H3K4me3 and H3K27ac Activated lncRPL34-AS1 Acts as a Suppressor of Esophageal Squamous Cell Carcinoma via Targeting miR-575/ACAA2 Axis and Binding to ALOX12B and CAT

Background: Long noncoding RNAs (lncRNAs) are abnormally expressed in a broad type of cancers and play signicant roles that regulate tumor development and metastasis. However, the pathological roles of lncRNAs in esophageal squamous cell carcinoma (ESCC) remain largely unknown. Here we aimed to investigate the role and regulatory mechanism of the novel lncRPL34-AS1 in the development and progression of ESCC. Methods: The expression level of lncRPL34-AS1 in ESCC tissues and different cell lines was determined by quantitative real-time PCR (RT-qPCR). Chromatin immunoprecipitation (ChIP) assay was used to evaluate the regulatory effect of histone modication on lncRPL34-AS1. Then, functional experiments in vitro and in vivo were employed to explore the effects of lncRPL34-AS1 on tumor growth and metastasis in ESCC. Mechanistically, uorescence in situ hybridization (FISH), bioinformatics analyses, luciferase reporter assay, RNA immunoprecipitation (RIP) assay and western blot assays were used to detect the regulatory relationship between lncRPL34-AS1, miR-575 and ACAA2. In addition, comprehensive identication of RNA binding proteins (ChIRP), mass spectrometry, and RIP assay were used to identify lncRPL34-AS1-interacting proteins. Results: LncRPL34-AS1 was signicantly down-regulated in ESCC tissues and cells, which was negatively correlated with overall survival in ESCC patients. The chromatin immunoprecipitation (ChIP) assays indicated that gain of H3K4me3 and H3K27 acetylation-activated lncRPL34-AS1 was down-regulated in ESCC. Functionally, upregulation of lncRPL34-AS1 dramatically suppressed ESCC cell proliferation, colony formation, cell cycle progression and induced apoptosis in vitro, whereas knockdown of lncRPL34-AS1 elicited the opposite function. Consistently, overexpression of lncRPL34-AS1 inhibited tumor growth and metastasis in vivo. Mechanistically, lncRPL34-AS1 acted as competing endogenous RNA (ceRNA) of miR-575 to relieve the repressive effect of miR-575 on its target ACAA2, then suppressed the tumorigenesis of ESCC. In addition, protein ALOX12B and CAT resulted direct binding targets of lncRPL34 ‐ AS1 and affected biological process in ESCC. Conclusions: Together, our results reveal a role for lncRPL34-AS1 in ESCC tumorigenesis and may provide a strategy for using lncRPL34-AS1 as a potential biomarker and a therapeutic target for patients with ESCC.


RNA extraction and real-time quantitative polymerase chain reaction (RT-qPCR)
The total RNA extraction, reverse transcription and real-time quantitative PCR of lncRPL34-AS1, miRNA and mRNA expression were performed as previously described [21]. The detailed sequences of primers were listed in Additional le 1: Table 1. siRNAs (si-RPL34-AS1 1#, 2# and 3#) and plasmid vector (pcDNA-lncRPL34-AS1, pcDNA3.1 empty vectors) were purchased from KeyGEN BioTECH. The miR-575 mimics, miR-575 inhibitor and pcDNA-ACAA2 were provided by Genomeditech. The all nucleotide sequences were listed in Additional le 1: Table 2. After 48 hours post-transfection, cells were harvested for RT-qPCR or Western blot analysis.

Cell proliferation, invasion and migration assays
The in vitro invasion, migration, CCK8, colony formation and EdU proliferation assay were detected as previous described [21,22].

Western blot assay
Western blot assay was performed as previous described [22]. Dual-luciferase reporter assay The sequences of lncRPL34-AS1 and ACAA2-3'UTR and their corresponding mutations were designed, synthesized and inserted into luciferase reporter vector PGL3-CMV-LUC-MCS (Genomeditech, Shanghai). The relative ratio of re y luciferase (Luc) activity/Renilla luciferase (Rena) activity was applied to evaluate uorescence intensity.

RNA uorescent in situ hybridization (FISH)
To localize the cellular distribution of lncRPL34-AS1, the FISH assay was performed using the lncRNA FISH Probe and Fluorescent in Situ Hybridization Kit (Ribobio, China) according to the manufacturer's guidelines. The probe cocktail included the 18S/cytoplasm probe, U6/nuclear probe, and lncRPL34-AS1 probe (DAPI and Cy3-labeled probes were synthesized for uorescence signals). The results were captured by the OLYMPUS laser confocal microscope FV1000 (Olympus, Tokyo, Japan).
Chromatin immunoprecipitation (ChIP) assay The ChIP assays were performed using ChIP Assay Kit (Cell Signaling Technology). EC109 and Het-1A cells with 37% formaldehyde were incubated for 10 minutes at room temperature to form DNA-protein crosslinks, and then cross-linked chromatin DNAs were broken into 200 to 1000 bp-sized segments by using an ultrasound breaker. Anti-H3K27ac antibody, anti-H3K4Me3 antibody or IgG antibody (Cell Signaling Technology) was used to precipitate the chromatin the lysate. After de-crosslinked enriched target of protein-DNA complex, and nally the enriched DNA fragments were puri ed and analyzed via RT-qPCR.

Detection of MDA and SOD levels
The malondialdehyde (MDA) content and superoxide dismutase (SOD) activity were examined by commercial detection kits (Beyotime, Shanghai, China) following the manufacturer's instructions.
In vivo transcription assays and ChIRP-Mass spectrometry (LC-MS/MS) assays 12-20 15 cm dishes of cells were used per ChIRP-MS experiment (200 million EC109 cells). The lncRPL34-AS1 in vivo probe labeled with biotin was provided by Akimics. Cell harvesting, lysis, disruption, pre-bind probe to streptavidin beads for ChIRP. After tryptic digestion and peptide desalting, for each sample, peptides were separated and analyzed with a nano-UPLC (EASY-nLC1200) coupled to Q-Exactive mass spectrometry (Thermo Finnigan). The raw MS les were processed with MaxQuant (Version 1.5.6.0) database search and label free quantitative analysis. Subsequently, the quantitative results of the samples were statistically analyzed, and the corresponding enriched proteins were obtained based on the FCa value and number of unique peptides. Bioinformatics analysis such as GO, KEGG pathway, and protein interaction of comprehensive RNA-Binding proteins-Mass Spectrometry results were analyzed and displayed.
Transcriptome sequencing (RNA-seq) and computational analysis EC109 cells transfected by lncRPL34-AS1 plasmid and scramble pcDNA were used for RNA-sEq. RNA quanti cation and quality assurance were evaluated by NanoDrop ND-1000, and RNA integrity and gDNA contamination tested by standard denaturing agarose gel electrophoresis. The mixed different sample libraries were sequenced by IlluminaNovaSeq6000 sequencer; raw sequencing data performed sequencing quality control (QC) to evaluate sequencing data analysis. The R software Ballgown was used to calculate the FPKM at gene level and transcript level, and the differences of expression at gene level and transcript level were calculated respectively. Based on the PCA analysis and correlation analysis of gene and transcript expression level, the differentially expressed genes were analyzed by scatter plot, volcano diagram, cluster diagram, GO function and KEGG pathway signi cant enrichment. Accession ID for the RNA-seq data is GSE154450.

Tumor xenograft model
Page 9/20 To study the effect of lncRPL34-AS1 on tumor growth and metastasis, BALB/c nude mice (male, 4-weekold) were obtained for tumor xenografts experiments (Jiangsu GemPharmatech, China). In tumor growth assay in vivo, EC109 cells transfected with lncRPL34-AS1-overexpressing or control vector were subcutaneously injected into the right ank of the mice (5 × 10 6 , 200 µl). Volumes of tumors were measured every 3 days after being apparently observed and calculated with the following formula: Volume= (length × width 2 )/2. After 2 weeks of tumor generated, mice were sacri ced and subcutaneous tumor tissues were detected for tumor weight, RT-qPCR and Hematoxylin and eosin (H&E) staining. The antibodies against Ki-67 (Abcam, Cambridge, MA, USA) were applied for immunohistochemistry (IHC). In tumor metastasis in vivo, lncRPL34-AS1-overexpressing or control EC109 cells (1 × 10 6 , 200 µl) were intravenously injected into tail vein of nude mice. After 34 days, the livers and lungs were removed, para n-embedded and nally validated by H&E staining. All animal procedures were approved by the Animal Care and Use Committee of Southeast University.

Statistical analysis
Data were analyzed using SPSS 23.0 (IBM, USA) and GraphPad Prism8.1 software. All results were expressed as the mean ± SD. The Student's t-test and one-way ANOVA were performed to analyze whether two or more groups had statistical signi cance. Chi-square test was used to estimate the correlation between the lncRPL34-AS1 expression and clinicopathologic features. Overall survival curves were protracted using the Kaplan-Meier method and estimated by the log-rank test. The gene expression correlation was analyzed using the Pearson correlation test. A two-sided P value < 0.05 was considered as statistically signi cant.

Results
Gain of H3K4me3 and H3K27 acetylation-activated lncRPL34-AS1 is down-regulated in ESCC and negatively correlated with poor prognosis To explore the role of RPL34-AS1 in ESCC, we rst assessed RPL34-AS1 expression level in 75 paired primary ESCC and matched adjacent nontumor tissues by RT-qPCR. The results showed that the expression of RPL34-AS1 was signi cantly reduced in the ESCC tissue samples (Fig. 1a). Moreover, RPL34-AS1 expression was down-regulated in ESCC cell lines compared with Het-1A cells, Among ESCC cell lines, EC109 cells showed the signi cant downregulation. Thus, we selected EC109 cell line to investigate the downstream regulatory pathway of RPL34-AS1 (Fig. 1b). Furthermore, the results of FISH assay showed that the RPL34-AS1 transcripts were distributed both in the cytoplasm and nucleus of EC109 cells (Fig. 1c), which indicated RPL34-AS1 might function in both cytoplasm and nucleus. To further investigate the expression pattern of RPL34-AS1 in ESCC, we also performed an analysis of RPL34-AS1 expression in a public microarray pro le dataset from the Cancer Genome Atlas (TCGA). Consistent with our previous results, RPL34-AS1 expression was downregulated in ESCC tissues (Fig. 1d). Then we divided all ESCC patients into high and low RPL34-AS1 expression level groups according to the median value. Kaplan-Meier survival analysis of ESCC patients with low RPL34-AS1 expression had a signi cant poorer overall survival than those with high RPL34-AS1 expression (P = 0.01788, Fig. 1e). As shown in Table 1, the correlation analysis between RPL34-AS1 expression and clinicopathologic characteristics of these ESCC patients indicated that low expression of RPL34-AS1 was positively correlated with age (P = 0.008) and hard food (P = 0.042).
To explore the epigenetic modi cation mechanism of lncRPL34-AS1 in ESCC, rstly, using UCSC Genome Bioinformatics Site (http://genome.ucsc.edu/), we found high enrichment of H3K4me3 and H3K27ac at the promoter of RPL34-AS1. Hence, we speculated that the downregulation of RPL34-AS1 could be attributed to H3K4me3 and H3K27ac at its promoter region. To con rm this hypothesis, we detected the gain of H3K4me3 and H3K27Ac in EC109 cells compared with Het-1A cells at the promoter of RPL34-AS1 (Fig. 1f). Together, the data above con rmed that RPL34-AS1 was frequently reduced in ESCC, histone methylation and acetylation activation of promoter may partly account for this dysregulation.
LncRPL34-AS1 suppresses ESCC cell proliferation, cellcycle progression, migration, invasion and promotes cell apoptosis in vitro Given that RPL34-AS1 was down-regulated in ESCC, loss-and gain-of-function approaches were employed to determine the biological function of RPL34-AS1 in ESCC cells. RPL34-AS1-overexpressing EC109 cell line was established by the transfection of pcDNA3.1-RPL34-AS1. In contrast, three siRNAs were designed to silence RPL34-AS1. Real-time PCR analysis con rmed that RPL34-AS1 expression was successfully down-regulated or up-regulated in EC109 cells (Fig. 2a). CCK-8 assays demonstrated that downregulation of RPL34-AS1 signi cantly enhanced the proliferation viability, whereas the upregulation of RPL34-AS1 exerted opposite effects (Fig. 2b). Colony formation assays further demonstrated that the cell cloning capabilities of EC109 were obviously enhanced by the downregulation of RPL34-AS1 and markedly impaired by the overexpression of RPL34-AS1 (Fig. 2c). Similarly, EdU assays revealed that knockdown of lncRP34-AS1 greatly increased the percentages of EdU-positive cells, which considerably decreased at overexpression of RPL34-AS1 (Fig. 2d). Moreover, ow cytometry assays revealed that the percentage of apoptotic ESCC cells was reduced by RPL34-AS1 knockdown and the overexpression of RPL34-AS1 promoted the ESCC cells apoptosis (Fig. 2e). Depletion of RPL34-AS1 promoted cell cycle progression and upregulation of RPL34-AS1 induced cell cycle arrest at G1/S phase in EC109 cells (Fig. 2f). Furthermore, western blot results showed the consistent trend of cell apoptosis and cycle that RPL34-AS1 knockdown led to increase the levels of Bcl-2, Cyclin D1 and decrease BAX expression, as well as the upregulation of RPL34-AS1 led to opposite effects (Fig. 2g).
Next, transwell assays were carried out to examine the effects of RPL34-AS1 on migration and invasion of EC109 cells. The results indicated that the migratory and invasive capabilities of EC109 cells were remarkably enhanced by downregulation of RPL34-AS1 but signi cantly suppressed by upregulation of RPL34-AS1 (Fig. 2h, i). These experiments suggested that lncRPL34-AS1 suppressed migration and invasion of EC109 cells.
Overexpression of lncRPL34-AS1 restrains growth and metastasis of ESCC in vivo To further determine the effects of lncRPL34-AS1 on tumor growth in vivo, EC109 cells transfected with pc-RPL34-AS1 or control vector were subcutaneously injected into BALB/c nude mice. 34 days after injection, the tumors were collected. The tumors formed in the pc-RPL34-AS1 group were substantially smaller than those in the control group (Fig. 3a). The results of tumor growth curves and weights indicated that RPL34-AS1 overexpression obviously reduced tumor growth in mice (Fig. 3b, c). Tumor tissues were harvested for RT-qPCR analysis of RPL34-AS1 and ACAA2. We con rmed that higher expression of RPL34-AS1 and ACAA2 were detected in tumor tissues arising from RPL34-AS1 overexpression group compared to control group (Fig. 3d, e). Furthermore, H&E and IHC for Ki67 were performed to detect the expression of Ki67, and results showed that RPL34-AS1 overexpression caused decreased Ki67 expression (Fig. 3f). To investigate the role of lncRPL34-AS1 in tumor metastasis, EC109 cells transfected with pcRPL34-AS1 or control vector were injected into the tail vein of nude mice. Compared with the control group, the lncRPL34-AS1 over-expression group blunted lower lung metastasis ( Fig. 3g). Altogether, these results suggested that RPL34-AS1 upregulation suppressed ESCC tumorigenesis in vivo.
LncRPL34-AS1 serves as a miRNA sponge of miR-575 to regulate ACAA2 expression To investigate the mechanisms underlying the role of RPL34-AS1 in ESCC, we examined the mRNA expression pro les in EC109 cells after overexpression of lncRPL34-AS1. The cluster analysis, volcano plot, GO and KEGG enrichment of differentially expressed genes show in Additional le 2: Figure. S1. The miRNA information in miRBase was used to perform target prediction based on the RPL34-AS1 sequence and the candidate miRNAs were screened by miRanda and TargetScan algorithm to display the ternary relationship of lncRNA-miRNA-mRNA in the form of a network diagram (Fig. 4a). We used differentially expressed genes of RNA-seq results and competing endogenous RNAs (ceRNA) targets to take the intersection in Venny (Fig. 4b). After DAVID bioinformatics resources functional notes, three metabolic pathways were enriched (Fig. 4c) and 6 targeted genes were selected for subsequent mechanism research in line with fold change and p value. The results of ACAA2 gene expression were proved to show consistent trend with RPL34-AS1 after knockdown and overexpression of RPL34-AS1 in EC109 cells (Fig. 4d, e). Correlated with previous ceRNAs prediction, the candidate miR-575 was selected for targeting ACAA2 and lncRPL34-AS1 via miRbase predicting (Fig. 4f). In brief, the lncRPL34-AS1/miR-575/ACAA2 regulatory network was established for verifying in ESCC.
Next, we detected the expression of miR-575 and ACAA2 in ESCC and matched adjacent normal tissues. Results of RT-qPCR showed signi cant upregulation of miR-575 and downregulation of ACAA2 in ESCC tissues relative to adjacent normal tissues (Fig. 4g). Furthermore, the results of RT-qPCR in EC109 cells also showed signi cant upregulation of miR-575 relative to Het-1A cells (Fig. 4h). Correspondingly, we then exempli ed the binding relationship between lncRPL34-AS1 and miR-575; we conducted a wild-type (WT) and mutant (MuT) PGL3-CMV-lncRPL34-AS1 vector containing the binding sites of miR-575. The results showed that the luciferase activity of WT lncRPL34-AS1 reporter vector was signi cantly reduced by miR-575 mimics, compared with the empty vector and mutant reporter vector (Fig. 4i). Followed by an anti-AGO2 RIP assay was implemented to validate the binding relationship between lncRPL34-AS1 and miR-575, the results of RIP implied that the expression of miR-575 in the lncRPL34-AS1 overexpression group was speci cally higher than the NC group (Fig. 4j).
To decipher the regulatory mechanisms of miR-575 on ACAA2, we transfected luciferase reporter vector harboring 3′ UTR (WT and MuT) of ACAA2 into EC109 cells and luciferase activity was then evaluated in the transfection of miR-575 mimics. As compared to the control vector, miR-575 mimics signi cantly reduced the luciferase activity of the ACAA2 reporter vector (ACAA2 3′ UTR-WT) (Fig. 4k). Furthermore, to con rm the role of lncRPL34-AS1 on regulation of miR-575/ACAA2, we then set up another dualluciferase (DLR) analysis and divided into two groups: Group 1 (RPL34-AS1 + ACAA2 WT + miR-575 mimics) and Group 2 (NC + ACAA2 WT + miR-575 mimics), the uorescence intensity in Group 2 was reduced by 31% compared with Group 1 (Fig. 4l). These results enlightened that lncRPL34-AS1 may regulate ACAA2 expression by competitively interacting with miR-575.
LncRPL34-AS1 suppresses ESCC cell growth and metastasis through lncRPL34-AS1/miR-575/ACAA2 axis Accordingly, to verify whether RPL34-AS1 served its tumor suppressor function through RPL34-AS1/miR-575/ACAA2 axis, rescue experiments were designed using inhibitors and mimics. As shown in Fig. 5a, knockdown of RPL34-AS1 decreased protein level of ACAA2 in EC109 cells, while upregulation of RPL34-AS1 enhanced the level of ACAA2 in EC109 cells. As few studies had explored the role of miR-575 in ESCC, we began to clarify the mechanism and biological function of miR-575 in EC109 cells. The results indicated that upregulation of miR-575 signi cantly enhanced the proliferation viability, migratory and invasive capabilities of EC109 cells (Fig. 5b, c).Simultaneously, the target mRNA ACAA2 expression caused by silencing or overexpressing RPL34-AS1 were reversed by miR-575 inhibitor or mimics, respectively (Fig. 5d). Moreover, we attempted to explore whether the biological function of RPL34-AS1 in EC109 cells could also be reversed by miR-575 inhibitor or mimics. The results indicated that the miR-575 mimics reversed the proliferation, migration and invasion inhibiting effects induced by overexpression of RPL34-AS1 in EC109 cells, whereas miR-575 inhibitor counteracted the promoting effects induced by knockdown of RPL34-AS1 in EC109 cells (Fig. 5e-h).
Also we performed the mechanistical function of miR-575 by targeting ACAA2. The results of RT-qPCR and western blot displayed that substantially increased expression of ACAA2 in miR-575 mimics group in EC109 cells (Fig. 5i). Furthermore, the overexpression of ACAA2 (pcACAA2) repressed the proliferation viability, migratory and invasive capabilities of EC109 cells, whereas the above effects were reversed by miR-575 mimics (Fig. 5j-m). In addition, the ow cytometry assays also revealed that the cell cycle progression was promoted after co-transfected by pcACAA2 and miR-575 mimics, as well as the same trend that co-transfected by RPL34-AS1 upregulation and miR-575 mimics (Fig. 5n). In summary, these data strongly suggest that lncRPL34-AS1 suppresses ESCC cell growth and metastasis through lncRPL34-AS1/miR-575/ACAA2 axis.

LncRPL34-AS1 affects biological processes via binding to protein ALOX12B and CAT
To search for the potential interacting molecules of RPL34-AS1 to regulate target genes at distal genomic loci, we puri ed endogenous RPL34-AS1 complexes using modi ed ChIRP [23] that allowed unbiased high-throughput discovery of RPL34-AS1 associated binding proteins, ChIRP-MS was optimized to identify lncRNA-associated proteins (Fig. 6a). We designed 38 probes (even/odd group) against human RPL34-AS1 RNA (Additional le 1: Table 3). The total TIC peaks of Positive Control (U1 snRNA), Negative control (Ctrl) and Test sample (Lnc) resolved components displayed in Fig. 6b. Protein enrichment classi cation information for comparison Lnc-Ctrl were displayed as Venn diagram, and the results showed a number of 73 proteins that can bind to RPL34-AS1 (Fig. 6c). The results of STRINGdb proteinprotein network enrichment analysis indicated the correlation between differential proteins (Fig. 6d). The KEGG pathways were enriched that Ribosome and Glycolysis/ Gluconeogenesis exerted signi cantly effect on interaction between lncRPL34-AS1 and binding proteins (Fig. 6e). According to the iBAQ and fold change, we selected 23 functional proteins for subsequent study (Fig. 6f). As shown in Fig. 6g, the functional protein pathway Glycolysis/ Gluconeogenesis was related to lncRPL34-AS1 in ESCC.
Among all the ChIRP-retrieved proteins, catalase (CAT) and Arachidonate 12-lipoxygenase, 12R-type (ALOX12B) came into notice. The protein ALOX12B catalyzes the region and stereo-speci c incorporation of a single molecule of dioxygen into free and esteri ed polyunsaturated fatty acids generating lipid hydroperoxides [24]. The catalase occurs in almost all aerobically respiring organisms and serves to protect cells from the toxic effects of hydrogen peroxide [25]. The results of MDA and SOD detection con rmed the correlation between lncRPL34-AS1 and cellular oxidative stress (Fig. 6h). To further con rm the relationship, we examined the interaction between RPL34-AS1 and the two proteins by RNA immunoprecipitation (RIP), the results showed signi cant enrichments of RPL34-AS1 bound to CAT and ALOX12B, compared with the non-speci c IgG control (Fig. 6i). Also, western blot showed that CAT and ALOX12B protein were positively proportional to RPL34-AS1 at a posttranscriptional level (Fig. 6j). As previous results shown, the lncRPL34-AS1 regulated target gene ACAA2 was a part of lipid metabolism as well as ALOX12B. In addition, correlation was found between ALOX12B and ACAA2 mRNA in ESCC of TCGA database (Fig. 6k). Altogether, these data showed that RPL34-AS1 affected biological processes via protein ALOX12B and CAT in ESCC.

Discussion
LncRNA is encoded by a less explored region of the human genome and may retain impairing cancer drivers. Recently, it has attracted attention as a potential key layer of cancer cell regulation [26]. LncRNA determines its molecular function: cytoplasmic lncRNA plays an important functional role in regulating the translation and decay of messenger RNA, relying on competitive endogenous RNAs (ceRNAs) to sequester miRNAs from binding to their cognate mRNA targets and protein modi cation; nuclear lncRNA has functions at the transcriptional level such as histone modi cation, alternative splicing, or direct transcriptional regulation [27,28]. Meanwhile, increasing studies have revealed that lncRNAs exerted biological and clinical relevance of metastasis-associated function in ESCC [29]. In the present study, we con rmed that the novel lncRPL34-AS1 was obviously downregulated in ESCC and signi cantly correlated with age and hard food as well as overall survival of ESCC patients. TCGA data also revealed that low expression of RPL34-AS1 indicated poor survival of patients with ESCC. Next, loss-and gain-offunction assays indicated that lncRPL34-AS1 acted as a tumor suppressor in vivo and vitro of ESCC. Given the cellular distribution of lncRPL34-AS1 in both nucleus and cytoplasm, we utilized histone modi cation, ceRNA regulatory network and RNA-binding proteins to explore the underlying molecular mechanisms of lncRPL34-AS1.
One of the main mechanisms of lncRNAs is their ability to function by binding to histone modi cation complexes and relaxing the structure of chromatins and improving transcriptional activity [30]. LncRNA can recruit chromatin modi ers to regulate the chromatin state, or directly regulate the transcription process through the chromosome loop, thereby bridging the distal enhancer to the promoter [31]. For instance, the gain of H3K4me3 and H3K27ac activation of the promoter facilitate activation of HOXCAS3 in gastric cancer [32]. Herein, we found through UCSC genome browser that H3K4me3 and H3K27ac were highly enriched at the promoter region of RPL34-AS1. Our results revealed that H3K4me3 and H3K27ac activation of promoter partly contributed to downregulation of RPL34-AS1 in ESCC through ChIP assays.
Recent studies also demonstrated that H3K4me3 and H3K27ac were related to gene activation, which mainly enriched in the promoter region near the transcription start site [33,34].
It is emerging that one of the most popular functional models by which lncRNAs regulate gene expression is to interact with miRNA as ceRNAs that bind to miRNA response elements (MREs) and repress targeting mRNAs [35,36]. Here, analyzed by bioinformatics analysis, lncRPL34-AS1 was predicted to harbor miRNA-binding sites of miR-575 in the lncRPL34-AS1 sequence, which was further con rmed by luciferase reporter assay and RIP assay. Furthermore, the expression of lncRPL34-AS1 was negatively associated with miR-575, and a signi cant reciprocal repression feedback loop present in ESCC cells. Importantly, miR-575 acted as a tumor promoter in ESCC, and subsequent rescue experiments further con rmed that miR-575 mimics or inhibitor reversed the tumor suppressor roles of lncRPL34-AS1. Together, our results revealed that lncRPL34-AS1 could serve as a ceRNA by sponging miR-575 in ESCC.
Acetyl CoA acyltransferase 2 (ACAA2) is a key enzyme in the fatty acid oxidation pathway which catalyzes the last step of mitochondrial beta-oxidation and participates in various pathways related to lipid metabolism [37]. To date, whether and how lncRNA contributed to ACAA2-induced progression in ESCC remains elusive. In the present study, our results indicated that lncRPL34-AS1 interacted with miR-575 to promote the expression of ACAA2, and inferred a novel mechanistic role of lncRPL34-AS1/miR-575/ACAA2 axis in regulating the progression of ESCC. Furthermore, we revealed that ACAA2 was downregulated in ESCC and ACAA2 overexpression clearly inhibited the proliferation, migration and invasion of ESCC cells. Hence, we further demonstrated the oncogenic role of ACAA2 and provided evidence for the posttranscriptional regulation of ACAA2 by lncRPL34-AS1 in ESCC.
Up to now, accumulating evidence has illustrated that lncRNAs participate in enabling scaffolding functions and combinatorial control by binding with speci c proteins [32]. Regulation roles of lncRNAs in the RNA-binding proteins interaction involve various aspects, such as altering protein localization, modulating protein expression and activity or serving as structural components [35,38]. In our ChIRP and MS assays, lncRPL34-AS1 could interact with ALOX12B and CAT proteins, which was speci cally veri ed by RIP assays. The correlation between lncRPL34-AS1 and CAT proved to be involved in cellular oxidative stress. Also, the target gene ACAA2 and binding protein ALOX12B of lncRPL34-AS1 all participate in lipid metabolism. Recent observations indicate that certain lncRNAs can bind directly or indirectly to lipids and play an important role in regulating lipid metabolism, lipid signaling and mitochondrial function [39][40][41].
Moreover, we found that the correlation between ACAA2 and ALOX12B was signi cantly statistic in TCGA database. As a result, lncRPL34-AS1 can be involved in the regulation of multiple lipid metabolism-related protein and oxidative stress protein in ESCC.
To our knowledge, this study is giving us a novel point of view for speci c molecular targets in ESCC, and it deepened our understanding of the relationship between lncRNA, histone modi cation, miRNA and proteins in disease progression, whereas lncRPL34-AS1 regulated the development of ESCC through other mechanisms such as post-transcriptional regulation and signaling pathway feedback required further investigation. Therefore, a deeper understanding of the therapeutic potential of lncRPL34-AS1 in ESCC which warrants additional detailed studies.

Conclusions
In summary, we identi ed abnormal histone modi cation-mediated activation of a novel lncRNA RPL34-AS1, which was obviously downregulated in ESCC and functioned as a tumor suppressor. Furthermore, we also demonstrated that lncRPL34-AS1 acted as an endogenous sponge of miR-575 and subsequently promoted ACAA2 expression to inhibit ESCC cell growth and metastasis. LncRPL34-AS1 may affect the ESCC progression via interacting with proteins, thus promoting transcription and translation process. As shown in Fig. 6l, our results reveal a role for lncRPL34-AS1 in ESCC tumorigenesis and may provide a strategy for using lncRPL34-AS1 as a potential biomarker and a therapeutic target for patients with ESCC. Availability of data and materials

Abbreviations
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request

Competing Interests
The authors declare no competing nancial interests.