LncRNA KCNQ1OT1 regulates microRNA-9-LMX1A expression and inhibits gastric cancer cell progression

LMX1A (LIM homeobox transcription factor 1α) is a tumor suppressor protein. Our previous study has shown that microRNA-9 (“miR-9”), being upregulated in human gastric cancer (GC), targets LMX1A to promote GC cell progression. Through searching long non-coding RNA (LncRNA) database, we identified that LncRNA KCNQ1OT1 is the competing endogenous RNA (ceRNA) of miR-9. KCNQ1OT1 putatively targets miR-9. Its level is downregulated in human GC tissues. In AGS cells and primary human GC cells, forced overexpression of KCNQ1OT1, by a lentiviral construct, induced miR-9 downregulation and LMX1A upregulation. Furthermore, KCNQ1OT1 overexpression inhibited GC cell survival, proliferation, migration and invasion, but inducing apoptosis activation. Contrarily, KCNQ1OT1 silencing, by targeted siRNAs, induced miR-9 accumulation and LMX1A downregulation. Consequently, GC cell proliferation, migration and invasion were enhanced. Importantly, KCNQ1OT1 overexpression or silencing was ineffective in LMX1A knockout AGC cells. Taken together, KCNQ1OT1 inhibits GC cell progression via regulating miR-9 and LMX1A expression.


INTRODUCTION
Gastric cancer (GC) is a common cancer, causing significant human mortalities each year [1]. In the past decades, due to the developments of sanitation, control of Helicobacter pylori infection and GC screenings, its incidence is declined [2,3]. However, the prognosis for the metastatic, recurrent, and other advanced GCs is not optimistic [1]. Identification of novel therapeutic targets and biomarkers for GC is extremely important [4,5]. It is the research focus of our group [6,7].
LMX1A (LIM homeobox transcription factor 1α) is the primary member of LIM-homeodomain conserved family transcription factors [8]. LMX1A regulates a number of physiological and cancerous behaviors [8]. LMX1A is downregulated in GC [7] and many other cancers [9,10]. Our group has shown that LMX1A exerted tumor-suppressive functions in GC cells [6,7,11]. We found that LMX1A inhibited c-Myc expression to exert tumor suppressive function in GC cells [6]. Also, LMX1A inhibited GC cell metastasis by inhibiting betacatenin-dependent genes [11]. microRNAs (miRNAs) are small, noncoding RNAs, regulating target gene expression by suppressing mRNAs translation and/or inducing their degradation. Our previous study has identified a LMX1A-targeting miRNA: microRNA-9 (miR-9). We show that miR-9 selectively targets and negatively regulates LMX1A to promote GC cell progression [7]. In human GC tissues, miR-9 upregulation is correlated with LMX1A downregulation [7]. The mechanism of miR-9 upregulation in human GC is still elusive.
Long non-coding RNAs (LncRNAs) are a large family of conserved non-coding RNAs (ncRNAs) with over AGING 200-nucleotide long [12][13][14]. LncRNA can decrease target miRNA expression by acting as their sponges [15]. LncRNAs regulate key cellular activities, including genomic imprinting, cell proliferation and survival, cell cycle control and differentiation as well as apoptosis and transformation [12][13][14]. Accumulative studies have confirmed that dysregulation of LncRNA plays a pivotal role in the progression of GC [15,16] and other human cancers [17,18]. One aim of this study is to identify possible miR-9-targeting LncRNA, regulating LMX1A expression and GC cell functions.
LncRNA KCNQ1OT1, or potassium voltage-gated channel subfamily Q member 1 (Kcnq1) overlapping transcript 1, is an imprinted antisense LncRNA [19,20]. It is an overlapping transcript of Kcnq1, locating at Kcnq1 loci on chromosome 11p15.5, being exclusively transcribed from the paternal chromosome [19,20]. Recent studies have shown that KCNQ1OT1 dysregulation participates in carcinogenesis and progression of human cancers [21,22]. Its expression and potential functions in GC are largely unknown. In the present study, we will show that LncRNA KCNQ1OT1 inhibits GC cell progression possibly via regulating miR-9 and LMX1A expression.

LncRNA KCNQ1OT1 is downregulated in human GC tissues
It is possible that miR-9 upregulation in GC (see our previous study [7]) could be due to downregulation of certain LncRNAs. As they could function as the ceRNAs of miR-9. To explore this possibility, the human LncRNA database, LncBase (Predicted v.2), was consulted to identify possible miR-9-targeting LncRNAs. These potential LncRNAs were further verified by other LncRNAs/miR databases, including StarBase and miRbase. The bioinformatic analyses identified that one LncRNA, KCNQ1OT1, putatively targeted miR-9. Its expression in GC tissues was tested next. Total RNA was extracted from fresh GC tissues and the adjacent normal epithelial tissues from twelve (12) primary GC patients [7]. KCNQ1OT1 was determined by qPCR assay, using the previouslydescribed primers [23].
The results show that KCNQ1OT1 levels are significantly downregulated in cancer tissues ("Can") ( Figure 1A), as compared to its levels in the surrounding epithelial ("Epi") tissues ( Figure 1A). Therefore, in GC cancer tissues KCNQ1OT1 downregulation correlated with miR-9 upregulation (see the previous results from same set of tissue samples [7]).

Forced overexpression of LncRNA KCNQ1OT1 inhibits AGS cell progression in vitro
LMX1A regulates a number of physiological and pathological processes [8]. It is considered as a tumor  (12) different human gastric cancer (GC) tissues ("Can") and matched surrounding normal epithelial tissues ("Epi") was shown, results were normalized to U6 RNA. All values were expressed as the mean ± standard deviation (Same for all Figures). *P <0.05 vs. "Epi" tissues. AGING suppressor [8,9,11]. LMX1A downregulation is detected in GC and many other cancers [9,10]. Our previous study has shown that LMX1A overexpression can inhibit AGS cell progression in vitro. We therefore analyzed the functional consequences of KCNQ1OT1 overexpression.

Forced overexpression of LncRNA KCNQ1OT1 induces miR-9 reduction and LMX1A upregulation in primary human GC cells
Next, we tested whether LncRNA KCNQ1OT1 exerted similar functions in primary human GC cells. As described previously [7], the primary human GC cells, derived from two independent patients (namely "GC-1/GC-2"), were infected with LV-KCNQ1OT1, which resulted in 5-6 folds increase of KCNQ1OT1 expression ( Figure 3A). On the contrary, miR-9 levels were significantly decreased ( Figure 3B), with LMX1A mRNA ( Figure 3C) and protein ( Figure 3D) levels increased in KCNQ1OT1-overexpressed GC cells. LMX1B protein expression was unaffected by KCNQ1OT1 overexpression ( Figure 3D).
Functional studies show that KCNQ1OT1 overexpression by LV-KCNQ1OT1 inhibited EdU incorporation ( Figure 3E) and cell viability ( Figure 3F) in primary human GC cells. Furthermore, LV-KCNQ1OT1-expressing primary GC cells presented with decreased cell migration and invasion (vs. Vector control cells, Figure 3G, 3H). Contrarily, the increased TUNEL staining ( Figure 3I, vs control cells) and Annexin V ratio ( Figure 3J) were detected in LV-KCNQ1OT1-expressing primary GC cells, indicating apoptosis activation. Therefore, ectopic KCNQ1OT1 overexpression inhibited viability, proliferation, migration and invasion, but inducing apoptosis activation in primary GC cells.

In LMX1A-knockout AGS cells altering KCNQ1OT1 expression fails to change cell viability and proliferation
To verify that LMX1A is one target of KCNQ1OT1, we utilized the CRISPR/Cas9 gene-editing method to knockout LMX1A (see our previous study [7]). Two stable LMX1A knockout ("Cas9-LMX1A-ko") AGS cell lines were established (namely "S1/S2") [7]. Results in Figure 5A confirmed LMX1A protein depletion in the stable cells, with LMX1B unaffected ( Figure 5A). LV-KCNQ1OT1 (see Figure 2) or KCNQ1OT1 siRNA ("seq2", see Figure 3) were tranduced to the Cas9-LMX1A-ko cells, both significantly altered KCNQ1OT1 expression ( Figure 5B). Significantly, neither LV-KCNQ1OT1 nor KCNQ1OT1 siRNA altered viability ( Figure 5C) and proliferation (EdU ratio, Figure 5D) in Cas9-LMX1A-ko AGS cells. Thus, in LMX1A-knockout AGS cells KCNQ1OT1 did not change cell viability and proliferation, indicating that LMX1A is one target of KCNQ1OT1. AGING AGS cells were infected with KCNQ1OT1-expressing lentivirus ("LV-KCNQ1OT1") or scramble control vector lentivirus ("Vec") for 24h, following puromycin selection two stable lines ("sL1/sL-2") with LV-KCNQ1OT1 established, expression of KCNQ1OT1 (A), miR-9 (B) and LMX1A mRNA (D) were tested by qPCR assays; The LMX1A 3'-UTR luciferase activity was shown (C); Expression of the listed proteins in total cell lysates were tested by Western blotting assay (E, results quantified in F); Cells were further cultured for the indicated time periods, cell survival and proliferation were tested by CCK-8 assay (G) and EdU staining (H), respectively; Cell migration and invasion were tested by "Transwell" (I) and "Matrigel Transwell" assay (J), respectively; Cell apoptosis was tested by Western blotting (testing apoptosis-associated proteins,   miR-9 (B) and LMX1A mRNA (C) were tested by qPCR assay; LMX1A and LMX1B protein expression in total cell lysates was tested by Western blotting assay (D, LMX1A protein results were quantified); Cells were further cultured for the indicated time periods, cell proliferation and viability were tested by EdU staining assay (E) and CCK-8 (F), respectively, with cell migration and invasion tested by "Transwell" and "Matrigel Transwell" assay (G and H); Cell apoptosis was quantified via the TUNEL staining assay (I) and Annexin V-FACS assay (J). For each assay, n=5. *P <0.05 vs. "Vec" cells. Experiments in this figure were repeated three times, and similar results were obtained. Bar=100 μm (E and G). siRNA ("seq1/seq2", two rounds, total 48h) or the scramble control siRNA ("si-C", two rounds, total 48h), expression of KCNQ1OT1 (A and H), miR-9 (B and I) and LMX1A mRNA (C and J) were tested by qPCR assays; LMX1A and LMX1B protein expression was tested by Western blotting assay (C and K, LMX1A protein results were quantified); Cells were further cultured for the indicated time periods, cell viability (D) and proliferation (E and L) were tested by the listed assays; Cell migration and invasion were tested by "Transwell" (F) and "Matrigel Transwell" (G) assays respectively. For each assay, n=5. *P <0.05 vs. "si-C" cells. Experiments in this figure were repeated three times, and similar results were obtained. Bar=100 μm ( E-G). AGING

DISCUSSION
Studies have proposed a possible role of KCNQ1OT1 in carcinogenesis and progression of multiple cancers [21][22][23]. Recently, a high level of KCNQ1OT1 was detected in lung cancer, promoting proliferation and invasion as well as chemoresistance of lung cancer cells [23,25]. Similarly, KCNQ1OT1 upregulation and mutation are associated with progression of hepatocellular carcinoma (HCC) and glioma [26,27]. Feng et al., reported that KCNQ1OT1 upregulation in breast cancer is important for cancer cell growth in vitro and in vivo [21]. Li et al., however demonstrated that KCNQ1OT1 downregulation attenuated myocardial ischemia/reperfusion injury from acute myocardial infarction, possibly by regulating adiponectin receptor 1 (AdipoR1)-p38 /NF-κB signaling cascade [28]. Here, we show that KCNQ1OT1 expression is downregulated in human GC tissues, which correlated with miR-9 upregulation.

Chemicals and reagents
Puromycin, neomycin and polybrene were obtained from Sigma-Aldrich (St. Louis, Mo). All antibodies were purchased from Abcam (Cambridge, MA). RNA reagents, Lipofectamine 2000 and other transcription reagents were obtained from Invitrogen (Shanghai, China). All sequences and plasmids were provided by Genechem (Shanghai, China).

Cell culture
AGS cells were cultured as previously described [7]. The primary human GC cells, derived from two independent written informed-consent GC patients ("GC-1" and "GC-2"), were cultured in the described medium [31]. The protocols were approved by the Ethics Board of Minhang Hospital, Fudan University (Ethics number: 2015-052, Principle Investigator, Dr. Feng Li), in according to Declaration of Helsinki.

Human tissues
As described previously [7], fresh GC cancer tissues along with matched surrounding normal gastric epithelial tissues from twelve (12) primary informedconsent GC patients (all male, 41-67 years old, with no prior chemotherapy and radiotherapy) were acquired. Tissues were washed, minced, and homogenized before further analysis.

KCNQ1OT1 siRNA
GC cells were seeded at 50% confluence. Two different siRNAs (designed and verified by RIBOBIO, Guangzhou, China [24]) against non-overlapping sequence of KCNQ1OT1 were individually transfected (siRNA concentration at 500 nM) by Lipofectamine 2000 reagent for 24h, repeated for a second round for another 24h (total 48h). KCNQ1OT1 knockdown was confirmed by qPCR assay. Control cells were transfected with the scramble non-sense control siRNA ("si-C").

Luciferase reporter assay
As reported [7], the human LMX1A 3'-UTR with the putative binding sites of miR-9 (see the sites at [7]) was amplified by PCR and then inserted into the firefly luciferase reporter vector, pGL4.13 (luc2/SV40) (Promega), at the XbaI site and downstream from the stop codon of the luciferase gene. AGS cells were plated at 1 × 10 5 cells per well into six-well plates (60% confluence), transfected with pGL4.13 construct along with the Renilla luciferase reporter vector and pRL-SV40 (Promega). Following the indicated genetic treatments, LMX1A 3'-UTR luciferase activity value was tested, and results were normalized to that in the control cells.

In vitro cell migration and invasion assays
As described [33], GC cells were seeded on "Transwell" upper chamber (at 3 × 4000 cells per chamber, BD Biosciences). The complete medium (with 10% FBS) was added to the lower compartments. After 24h the migrated cells on the lower surface were stained. Matrigel (Sigma) was added in the chamber surface when analyzing cell invasion. Five repeated views in each condition were included to calculate the average number of migrated/invasive cells.

EdU assay
EdU (5-ethynyl-20-deoxyuridine) Apollo-567 Kit (RIBOBIO, Shanghai, China) was employed to test cell proliferation. Briefly, cells were seeded onto the twelvewell plates. Following treatment, EdU and DAPI dyes were added, and cells were incubated for additional 8h. Cell nuclei were then visualized under a fluorescent microscope. For each condition at least 500 cells in five random views were included to calculate EdU ratio (EdU/DAPI×100%).

Apoptosis assays
Testing cell apoptosis, by the Annexin V FACS and TUNEL staining assay, was described in detail in our previous study [7].

LMX1A knockout
The two different lentiCRISPR-GFP-puro LMX1A knockout constructs (see our previous study [7]) were independently transfected to AGS cells by Lipofectamine 2000. GFP-positive cells were FACS-sorted, cultured for another three weeks. Stable cells were achieved by puromycin selection, named as "Cas9-LMX1A-ko" cells.

Western blotting
Total cellular lysates were resolved by SDS-PAGE (10%) gels, transferred to a PVDF membrane. The latter was incubated with blocking buffer and following primary and secondary antibodies. ECL method was applied to detect the immuno-reactive bands.

Statistical analyses
The GraphPad Prism software was employed for statistical analyses. All values were expressed as the mean ± standard deviation (SD). Repeated-measures analysis of variance (RMANOVA) followed by Dunnett's post hoc test for multiple comparisons were applied to evaluate statistical significance. All differences were considered significant at P < 0.05. To determine significance between two treatment groups, the two-tailed t tests were carried out.

AUTHOR CONTRIBUTIONS
All listed authors designed the study, performed the experiments and the statistical analysis, and wrote the manuscript. All authors have read the manuscript and approved the final version.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interests.

FUNDING
This project was supported by Minhang District University Building Project (2017MWDXK03). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.