lncRNA-GPAND Promotes Gastric Cancer Progression via RUNX2 and MMP13

Background: Gastric cancer (GC) is still one major reason for cancer-related death worldwide and in China, which ranks the second highest common cancer death rate. It is of great importance to study the molecular mechanisms by which gastric cancer develops. Methods: In this study, in situ hybridization histochemistry (ISHH) was used to examine the lncRNA-GPAND expression levels using gastric cancer tissue array. The real-time live-cell imaging system was used to investigate the effect of GPAND on cell proliferation and apoptosis of GC cell lines. Cell cycle of AGS cell line was examined after GPAND was suppressed using the ow cytometry (FCM). Transwell method was used to study the effect of GPAND on the invasion characteristics of GC cell line. Then the next generation sequencing (NGS) was used to study the potential molecular mechanism and the pathway, and the RT-qPCR was performed to verify the potential targets found by NGS method. Results: It was shown that GPAND was signicantly over-expressed in the gastric cancer (GC) tissues (n=215) compared with the paired non-cancerous tissues (n=215), the expression levels of GPAND of GC tissues of TNM stage I-II (n=45) were signicantly higher than that of stage III-IV (n=147). It has shown that knockdown of GPAND inhibited the AGS and N87 cell proliferation and promoted the cell apoptosis of AGS and N87 cell lines signicantly, and the G1 phase percentage was remarkably increased in GPAND knockdown group of AGS cell line compared with control group. Moreover, suppression of GPAND inhibited the AGS cell invasion signicantly. It was found via the NGS method that RUNX2 and MMP13 were signicantly up-regulated when the GPAND was over-expressed. Conclusions: These observations suggest the lncRNA-GPAND/RUNX2/MMP13 axis to be a viable therapeutic target for gastric cancer. MMP13 were signicantly up-regulated in GPAND overexpression groups compared with control group via NGS technique and western blotting experiment. Our ndings suggest that GPAND/RUNX2/MMP13 might be the potential therapeutic targets of the gastric cancer. proliferation difference between GPAND siRNA transfected groups and the control groups. There was not remarkable difference between 50 nM and 100 nM GPAND-siRNA transfection groups. by system Proliferation, showed the signicant apoptosis cell difference between GPAND siRNA transfected There was not signicant difference GPAND-siRNA transfection


Introduction
As it has been reported, there were about 27,600 newly diagnosed cases and 11,010 cancer-related deaths for gastric cancer in the United States in 2020 [1]. Although the incidence and mortality of gastric cancer has steadily declined since 2000, it remains one of main important cause of cancer mortality in China, e.g. 15.1% for Chinese males, and 11.1% among females [2]. The current therapy, especially the target therapy for gastric cancer is rather unsatisfactory. Hence, to investigate the molecular mechanism promoting the tumorigenesis, progression, invasion and metastasis of gastric cancer is of great signi cance.
Long non-coding RNAs (lncRNAs) has been the research highlights for these years [3,4]. Previous studies have showed that lncRNAs are able to play important roles in gastric cancer [5] development via transcriptional and translational regulation of the oncogenes and cancer suppressor genes, which usually mediated by other non-coding RNAs such as microRNAs and circRNAs [6,7].
The runt-related transcription factor 2 (RUNX2) is an important transcription factor for embryogenesis [8], plays a major role at the late stage of chondrocyte differentiation [9], and impacts skeletal development via regulating the differentiation of chondrocytes and osteoblasts and the expression of many extracellular matrix protein genes including those of Col1a1, Col1a2, Spp1, Bglap, and Mmp13 during chondrocyte and osteoblast differentiation [10]. RUNX2 is abnormally up-regulated in prostate cancer [11,12] and breast cancer cell lines [13,14] and tissues [15]. High RUNX2 expression correlates with the triplenegative breast cancer subtype and with lower patient survival rates compared with samples displaying low RUNX2 expression [16]. Furthermore, it was also correlated with the drug-sensitivity of pancreatic cancer [17]. The high expression of Runx2 indicates a trend toward a poor survival rate of osteosarcoma [18] partly due to its oncogenesis role of osteosarcoma [19].
The matrix metalloproteinases (MMPs) are a family of structurally related enzymes playing a role in multiple physiological processes, i.e., embryonic, adult tissue development and remodeling, evaluation, uterine implantation, wound healing, etc, and MMPs may have the function in some pathological conditions, such as arthritis, atherosclerosis and cancer invasion and metastasis. Human collagenase-3 (MMP-13) is a matrix metalloproteinase involved in a variety of malignant tumors including gastric cancer [20,21], cervical cancer [22,23], skin melanoma, breast cancer [24], colon cancer [25], head and neck carcinomas [26], chondrosarcomas [27] and basal cell carcinomas of the skin [28]. The increased expression of MMP13 usually predicts metastasis and poorer outcome of malignant tumors. GPAND, i.e., the abbreviation for "gastric progression associated non-coding RNA", refers to the long noncoding RNA, AK125207. Here, we found that GPAND was signi cantly over-expressed in the gastric cancer tissues compared with the paired non-cancerous tissues. The depletion of GPAND inhibited the cell proliferation and invasion, and increased the cell apoptosis and G1 phase percentage of gastric cancer cells signi cantly. RUNX2 and MMP13 were signi cantly up-regulated in GPAND overexpression groups compared with control group via NGS technique and western blotting experiment. Our ndings suggest that GPAND/RUNX2/MMP13 might be the potential therapeutic targets of the gastric cancer.

Materials And Methods
Ethics.
The procedures performed in this study were approved by the Ethics Committee of the Shanghai Outdo Biotech Company, the member of National Human Genetic Resources Sharing Service Platform (Shanghai, China). The present study was performed in accordance with the ethical standards of The Institutional and National Research Committee and with The Declaration of Helsinki.

Tissue arrays and interpretation criteria for In Situ Hybridization ISH results
HStmA180Su15, HStmA180Su17 and HStm-Ade180Sur-03 tissue arrays (Outdo Biotech Co. Ltd., Shanghai, China) are used in this study to examine the GPAND expression levels. The hematoxylin-eosin staining and GPAND in situ hybridization of these tissue arrays can be seen in the S. le-1. The ISH results were interpreted by senior pathologists. The rating criteria for staining intensity was as follows: no staining was negative (-or 0), weak positive was light yellow (+or 1), moderate positive was brown (++ or 2), and strong positive was brown (+++ or 3). (Note: If there are both moderate and strong positive tissues, we usually record 2-3). If located in the nucleus, the positive rate of staining was calculated by recording 100 cells randomly in each eld of vision, and then the percentage of positive cells in 100 cells was recorded as ×1%. According to the same principle, the percentage of positive cells in the other two visual elds was×2% and ×3%, and the average of positive staining rates in the tissues was ×1%, ×2% and ×3%. If located in the cytoplasm or membrane, we also chose three different visual elds of staining intensity to calculate the positive rate and then take the average value. The standardization scheme of original experimental data is as follows: 1) Staining intensity score: 0 (negative), 1 (1+), 2 (2+), 3 (3+); 2) Positive staining percentage score: 0 (negative), 1 (1-25%), 2 (26-50%), 3 (51-75%) and 4 (76-100%); 3) Total score and grouping: grouping with the product of "staining intensity score" and "staining positive rate score" as the total score.

Constructs and transfection
The PCR product of GPAND was ampli ed using the GeneAmp PCR System 2400 thermal cycler rstly  Imaging System (Incucyte Zoom, Essen BioScience, USA). Cells were cultured with medium supplemented with siRNAs at the indicated doses for 4 hours and then the culture medium was replaced with the culture medium. Cell con uence was monitored and measured using the IncuCyte Zoom system (Essen Bioscience, Ann Arbor, MI, USA). The dynamic pictures were taken once per 2h to capture 3-4 elds of view/well/time for recording and was analyzed by IncuCyte ZOOM automatic software (Version: IncuCyte ZOOM 2016B, USA) and database.

Cell apoptosis analysis
The cell apoptosis was analyzed using the Real-Time Live-Cell Imaging System (Incucyte Zoom, Essen BioScience, USA) as mentioned above. In this section, CellEvent® Caspase-3/7 Green ReadyProbes® Reagent Thermo sher/MP, Cat.No.R37111, USA , e.g., Apo-488, a uorogenic, no-wash indicator of activated caspase-3/7 for live-and xed-cell applications, is used. Activation of caspase-3 is an early indicator of apoptosis and CellEvent® Caspase-3/7 Green reagent allows rapid and sensitive detection of cells destined for cell death. CellEvent® Caspase-3/7 Green reagent is a four amino acid peptide (DEVD conjugated to a nucleic acid-binding dye that is non-uorescent when not bound to DNA. The CellEvent® Caspase-3/7 Green reagent is intrinsically non-uorescent, as the DEVD peptide inhibits binding of the dye to DNA. Upon activation of caspase-3/7 in apoptotic cells, the DEVD peptide is cleaved and the free dye can bind DNA, generating a bright green uorescence. The uorescence emission of the dye when bound to DNA is 530 nm and can be observed using a standard FITC lter set.

Cell cycle analysis
Here, the Cell cycle detection kit KeyGEN BioTECH Cat.No.KGA511 China was used. 1×10 6 cells were collected, centrifugated, discarded the supernatant, washed with pre-cooled PBS, added with pre-cooled 70% ethanol, and xed overnight at 4°C, for each sample 48 hours after cell transfection. Then the cells were collected by centrifugation, washed once with 1mL PBS, added 500uLPBS containing 50ug/mL propidium bromide (PI), 100ug/mL RNase A, 0.2% Triton X-100, and were incubated at 4℃ in the dark for standard procedures and the results were analyzed with the cell cycle tting ModFitLT™ software (version 3.0; Verity Software House, Inc.).

Cell invasion analysis
The Matrigel (Cat.No.356234, BD Co.Ltd, USA) was dissolved overnight at 4°C, diluted with a pre-cooled serum-free medium at a volume ratio of 1:3, and then 40ul matrigel was added into the pre-cooled transwell chamber (Cat#REF353097, Co.Ltd USA), and was incubated at 37°C for 2 hours to be solidi ed. The excess liquid in the chamber was aspirated, 100ul and 600ul serum-free medium was added to the upper and lower chambers respectively, and was equilibrated overnight at 37°C. On the second day of cell transfection with siRNA, 1×105 cells were counted, resuspended with 100ul serum-free DMEM Hyclone Fluorometer was used to detect the concentration of the library, and the Agilent 4200 was used to detect the size of the library. The quality control standard of sequencing data is as follows: the amount of data is 6G/sample, and the ratio of the quality of each base greater than 20 (Q20) is not less than 85%. RNA-seq reads quality was evaluated using FastQC and mapped to the human genome reference assembly (hg38) using Bowtie 2. Signi cant differential expressed genes (DEGs) were identi ed as those with a False Discovery Rate (FDR) value above the threshold (q-value ≤ 0.05), fold-change ≥ 2 and FPKM > 5 using edgeR software. For functional analysis, the selected DEGs were mapped to each term of the GO and KEGG databases, the number of genes of each item was calculated, and then a hypergeometric test was applied to screen the GO or KEGG pathway terms with signi cant enrichment. In addition, the data were analyzed using Ingenuity Pathway Analysis® (IPA, www. ingenuity.com) to further identify biological processes enriched for the DEGs. For pathway analysis, MapMan package was used, and Volcano plots Ggrepel packges was used to make volcano plot.
Reverse transcription-quantitative PCR (RT-qPCR) Total RNA was extracted from the gastric and normal gastric cell lines using the TRIzol® reagent (Invitrogen; Thermo Fisher Scienti c, Inc.) and examined on a NanoDrop machine according to the manufacturer's protocols (Eppendorf). Electrophoresis and spectrophotometric methods were used to examine the concentrations and purity of the RNA samples. cDNA was prepared using ImProm-II™ Reverse Transcription System according to manufacturer's protocols (cat. no. A3800; Promega Corporation). The temperature protocol of reverse transcription reactions was as follows: i) mRNA, 85˚C for 5 min, 30˚C for 10 min, 42˚C for 60 min and 85 ˚C for 10 min; and ii) 85˚C for 5 min, 42˚C for 60 min and 85˚C 10 min for miRNA. The Taq plus Master mix (Vazyme P212-01 China) was used for qPCR and each reaction was performed in triplicate. The thermocycling conditions used for RT-qPCR were as follows: Initial denaturation at 95˚C for 5 min, followed by 40 cycles of 95 ˚C for 15 sec and 60˚C for 34 sec. The 2-∆∆Cq method [29] was utilized to calculate the relative expression levels of mRNA and miRNA (primers used can be seen in Table S3), using GAPDH and U6 as reference genes for mRNA and miRNA, respectively(Primers used were listed in Table S7-S8).

miRNA target prediction
In order to predict potential miRNA target genes, microRNA.org website (http://www.microrna.org/microrna/home.do) was used, the binding free energy was calculated and biding sites were analyzed using http://bibiserv.techfak.uni-bielefeld.de/rnahybrid website.

Statistics
Data were expressed as Mean±Sd or Mean±SEM (n=3) (According to the statistics method used, see in the legends). SPSS statistical software package, version17.0 (SPSS, Chicago, IL, USA) was used. The wilcoxon signed-rank test was used to compare the immunohistochemistry scores of cancer and matched adjacent non-cancerous tissues. The one-way ANOVA followed by Tukey's post hoc test were performed for cell proliferation and apoptosis experiments. P values < 0.05 were considered statistically signi cant.

Up-regulation of GPAND in GC tissues and cell lines
In Situ Hybridization (ISH) method was used to evaluate the expression levels of GPAND of gastric cancer (GC) tissues using GC tissue arrays (Fig.1, the hematoxylin-eosin staining and GAPDH in situ hybridization of tissue arrays can be seen in the S. Fig-1). It was shown that GPAND was expressed in the nucleus, membrane and cytoplasm of gastric cancer cell and was signi cantly up-regulated in the gastric cancer tissues (n=215) compared with the paired non-cancerous tissues (n=215) (Fig.2A, p < 0.0001). Further analysis revealed that the expression levels of GPAND of GC tissues of TNM stage I-II (n=45) were signi cantly higher than that of stage III-IV (n=147) (Fig.2B, p < 0.001). Tumor volume subgroup analysis showed that GPAND was remarkably up-regulated in the tissues of tumor volume ≥100cm 3 (n=46) compared with that of ≤15cm 3 (n=73) (Fig.2C, p < 0.05). However, it was not showed that GPAND expression levels of patients aged ≤55y (n=44) was signi cantly different than that aged ≥75y (n=50) (Fig.2D). Interestingly, when compared with the survival of patients with GPAND staining score of 10-12 (n=23) (Fig.2E), the survival of those scored 0 (n=15) was not remarkably different with them. Furthermore, there was not signi cant difference of the GPAND staining score between male (n=161) and female (n=75) (Fig.2F). We then examined the expression levels of GPAND using several GC cell lines and found that it was signi cantly up-regulated in the MGC803, NCI-N87, AGS, OACP4C, and SK-GT2 cell lines compared with normal gastric mucosa cell line, GES-1 (Fig.2G). All the staining score of Fig.2 can be seen in the Table S1-S6, respectively.

Depletion of GPAND inhibited the cell proliferation and promoted apoptosis of GC cell lines
In this section, to assess the role played by GPAND on the proliferation of GC cell lines, two methods i.e., the real-time live-cell imaging system analysis and the cell counting kit-8 were used. Here, AGS and SK-GT2 cell line were analyzed after GPAND was knocked down. To evaluate the proliferation difference, cells transfected with negative control (NC) siRNA and the cells cultured using culture medium supplemented with the same amount of transfection media, i.e., the blank control, were used as the control groups. We found that depletion of GPAND inhibited signi cantly the cell proliferation of AGS

GPAND promoted cell cycle of AGS cell line
According to the nding that depletion of GPAND could inhibit the growth and promote the apoptosis of GC cell lines, we speculated that maybe GPAND could affect the cell cycle of these cell lines. Hence, we performed the ow cytometry experiment to study this speculation. We found that after GPAND was knockdown using two siRNAs, i.e., si-1 and si-2, which showed effective depletion of GPAND in the AGS cell line S. Fig-2 (Fig.5C-D). Furthermore, G1 phase percentage was also signi cantly increased in GPAND knockdown group (si-1: 31.24±0.599 vs 16.99±0.554, p < 0.001; si-2:32.36±0.320 vs 16.99±0.554, p < 0.001) compared with control group at 24h (Fig.5A-B).

Suppression of GPAND inhibited the cell invasion of AGS cell line
In addition to study the roles played by the GPAND on the proliferation, apoptosis and cell cycle of GC cell lines. It is of interesting that if the GPAND could affect the invasion of GC cell line. We then investigated the effect of suppression of GPAND using both siRNA-1 and siRNA-2 on the invasion characteristic of AGS cell line at 48h and 72h. It was found that the suppression of GPAND inhibited the AGS cell invasion signi cantly compared with the negative control group at 48h (si-1 vs NC: 130.250±8.498 vs 195.375±20.311 , p < 0.001; si-2 vs NC:105.125±6.334 vs 195.375±20.311, p < 0.001) (Fig.6A-B) (Fig.6C-D).

Downstream molecular change induced by depletion of GPAND
Given that the roles played by GPAND in the proliferation, apoptosis, cell cycle and invasion of AGS cell line, in order to explore the potential molecular mechanism by which GPAND, the next generation sequencing method was used to analyze the relative gene pro le change after the cDNA of GPAND was transfected compared with the negative control. Here, we found that MMP13 and RUNX2 gene were upregulated in the GPAND up-regulated group compared with the control group (Table 1 & 2, Fig.7A-B). The representative graph of the potential mechanism by which GPAND-RUNX2-MMP13 axis promoted the invasion and metastasis of gastric cancer is shown in the gure 8.

Discussion
AK125207 was screened by us using the commercial microarray (Arraystar, USA) 7 years ago according to the coding-potential predicting tool "http://lilab.research.bcm.edu/cpat/" , it was predicted as noncoding RNA. Furthermore, the source by which AK125207 was acquired is http://nred.matticklab.com/cgibin/ncrnadb.pl and the annotation of the AK125207 awas non-coding RNA. In this study, to explore the expression levels of AK125207 (named as GPAND) in the gastric cancer tissues, we used the commercial tissue arrays via in-situ hybridization histochemistry (ISHH) method other than using the gastric cancer tissues via the real-time quantitative PCR examination was rstly because of the di culty for us to acquire the enough gastric cancer samples with the survival follow-up outcome at the beginning of this study. In order to investigate the clinical meaning of GPAND, it is of great importance to use the samples with the survival data. Hence, we bought the commercial tissue arrays to meet our study needs. Secondly, ISHH technique makes it possible for us to further know the localization of GPAND on the cancer cells, however, RT-qPCR method only provide the quantitative analysis results. Thirdly, in our study positive staining score was evaluated by assessing the relative intensity and scope of positive staining (see in the method section) in the cancer cells area which combined both the staining intensity and the scope factors, providing more meaning and data than RT-qPCR method in our view. Of course, the shortcoming of tissue arrays is obvious for the random and uncertain sampling, hence, to investigate enough amount samples is very important and we here analyzed 215 samples and we were lucky to some extent that the expression level score of GPAND of gastric cancer tissues were found to be signi cantly up-regulated compared with the normal gastric tissues. However, we did not conclude that the survival was correlated with the expression levels of GPAND. Hence, it is of interesting that if more samples were used for analysis, could the remarkable survival difference be observed? We expect to answer this question in the following study.
After observing the difference of expression levels of GPAND in the gastric cancer tissues and normal tissues, we then investigate the roles played by the GPAND of the proliferation, apoptosis, cell cycle and invasion characteristics in the gastric cancer cells. To our surprise, the knock down of GPAND led to the inhibition of proliferation, promotion of apoptosis, increase of G1 phase, and stronger invasive capability. Especially, to make it sure that the depletion of GPAND may affect the proliferation of gastric cancer cells, two methods, i.e., CCK-8 and real-time live-cell imaging system experiments were performed for at least 3 times using two siRNAs targeted the GPAND, for this step is very critical for the further investigation of the roles of the GPAND. Although only SK-GT2 cell line was used to perform CCK-8 experiment, it also could support the view of the function of GPAND on the proliferation of gastric cancer cells to some extent in order to verify the results obtained by the live-cell imaging system.
Due to the observation that the GPAND maybe exert its oncogenic role in the development of gastric cancer, it is of great interest to explore its molecule mechanism by which the GPAND may function as the oncogene. Hence, the next generation sequencing was performed to study the potential mechanism. In this section, both siRNA-1 and siRNA-2 were used to knock down the GPAND, and it is of great meaning that MMP13 and RUNX2 were overexpressed compared with the negative control group. As far as we know, MMP13 functions as the oncogene of gastric cancer, colon cancer, breast cancer and etc. RUNX2 promotes the development of multiple tumors such as prostate cancer and breast cancer and is a negative prognostic indicator of the survival of osteosarcoma.
RUNX2, a mesenchymal stem marker [30], was reported by Chao Han et al that the high expression of RUNX2 and low expression of FN1 analyzed using the GEO2R tool could be the poor predictive marker of long survival in diffuse, poorly differentiated, and lymph node-positive GC [31]. Moreover, ZJ Guo et al also found that the levels of RUNX2 expression in gastric cancer tissues were correlated with the differentiation degrees, invasion depth and lymph node metastasis. COX regression analysis indicated that RUNX2 was an independent prognostic indicator for GC patients [32]. It was also found that the RUNX2/OPN axis regulated the ability of osteosarcoma cells to attach to pulmonary endothelial cells as a key step in metastasis of osteosarcoma cells to the lung [33]. These ndings supported to some extent our research that maybe the long non-coding RNA GPAND exerted its oncogene role via RUNX2 and MMP13.
More interestingly, we also examined the mRNA expression levels of MMP13 and RUNX2 in the gastric cancer cell line and found that both of them were signi cantly up-regulated in the gastric cancer cell line compared with the gastric mucosal cell line, GES-1. Furthermore, we also examined the roles played by the MMP13 and RUNX2 in the gastric cancer cell line (data not shown), the potential mechanism by which the GPAND affect the MMP13 and RUNX2 expression are being studied and these ndings will be reported in the next article.
In conclusion, in this study, we preliminarily reported that lncRNA GPAND may function as the oncogene in the development of gastric cancer via MMP13 and RUNX2. The GPAND might be a potential therapeutic target of gastric cancer if the regulatory mechanism by which it affected the MMP13 and RUNX2 were more elucidated.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
All authors agree to publish this paper.

Competing interests
No potential con ict of interest was reported by the authors.
Authors' contributions Y W initiated the project, designed the experiments, wrote the manuscript and interpreted the data. P X nished the data analysis. CB W, and ZB T performed the experiments. All authors read and approved the nal version of the manuscript.

Funding
The present study was supported by the National Natural Science Foundation of China (No. 81772631), Shenzhen Science and Technology Innovation Committee (Grant no. JCYJ20190814121001751, JCYJ20190814110203636).

Availability of data and materials
The data and materials are available.  Knockdown of GPAND inhibited the cell proliferation and promoted apoptosis of GC cell lines, SK-GT2.
One-way ANOVA followed by Tukey's post hoc test was performed in this section to compare the effect of GPAND knockdown on the proliferation and apoptosis of SK-GT2 cell line. In this section 100 nM and 50 nM siRNA-2 (Table.S1) was used to suppress the GPAND expression. A. The phase object con uence of SK-GT2 cells, measured by the real-time live-cell imaging system proliferation, showed the signi cant Page 22/27 proliferation difference between GPAND siRNA transfected groups and the control groups. There was not remarkable difference between 50 nM and 100 nM GPAND-siRNA transfection groups. B. The green object count, measured by the real-time live-cell imaging system Proliferation, showed the signi cant apoptosis (the green uorescence indicated cell apoptosis) difference between GPAND siRNA transfected groups and the control groups. There was not signi cant difference between 50 nM and 100 nM GPAND-siRNA transfection groups. C. Representative graphs of cell proliferation and apoptosis taken by the realtime live-cell imaging system (Bar: 300 μm). D. The CCK-8 experiment showed that the GPAND suppression inhibited the growth of SK-GT2 cell line from the third day after siRNA transfection was performed. NC: siRNA negative control; Blank ctrl: Blank control; GPAND si: GPAND siRNA; CDDP: cisplatin ***: P < 0.001; **: P < 0.01.   Molecular change induced by cDNA transfection of GPAND A.Heat map of the next generation sequencing of GPAND up-regulation after cDNA was transfected for 36h and negative control groups. B.
Heat map of the next generation sequencing of GPAND depletion after cDNA was transfected for 48h and negative control groups.

Figure 8
The representative graph showing the potential mechanism by which GPAND-RUNX2-MMP13 axis promoted the invasion and metastasis of gastric cancer.