Glo Kit (Promega, cat. no. V6612) according to the manufacturer's instructions.

Vectors, siRNAs, cell transfection, and lentivirus production

ME1 vectors including wild-type (cat. no. EX-T8139-Lv121) and silent mutation (cat. no. CS-T8139-Lv121-1) were purchased from GeneCopoeia, Inc. The siRNAs targeting ME1 (targeting sequences: GGGCATATTGCTTCAGTTC, GAGAGACAGCAATTGAACA) and ME2 (targeting sequence: CCCAGTATGGACACATCTTTA) were synthesized by RiboBio. The siRNAs targeting G6PD and PHGDH were purchased from RiboBio. All the transfection experiments were conducted with Lipofectamine 3000 (Thermo Fisher Scientific, cat. no. L3000015) as recommended.

SGC7901, MGC803, and HGC27 cells were transfected with lentiviruses containing ME1 shRNA (GenePharma, cat. no. 2016 13971), and stable cell lines were obtained after treatment with 3 μg/mL puromycin for 3 to 5 days. Knockdown lentivirus targeting ME2 was constructed as reported previously (16) and transfected into HGC27 cells.

Anoikis and soft agar colony formation assay

Anoikis was induced by plating cells (2 × 106) on ultralow attachment 6-well plate (Sigma, cat. no. CLS3471-24EA). For soft agar colony formation assay, cell suspension was mixed with 0.7% soft agar in 2×DMEM containing 20% FBS in equal volume and layered in triplicate onto 1.4% solidified agar in 2×DMEM containing 20% FBS. After 10- to 14-day culture, colonies were counted under microscopy and photographed.

Reporter assay

The dual reporter construct expressing Gaussia luciferase under the human ME1 promoter, and secreted alkaline phosphatase (SEAP) under the CMV promoter (used for transfection normalization) was from GeneCopoeia, Inc (cat. no. HPRM23418-PG04). The indicated cells were plated 18 hours before transfection in 24-well plates and transiently transfected with 500 ng of the reporter plasmid using Lipofectamine 3000. Plasmids containing ETV4 open reading frame or siRNAs targeting ETV4 were transfected 24 hours later using Lipofectamine 3000. The luciferase activity was determined according to the manufacturer's instructions (GeneCopoeia, Inc., cat. no. LF032) and normalized to that of the SEAP activity.

Chromatin immunoprecipitation assay

The chromatin immunoprecipitation (ChIP) assay was performed with an EZ-Chip Kit (Millipore, cat. no. 17-371) following the manufacturer's instruction as described previously (16). 293T cells were grown to 80% confluence, and crosslinking was performed with 1% formaldehyde for 10 minutes. The cell lysates were sonicated to shear DNA to sizes of 300 to 1,000 bp. Equal aliquots of chromatin supernatants were incubated with anti-ETV4 or anti-IgG antibody (Millipore) overnight at 4°C with rotation. After reverse cross-link of protein/DNA complexes to free DNA, RT-PCR was carried out using the specific primer (forward: 5′-ACACCTGTCAGTTTCTACAGA-3′, reverse: 5′-CATTATTCAGAGAGAGCAGTGG-3′) detecting the ETV4-binding site on ME1 promoter region.

In vivo tumorigenesis and metastasis assays

All female BALB/c nude mice (4–5 weeks old) used in our study were purchased from the Beijing Vital River Laboratory Animal Technology Co., Ltd. and housed in specific pathogen-free units. Subcutaneous mice model was performed as reported previously (22, 23).

For lung metastasis model, 5 × 106 cells resuspended in 100 μL of sterile PBS were injected into the tail veins of nude mice. Lung colonization was monitored at the indicated time point after intraperitoneal injection of d-luciferin (Goldbio, cat. no. LUCK-1) with a Xenogen IVIS 100 bioluminescent imaging system. Sixty days later, mice were sacrificed with cervical dislocation, and the lungs were dissected out and paraffin embedded to histopathologically examine the metastatic locus.

Peritoneal dissemination ability of gastric cancer cells was evaluated through intraperitoneal injection. In brief, SGC7901 (NC, sh#1, sh#2) cells (5 × 106) in 0.5 mL of PBS were injected into the peritoneal cavity of BALB/c nude mice. Mice were carefully monitored until they were killed at 60 days after injection. Colon metastasis was examined and recorded.

Patient-derived xenograft models and in vivo siRNA treatment

The patient-derived xenograft (PDX)–bearing male nude mice model was raised and passaged as described previously (23, 25). In brief, patient-derived tumor materials were collected in culture medium and transferred to the animal houses on wet ice within 1 hour after resection. Upon arrival, necrotic and supporting tissues were carefully removed using sterilized surgical blades. The tumor gross was cut into different fragments for several purposes, flash frozen, paraffin embedding for histopathologic analysis. One third was flash frozen for protein extraction or stored at −80°C for genomic profiling genomic typing and one third was fixed in 10% neutral-buffered formalin for histopathologic examination. The rest was implanted subcutaneously into the flank region of female nude mice, and the incision was closed with surgical suture. Successfully engrafted tumor models were then passaged and banked in liquid nitrogen after three passages in mice.

For siRNA treatment analysis, PDX-bearing mice were prepared after subcutaneous incubation of gastric cancer tumor mass. Cholesterol-modified ME1 siRNA or control siRNA (RiboBio, 5 nmol/kg) dissolved in diluted water were intratumorally injected every 3 days for 18 days. All animal experiments were carried out in accordance with the NIH Guide for the Care and Use of Laboratory Animals with the approval from the Institutional Animal Care and Use Committee of Sun Yat-Sen University.

Statistical analysis

All in vitro experiments were repeated three times or more, and data are presented as mean ± SD unless otherwise indicated. The Student t test assumed two-tailed distributions to calculate statistical significance between groups. Survival curves were generated using the Kaplan–Meier method and compared using the log-rank tests. The independent prognostic factors were identified by the Cox proportional hazards regression model. ROC curve was generated with Medcalc software. Differences were analyzed by GraphPad Prism 5 and P values less than 0.05 were considered to reach statistical significance.

Downregulation of ME2 in gastric cancer due to genomic alterations

Genomic alteration of SMAD4 occurs frequently in human cancers (Supplementary Fig. S1A). We screened nearby protein-coding genes located with SMAD4 and found that ME2 is positioned approximately 100 kb upstream to SMAD4 locus (Fig. 1A). Consistently, frequent genomic deletion of ME2 was observed in human cancers (Supplementary Fig. S1B). Although homozygous deletion of SMAD4 and ME2 was found only in 4.76% of TCGA stomach tissues, 162 of 414 (39.23%) cases bear hemizygous deletion of SMAD4 and ME2 (Fig. 1A and B). Moreover, copy number and transcriptomic analyses in TCGA gastric cancer database showed positive correlations between SMAD4 and ME2 at both the DNA and mRNA level (Fig. 1C and D). These positive correlations were also validated in pan-cancer cell lines from the CCLE database (Fig. 1E and F). To address whether genomic loss could lead to alteration of gene expression, we compared copy numbers of SMAD4 and ME2 with corresponding mRNA level. Analysis of CCLE databases revealed that expression of SMAD4 and ME2 was tightly correlated with their gene copy number (Fig. 1G and H). In a panel of gastric cancer cell lines, expression of ME2 mirrored that of SMAD4 at mRNA and protein level (Fig. 1I and J). Importantly, ME2 and SMAD4 were downregulated in the majority of gastric cancer cells, except for that in HGC27 cells, compared with that of GES1 cells (Fig. 1I and J). Our data indicated that ME2 was suppressed in gastric cancer cells, probably due to hemizygous codeletion with SMAD4.

Figure 1.

ME2 is suppressed in gastric cancer due to codeletion with SMAD4. A, Ideogram of chromosome 18 showing close proximity (<200 kb) of ME2 to SMAD4. Copy number alteration (CNA) of SMAD4 in gastric cancer. B, Frequencies of ME2 and SMAD4 copy number alteration in TCGA gastric cancer samples. C, Correlation of SMAD4 and ME2 copy number alteration in TCGA gastric cancer samples. D, Correlation of SMAD4 and ME2 mRNA level in TCGA gastric cancer samples. E, Correlation of SMAD4 and ME2 copy number alteration in CCLE cancer cell lines. F, Correlation of SMAD4 and ME2 mRNA level in CCLE cancer cell lines. G, Correlation of SMAD4 CNA and mRNA level in CCLE cancer cell lines as well as in gastric cancer cell lines. H, Correlation of ME2 copy number alteration and mRNA level in CCLE cancer cell lines as well as in gastric cancer cell lines. I, qPCR assays of ME2 and SMAD4 in a panel of gastric cancer cells and GES1 epithelial cells. J, Immunoblots of ME2 and SMAD4 in a panel of gastric cancer cells and GES1 epithelial cells. Pearson correlation coefficient (r) and P values are displayed in B–H.

Figure 1.

ME2 is suppressed in gastric cancer due to codeletion with SMAD4. A, Ideogram of chromosome 18 showing close proximity (<200 kb) of ME2 to SMAD4. Copy number alteration (CNA) of SMAD4 in gastric cancer. B, Frequencies of ME2 and SMAD4 copy number alteration in TCGA gastric cancer samples. C, Correlation of SMAD4 and ME2 copy number alteration in TCGA gastric cancer samples. D, Correlation of SMAD4 and ME2 mRNA level in TCGA gastric cancer samples. E, Correlation of SMAD4 and ME2 copy number alteration in CCLE cancer cell lines. F, Correlation of SMAD4 and ME2 mRNA level in CCLE cancer cell lines. G, Correlation of SMAD4 CNA and mRNA level in CCLE cancer cell lines as well as in gastric cancer cell lines. H, Correlation of ME2 copy number alteration and mRNA level in CCLE cancer cell lines as well as in gastric cancer cell lines. I, qPCR assays of ME2 and SMAD4 in a panel of gastric cancer cells and GES1 epithelial cells. J, Immunoblots of ME2 and SMAD4 in a panel of gastric cancer cells and GES1 epithelial cells. Pearson correlation coefficient (r) and P values are displayed in B–H.

Close modal

ME1 is essential for survival of gastric cancer cells under glucose deprivation

Malic enzymes replenish the TCA cycle and produce the major antioxidant NADPH via oxidation of malate to pyruvate (16, 19) and are redundant metabolic pathway due to three isoenzymes (Fig. 2A). Because ME2 was suppressed in gastric cancer, we therefore focused on its paralogs including ME1 and ME3. Western blot assays showed remarkable increase in protein levels of ME1 in the majority of gastric cancer cell lines compared with that in GES1 cells, whereas ME3 was only detected in MKN74 and BGC823 cells (Fig. 2B). To explore collateral lethality of ME2 genomic alteration in gastric cancer, lentiviruses containing short hairpin RNAs targeting ME1 were introduced into ME2-downregulated SGC7901 and MGC803 cells (Supplementary Fig. S2A). Surprisingly, intracellular NADPH as well as GSH levels were not affected after ME1 knockdown (Fig. 2C; Supplementary Fig. S2B). We hypothesized that the resting ME2 was sufficient for NADPH production in this condition. However, cancer cells often experience nutrition stress due to insufficient vascularization (26, 27). To mimic in vivo situation, we cultured cells in glucose deprivation medium and found that NADPH level in knockdown cells was significantly decreased compared with that in control cells (Fig. 2D). Glucose deprivation induced significantly elevated apoptotic percentage of SGC7901 and MGC803 cells after knockdown of ME1 (Fig. 2E), which could be restored by reintroduction of ME1 vector with silent mutation (Supplementary Fig. S2C and S2D) or pretreatment with NAC (Fig. 2F; Supplementary Fig. S2E). However, HGC27 cells with elevated ME2 expression (Fig. 1J) were resistant to glucose deprivation, which could be completely abrogated by simultaneous silence of ME1 and ME2 (Fig. 2G; Supplementary Fig. S2F). Intracellular ROS level was remarkably elevated after knockdown of ME1 in SGC7901 and MGC803 cells (Fig. 2H), whereas increase of ROS (Fig. 2I) and decrease of NADPH level (Fig. 2J) in HGC27 cells was only observed after coinhibition of ME1 and ME2. These data showed that during glucose deprivation, ME1 enables survival of gastric cancer cells with underexpressed ME2.

Figure 2.

ME1 is essential for cell survival during glucose deprivation. A, Overview of malic enzyme reaction. B, Immunoblots of ME1 and ME3 in a panel of gastric cancer cells and GES1 epithelial cells. Vinculin was used as a loading control. C, Measurement of NADPH/NADP+ in the indicated cells (control, sh#NC; ME1 knockdown, sh#1 or sh#2; knockdown-resistant ME1 vector, sh#1+R) cultured in normal medium. D, Measurement of NADPH/NADP+ in the indicated cells cultured in glucose deprivation medium. E, Brightfield images of SGC7901 and MGC803 cells cultured in glucose deprivation medium after knockdown of ME1. F, Cell apoptosis of indicated cells cultured in glucose deprivation medium. Representative images and quantification data are shown. G, Brightfield images and apoptotic percentage of HGC27 cells after knockdown of ME1 and ME2 cultured in glucose deprivation medium. H, Measurement of ROS level in SGC7901 and MGC803 cells cultured in glucose deprivation medium after knockdown of ME1. Measurement of ROS (I) and NADPH/NADP+ (J) in HGC27 cells after knockdown of ME1 and ME2 cultured in glucose deprivation medium. All error bars represent the SD of at least three replicates from two independent experiments. P values were determined by two-tailed t test.

Figure 2.

ME1 is essential for cell survival during glucose deprivation. A, Overview of malic enzyme reaction. B, Immunoblots of ME1 and ME3 in a panel of gastric cancer cells and GES1 epithelial cells. Vinculin was used as a loading control. C, Measurement of NADPH/NADP+ in the indicated cells (control, sh#NC; ME1 knockdown, sh#1 or sh#2; knockdown-resistant ME1 vector, sh#1+R) cultured in normal medium. D, Measurement of NADPH/NADP+ in the indicated cells cultured in glucose deprivation medium. E, Brightfield images of SGC7901 and MGC803 cells cultured in glucose deprivation medium after knockdown of ME1. F, Cell apoptosis of indicated cells cultured in glucose deprivation medium. Representative images and quantification data are shown. G, Brightfield images and apoptotic percentage of HGC27 cells after knockdown of ME1 and ME2 cultured in glucose deprivation medium. H, Measurement of ROS level in SGC7901 and MGC803 cells cultured in glucose deprivation medium after knockdown of ME1. Measurement of ROS (I) and NADPH/NADP+ (J) in HGC27 cells after knockdown of ME1 and ME2 cultured in glucose deprivation medium. All error bars represent the SD of at least three replicates from two independent experiments. P values were determined by two-tailed t test.

Close modal

ME1 mediates anoikis resistance of gastric cancer

Like glucose starvation, matrix detachment elicits energy stress (9) evidenced by elevated H2O2 levels (Fig. 3A), which were correlated to the extent of NADPH depletion (Fig. 3B). We therefore tested whether ME1 was required for redox regulation during anchorage-independent growth, a hallmark of cancer metastasis (28). Matrix detachment significantly increased intracellular H2O2 and decreased the NADPH level in ME1 knockdown SGC7901 and MGC803 cells in comparison with control cells (Fig. 3C and D). Knockdown of ME1 in SGC7901 and MGC803 cells significantly suppressed colonies in soft agar (Fig. 3E). Apoptotic assays further confirmed ME1 mediated anoikis resistance (Fig. 3F). Enforced expression with silent mutated ME1 or pretreatment with NAC significantly attenuated anoikis of ME1 knockdown cells (Fig. 3F). During suspension culture, HGC27 cells with ME1 knockdown showed resistance to anoikis, and significantly elevated apoptosis was observed after concomitant silence of ME2 (Fig. 3G–I). However, neither migration ability nor epithelial–mesenchymal transition markers of gastric cancer cells showed differences following ME1 knockdown (Supplementary Fig. S3A and S3B). As serine metabolism and pentose phosphate shunt also provided substantial NADPH in cancer cells, we next examined the possible influence of these pathways on gastric cancer cell survival under glucose-deprived or anchorage-independent conditions. Knockdown of PHGDH or G6PD resulted in elevated apoptosis percentage in SGC7901 and MGC803 cells under glucose-deprived or anchorage-independent conditions (Supplementary Fig. S3C and S3D). However, changes in apoptosis were not as obvious as that in ME1 knockdown cells (Supplementary Fig. S3D). These data clearly supported the notion that ME1 could protect gastric cancer cells from anchorage-independent growth.

Figure 3.

ME1 mediates anoikis resistance in gastric cancer. Intracellular ROS (A) or NADPH/NADP+ (B) level of indicated cells cultured in attached or detached conditions. C, Intracellular ROS level in SGC7901 and MGC803 cells cultured in detached conditions was measured after knockdown of ME1. D, NADPH/NADP+ level in the indicated cells cultured in detached conditions was measured. E, Soft agar colony formation assays in SGC7901 and MGC803 cells after knockdown of ME1. F, Representative histograms depicting apoptosis and apoptotic rate of indicated cells after 48 hours of suspension, as determined by flow cytometry. G, Apoptotic rate of HGC27 cells after 72 hours of suspension. H, Immunoblots of cleaved PARP (c-PARP) in HGC27 cells after 72 hours of suspension. I, Intracellular ROS level in HGC27 cells cultured in detached conditions. All error bars represent the SD of at least three replicates from two independent experiments. P values were determined by two-tailed t test.

Figure 3.

ME1 mediates anoikis resistance in gastric cancer. Intracellular ROS (A) or NADPH/NADP+ (B) level of indicated cells cultured in attached or detached conditions. C, Intracellular ROS level in SGC7901 and MGC803 cells cultured in detached conditions was measured after knockdown of ME1. D, NADPH/NADP+ level in the indicated cells cultured in detached conditions was measured. E, Soft agar colony formation assays in SGC7901 and MGC803 cells after knockdown of ME1. F, Representative histograms depicting apoptosis and apoptotic rate of indicated cells after 48 hours of suspension, as determined by flow cytometry. G, Apoptotic rate of HGC27 cells after 72 hours of suspension. H, Immunoblots of cleaved PARP (c-PARP) in HGC27 cells after 72 hours of suspension. I, Intracellular ROS level in HGC27 cells cultured in detached conditions. All error bars represent the SD of at least three replicates from two independent experiments. P values were determined by two-tailed t test.

Close modal

ME1 is required for tumor growth, lung metastasis, and peritoneal dissemination of gastric cancer in vivo

Knockdown of ME1 in SGC7901 cells significantly suppressed tumor growth in vivo as evidenced by slowed growth curve and reduced xenograft weight (Fig. 4A–C). Moreover, immunostaining assays of Ki-67, TUNEL, and cleaved caspase-3 indicated that tumors formed by negative control cells showed characteristic of rapid proliferation and less apoptosis than that formed by ME1 knockdown cells (Fig. 4D). However, knockdown of ME1 or ME2 alone in HGC27 cells showed no effects on tumor growth in the mice, whereas simultaneous ablation of ME1 and ME2 significantly suppressed tumor growth in the subcutaneous mice model (Supplementary Fig. S4A and S4D), which was consistent with the in vitro results. To analyze anoikis in vivo, gastric cancer cells were injected into the tail vein, and fluorescence imaging was used to monitor lung metastasis. The majority of gastric cancer cells diminished 48 hours after injection (Fig. 4E). However, SGC7901/sh#NC cells formed remarkably large and excessive lung metastatic diseases than SGC7901/sh#1 or SGC7901/sh#2 cells (Fig. 4E). Hematoxylin and eosin staining of dissected lungs showed significantly more metastasis nodules in the control group compared with that in the knockdown groups (Fig. 4F). IHC analysis of the lung metastases confirmed effective knockdown of ME1 with no effects on ME2 (Supplementary Fig. S5A). Moreover, knockdown of ME1 significantly suppressed metastasis in the intestinal wall after peritoneal injection of gastric cancer cells (Fig. 4G and H). Altogether, our results indicated that ME1 is critical for gastric cancer growth and metastasis.

Figure 4.

Knockdown of ME1 inhibits growth and metastasis of gastric cancer in vivo. A, Tumor volume progression of xenografted subcutaneous SGC7901 (sh#NC, sh#1, sh#2) cells (n = 6). Tumor growth curves were measured after injection, and tumor diameters were measured every 3 days. The values were given as mean ± SD. B, Photograph of dissected xenografts. C, Weight of dissected xenografts was recorded. D, Representative immunostaining of Ki-67, TUNEL, and cleaved caspase-3 (c-Cas3) in xenografted tumors. Scale bar, 50 μm. E, Representative luciferase imaging of lung metastatic cells in nude mice after knockdown of ME1. F, Representative results of hematoxylin and eosin staining (left) of metastatic lung nodules from mice injected with ME1 knockdown and control SGC7901 cells via the tail vein. Metastatic nodules under naked eyes or microscope were counted and recorded (right). G, SGC7901 cells were injected intraperitoneally and metastases in the colonic wall was recorded 60 days later. H, Dissected colons were photographed and metastatic nodules are indicated (arrows). Hematoxylin and eosin staining of colon was performed and metastatic numbers were recorded. All error bars represent the SD of results from 6 mice. P values were determined by two-tailed t test.

Figure 4.

Knockdown of ME1 inhibits growth and metastasis of gastric cancer in vivo. A, Tumor volume progression of xenografted subcutaneous SGC7901 (sh#NC, sh#1, sh#2) cells (n = 6). Tumor growth curves were measured after injection, and tumor diameters were measured every 3 days. The values were given as mean ± SD. B, Photograph of dissected xenografts. C, Weight of dissected xenografts was recorded. D, Representative immunostaining of Ki-67, TUNEL, and cleaved caspase-3 (c-Cas3) in xenografted tumors. Scale bar, 50 μm. E, Representative luciferase imaging of lung metastatic cells in nude mice after knockdown of ME1. F, Representative results of hematoxylin and eosin staining (left) of metastatic lung nodules from mice injected with ME1 knockdown and control SGC7901 cells via the tail vein. Metastatic nodules under naked eyes or microscope were counted and recorded (right). G, SGC7901 cells were injected intraperitoneally and metastases in the colonic wall was recorded 60 days later. H, Dissected colons were photographed and metastatic nodules are indicated (arrows). Hematoxylin and eosin staining of colon was performed and metastatic numbers were recorded. All error bars represent the SD of results from 6 mice. P values were determined by two-tailed t test.

Close modal

Intratumoral knockdown of ME1 suppresses growth of gastric cancer in vivo

To further explore whether ME1 could be used as a therapeutic target in gastric cancer, we assessed the antitumor activity of ME1 targeting siRNA in mice bearing SGC7901 cells xenografted tumors and three PDXs (PDX#1–3). These PDX models were characterized with decreased ME2 and increased ME1 expression in the tumor tissues compared with corresponding normal tissues (Supplementary Fig. S5B). When the tumor volume reached approximately 50 mm3, siRNAs targeting ME1 were injected intratumorally once every other day. The growth of tumors treated with siRNAs was significantly suppressed in SGC7901 cell–based xenografts and in three PDXs compared with control group (Fig. 5A–F). The tumors developed from the ME1 siRNAs treatment group displayed lower Ki-67 and ME1 staining, elevated TUNEL signal, and cleaved caspase-3 expression than that in control group (Fig. 5G). As PDX models closely resemble the biological characteristics and genomic landscape of human cancers at the population level (25), our study provides clear preclinical clues for developing ME1 inhibitors in gastric cancer.

Figure 5.

Intratumoral silencing of ME1 suppressed gastric cancer growth in vivo. A, Effect of intratumoral ME1 knockdown on tumor growth in mice after injection of SGC7901 cells. B, Dissected xenografts after intratumoral silence of ME1 were photographed. C, Tumor weights of xenografts were recorded. D–F, Effects of intratumoral ME1 knockdown on three PDX models. G, Hematoxylin and eosin (H&E) and immunostaining of Ki-67, ME1, TUNEL, and c-Cas3 in cell line or PDX-based xenografts. Scale bar, 50 μm. All error bars represent the SD of results from 5 mice. P values were determined by two-tailed t test.

Figure 5.

Intratumoral silencing of ME1 suppressed gastric cancer growth in vivo. A, Effect of intratumoral ME1 knockdown on tumor growth in mice after injection of SGC7901 cells. B, Dissected xenografts after intratumoral silence of ME1 were photographed. C, Tumor weights of xenografts were recorded. D–F, Effects of intratumoral ME1 knockdown on three PDX models. G, Hematoxylin and eosin (H&E) and immunostaining of Ki-67, ME1, TUNEL, and c-Cas3 in cell line or PDX-based xenografts. Scale bar, 50 μm. All error bars represent the SD of results from 5 mice. P values were determined by two-tailed t test.

Close modal

ETV4 upregulates ME1 expression in gastric cancer

Tumor cells often experience oxidative stress, which could facilitate tumor growth by causing genomic instability and reprogramming cancer cell metabolism (29). Glucose restriction and anchorage-independent growth induced elevated H2O2 levels (Fig. 6A). We therefore asked whether glutaminolysis pathway mediated by ME1 was activated by ROS overproduction. As shown in Fig. 6B and C, expression of ME1 was upregulated under glucose deprivation medium or in matrix-detached conditions at both the mRNA and protein level, and this effect was reversed by the antioxidant NAC, indicating a transcriptional regulation manner of ME1 by ROS. To this end, we searched transcription factors with potential binding capacity with ME1 in JASPAR database and found ETV4 response elements in the ME1 promoter (Fig. 6D). Interestingly, ETV4 mRNA and protein levels mirrored that of ME1 during ROS stress (Fig. 6C; Supplementary Fig. S6A). Overexpression of ETV4 in SGC7901 and MGC803 cells increased ME1 protein levels (Fig. 6E), whereas knockdown of ETV4 in SGC7901 and MGC803 cells by siRNAs reduced both mRNA and protein levels of ME1 (Fig. 6F; Supplementary Fig. S6B). Upregulation of ME1 induced by ROS was blocked by ETV4 depletion in SGC7901 and MGC803 cells (Fig. 6G). Importantly, as in the case with ME1, knockdown of ETV4 sensitized SGC7901 and MGC803 cells to glucose deprivation and detached conditions (Supplementary Fig. S6C). These results suggest that ME1 was transcriptionally upregulated by ROS/ETV4 during energy stress conditions in gastric cancer cells.

Figure 6.

ETV4 upregulates ME1 expression in gastric cancer. A, Intracellular ROS level in SGC7901 and MGC803 cells cultured in glucose deprivation medium or in detached conditions. B, The ME1 mRNA level in SGC7901 and MGC803 cells cultured in glucose deprivation medium or in detached conditions was measured. C, Immunoblots of ETV4 and ME1 in SGC7901 and MGC803 cells cultured in glucose deprivation medium or in detached conditions. D, ETV4 DNA-binding sites are presented in the human ME1 promoter region. E, Immunoblots of ETV4 and ME1 in SGC7901 and MGC803 cells after enforced expression of ETV4. F, Immunoblots of ETV4 and ME1 in SGC7901 and MGC803 cells after siRNA-mediated knockdown of ETV4. G, Immunoblots of ETV4 and ME1 in SGC7901 and MGC803 cells cultured in glucose deprivation medium after siRNA-mediated knockdown of ETV4. H, Dual-luciferase reporter assays in SGC7901 and MGC803 cells cultured in glucose deprivation medium or in detached conditions. I, Dual-luciferase reporter assays in SGC7901 and MGC803 cells after enforced expression of ETV4. J, Dual-luciferase reporter assays in SGC7901 and MGC803 cells after knockdown of ETV4. K, ChIP-PCR in 293T cells demonstrating ME1 promoter occupancy by ETV4. Data are representative of two independent experiments. All error bars represent the SD of at least three replicates from two independent experiments. P values were determined by two-tailed t test.

Figure 6.

ETV4 upregulates ME1 expression in gastric cancer. A, Intracellular ROS level in SGC7901 and MGC803 cells cultured in glucose deprivation medium or in detached conditions. B, The ME1 mRNA level in SGC7901 and MGC803 cells cultured in glucose deprivation medium or in detached conditions was measured. C, Immunoblots of ETV4 and ME1 in SGC7901 and MGC803 cells cultured in glucose deprivation medium or in detached conditions. D, ETV4 DNA-binding sites are presented in the human ME1 promoter region. E, Immunoblots of ETV4 and ME1 in SGC7901 and MGC803 cells after enforced expression of ETV4. F, Immunoblots of ETV4 and ME1 in SGC7901 and MGC803 cells after siRNA-mediated knockdown of ETV4. G, Immunoblots of ETV4 and ME1 in SGC7901 and MGC803 cells cultured in glucose deprivation medium after siRNA-mediated knockdown of ETV4. H, Dual-luciferase reporter assays in SGC7901 and MGC803 cells cultured in glucose deprivation medium or in detached conditions. I, Dual-luciferase reporter assays in SGC7901 and MGC803 cells after enforced expression of ETV4. J, Dual-luciferase reporter assays in SGC7901 and MGC803 cells after knockdown of ETV4. K, ChIP-PCR in 293T cells demonstrating ME1 promoter occupancy by ETV4. Data are representative of two independent experiments. All error bars represent the SD of at least three replicates from two independent experiments. P values were determined by two-tailed t test.

Close modal

To determine whether ROS/ETV4 axis upregulated ME1 expression transcriptionally, we cloned 1.2 kb of genomic DNA upstream of the transcription start site of the ME1 gene into a luciferase reporter plasmid. Redox stress significantly increased the luciferase activity of ME1 promoter in SGC7901 and MGC803 cells, which could be abrogated by NAC pretreatment (Fig. 6H). As shown in Fig. 6I, enforced ETV4 expression induced elevated reporter activity in SGC7901 and MGC803 cells. Knockdown of ETV4 in SGC7901 and MGC803 cells reduced transcriptional activity of ME1 reporter (Fig. 6J). ChIP-PCR analysis further showed direct binding of ETV4 with ME1 promotor (Fig. 6K).

Combination of ETV4 and ME1 predicts poor prognosis in gastric cancer

To investigate the biological role of ME1 in human gastric cancer progression, IHC was performed to examine the protein expression level of ME1 in 207 cases of paraffin-embedded gastric tissues (Supplementary Fig. S7A). ME1 expression was significantly increased in distant organ metastasis (M) and lymph node metastasis (Ln) tissues compared with adjacent normal tissues (ANT) and paired primary tumor tissues (T; Fig. 7A and B), supporting potential link between ME1 expression and gastric cancer metastasis. Consistently, ME1 expression was markedly overexpressed in the gastric cancer tissues compared with paired normal gastric tissues at protein (Fig. 7A and B) and mRNA level (Fig. 7C).

Figure 7.

ETV4/ME1 axis is overactivated in gastric cancer. A, Representative staining showed upregulated expression of ME1 protein in gastric tumor tissues compared with corresponding nontumorous tissues. Scale bar, 100 μm. B, Immunoscoring of ME1 in adjacent normal tissues (ANT), gastric cancer tissues (T), distant organ metastasis (M), and lymph node metastasis (Ln). C, The ME1 mRNA expression in 50 paired gastric tumor samples and corresponding nontumorous tissues. D, Kaplan–Meier analysis of overall survival or disease-free survival curves for gastric cancer patients with low versus high expression of ME1. E, Immunoblots of ETV4 and ME1 in eight freshly collected gastric cancer samples. F, The relative protein expression levels in eight freshly collected gastric cancer samples were quantified by comparing the gray level of each band using ImageJ Software. G, Correlations between mRNA level of ETV4 and ME1 in 50 freshly collected gastric cancer samples. H, Correlations of ETV4 and ME1 protein expression in gastric cancer tissues based on immunoscoring. I, Gastric cancer patients were divided into three groups according to immunoscoring of ETV4 and ME1, and overall survival curve was generated with Kaplan–Meier methods. J, Receiver operating characteristic (ROC) curve analysis of ME1 [area under a curve (AUC) = 0.587 (95% CI, 0.517–0.655)] or ETV4 [AUC = 0.672 (95% CI, 0.603–0.735)] single scoring or combinational scoring [AUC = 0.785 (95% CI, 0.722–0.839)]. P values were determined by two-tailed t test. K, Proposed working model of the current study.

Figure 7.

ETV4/ME1 axis is overactivated in gastric cancer. A, Representative staining showed upregulated expression of ME1 protein in gastric tumor tissues compared with corresponding nontumorous tissues. Scale bar, 100 μm. B, Immunoscoring of ME1 in adjacent normal tissues (ANT), gastric cancer tissues (T), distant organ metastasis (M), and lymph node metastasis (Ln). C, The ME1 mRNA expression in 50 paired gastric tumor samples and corresponding nontumorous tissues. D, Kaplan–Meier analysis of overall survival or disease-free survival curves for gastric cancer patients with low versus high expression of ME1. E, Immunoblots of ETV4 and ME1 in eight freshly collected gastric cancer samples. F, The relative protein expression levels in eight freshly collected gastric cancer samples were quantified by comparing the gray level of each band using ImageJ Software. G, Correlations between mRNA level of ETV4 and ME1 in 50 freshly collected gastric cancer samples. H, Correlations of ETV4 and ME1 protein expression in gastric cancer tissues based on immunoscoring. I, Gastric cancer patients were divided into three groups according to immunoscoring of ETV4 and ME1, and overall survival curve was generated with Kaplan–Meier methods. J, Receiver operating characteristic (ROC) curve analysis of ME1 [area under a curve (AUC) = 0.587 (95% CI, 0.517–0.655)] or ETV4 [AUC = 0.672 (95% CI, 0.603–0.735)] single scoring or combinational scoring [AUC = 0.785 (95% CI, 0.722–0.839)]. P values were determined by two-tailed t test. K, Proposed working model of the current study.

Close modal

To determine the clinical relevance of ME1 in gastric cancer, archived patients were divided into high expression and low expression group according to immunoscoring of ME1. Statistical analyses revealed that expression of ME1 was significantly correlated with differentiation state (P = 0.046), but not with other clinical parameters, including age, gender, tumor size, lymph node metastasis, venous invasion, perineural invasion, and tumor–node–metastasis (TNM) stage (Supplementary Table S2). In addition, Kaplan–Meier survival analysis and log-rank test showed that ME1 overexpression was correlated with shorter overall survival and disease-free survival (P < 0.001; Fig. 7D). Univariate and multivariate analyses indicated that only upregulated ME1 expression and TNM stage were independent prognostic factors for outcome in gastric cancer (P < 0.001; Supplementary Table S3).

We further analyzed the expression of ETV4 and ME1 in eight freshly collected gastric cancer samples. Western blot analysis indicated that both ETV4 and ME1 were significantly upregulated in the eight tumor samples examined, compared with the paired adjacent noncancerous tissues from the same patients (Fig. 7E). In addition, ETV4 expression was positively correlated with ME1 expression at protein level (P = 0.0043, r = 0.6727) as analyzed in the eight samples (Figs. 7E and F; Supplementary Fig. S7B). In 50 freshly collected clinical gastric cancer samples, ETV4 mRNA expression was statistically correlated with the mRNA levels of ME1 (P < 0.001, r = 0.6298; Fig. 7G). Figure 7H showed that there was a significant positive correlation between ETV4 expression and ME1 (P = 0.035) in the 207 gastric cancer samples. Samples that had lower level of ETV4 expression also had a lower ME1 expression, whereas samples that had higher level of ETV4 expression had a higher ME1 expression (Supplementary Fig. S7C). In our patient cohort, overexpression of ETV4 was significantly associated with outcome of gastric cancer (Supplementary Fig. S7D). On the basis of ETV4 and ME1 expression, gastric cancer patients were categorized into three groups with different risks of disease progression or death. Patients with high expression of both ETV4 and ME1 showed the worst outcome (Fig. 7I; Supplementary Fig. S7E). Moreover, combination of ETV4 and ME1 immunostaining showed higher predictive value than either parameter alone of gastric cancer patients survival in ROC curve analysis (Fig. 7J).

Homozygous deletion of SMAD4 was frequently identified in nearly one third of pancreatic cancer cases (30), and loss of neighboring housekeeping genes often leads to collateral lethality (4, 6). However, neither homozygous deletion (31) nor inactivation mutation (32) of SMAD4 was reported to be tightly associated with gastric cancer, although knockout mice have clearly demonstrated its tumor-suppressive functions in the gastrointestinal tract (33). Our analysis identified frequent hemizygous deletion of SMAD4 in gastric cancer and concurrent underexpression of ME2, which leads to high dependency of cancer cells to ME1 in energy stress conditions.

Cancer cells often require much more NADPH supplementation for redox hemostasis, lipid oxidation, and biomolecular synthesis than their normal counterparts (19, 34, 35). It is predicted that intracellular NADPH comes from the oxidative pentose phosphate pathway (∼30%), the glutaminolysis flux through malic enzymes (∼30%), and the methylenetetrahydrofolate dehydrogenase–mediated folic metabolism (∼40%) in proliferating cells (18). We previously reported that disruption of G6PD-gated pentose phosphate pathway resulted in marked reduction in NADPH and enhanced sensitivity to ROS stresses (23). In this study, we focused on malic enzymes family and found that ME2 was downregulated in gastric cancer due to concomitant genomic deletion with SMAD4 (6, 32). Cells could tolerate hemizygous deletion of ME2 as the redundant paralog ME1 provided metabolic anaplerosis for reducing equivalents and TCA substrates. However, metabolic stress developed when tumor growth exceeded the ability of available vasculature to supply tumor cells with oxygen and nutrients, which was a common impediment to tumor growth (9) and totally distinguished them from in vitro culture system. When its paralog ME2 was suppressed, ME1 showed a key function in providing NADPH for glutathione regeneration and ROS elimination, which was critical for gastric cancer cell survival under energy stress conditions, such as glucose limitations, anchorage-independent growth, and solid tumor formation in vivo. Moreover, ME1 was transcriptionally upregulated by ROS in an ETV4-dependent manner. Our results provided comprehensive insights into the redundant roles of ME1 in gastric cancer tumorigenesis and metastasis (Fig. 7K).

Expression of ME1 was known to be regulated by well-known oncogenes or tumor suppressors such as KRAS (20, 35) or TP53 (16). Overexpression of ME1 was reported to predict poor prognosis of hepatocellular carcinoma (36) and to confer radiation resistance in lung cancer (21). Suppression of ME1 led to glucose addiction of nasopharyngeal cancer (13) and colorectal cancer cells (14), which was further confirmed in our study. However, Zheng FJ and colleagues found that enzymic activity and protein level of ME1 was induced by excessive carbohydrate supplementation, including glucose and pyruvate (13), whereas upregulation of ME1 was observed after glucose deprivation in our study, indicating a tissue-specific regulation pattern. Moreover, aberrant expression of ME1 was associated with poor prognosis in our patient cohort. Taken together, these findings suggested that ME1 has an important function in the growth and survival of cancer cells and that it could be used as a drug target in cancer therapy. We next explored therapeutic potential of ME1 inhibition in cell line–based as well as PDX models via in vivo siRNA treatment. Silence of ME1 significantly suppressed tumor growth and induced elevated cell apoptosis. Importantly, a recent report has revealed a panel of small molecules that could inhibit activities of malic enzymes (17). However, further studies are needed to confirm the clinical benefit of these inhibitors.

In this study, we provide the genetic and pharmacologic evidences that ME1 inhibition is lethal in cells with collateral loss of ME2 due to hemizygous deletion of SMAD4, whereas ME2-intact cells could rely on ME2 to undergo glutaminolysis and to provide NADPH for cell survival under redox stress conditions. Inhibition of ME1 would be a promising therapeutic alternative in gastric cancer treatment.

No potential conflicts of interest were disclosed.

Conception and design: H.-Q. Ju, R.-H. Xu

Development of methodology: Y.-X. Lu, H.-Q. Ju, Y. Wang, P.-S. Hu, D.-S. Zhang, F. Wang, R.-H. Xu

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y.-X. Lu, H.-Q. Ju, D.-L. Chen, Y. Wang, Q.-N. Wu, H.-B. Qiu, Z.-Q. Wang, D.-S. Zhang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y.-X. Lu, H.-Q. Ju, Z.-X. Liu, Q. Zhao, Q.-N. Wu, R.-H. Xu

Writing, review, and/or revision of the manuscript: Y.-X. Lu, Z.-X. Liu, R.-H. Xu

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y.-X. Lu, Y. Wang, Q. Zhao, Z.-l. Zeng, Z.-Q. Wang, D.-S. Zhang, F. Wang, R.-H. Xu

Study supervision: F. Wang, R.-H. Xu

Other (assistance in generation of PDX models): Y. Wang

This research was supported by National High Technology Research and Development Program of China (863 Program), China (No. 2015AA020103 to R.-H. Xu), National Natural Science Foundation of China (nos. 81602137 to H.-Q. Ju; 81572392 to Z.-L. Zeng; 31501069 to Z.-X. Liu), Natural Science Foundation of Guangdong Province (nos. 2017A030313485 to H.-Q. Ju; 2014A030312015 to R.-H. Xu), and Science and Technology Program of Guangdong (no. 2015B020232008 to R.-H. Xu).

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.
Cancer Genome Atlas Research Network
. 
Comprehensive genomic characterization defines human glioblastoma genes and core pathways
.
Nature
2008
;
455
:
1061
8
.
2.
Druker
BJ
. 
Translation of the Philadelphia chromosome into therapy for CML
.
Blood
2008
;
112
:
4808
17
.
3.
Gerber
DE
,
Minna
JD
. 
ALK inhibition for non-small cell lung cancer: from discovery to therapy in record time
.
Cancer Cell
2010
;
18
:
548
51
.
4.
Muller
FL
,
Colla
S
,
Aquilanti
E
,
Manzo
VE
,
Genovese
G
,
Lee
J
, et al
Passenger deletions generate therapeutic vulnerabilities in cancer
.
Nature
2012
;
488
:
337
42
.
5.
Liu
Y
,
Zhang
X
,
Han
C
,
Wan
G
,
Huang
X
,
Ivan
C
, et al
TP53 loss creates therapeutic vulnerability in colorectal cancer
.
Nature
2015
;
520
:
697
701
.
6.
Dey
P
,
Baddour
J
,
Muller
F
,
Wu
CC
,
Wang
H
,
Liao
WT
, et al
Genomic deletion of malic enzyme 2 confers collateral lethality in pancreatic cancer
.
Nature
2017
;
542
:
119
23
.
7.
Kim
KH
,
Kim
W
,
Howard
TP
,
Vazquez
F
,
Tsherniak
A
,
Wu
JN
, et al
SWI/SNF-mutant cancers depend on catalytic and non-catalytic activity of EZH2
.
Nat Med
2015
;
21
:
1491
6
.
8.
Wong
CC
,
Qian
Y
,
Li
X
,
Xu
J
,
Kang
W
,
Tong
JH
, et al
SLC25A22 promotes proliferation and survival of colorectal cancer cells with KRAS mutations and xenograft tumor progression in mice via intracellular synthesis of aspartate
.
Gastroenterology
2016
;
151
:
945
60
.
9.
Jeon
SM
,
Chandel
NS
,
Hay
N
. 
AMPK regulates NADPH homeostasis to promote tumour cell survival during energy stress
.
Nature
2012
;
485
:
661
5
.
10.
Bryant
HE
,
Schultz
N
,
Thomas
HD
,
Parker
KM
,
Flower
D
,
Lopez
E
, et al
Specific killing of BRCA2-deficient tumours with inhibitors of poly(ADP-ribose) polymerase
.
Nature
2005
;
434
:
913
7
.
11.
Nijhawan
D
,
Zack
TI
,
Ren
Y
,
Strickland
MR
,
Lamothe
R
,
Schumacher
SE
, et al
Cancer vulnerabilities unveiled by genomic loss
.
Cell
2012
;
150
:
842
54
.
12.
Murata
S
,
Zhang
C
,
Finch
N
,
Zhang
K
,
Campo
L
,
Breuer
EK
. 
Predictors and modulators of synthetic lethality: an update on PARP inhibitors and personalized medicine
.
BioMed Res Int
2016
;
2016
:
2346585
.
13.
Zheng
FJ
,
Ye
HB
,
Wu
MS
,
Lian
YF
,
Qian
CN
,
Zeng
YX
. 
Repressing malic enzyme 1 redirects glucose metabolism, unbalances the redox state, and attenuates migratory and invasive abilities in nasopharyngeal carcinoma cell lines
.
Chinese J Cancer
2012
;
31
:
519
31
.
14.
Murai
S
,
Ando
A
,
Ebara
S
,
Hirayama
M
,
Satomi
Y
,
Hara
T
. 
Inhibition of malic enzyme 1 disrupts cellular metabolism and leads to vulnerability in cancer cells in glucose-restricted conditions
.
Oncogenesis
2017
;
6
:
e329
.
15.
Pongratz
RL
,
Kibbey
RG
,
Shulman
GI
,
Cline
GW
. 
Cytosolic and mitochondrial malic enzyme isoforms differentially control insulin secretion
.
J Biol Chem
2007
;
282
:
200
7
.
16.
Jiang
P
,
Du
W
,
Mancuso
A
,
Wellen
KE
,
Yang
X
. 
Reciprocal regulation of p53 and malic enzymes modulates metabolism and senescence
.
Nature
2013
;
493
:
689
93
.
17.
Ranzani
AT
,
Nowicki
C
,
Wilkinson
SR
,
Cordeiro
AT
. 
Identification of specific inhibitors of trypanosoma cruzi malic enzyme isoforms by target-based HTS
.
SLAS Disco
2017
:
2472555217706649
.
18.
Fan
J
,
Ye
J
,
Kamphorst
JJ
,
Shlomi
T
,
Thompson
CB
,
Rabinowitz
JD
. 
Quantitative flux analysis reveals folate-dependent NADPH production
.
Nature
2014
;
510
:
298
302
.
19.
Pavlova
NN
,
Thompson
CB
. 
The emerging hallmarks of cancer metabolism
.
Cell Metab
2016
;
23
:
27
47
.
20.
Son
J
,
Lyssiotis
CA
,
Ying
H
,
Wang
X
,
Hua
S
,
Ligorio
M
, et al
Glutamine supports pancreatic cancer growth through a KRAS-regulated metabolic pathway
.
Nature
2013
;
496
:
101
5
.
21.
Chakrabarti
G
. 
Mutant KRAS associated malic enzyme 1 expression is a predictive marker for radiation therapy response in non-small cell lung cancer
.
Rad Oncol
2015
;
10
:
145
.
22.
Lu
YX
,
Ju
HQ
,
Wang
F
,
Chen
LZ
,
Wu
QN
,
Sheng
H
, et al
Inhibition of the NF-kappaB pathway by nafamostat mesilate suppresses colorectal cancer growth and metastasis
.
Cancer Lett
2016
;
380
:
87
97
.
23.
Ju
HQ
,
Lu
YX
,
Wu
QN
,
Liu
J
,
Zeng
ZL
,
Mo
HY
, et al
Disrupting G6PD-mediated Redox homeostasis enhances chemosensitivity in colorectal cancer
.
Oncogene
2017
;
36
:
6282
92
.
24.
Lu
YX
,
Chen
DL
,
Wang
DS
,
Chen
LZ
,
Mo
HY
,
Sheng
H
, et al
Melatonin enhances sensitivity to fluorouracil in oesophageal squamous cell carcinoma through inhibition of Erk and Akt pathway
.
Cell Death Dis
2016
;
7
:
e2432
.
25.
Gao
H
,
Korn
JM
,
Ferretti
S
,
Monahan
JE
,
Wang
Y
,
Singh
M
, et al
High-throughput screening using patient-derived tumor xenografts to predict clinical trial drug response
.
Nat Med
2015
;
21
:
1318
25
.
26.
Kroemer
G
,
Pouyssegur
J
. 
Tumor cell metabolism: cancer's Achilles' heel
.
Cancer Cell
2008
;
13
:
472
82
.
27.
Martinez-Outschoorn
UE
,
Peiris-Pages
M
,
Pestell
RG
,
Sotgia
F
,
Lisanti
MP
. 
Cancer metabolism: a therapeutic perspective
.
Nat Rev Clin Oncol
2017
;
14
:
113
.
28.
Taddei
ML
,
Giannoni
E
,
Fiaschi
T
,
Chiarugi
P
. 
Anoikis: an emerging hallmark in health and diseases
.
J Pathol
2012
;
226
:
380
93
.
29.
Panieri
E
,
Santoro
MM
. 
ROS homeostasis and metabolism: a dangerous liason in cancer cells
.
Cell Death Dis
2016
;
7
:
e2253
.
30.
Bardeesy
N
,
Cheng
KH
,
Berger
JH
,
Chu
GC
,
Pahler
J
,
Olson
P
, et al
Smad4 is dispensable for normal pancreas development yet critical in progression and tumor biology of pancreas cancer
.
Genes Dev
2006
;
20
:
3130
46
.
31.
Lei
J
,
Zou
TT
,
Shi
YQ
,
Zhou
X
,
Smolinski
KN
,
Yin
J
, et al
Infrequent DPC4 gene mutation in esophageal cancer, gastric cancer and ulcerative colitis-associated neoplasms
.
Oncogene
1996
;
13
:
2459
62
.
32.
Powell
SM
,
Harper
JC
,
Hamilton
SR
,
Robinson
CR
,
Cummings
OW
. 
Inactivation of Smad4 in gastric carcinomas
.
Cancer Res
1997
;
57
:
4221
4
.
33.
Xu
X
,
Brodie
SG
,
Yang
X
,
Im
YH
,
Parks
WT
,
Chen
L
, et al
Haploid loss of the tumor suppressor Smad4/Dpc4 initiates gastric polyposis and cancer in mice
.
Oncogene
2000
;
19
:
1868
74
.
34.
Lewis
CA
,
Parker
SJ
,
Fiske
BP
,
McCloskey
D
,
Gui
DY
,
Green
CR
, et al
Tracing compartmentalized NADPH metabolism in the cytosol and mitochondria of mammalian cells
.
Mol Cell
2014
;
55
:
253
63
.
35.
Ying
H
,
Kimmelman
AC
,
Lyssiotis
CA
,
Hua
S
,
Chu
GC
,
Fletcher-Sananikone
E
, et al
Oncogenic Kras maintains pancreatic tumors through regulation of anabolic glucose metabolism
.
Cell
2012
;
149
:
656
70
.
36.
Wen
D
,
Liu
D
,
Tang
J
,
Dong
L
,
Liu
Y
,
Tao
Z
, et al
Malic enzyme 1 induces epithelial-mesenchymal transition and indicates poor prognosis in hepatocellular carcinoma
.
Tumour Biol
2015
;
36
:
6211
21
.