Abstract

   In previous study we characterized the oncogenic role of long
   non-coding RNA MALAT1 in esophageal squamous cell carcinoma (ESCC), but
   the detailed mechanism remains obscure. Here we identified glyoxalase 1
   (GLO1) as the most possible executor of MALAT1 by microarray screening.
   GLO1 is responsible for degradation of cytotoxic methylglyoxal (MGO),
   which is by-product of tumor glycolysis. Accumulated MGO may lead to
   glycation of DNA and protein, resulting in elevated advanced glycation
   end products (AGEs), while glyoxalase 1 detoxify MGO to alleviate its
   cytotoxic effect to tumor cells. GLO1 interfering led to accumulation
   of AGEs and following activation of DNA injury biomarkers, which lead
   to cell cycle arrest and growth inhibition. In silico analysis based on
   online database revealed abundant enrichment of histone acetylation
   marker H3K27ac in GLO1 promotor, and acetyltransferase inhibitor C646
   declined GLO1 expression. Acetyltransferase KAT2B, which was also
   identified as a target of MALAT, mediated histone lysine acetylation of
   GLO1 promotor, which was confirmed by ChIP-qPCR experiment. Shared
   binding sites of miR-206 were found on MALAT1 and KAT2B mRNA.
   Dual-luciferase reporter assays confirmed interaction within
   MALAT1-miR-206-GLO1. Finally, we identified MALAT1 encapsuled by
   exosome from donor cells, and transferred malignant behaviors to
   recipient cells. The secreted exosomes may enter circulation, and serum
   MALAT1 level combined with traditional tumor markers showed potential
   power for ESCC diagnosis.

   Keywords: Esophageal squamous cell carcinoma, Exosome, MALAT1,
   Glyoxalase I, KAT2B

1. Introduction

   Esophageal cancer was ranked as the sixth cause of
   malignancy-associated mortality [[41]1,[42]2], and esophageal squamous
   cell cancer (ESCC) was the predominant pathological type in China
   [[43]3]. The role of long noncoding RNAs (lncRNAs) in ESCC has
   attracted considerable attention [[44]4]. In addition, in a previous
   study, we characterized the involvement of lncRNA MALAT1 in ESCC
   proliferation and metastasis [[45]5], which has been further confirmed
   by other groups [[46][6], [47][7], [48][8]]. However, there is still a
   lack of evidence to elucidate the underlying mechanisms.

   Using mRNA array screening, we found that glyoxalase I (GLO1) may be
   the most possible downstream effector of MALAT1. As reported, GLO1 is a
   ubiquitous enzyme involved in the detoxification of methylglyoxal (MGO)
   produced during tumor aerobic glycolysis [[49]9], Methylglyoxal is an
   extremely cytotoxic metabolite that spontaneously glycates proteins and
   DNA, irreversibly forming toxic compounds called advanced glycation end
   products (AGEs) [[50]10]. Glycation of DNA contributes to
   antiproliferative and apoptotic effects, and activates the DNA injury
   repair pathway to arrest the cell cycle, thus impairing tumor growth.
   GLO1-mediated degradation is the major method of methylglyoxal
   detoxification [[51]11], and it antagonizes the adverse effects on
   tumor growth caused by MGO accumulation, leading to accelerated tumor
   proliferation.

   Being widely accepted as an oncogene, GLO1 in malignant biology has
   attracted considerable attention [[52]12]. Previous studies have
   revealed the involvement of GLO1 in gastric [[53]13], breast [[54]14],
   skin [[55]15], and prostate cancers [[56][16], [57][17], [58][18]], and
   efforts have been made to develop GLO1-based anticancer therapies
   [[59]19,[60]20]. To the best of our knowledge, this study is the first
   to uncover the role of GLO1 in ESCC and to characterize the regulatory
   mechanism of MALAT1 in enhancing GLO1 expression.

   Previous studies have found that excess lncRNAs are encapsulated in
   exosomes and secreted outside tumor cells, resulting in the
   transcellular spread of malignant phenotypes [[61]21,[62]22].
   Furthermore, tumor cell-derived exosomes may be released into
   circulation and be used as diagnostic biomarkers [[63]23,[64]24]. In
   the present study, we verified that intercellular transmission of
   ESCC-derived exosomes contributes to the spread of the malignant
   phenotype and explored the potential of serum MALAT1 as a diagnostic
   biomarker using receiver operating characteristic (ROC) curve analysis.

2. Materials and methods

2.1. Samples and cell lines

   The samples used for detection consisted of two independent cohorts. In
   cohort 1, cancerous and corresponding normal esophageal epithelial
   tissues were obtained from patients who underwent surgery between July
   2016 and June 2019 at the Department of Cardiothoracic Surgery, Jinling
   Hospital. Patients who received radiotherapy or chemotherapy before
   surgery were excluded from the study. In cohort 2, whole blood samples
   were obtained from 98 ESCC patients and 100 healthy controls at Jinling
   Hospital during the same period for MALAT1 quantification, each of whom
   donated 10 ml of blood samples. Whole blood samples were centrifuged at
   3000×g for 10 min to obtain serum samples.

   Four esophageal squamous cell carcinoma cell lines (EC109, EC9706,
   KYSE150, and KYSE450) were obtained from the Cell Bank of the Chinese
   Academy of Sciences (Shanghai, China) and the Cancer Institute and
   Hospital, Chinese Academy of Medical Sciences (Beijing, China). The
   cells were cultured in RPMI-1640 medium (Hyclone, USA) supplemented
   with 10 % fetal bovine serum (FBS) and maintained in a humidified
   incubator at 37 °C with 5 % CO[2].

   This study was approved by the Research Ethics Committee of the Jinling
   Hospital. Written informed consent for biological research was obtained
   from all participants. The study was conducted in accordance with the
   1975 Declaration of Helsinki.

2.2. Isolation and identification of exosomes

   The serum or cell culture medium (1.5 ml) was centrifuged at 2000×g at
   4 °C for 30 min, and the supernatant was transferred and further
   centrifuged at 10,000×g at 4 °C for 45 min. The supernatant was then
   collected and filtered through a 0.22 μm filter (Millipore, USA). The
   filtrate was collected and subjected to ultracentrifugation at 10,000×g
   and 4 °C for 70 min. The precipitate was resuspended in 10 ml of
   pre-cooled 1 × PBS and ultracentrifugation was repeated. Finally, the
   isolated exosomes were resuspended in 1.5 ml of pre-cooled 1 × PBS and
   characterized using transmission electron microscopy. Exosome
   concentration-particle size testing (NTA) was performed using ZetaView
   (Particle Metrix Ltd., Germany).

2.3. RNA extraction and qRT-PCR

   Total RNA was isolated from the tissues, serum, and cell cultures using
   TRIzol reagent (Takara, Japan) according to the manufacturer's
   instructions. cDNA was synthesized from 200 ng of RNA using the
   PrimeScript RT reagent Kit (Takara, Japan) and amplified by
   quantitative real-time PCR with SYBR Green Kit (Takara, Japan) on an
   Illumina Eco™ (Illumina, USA). GAPDH was used as an internal reference,
   and the relative expression levels of mRNAs were normalized to GAPDH
   using 2^−△△t. Primers are listed in [65]Supplementary Table S1.

2.4. RNA interfering and overexpression

   The cells were cultured in six-well plates at a density of
   3 × 10^5/well overnight. siRNA or the negative control at a final
   concentration of 100 nM was separately transfected into cells using
   Lipofectamine 2000 (Invitrogen, USA) The sequences of the interfering
   RNA and negative control are provided in [66]Supplementary Table S1
   pcDNA3.1, was used for GLO1 overexpression. The interfering or
   overexpression efficiency was determined by qRT-PCR or western blotting
   48 h post-transfection as described [[67]5].

2.5. Expression profile screening

   MALAT1 was knocked down using siRNA in EC109 cells for further
   microarray analysis. Total RNA was extracted as described above, and
   RNA quality was determined using standard denaturing agarose gel
   electrophoresis and NanoDrop ND-1000 (Thermo, USA). mRNA microarray
   screening was performed using GeneChip® Human Transcriptome Array 2.0
   (Affymetrix, USA), and miRNA microarray screening was performed using
   Affymetrix GeneChip miRNA 4.0 Array.

2.6. CCK8 and edu assays

   A Cell Counting Kit 8 (Djingo, Japan) was used for cell viability
   analysis. Cells transfected with siRNAs or a negative control (NC) were
   seeded in 96-well plates (5 × 10^3/well). Viable cells were quantified
   by measuring the absorbance of reduced WST-8 at 450 nm every 24 h,
   according to the manufacturer's protocol. Cell proliferation was
   detected using the Cell-Light EdU Apollo 488 In Vitro Kit (Ribobio,
   Guangzhou, China) 48 h after the indicated treatments.

2.7. Transwell assays

   Cell migration and invasion assays were performed using Costar chambers
   containing transwell inserts with a pore size of 8 μm (Corning Inc.,
   USA). Cells (5 × 10^4) transfected with siRNAs or NC were seeded in the
   upper chamber, coated with or without Matrigel (Invitrogen, USA).
   Serum-free medium was added to the upper chamber, whereas medium mixed
   with 20 % FBS was maintained in the bottom chamber. Two days
   post-incubation, cells in the upper chamber were removed, while cells
   invading the membrane were fixed with methanol and stained with 0.1 %
   crystal violet. The invading cells were imaged and counted under a
   microscope in five random fields (Olympus, Japan).

2.8. Flow cytometric analysis

   Cells were seeded on six-well plates at a density of
   1.5 × 10^5 cells/well overnight and then transfected with siRNA or NC
   as previously described. Two days post-transfection, cells were
   harvested, fixed with 70 % ethanol for 12 h, and stained with propidium
   iodide (Beyotime, China). Finally, the stained cells were analyzed by
   FACS FACSCalibur (BD Bioscience, USA) and Modlfit software (Verity
   Software House, USA). For apoptosis analysis, the harvested cells were
   stained with FITC-Annexin V and propidium iodide (Beyotime, China), and
   the binding of Annexin V-FITC and PI to the cells was read using FACS
   Calibur (BD Bioscience, USA) and analyzed using Modlfit software
   (Verity Software House, USA).

2.9. Xenografts mouse model and immunohistochemistry staining

   Four-week-old female nude BALB/c mice were housed in laminar flow
   cabinets under specific pathogen-free conditions and were randomly
   allocated into two groups (n = 6 per group). EC109 cells were
   transfected with lentiviral vectors for si-GLO1 or NC separately, and
   suspended in PBS. Suspended cells (5 × 10^6) were subcutaneously
   injected into a single side of each mouse. Tumor size was measured
   every three days, and tumor volume was calculated using the formula
   TV = ½ × a × b^2 (a = length; b = width). At day 28 post-injection, the
   mice were sacrificed by cervical dislocation under anesthesia, and the
   tumors were removed and weighed.

   The dissected tumors were embedded in paraffin, and
   immunohistochemistry was performed using rabbit anti-β-actin (Abcam
   ab124964, UK) and rabbit-anti-GLO1 (Abcam ab171121, UK). The results
   were observed and recorded under a microscope. Animal experiments were
   approved by the Ethics Committees of Jinling Hospital, and the
   experimental procedures complied with ethical guidelines.

2.10. Western blotting

   Cells were lysed using RIPA reagent (Beyotime, China) mixed with
   protease inhibitor PMSF (Roche, Switzerland) 48 h after siRNA
   transfection, and the protein extracts were separated by 10 %
   SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred
   to PVDF membranes (Sigma, USA). The membranes were incubated with
   specific primary antibodies overnight, followed by incubation with
   horseradish peroxidase (HRP)-conjugated anti-rabbit secondary
   antibodies (anti-rabbit) for 1 h. Signals were obtained using an
   enhanced chemiluminescence chromogenic substrate and visualized on an
   X-ray film (Fujifilm, Tokyo, Japan). GAPDH was used as a control.

2.11. ELISA

   GLO1 knockdown cells and negative controls were collected 48 h after
   oligo-nucleic acid transfection and subjected to sonication. The cell
   lysates were collected and sent for ELISA quantification using a
   commercially available AGE assay kit (MLBio, Shanghai, China),
   according to the manufacturer's instructions. The AGE
   concentration-OD450 regression model was acquired by testing the OD450
   value of standard samples with different concentrations, and the actual
   AGE content in each cell line was further calculated based on the
   resultant regression model.

2.12. Pharmaceutical treatment of ESCC cells

   The commercial histone acetyltransferase inhibitor, C646, was used for
   histone deacetylation (Medchem Express, USA). Cells were seeded in
   six-well plates (3 × 10^5/well) overnight and treated with C646 at a
   final concentration of 20 μM. Twenty-four hours later, the cells were
   collected and western blotting was performed to test GLO1 protein
   levels. For the cytotoxicity assay of methylglyoxal, cells were treated
   with methylglyoxal (0.025 mM) (Sigma Aldrich, USA) for 72 h, and
   dimethyl sulfoxide (DMSO)-treated cells were used as controls.

2.13. ChIP-qPCR

   For chromatin immunoprecipitation (ChIP)-qPCR experiments, cells were
   seeded in six-well plates (3 × 10^5 cells/well) and transfected with
   siRNAs or negative control oligo-nucleic acids. One day
   post-transfection, the proteins were cross-linked to DNA using 1 %
   formaldehyde for 15 min. The harvested cells were washed with 1 × PBS
   and treated with lysis buffer, and the lysates were subjected to
   ultrasonic vibration to fragment chromatin. Thereafter, 20 μg of
   cross-linked fragmented chromatin was immunoprecipitated using a
   commercial EZ-Magna ChIPTM G kit (Millipore, Billerica, MA, USA).
   Anti-histone H3K27ac (Abcam ab4729, UK) was used, and goat anti-rabbit
   IgG was used as the negative control. The retrieved DNA was quantified
   using real-time qPCR analysis. Three target regions of the GLO1
   promoter were quantified using the specific primers listed in
   [68]Supplementary Table S1. The results were calculated as a percentage
   of the input DNA using the following formula: percentage
   input = 2 % × 2 (C[T] 2 % input sample − C[T] IP sample).

2.14. Dual-luciferase reporter experiment

   Wild-type (wt) or mutant (mut) versions of MALAT1 and GLO1 were
   synthesized de novo and cloned into a Pezx-FR02 vector using digestion
   and ligation enzymes. The reorganized vectors were named Pezx-FR02-WT
   and Pezx-FR02-MUT. The reorganized plasmids were verified by DNA
   sequencing and then co-transfected with miR-206 mimic or miR-NC into
   HEK293T cells using Lipofectamine™ 2000. Two days post-transfection,
   cells were lysed and centrifuged for 3–5 min. The Luc-Pair™
   Duo-Luciferase HS Assay Kit (GeneCopoeia, Rockville, MD, USA) was used
   to determine luciferase activity, and internal luciferase activity was
   used as a reference.

2.15. Exosome labeling and tracing

   EC109-derived exosomes were incubated with DiO dye (Rengen Biosciences,
   China) at 37 °C for 30 min and then added to the culture media of
   recipient KYSE150 and TE-1 cells. The mixture was incubated at 37 °C,
   and the cells were stained with DAPI 12 h later (Boster Biological
   Technology Co., Ltd., China). The cells were observed and imaged using
   a confocal microscope (Leica, Germany).

2.16. Statistical analysis

   Results are presented as the mean ± standard deviation for independent
   experiments performed in triplicate. Statistical data analysis was
   performed using R software version 3.6.3. Student's t-test, chi-square
   test, or one-way ANOVA were used to assess statistical significance, as
   appropriate. Kaplan–Meier analysis and log-rank tests were used for
   survival analysis. Receiver operating characteristic (ROC) curves were
   used to assess the diagnostic performance of the selected biomarkers. A
   value of P < 0.05 was considered statistically significant.

3. Results

3.1. GLO1 acts as a downstream effector of MALAT1

   A total of 315 upregulated and 216 downregulated mRNAs were identified
   as differentially expressed between the negative control (NC) and
   MALAT1 knockdown (si-MALAT1) EC109 cells (fold change >1.5, P < 0.05)
   ([69]Fig. 1A and B). KEGG pathway analysis suggested that MALAT1
   participates in most carcinogenesis-related pathways, including
   metabolism, extracellular matrix (ECM)-receptor interaction, and
   PI3K-AKT pathways ([70]Fig. 1C and D). Screening results suggested that
   GLO1 was the most downregulated gene upon MALAT1 knockdown ([71]Fig.
   1A). Data based on GEPIA 2.0 ([72]http://gepia2.cancer-pku.cn/#index)
   suggested a positive correlation between MALAT1 and GLO1 ([73]Fig. 1E),
   while data from our previous gene expression analysis also found a
   positive association between MALAT1 and GLO1 ([74]Fig. 1F, GEO No.:
   [75]GSE89102
   ). Furthermore, our qRT-PCR results based on 40 ESCC tissues also
   suggested a roughly linear association ([76]Fig. 1G). We then verified
   the decreased expression of GLO1 in MALAT1 knockdown ESCC cell lines by
   qRT-PCR and western blotting ([77]Fig. 1H–J).

Fig. 1.

   [78]Fig. 1
   [79]Open in a new tab

   Microarray screening and validation of MALAT1 downstream target GLO1

   A: Hierarchical clustering of MALAT1 target mRNAs derived from three
   biological replicates. B: Volcano plots analysis of differentially
   expressed mRNAs upon MALAT1 interfering. C–D: KEGG pathway enrichment
   analysis of the top up- and down-regulated genes upon MALAT1 knockdown.
   E: Co-linear association between MALAT1 and GLO1 levels revealed by
   normalized expression data from TCGA (ESCA tumor) and GTEx (esophagus
   mucosa) databases. F: Co-expression pattern between MALAT1 and GLO1
   derived from [80]GSE89102. G: Person correlation analysis between
   MALAT1 and GLO1 from forty ESCC tissues. H: Interfering efficacy of
   MALAT1 by siRNA was detected by qRT-PCR. I: GLO1 mRNA level decreased
   upon MALAT1 knockdown revealed by qRT-PCR. J: western blotting results
   suggest decreased GLO1 protein levels induced by MALAT1 knockdown.Data
   are presented as mean ± SD of triplicate experiments (*p < 0.05,
   **p < 0.01, ***p < 0.001).

3.2. Roles of GLO1 in ESCC malignant phenotypes and MGO detoxification

   Data mining showed that GLO1 was significantly elevated in ESCC tissues
   compared with normal tissues in unpaired tissues ([81]Figs. S1A–B), and
   gradually increased with lymph node metastasis and TNM stage ([82]Figs.
   S1C–D). Kaplan–Meier analysis of The Cancer Genome Atlas (TCGA) data
   demonstrated that patients with higher GLO1 expression predicted worse
   overall, disease-free, and progression-free survival (OS, DFS, and PFS,
   respectively; [83]Figs. S1E–G). A paired comparison based on TCGA data
   also suggested GLO1 overexpression in ESCC tissues ([84]Fig. S1H). Our
   qRT-PCR results from 70 pairs of tumor and corresponding normal tissues
   further confirmed the overexpression of GLO1 in ESCC tissues ([85]Fig.
   2A), and high levels of GLO1 correlated with deep invasion, lymph node
   metastasis, and advanced stage of ESCC ([86]Supplementary Table S2). We
   also found GLO1 overexpression in the majority of pan-cancers,
   indicating that its oncogenic role was prevalent ([87]Fig. S1I).

Fig. 2.

   [88]Fig. 2
   [89]Open in a new tab

   GLO1 promotes ESCC malignant phenotypes in vitro

   A: RT-qPCR analysis of GLO1 mRNA levels in 70 paired ESCC and normal
   esophageal epithelial tissue samples. B: GLO1 protein levels in cells
   transfected with siRNA or negative control. C–D: Knockdown of GLO1
   hampered ESCC cells growth. E: EdU assays performed on cells with GLO1
   interfering or negative control. Original magnification, × 200. F:
   Effects of GLO1 knockdown on the migration and invasion abilities of
   ESCC cells. The number of cells that migrated or invaded was counted in
   three different fields. Original magnification, × 200. G–H: Effects of
   GLO1 knockdown on the cell cycle as assessed by flow cytometry. I–J:
   Apoptosis analysis of cells with GLO1 interference and negative
   control, detected by flow cytometric analysis of Annexin-V-FITC/PI
   staining. Results are presented as the mean ± SD for three biological
   replicates. Student's t-test was used to analyze the data. *P < 0.05;
   **P < 0.01; ***P < 0.001.

   Knockdown of GLO1 by siRNA ([90]Fig. 2B) dramatically suppressed ESCC
   cell growth ([91]Fig. 2C and D), proliferation ([92]Fig. 2E),
   migration, and invasion ([93]Fig. 2F). GLO1 inhibition led to cell
   cycle arrest at the G2/M phase and increased apoptosis ([94]Fig. 2G–J).
   Additionally, a xenogeneic mouse model verified the effects of GLO1
   ([95]Fig. 3A–F). The rescue experiment showed that ectopic GLO1
   expression might reverse the growth inhibition induced by MALAT
   interference ([96]Fig. 3G–J), supporting the hypothesis that GLO1 acts
   as a downstream effector.

Fig. 3.

   [97]Fig. 3
   [98]Open in a new tab

   GLO1 promotes ESCC tumor formation in vivo and rescues cell growth
   inhibition induced by MALAT1 interfering. A: Interfering efficiency of
   sh-GLO1 lentivirus. B: Images of mice from different treatment groups
   four weeks after tumor injection. C–D: Quantification of tumor volume
   and mouse weight in subjects injected with EC109 cells transfected with
   sh-NC or sh-GLO1 lentivirus vector during the four weeks period. E: The
   dissected tumors were photographed (left), and tumor weight decreased
   in the GLO1 knockdown mice group (right). F: Representative images for
   immunohistochemical staining of GLO1 in tumor nodes dissected from
   different groups (left). IHC score of GLO1 staining in sh-GLO1 and
   sh-NC samples (right). G–H: Analysis of ectopic expression of GLO1 in
   EC109 and TE1 cells at the mRNA (G) and protein (H) levels. I–J:
   Ectopic expression of GLO1 promoted tumor growth and restored the
   inhibitory effect of si-MALAT1 in EC109 and TE1 cellsData are presented
   as the mean ± SD of triplicate experiments; *P < 0.05; **P < 0.01;
   ***P < 0.001; ns not significant.

   We next sought to determine AGEs levels and DNA-injury biomarkers 24 h
   after GLO1 knockdown. The viability of ESCC cells was positively
   correlated with the concentration of AGEs ([99]Fig. 4A), and MGO
   significantly inhibited ESCC cell proliferation ([100]Fig. 4B and C).
   As expected, GLO1 inhibition led to increased levels of AGEs ([101]Fig.
   4D and E), coupled with the activation of DNA injury biomarkers
   ([102]Fig. 4F). Moreover, MALAT1 knockdown activated DNA
   injury-associated biomarkers and ectopic expression of GLO1 reversed
   these effects ([103]Fig. 4F), further supporting GLO1 as a downstream
   effector of MALAT1. MGO treatment of ESCC cells led to the activation
   of DNA injury biomarkers, which were reversed by GLO1 ectopic
   expression ([104]Fig. 4G). Furthermore, we found that MGO treatment led
   to a smaller cell morphology and decreased cell proliferation, which
   could be reversed by ectopic expression of GLO1 ([105]Fig. 4H and I).

Fig. 4.

   [106]Fig. 4
   [107]Open in a new tab

   GLO1 knockdown led to cytotoxic effects and DNA injury marker
   activation

   A: Standard AGEs-OD450 value correlation curve. B-C: Effects of
   exogenous methylglyoxal on ESCC growth were determined by CCK8
   analysis. D: Interfering efficacy of GLO1 by siRNA. E: ELISA analysis
   showed that GLO1 knockdown resulted in elevated AGEs concentration in
   the culture medium. F: MALAT1 knockdown activated DNA injury-associated
   biomarkers, and ectopic expression of GLO1 reversed the effects induced
   by MALAT1 knockdown. G: DNA injury biomarkers are activated by MGO
   treatment and reversed by ectopic GLO1 expression. H: MGO treatment led
   to cell shrinkage in morphology, as detected by cell microscopy,
   whereas ectopic expression of GLO1 reversed these effects. I: The MGO
   treatment-induced decrease in cell proliferation was reversed by
   ectopic expression of GLO1. *P < 0.05; **P < 0.01; ***P < 0.001; ns not
   significant.

3.3. KAT2B-mediated histone acetylation activates GLO1 expression

   Analysis of epigenetic modification biomarkers using the UCSC genome
   browser ([108]http://genome.ucsc.edu/) revealed abundant enrichment of
   histone H3K27ac signals around the GLO1 promoter ([109]Fig. 5A). Next,
   we conducted pharmacological inhibition in ESCC cell lines using the
   acetyltransferase inhibitor C646, and GLO1 expression was significantly
   attenuated ([110]Fig. 5B). By re-analyzing the mRNA microarray results,
   we identified the acetyltransferase KAT2B as a direct target of MALAT1.
   Downregulation of KAT2B by siRNA abrogated GLO1 protein levels
   ([111]Fig. 5C and D), and the expression of KAT2B and GLO1 was
   positively correlated in ESCC tissues ([112]Fig. 5E), indicating the
   regulatory role of KAT2B in GLO1 expression. We then detected H3K27ac
   enrichment in the GLO1 promoter using CHIP-qPCR, and the results
   indicated that H3K27ac was weaker upon KAT2B downregulation ([113]Fig.
   5F–I).

Fig. 5.

   [114]Fig. 5
   [115]Open in a new tab

   GLO1 expression was regulated by histone H3K27ac modification on its
   promotor

   A: Screenshot of the UCSC genome browser view showing genomic features
   of the GLO1 promoter. B: Histone acetyltransferase inhibitor C646
   suppressed GLO1 protein expression. C: Efficiency of knockdown for
   acetyltransferase KAT2B by siRNA. D: Knockdown of acetyltransferase
   KAT2B led to the repression of GLO1 expression according to western
   blotting analysis. E: Pearson correlation analysis between KAT2B and
   GLO1 in forty ESCC tissues. F: Schematic illustration of indicated
   targets for qPCR following CHIP assays. G–I: ChIP-qPCR revealed that
   KAT2B knockdown decreased H3K27ac acetylation at the promoter region of
   GLO1 at two separate locations. *P < 0.05; **P < 0.01; ***P < 0.001; ns
   not significant.

3.4. MALAT1 regulates KAT2B expression through a competing endogenous RNA
(ceRNA) mechanism via miR-206

   First, we verified that knockdown of MALAT1 suppressed KAT2B expression
   by western blotting ([116]Fig. 6A), and data from the GEPIA 2.0
   database suggested that MALAT1 positively correlated with KAT2B in ESCC
   (R = 0.17, P = 0.022; [117]Fig. 6B). We started with profile screening
   of miRNA targets for MALAT1 using three independent sets of microarrays
   ([118]Fig. 6C), followed by prediction of lncRNA-miRNA-mRNA interaction
   on ENCORI ([119]https://starbase.sysu.edu.cn/, [120]Figs. S2A–C).
   Microarray analysis revealed that miR-206 is a MALAT1 target, while
   MALAT1 and KAT2B shared common binding sites of miR-206 ([121]Fig. 6D),
   and MALAT1 knockdown resulted in elevated miR-206 levels ([122]Fig.
   6E). Ectopic expression of miR-206 using mimics suppressed MALAT1 and
   KAT2B expression ([123]Fig. 6F–H), and we confirmed the degradation
   effects of miR-206 on MALAT1 and KAT2B using a luciferase reporter
   system ([124]Fig. 6I and J). Furthermore, ectopic expression of miR-206
   reversed the inhibitory effect of KAT2B on MALAT1 knockdown ([125]Fig.
   6K). qRT-PCR results from ESCC tissues indicated an obvious positive
   correlation for MALAT1 and KAT2B ([126]Fig. 6L), as well as an inverse
   correlation for MALAT1-miR-206 and KAT2B-miR-206 ([127]Fig. 6M − N).
   Using data mining, we found that miR-206 expression was relatively
   lower in ESCC tissues ([128]Fig. S3A), especially in advanced stages
   ([129]Figs. S3B–C), and ectopic expression of miR-206 by mimics
   inhibited ESCC growth ([130]Figs. S3D–E), migration, and invasion
   ([131]Figs. S3F–G). These findings suggest that miR-206 acts as a tumor
   suppressor in ESCC, which is supported by its antagonistic role in
   MALAT1.

Fig. 6.

   [132]Fig. 6
   [133]Open in a new tab

   MALAT1 regulates KAT2B expression through ceRNA mechanism

   A: Western-blotting results confirmed knockdown of MALAT1 decreased
   KAT2B protein level. B: Co-expression between MALAT1 and GLO1 mRNA
   revealed by TCGA database and analysis performed using GEPIA. C:
   Heatmap for microarray analysis of MALAT1 target miRNAs. D: Sequences
   of reporter constructs containing the wild-type or mutant binding site
   (BS) of miR-206. E: qRT-PCR confirmation of miR-206 in MALAT1 knockdown
   cells. U6 was used as an internal control for miRNA quantification. F:
   The efficiency of ectopic expression of miR-206 by mimics. G–H: Effects
   of miR-206 mimics on KAT2B and MALAT1 RNA level. I–J: Luciferase
   reporter activity of wild-type (WT) or mutated (MUT) reporter in 293T
   cells co-transfected with miR-206 mimics or NC. K: Ectopic expression
   of miR-206 rescued the inhibition of the KAT2B protein induced by
   MALAT1 knockdown. L–N: Pearson correlation analysis for MALAT1-KAT2B,
   MALAT1-miR206, and KAT2B-miR206 pairs in the same set of 40 ESCC
   tissues.*P < 0.05, **P < 0.01, ***P < 0.001, n.s, no significance.

3.5. Transcellular spread of malignant phenotypes induced by exosomal MALAT1
and liquid detection based on serum MALAT1

   We separated exosomes from the culture medium of ESCC cell lines and
   quantified MALAT1 expression in the exosomes. Electron microscopy and
   concentration-particle size testing confirmed the morphological
   patterns ([134]Fig. 7A and B). The qRT-PCR results showed expression
   levels among the four ESCC cell lines ([135]Fig. 7C). These findings
   showed that exosomal MALAT1 from EC109 surpasses all others. We then
   isolated exosomes from EC109 cells and co-cultured them with KYSE150
   and TE-1 cells ([136]Fig. 7D). Twenty-four hours later, we detected
   upregulated MALAT1 levels in recipient cells ([137]Fig. 7E), which
   exhibited accelerated growth ([138]Fig. 7F and G). To evaluate the
   clinical significance of exosomal MALAT1, we measured serum exosomal
   MALAT1 levels in 98 ESCC serum samples and 100 healthy donors
   ([139]Fig. 7H). The distribution of demographic characteristics for the
   cases and controls is listed in [140]Supplementary Table S3. Serum
   MALAT1 was elevated in an orderly manner according to the tumor grade.
   High exosomal MALAT1 levels were mostly found in middle- and late-stage
   ESCC patients but not in healthy and early stage patients ([141]Fig.
   7I). Furthermore, we found no significant difference between serum
   total MALAT1 and serum exosomal MALAT1 levels, suggesting that serum
   MALAT1 exists mainly in exosomes ([142]Fig. 7J). A diagnostic model
   based on serum MALAT1 achieved moderate performance in ESCC diagnosis,
   with an area under the curve (AUC) value of 0.755 (CI = 0.686–0.824),
   which is superior to that of the classic tumor marker ([143]Fig. 7K).
   Combined ROC analysis integrating MALAT1 and the three biomarkers
   improved diagnostic performance, with an AUC value of 0.834
   (CI = 0.775–0.893; [144]Fig. 7L).

Fig. 7.

   [145]Fig. 7
   [146]Open in a new tab

   Exosomal MALAT1 mediated intercellular transmission of malignant
   phenotype and its application in ESCC non-invasive diagnosis. A:
   Identification of exosomes from different ESCC cell lines by
   transmission electron microscope. Scale bar = 200 nm. B:
   Concentration-particle size testing of isolated exosomes. C: The levels
   of exosomal MALAT1 in the four ESCC cell lines determined by
   quantitative RT-qPCR. D: Representative confocal images showing the
   delivery of Dio-labeled exosomes (EC109-derived) to DAPI-labeled
   KYSE150 and TE1 cells. Scale bar = 25 μm. E: Uptake of EC109-derived
   exosomes elevates MALAT1 levels in recipient cells by quantitative
   RT-PCR. F–G: Treatment of EC109-derived exosomes enhanced recipient
   cells growth capability; H: Unpaired comparison of serum MALAT1 level
   between ESCC patients and healthy controls. I: Serum MALAT1 level
   ascended as ESCC stages. J: No significant difference was found between
   serum total and exosomal MALAT1 level. K: ROC analysis for serum MALAT1
   and classic tumor biomarkers in ESCC diagnosis L: Combined ROC analysis
   integrating MALAT1 and three biomarkers in ESCC diagnosis.*P < 0.05;
   **P < 0.01; ***P < 0.00.

4. Discussion

   In a previous study, we demonstrated the oncogenic role of MALAT1 in
   ESCC, revealing an association between MALAT1 knockdown and DNA injury
   repair pathway activation. This association is considered to correlate
   with cell cycle arrest and growth inhibition [[147]25]. Another study
   has also provided evidence supporting the depletion of MALAT1-induced
   DNA damage [[148]26], we wondered how MALAT1 participating in the DNA
   injury repair pathway was able to promote ESCC growth.

   The microarray screening results suggested that GLO1 is a possible
   target of MALAT1. GLO1 is a member of the glyoxalase system that
   participates mainly in the degradation of cytotoxic MGO, which is a
   by-product of tumor glycolysis. MGO binds to proteins and DNAs, leading
   to glycation of products (AGEs). AGEs can directly produce deleterious
   effects or exert their damaging effects through the receptors of AGEs
   (RAGEs). Both MGO and AGEs can induce the most deleterious consequences
   directly or by stimulating RAGE signaling [[149]27,[150]28].

   As the rate-limiting enzyme, glyoxalase 1 catalyzes the conversion of
   methylglyoxal to S-D-lactyl glutathione in a glutathione
   (GSH)-dependent manner, followed by the hydrolysis of S-D-lactyl
   glutathione to non-toxic d-lactic acid performed by glyoxalase 2
   [[151]11]. The product S-D-lactyl glutathione promotes microtubule
   assembly during cell division and is thought to facilitate cell
   proliferation [[152]11].

   MGO metabolism depends mainly on the expression of GLO1
   [[153]29,[154]30]. Approximately 2 % of the population had an increased
   copy number of GLO1. These extra copies can increase the expression of
   GLO1 and the activity of the enzyme by 2–4 fold [[155]31].

   Previous studies have shown that GLO1 expression is regulated at
   transcriptional [[156]32,[157]33] and post-translational levels
   [[158]34]. Here, we present the previously unappreciated importance of
   epigenetic modulation of GLO1 expression. Studies have shown that KAT2B
   catalyzes histone acetylation modifications of PLK4 [[159]35], TBX5
   [[160]36], and CXCL12 [[161]37]. We found that KAT2B knockdown led to
   weaker H3K27ac deposits and decreased GLO1 expression, confirming that
   epigenetic modifications of GLO1 contribute to its expression.

   Using miRNA array screening and bioinformatics prediction, we
   identified miR-206 as a mediator of MALAT1 and KAT2B. miR-206 is widely
   recognized as a tumor suppressor [[162][38], [163][39], [164][40],
   [165][41]], and analysis based on a public database also supports its
   suppressor role in ESCC. The interplay between MALAT1 and miR-206 has
   been reported in previous studies [[166]42,[167]43], further supporting
   the ceRNA mechanism for the MALAT1-miR-206-KAT2B axis.

   Exosomes are small vesicles secreted by different cell types and
   covered by a lipid bilayer, with sizes ranging from 50 to 150 nm.
   Exosomes act in the pathological and physiological environments of
   target cells by the transcellular transport of different molecules,
   such as proteins, RNA, and DNA. Previous studies have shown that
   exosomes from tumor tissues may be released into circulation, and
   exosomal lncRNAs could be used as liquid biopsy biomarkers. These
   findings show a promising future for using exosomes for cancer
   non-invasive screening, especially for gastrointestinal cancers that
   rely on gastroscopic biopsy. Here, we developed a serum diagnostic
   model based on serum MALAT1 and traditional tumor biomarkers that could
   distinguish patients with ESCC from others, suggesting the potential
   value of serum MALAT1 in ESCC monitoring and follow-up. Our results
   showed that serum MALAT1 mainly exists in exosomes, which is consistent
   with a previous study [[168]44], indicating a protective effect of
   exosomes on lncRNAs, making clinical application possible.

   Thus, based on existing data, we concluded that MALAT1 upregulated
   acetyltransferase KAT2B expression through a competing endogenous RNA
   (ceRNA) mechanism, and KAT2B further altered the histone H3K27ac
   abundance on the GLO1 promoter in an epigenetic manner. High GLO1
   expression facilitates the degradation of deleterious MGO produced
   during glycolysis and reduces the concentration of AGEs, alleviating
   toxic effects on cell growth ([169]Fig. 8).

Fig. 8.

   [170]Fig. 8
   [171]Open in a new tab

   Schematic diagram of the MALAT1-KAT2B-GLO1 regulatory network and its
   implications in the progression of ESCC

   MALAT1 sponges miR-206 to alleviate its degradation of KAT2B mRNA,
   resulting in elevated KAT2B expression. KAT2B promotes GLO1 expression
   via epigenetic modification of the GLO1 promoter through histone
   acetylation. Overexpressed GLO1 detoxifies methylglyoxal, which is a
   by-product of tumor glycolysis and is cytotoxic to tumor cells.
   Finally, degradation of methylglyoxal by GLO1 enhanced tumor growth and
   metastasis. Moreover, MALAT1 is encapsulated by exosomes and released
   into circulation, resulting in malignant phenotypes that spread
   intercellularly, making liquid biopsy through exosomal MALAT1 possible.

   Our study has some limitations, such as the limited sample size, which
   requires multicenter participation and many more samples to confirm the
   clinical value of serum MALAT1 for ESCC diagnosis and monitoring. Given
   the advantages of targeting and biological compatibility, recent
   studies have attempted to reconstruct tumor-derived exosomes as drug
   cargos, such as siRNAs and chemotherapy drugs [[172]45], which have
   shown promising prospects. Whether exosomes may be ideal pharmaceutical
   vectors for MALAT1 or GLO1 target therapy requires further evaluation.
   Furthermore, the upstream regulator of MALAT1 remains unclear, and we
   are investigating this in our future work.

5. Conclusions

   MALAT1 regulates the expression of KAT2B through a competitive
   endogenous RNA (ceRNA) way, and changes the histone acetylation of the
   GLO1 promoter in a KAT2B dependent manner. High GLO1 expression
   alleviated the toxic effect of methylglyoxal on ESCC cells by
   facilitating methylglyoxal degradation. ESCC exosome-derived MALAT1
   mediates the transfer of oncogenic behavior, and exosomal MALAT1 in the
   serum may be used as a diagnostic indicator for late-stage ESCC.

Ethics approval and consent to participate

   This study was approved by the Research Ethics Committee of the Jinling
   Hospital. Written informed consent for biological research was obtained
   from all participants. The study was conducted in accordance with the
   1975 Declaration of Helsinki.

Consent for publication

   Not applicable.

Availability of data and materials

   Additional related data are available upon request.

Funding

   This work was supported by the National Natural Science Foundation of
   China (No. 81702444), China Postdoctoral Science Foundation
   2018M643884, and 2020T130128ZX, and Basic research project (Natural
   science funds) of Jiangsu Province [BK20160606].

CRediT authorship contribution statement

   Liwen Hu: Project administration, Funding acquisition,
   Conceptualization. Kai Xie: Writing – original draft, Investigation,
   Formal analysis, Data curation. Chao Zheng: Methodology, Data curation.
   Bingmei Qiu: Resources, Formal analysis. Zhisheng Jiang: Resources,
   Investigation. Chao Luo: Methodology, Data curation. Yifei Diao:
   Validation, Formal analysis. Jing Luo: Writing – original draft,
   Validation. Xinyue Yao: Supervision, Investigation, Conceptualization.
   Yi Shen: Supervision, Conceptualization.

Declaration of competing interest

   The authors declare that they have no known competing financial
   interests or personal relationships that could have appeared to
   influence the work reported in this paper.

Acknowledgments