H19 contributes to aerobic glycolysis, proliferation and immune escape of gastric cancer cells via the miR-519d-3p/LDHA axis

Background Long non-coding RNAs (lncRNAs) have been investigated in multiple human cancers including gastric cancer (GC). Our research aims to explore the role of H19 in aerobic glycolysis, proliferation, and immune escape of GC cells. Methods The expression of H19 in GC samples were analyzed using Gene Expression Proling Interactive Analysis (GEPIA), Gene Expression Omnibus (GEO) data, and real-time quantitative polymerase chain reaction (RT-qPCR) analysis. Glucose consumption and lactate production were applied to assess the aerobic glycolysis of GC cells. Subcellular fractionation, luciferase reporter, and western blot assays certied the binding between genes. CCK-8 and colony formation assays were used to determine GC cell proliferation. Flow cytometry, ELISA, and RT-qPCR assays were applied to analyze the immunosuppressive effect of H19. Results H19 was highly expressed in samples of patients with GC, and associated with tumor growth in vivo. H19 knockdown suppressed glucose consumption, lactate production, proliferation of GC cells by regulating miR-519d-3p/LDHA axis. Both miR-519d-3p depletion and LDHA overexpression could reverse the H19 knockdown-induced decrease in aerobic glycolysis and proliferation. Moreover, conditioned medium (CM) from stable knockdown H19 GC cells modulated the activity of immune cells including γδT cells, Jurkat cells, and tumor-associated macrophages (TAMs) in a lactate dependent manner. Conclusions The H19/miR-519d-3p/LDHA axis mainly contributed to aerobic glycolysis, proliferation, and immune escape of GC cells. Our results showed that the levels of H19 were increased signicantly in GC tissue specimens and associated with poor prognosis. We further demonstrated that H19 knockdown suppressed aerobic glycolysis and cell proliferation through the miR-519d-3p/LDHA axis in GC cells. on γδT cells. Our results indicated that CM from stable knockdown H19 GC cells elevated the population of IFN-γ + γδT cells, while lactate treatment counteracted this effect. In addition, we also asked whether CM from stable knockdown H19 GC cells could affect T cells and TAMs via lactate. The addition of lactate abolished the increase in IL-2 expression in PMA and lonomycin-activated Jurkat cells induced by CM from stable knockdown H19 GC cells. Moreover, treatment with lactate reversed the effect of CM from stable knockdown H19 GC cells on TAMs generation. These data indicated that H19 signaling-mediated extracellular lactate accumulation could modulate the activity of immune cells including γδT cells, T cells, and TAMs, which might contribute to the immune escape of GC cells. Nevertheless, the effect of the H19/lactate axis on other immune cells, such as NK cells, myeloid-derived suppressor cells (MDSC), and Foxp3 regulatory T cells (Treg), are needed to investigate in our future study. by RT-qPCR in SGC-7901 and MKN-45 cells after transfection with si-NC, si-H19-1, or si-H19-2. b LDHA protein level was detected by western blot in SGC-7901 and MKN-45 cells after transfection with si-NC, si-H19-1, or si-H19-2. β-actin served as a loading control. c LDHA protein level was detected by western blot in both stable knockdown H19 (sh-H19) SGC-7901 and MKN-45 cells after transfection with negative control plasmid (NC) or LDHA overexpression plasmid (LDHA). β-actin served as a loading control. d-g Glucose consumption (d, e) and lactate production (f, g) were examined in both sh-H19 SGC-7901 and MKN-45 cells after transfection with NC or LDHA. Each experiment was performed in triplicate. Data are presented as the mean ± SD and analyzed by student’s t-test (*p<0.05, **p<0.01, ***p<0.001, ns. not signicant) a loading d-g Glucose consumption (d, e) and lactate production (f, g) were examined in both sh-H19 SGC-7901 and MKN-45 cells after transfection with NC or LDHA. Each experiment was performed in triplicate. Data are presented as the mean ± SD and analyzed by student’s t-test (*p<0.05, **p<0.01, ns. not signicant) cytoplasmic RT-qPCR b Three H19 and LDHA predicted by three different algorithms targetscan7.2 LDHA, and miRcode for H19). The levels of miR-519d-3p, miR-216b-5p, and miR-17-5p were detected by RT-qPCR in SGC-7901 and MKN-45 cells after transfection with si-NC, si-H19-1 or si-H19-2. hybridization between H19 miR-519d-3p. Luciferase reporter assay showed the binding of miR-519d-3p and wide type (WT) H19 but not mutant (MUT) were measured by RT-qPCR and after transfection with H19 regulated LDHA expression and glycolysis via miR-519d-3p. a The hybridization models between LDHA and miR-519d-3p. b Luciferase reporter assay showed the binding of miR-519d-3p and wide type (WT) LDHA 3’UTR but not mutant (MUT) LDHA 3’UTR. c-d The mRNA (c) and protein (d) levels of LDHA were measured in both SGC-7901 and MKN-45 cells after transfection with inhibitor negative control (NC), miR-519d-3p inhibitor, mimic NC, or miR-519d-3p mimic. β-actin served as a loading control. e-f The mRNA (e) and protein (f) levels of LDHA were measured in both sh-H19 SGC-7901 and MKN-45 cells after transfection with inhibitor NC or miR-519d-3p inhibitor. β-actin served as a loading control. g-h Glucose consumption (g) and lactate production (h) were examined in both sh-H19 SGC-7901 and MKN-45 cells after transfection with inhibitor NC or miR-519d-3p inhibitor. Each experiment was performed in triplicate. Data are presented as the mean ± SD and analyzed by student’s t-test (*p<0.05, **p<0.01, ***p<0.001) and protein (f) levels of LDHA were measured in both sh-H19 SGC-7901 and MKN-45 cells after transfection with inhibitor NC or miR-519d-3p inhibitor. β-actin served as a loading control. g-h Glucose consumption (g) and lactate production (h) were examined in both sh-H19 SGC-7901 and MKN-45 cells after transfection with inhibitor NC or miR-519d-3p inhibitor. Each experiment was performed in triplicate. Data are presented as the mean ± SD and analyzed by student’s t-test (*p<0.05, and by ow cytometry. b The concentrations of IL-2 in the supernatants of PMA and lonimycin-activated Jurkat cells co-treated with CM from sh-H19 SGC-7901 and MKN-45 cells and lactate. c The expression of the markers of M1 and M2 macrophages in PMA-treated THP-1 macrophages co-treated with CM from sh-H19 SGC-7901 and MKN-45 cells and lactate. Each experiment was performed in triplicate. Data are presented as the mean ± SD and analyzed by student’s t-test (*p<0.05, **p<0.01, ***p<0.001) by ow cytometry. b The concentrations of IL-2 in the supernatants of PMA and lonimycin-activated Jurkat cells co-treated with CM from sh-H19 SGC-7901 and MKN-45 cells and lactate. c The expression of the markers of M1 and M2 macrophages in PMA-treated THP-1 macrophages co-treated with CM from sh-H19 SGC-7901 and MKN-45 cells and lactate. Each experiment was performed in triplicate. Data are presented as the mean ± SD and analyzed by student’s t-test (*p<0.05, **p<0.01, ***p<0.001)

migration, and invasion [14,15]. However, many unde ned molecular mechanisms by which H19 contributes to the progression of GC need to be thoroughly investigated.
As a distinctive hallmark of cancer, the Warburg effect or aerobic glycolysis confers on cancer cells a growth advantage by providing energy and biosynthesis building blocks even in the presence of abundant oxygen [16]. Aerobic glycolysis is now widely accepted and served as anti-tumor therapeutic target [17], but the molecular mechanisms controlling aerobic glycolysis have not been elucidated.
Accumulated evidence has indicated that lncRNAs could regulate glucose metabolism in cancer cells by directly regulating the glycolytic enzymes and glucose transporters, or indirectly modulating the signaling pathways [18]. For instance, lncRNA AGPG enhanced glycolysis activity and cell proliferation in esophageal squamous cell carcinoma by stabilizing PFKFB3 [19]. Zhao and his coworkers noted that lncRNA MACC1-AS1 promoted GC cell metabolic plasticity via AMPK/Lin28 mediated mRNA stability of MACC1 [20]. Although increased H19 got involved in glycolysis and stemness maintenance in breast cancer stem cells via the let-7/HIF-1α/PDK1 pathway signaling cascade [21], the effect of H19 on aerobic glycolysis in GC remains largely unknown.
Mounting evidence suggested that aerobic glycolysis in cancer cells has the ability to regulate the immune response in the tumor microenvironment [22,23]. Lactate, a kind of glycolytic metabolites, has been identi ed as a crucial component that contributes to the immunosuppressive tumor microenvironment [24]. Brand et al. showed that tumor-secreted lactate could dampen interferon-γ (IFN-γ) production by CD8 + T and NK cells and inhibiting their cytolytic activity [25]. Nonetheless, the link connecting H19, aerobic glycolysis, and immune escape in GC cells remains not fully characterized.
In this study, we investigated whether and how H19 modulated aerobic glycolysis, cell proliferation and immune evasion. Our results showed that the levels of H19 were increased signi cantly in GC tissue specimens and associated with poor prognosis. We further demonstrated that H19 knockdown suppressed aerobic glycolysis and cell proliferation through the miR-519d-3p/LDHA axis in GC cells.
Moreover, H19 knockdown modi ed the activity of γδ T cells, T cells and tumor-associated macrophages (TAMs), which participate in tumor immune response, in a lactate dependent manner.

Methods
Clinical tissue specimens Twelve paired fresh GC tissue specimens and adjacent normal tissue specimens from patients were obtained from the First A liated Hospital of Soochow University (Suzhou, China) between May 2017 and August 2018. All tissue specimens were snap-frozen in liquid nitrogen. This study was approved by the Institutional Review Board of the First A liated Hospital of Soochow University. Written informed consent was obtained from each patient. Basic information of clinical tissue specimens are provided in Supplementary Table 1. Cell culture Human GC lines AGS, SGC-7901, and MKN-45, normal gastric mucosa cell GES-1, Jurkat, and THP-1 cells were purchased from the Chinese Academy of Science Cell Bank (Shanghai, China). All cells were cultured in RPMI-1640 (Biological Industries, BeitHaemek, Israel) supplemented with 10% fetal bovine serum (FBS, Biological Industries) and 1% Penicillin -streptomycin (Beyotime, Shanghai, China, #C0222) in a humidi ed incubator with 5% CO 2 at 37°C.
The protein bands were visualized with ECL reagents (NCM Biotech, Suzhou, China, #10100) in a Chemi Doc TM MP Imaging System (Bio-Rad). The band density was analyzed using ImageJ software.

CCK-8 assay
The cell proliferation was assessed at 0, 24, 48 and 72 hours using Cell Counting Kit-8 (CCK-8) assay (Dojindo, Kumamoto, Japan, #CK04). Brie y, Cells were plated into 96-well plates at a density of 3´10 3 cells per well. Then, 10ul CCK8 solution was added to each well and cells were incubated for another 2 h at 37°C. The optical density (OD) at 450nm was measured. Each group contained 5 replicates and all experiments were performed in triplicate.

Colony formation assay
Cells were seeded into 6-well plates at 1´10 3 cells per well and cultured in a humidi ed incubator with 5% CO 2 at 37°C for 10-14 days. The cell colonies were xed in methanol for 30 min and then stained with 1% crystal violet (Sigma) for 30 min. Then, the colonies were imaged and counted.

Cytoplasm and nuclear localization
Cytoplasm and nuclear were separated from the cells using Minute TM Cytoplasmic and Nuclear Extraction Kit (inventbiotech, USA, #SC-003) according to the manufacturer's instructions.

Luciferase reporter assay
PmirGLO vectors containing the wild type (WT) or mutant (MUT) H19 or LDHA 3'UTR were obtained from Realgene Biotech (Nanjing, China). Luciferase reporter assay was performed following the previous protocols [26]. Brie y, MKN-45 cells were cultured in a 12-well plate and co-transfected with WT or MUT luciferase plasmids and miR-519d-3p mimic or control miRNA using Lipofectamine 2000. 24h after transfection, the cells were lysed with Passive Lysis Buffer (Promega, Madison, USA, #E1910), and the luciferase activity was measured with a Dual Luciferase Reporter Assay System (Promega, #E1910) according to the manufacturer's instructions.

Flow cytometry assay
For intracellular staining of IFN-γ, γδT cells were treated with CM or CM+lactate (Sanjia, Suzhou, China, #SL017769) for 24h, and stimulated with Cell Activation Cocktail with Brefeldin A (Biolegend, USA, #423304) for 10h. The cells were stained with PC7-conjugated antibody against CD3 (Biolegend) and FITC-conjugated antibody against γδT (Biolegend) at 4°C for 20 minutes. Then the cells were xed and permeabilized with a Fixation/Permeabilization Solution Kit (BD Biosciences, #554714) according to the manufacturer's instructions, and stained with PE-conjugated antibody against IFN-γ (Biolegend) allowing analysis by a ow cytometry analyzer (Beckman Coulter).

Enzyme-linked immunosorbent assay (ELISA)
The concentrations of human IL-2 in the supernatants of Jurkat cells were examined with ELISA Kit (NeoBioscience, Shenzhen, China, #EHC003.96) according to the manufacturer's instructions.

Xenograft tumor
Six-eight-weeks-old female BALB/c athymic nude mice were purchased from the Shanghai Laboratory Animal Center. All animal experiments were approved by the Institutional Animal Care and Use Committee of Soochow University (Suzhou, China). All the mice were randomly divided into two groups (n=4 each group). 5´10 6 MKN-45 cells with or without H19 knockdown were subcutaneously injected into the layer of the right ank of the mice. Tumor volume (mm 3 ) was calculated according to the following equation: V (mm 3 ) = S 2 (mm 2 ) × L(mm)/2, where S and L are the smallest and the largest perpendicular tumor diameters, respectively [28]. After 21 days, mice were sacri ced and the xenograft tumor tissues were removed, photographed and weighted.

Conditioned medium
Cells (3×10 5 ) were seeded in 6-well plates and cultured overnight. The culture medium was changed to fresh medium in each well. After 24 h, the conditioned medium (CM) was collected.
γδT cell separation and culture γδT cells were isolated from human peripheral blood mononuclear cells (PBMCs) and cultured as our previously described [29].

Statistical analysis
All data were presented as mean ± SD. Student's t-test was used for comparisons between two groups. For survival curves, Kaplan-Meier method with log-rank or Gehan-Breslow-Wilcoxon test was used. Signi cant differences were displayed as follows: *p<0.05, **p<0.01, ***p<0.001 and ns. not signi cant.

Results
H19 was overexpression in GC tissue specimens and promoted tumor growth in nude mice.
Based on a new database visualization website of GEPIA with TCGA data (http://gepia.cancer-pku.cn/), the expression of H19 was signi cantly increased in GC tissue samples compared to normal samples (Fig. 1a). GSE2685 and GSE13861 microarray data from Gene Expression Omnibus (GEO) databases (http://www.ncbi.nlm.nih.gov/gds/) also showed that the H19 level was up-regulated in GC tissue samples (Fig. 1b). To corroborate the above results, we investigated the H19 expression level in 12 pairs of fresh GC tissues and corresponding normal adjacent tissues by RT-qPCR assay, and found that the relative expression of H19 was higher in tumor than that in normal tissues (Fig. 1c). We further evaluated the prognostic value of H19 in GC using GEPIA with TCGA data and GSE26253 microarray data. As shown in Fig. 1d, GC patients with high H19 expression had a shorter overall survival in TCGA data. Moreover, high expression of H19 predicted shorter recurrence-free survival in GSE26253 microarray data (Fig. 1e). Next, the expression of H19 in normal gastric mucosa cell GES-1 and three GC cell lines AGS, MKN-45, and SGC-7901 cells was examined. The RT-qPCR results noted that H19 was overexpression in GC cell lines (Fig. 1f).
We also established a nude-mouse xenograft tumor model using stable knockdown H19 MKN-45 cells to demonstrate the effect of H19 knockdown on tumor growth in vivo (Fig. S1a-b and Fig. 1g-i). The tumor sizes, images, and mass revealed that H19 knockdown signi cantly suppressed MKN-45 tumor growth ( Fig. 1g-i). Collectively, these data demonstrated that H19 was highly expressed in a subset of patients with GC, and positively associated with tumor growth in vivo.
H19 knockdown inhibited glucose consumption and lactate production in gastric cancer cells.
Given that H19 could regulate aerobic glycolysis in breast cancer stem cells and ovarian cancer cells [21,30], we hypothesized that H19 might be involved in the glycolysis process in gastric cancer cells. Hence, the H19 knockdown cell model was established by transfecting SGC-7901 and MKN-45 cells with two siRNAs (Fig. 2a). As shown in Fig. 2b, H19 depletion dramatically reduced the glucose consumption in SGC-7901 and MKN-45 cells. Moreover, we also analyzed the effect of H19 suppression on lactate production and found that the silencing of H19 signi cantly inhibited lactate production (Fig. 2c). Consistent with the above results, the glucose consumption and lactate production were also obviously decreased in stable knockdown H19 SGC-7901 or MKN-45 cells (Fig. 2d-e). These results indicated that H19 knockdown inhibited aerobic glycolysis in gastric cancer cells.
To elucidate the mechanisms that H19 knockdown suppressed aerobic glycolysis, a spectrum of key glycolysis-related genes were examined in H19 knockdown SGC-7901 and MKN-45 cells by RT-qPCR. The results showed that compared with GLUT1, LDHB, HK2, PKM2, and PDK1, the mRNA expression level of LDHA was positively correlated with the expression level of H19 in SGC-7901 and MKN-45 cells (Fig. 3a). Moreover, H19 knockdown reduced the protein expression of LDHA in SGC-7901 and MKN-45 cells ( Fig. 3b and Fig. S2a).
To explore how H19 regulated LDHA expression in gastric cancer cells, we performed cytoplasmic and nuclear separation experiments with RT-qPCR to determine the subcellular localization of H19 and LDHA. The results showed that H19 was distributed in both the nuclear and cytoplasm, while LDHA was predominantly localized in the cytoplasm (Fig. 4a). Moreover, the predicted results of lncATLAS (http://lncatlas.crg.eu/) also indicated that H19 was localized in both the nuclear and cytoplasm (Fig.   S3). Therefore, we speculated that H19 may function as a competitive endogenous RNA (ceRNA) to regulate LDHA expression by sponging one or several speci c miRNAs. Hence, we predicted the possible binding miRNAs of H19 and LDHA by using three publicly available prediction tools starbase3.0 (for H19), targetscan7.2 (for LDHA), and miRcode (for H19). As shown in Fig. 4b, three miRNAs (miR-519d-3p, miR-216b-5p and miR-17-5p) could bind to both H19 and LDHA. Subsequently, the levels of the three candidate miRNAs were detected by RT-qPCR, and the results revealed that miR-519d-3p expression was signi cantly increased in H19 knockdown SGC-7901 and MKN-45 cells (Fig. 4c). As shown in Fig. 4d, we showed the hybridization models between H19 and miR-519d-3p. Next, we performed luciferase reporter assays to determine whether miR-519d-3p could directly regulate H19. miR-519d-3p overexpression impaired the luciferase activity of the pmirGLO-H19-WT vector but failed to reduce that of the pmirGLO-H19-MUT vector (Fig. 4e). Moreover, the results of RT-qPCR in SGC-7901 and MKN-45 cells also indicated that the down-regulation of miR-519d-3p signi cantly increased the levels of H19, whereas miR-519d-3p overexpression reduced the levels of H19 (Fig. 4f).
Notably, H19 has been reported to promote cancer cell proliferation via multiple mechanisms [31][32][33]. Herein, we investigated the effect of H19 knockdown on GC cell proliferation and explored the underlying molecular mechanisms. CCK-8 assay indicated that H19 knockdown signi cantly decreased the proliferation rate of GC cells (Fig. 6a-b). These results were further con rmed by clonogenic assay (Fig. 6c-d). Given that H19 modulated glycolysis through regulating LDHA expression in a miR-519d-3p dependent manner, we thus next determined whether H19 was involved in GC cell growth via the miR-519d-3p/LDHA axis. Rescue experiments validated that the effect of H19 knockdown on GC cell proliferation rate and colony formation was reversed by miR-519d-3p depletion and LDHA overexpression (Fig. 6).
Previous studies have shown that lactate, a major glycolytic metabolite, could modify tumor immune response [25,34]. Also, our results indicated that H19 knockdown markedly decreased the accumulation of extracellular level of lactate ( Fig. 2c and 2e). Therefore, we asked whether H19 was able to modulate tumor immune response in a lactate dependent manner. As shown in Fig. 7a, CM from stable knockdown H19 GC cells obviously elevated the population of IFN-γ-positive (IFN-γ+) γδT cells, an important antitumor immune cell [35]. Moreover, treatment with lactate counteracted the increases in the population of IFN-γ + γδT cells induced by conditioned medium from stable knockdown H19 GC cells (Fig. 7a). Besides, the addition of lactate also reduced the IL-2 expression in PMA and lonomycin-activated Jurkat cells, which was enhanced by CM from stable knockdown H19 GC cells (Fig. 7b).
We also investigated whether H19 signaling could modulate TAMs generation from human monocyte cell line THP-1. After treatment with PMA for 24 h, THP-1 cells were cultured with CM from stable knockdown H19 GC cells. The results of RT-qPCR showed that macrophages treated with CM from stable knockdown H19 GC cells exhibited higher levels of M1 marker NOS2, CXCL9, TNF-α and HLA-DR, while lower levels of M2 marker ARG1 and TGF-β (Fig. 7c). Treatment with lactate abolished the effect of conditioned medium from stable knockdown H19 GC cells on TAMs generation (Fig. 7c). The above results suggested that increased levels of extracellular lactate induced by H19 signaling modulated the activity of immune cells including γδT cells, Jurkat cells and TAMs, which might contribute to immune escape of GC cells.

Discussion
H19 has been identi ed to act as an important regulator in the tumorigenesis and progression of GC. For instance, up-regulated H19 was associated with lymph node metastasis and TNM stage of GC patients and promoted cell growth and metastasis via the miR-22-3p/Snail1 signaling pathway in GC [36]. H19derived miR-675 was demonstrated to enhance the proliferation and invasion of GC cells via RUNX1 [37].
Herein, ndings from our study were consistent with previous publications that H19 was highly expressed in GC tissues and associated with the poor prognosis of patients with GC. Using a nude-mouse xenograft tumor model, H19 knockdown was found to signi cantly suppress tumor growth in vivo. More importantly, we demonstrated that H19 modulated the aerobic glycolysis, proliferation, and immune escape of GC cells by sponging miR-519d-3p to induce the expression of LDHA (Fig. 8).
Previous studies noted that H19 participated in cancer cell aerobic glycolysis. The H19/let-7/HIF-1α signaling-mediated PDK1 could regulate glycolysis, and further contribute to breast cancer stem cell maintenance under hypoxic conditions [21]. In ovarian cancer, H19 overexpression increased glucose consumption, lactate production and PKM2 expression in SKOV3 and A2780 cells treated with ginsenoside 20(S)-Rg3 [30]. The above studies showed the important roles of H19 in the regulation of aerobic glycolysis in cancers, but the effect of H19 on aerobic glycolysis in GC remains largely unknown. In this study, we found that H19 knockdown decreased the glucose consumption and lactate production in SGC-7901 and MKN-45 cells. Aerobic glycolysis is driven by a spectrum of key glycolysis-related genes, including GLUT1, LDHB, LDHA, HK2, PKM2, and PDK1. Among these glycolysis-related genes, the mRNA and protein levels of LDHA were positively associated with H19 levels in both SGC-7901 and MKN-45 cells. Importantly, LDHA was highly expressed in GC samples [38,39], and involved in glucose metabolism, proliferation, and migration of GC cells [40,41]. Therefore, we inferred that H19 participated in GC aerobic glycolysis by regulating LDHA. The results of western blot indicated that LDHA overexpression abolished the effect of H19 suppression on the protein expression of LDHA in SGC-7901 and MKN-45 cells. More importantly, LDHA overexpression reversed the H19 depletion-induced decreases in glucose consumption and lactate production in SGC-7901 and MKN-45 cells. Our data suggested that the important effect of H19 on GC aerobic glycolysis is LDHA-dependent. Given that H19 was involved in regulating glycolysis in breast cancer stem cells under hypoxic conditions [21], we can not explain the roles of H19 in modulating aerobic glycolysis in GC cells under hypoxic conditions in the current study. Further investigations are required to answer this question.
It is well known that the lncRNAs' biological function is largely dependent on their subcellular localization [42]. Hence, to explore how H19 regulated LDHA expression in GC cells, we performed cytoplasmic and nuclear separation experiments with RT-qPCR analysis to determine the subcellular localization of H19 and LDHA. The results showed that H19 and LDHA were co-localized in the cytoplasm. Therefore, we speculated that H19 may function as a ceRNA to regulate LDHA expression by sponging one or several speci c miRNAs. Subsequently, bioinformatics analysis indicated that miR-519d-3p, miR-216b-5p, and miR-17-5p could bind to both H19 and LDHA. Moreover, there was a negative correlation between miR-519d-3p and H19 or LDHA in GC cells. Importantly, miR-519d-3p functioned as a tumor suppressor in multiple malignant tumors, such as colorectal cancer [43], breast cancer [44], pancreatic ductal adenocarcinoma [45], and GC [46]. Herein, in vitro luciferase assays indicated that both H19 and LDHA were direct targets of miR-519d-3p. Furthermore, we con rmed that miR-519d-3p overexpression decreased, while miR-519d-3p knockdown increased, the levels of H19 and LDHA in SGC-7901 and MKN-45 cells. Moreover, miR-519d-3p knockdown abolished the effect of H19 suppression on LDHA expression, glucose consumption, lactate production, and cell proliferation in GC cells. The above results revealed that H19 could regulate glycolysis and cell proliferation through the miR-519d-3p/LDHA axis.
As a major glycolytic metabolite, lactate functioned as a modulator of immune response and tumor progression [24,25,34]. LDHA-associated lactate accumulation blunted tumor surveillance by inhibiting the function and survival of T and NK cells [25]. Chen et al. indicated that lactate could activate macrophage G protein-coupled receptor 132 (Gpr132) to promote the alternatively activated macrophage (M2)-like phenotype [47]. Furthermore, lactate was markedly up-regulated in GC tumor-in ltrating T cells and associated with decreased T helper (Th)1 cell and cytotoxic T lymphocytes (CTLs) [48]. Based on our data that H19 affected the lactate accumulation, we concluded that H19 might modulate tumor immune response in a lactate dependent manner. Our and others' studies showed that γδT cells have important roles in tumor immunosurveillance against multiple malignancies such as colorectal cancer and leukemia [29,49]. Importantly, tumor-in ltrating γδT cells were an independent prognostic factor in GC patients and could predict the survival bene t of adjuvant chemotherapy in patients with TNM II and III diseases [50]. Therefore, we explored the effect of CM from stable knockdown H19 GC cells on γδT cells. Our results indicated that CM from stable knockdown H19 GC cells elevated the population of IFN-γ + γδT cells, while lactate treatment counteracted this effect. In addition, we also asked whether CM from stable knockdown H19 GC cells could affect T cells and TAMs via lactate. The addition of lactate abolished the increase in IL-2 expression in PMA and lonomycin-activated Jurkat cells induced by CM from stable knockdown H19 GC cells. Moreover, treatment with lactate reversed the effect of CM from stable knockdown H19 GC cells on TAMs generation. These data indicated that H19 signaling-mediated extracellular lactate accumulation could modulate the activity of immune cells including γδT cells, T cells, and TAMs, which might contribute to the immune escape of GC cells. Nevertheless, the effect of the H19/lactate axis on other immune cells, such as NK cells, myeloid-derived suppressor cells (MDSC), and Foxp3 regulatory T cells (Treg), are needed to investigate in our future study.

Conclusions
In summary, our results demonstrated the novel role of H19 in modulating aerobic glycolysis, proliferation, and immune escape of GC. Moreover, H19 mediated the effect on aerobic glycolysis, proliferation, and immune escape of GC cells via the miR-519d-3p/LDHA axis.     miR-519d-3p inhibitor, mimic NC, or miR-519d-3p mimic. β-actin served as a loading control. e-f The mRNA (e) and protein (f) levels of LDHA were measured in both sh-H19 SGC-7901 and MKN-45 cells after transfection with inhibitor NC or miR-519d-3p inhibitor. β-actin served as a loading control. g-h Glucose consumption (g) and lactate production (h) were examined in both sh-H19 SGC-7901 and MKN-45 cells after transfection with inhibitor NC or miR-519d-3p inhibitor. Each experiment was performed in triplicate.