Small nucleolar RNA host gene 1 promotes development and progression of colorectal cancer through negative regulation of miR‐137

Abstract Small nucleolar RNA host gene 1 (SNHG1) is critical in the progression of cancers. However, the mechanism by which SNHG1 regulates the progression of colorectal cancer (CRC) remains unclear. Expressions of SNHG1 and miR‐137 in CRC tissues and cell lines were evaluated by quantitative real‐time polymerase chain reaction. A luciferase reporter gene assay was conducted to investigate miR‐137 target. Additionally, RNA pull‐down assay was performed to explore the physical association between miR‐137, SNHG1, and RNA induced silencing complex (RISC). Cell cycling and invasion were examined by flow cytometry (FCM) and transwell assays. The in vivo carcinogenic activity of SNHG1 was examined using murine xenograft models. Expression of RICTOR, serine/threonine kinase 1 (AKT), serum and glucocorticoid‐inducible kinase 1 (SGK1), p70S6K1, and LC3II/LC3I ratio was examined by Western blot analysis. SNHG1 upregulation was observed in CRC tissues and cell lines, which was associated with the lymph node metastasis, advanced TNM stage and poorer prognosis. SNHG1 increased RICTOR level in CRC via sponging miR‐137. In addition, SNHG1 silencing inhibited CRC cell proliferation and migration in vitro and in vivo. SNHG1 regulated RICTOR expression by sponging miR‐137 and promoted tumorgenesis in CRC.


| INTRODUCTION
Colorectal cancer (CRC) is the third most common malignant tumor in humans, and is also the most common malignant gastrointestinal tumor. 1 Although considerable efforts have been made to clarify the etiology and pathogenesis of CRC, many questions remain unanswered. 2,3 It is important to elucidate the molecular mechanisms of CRC tumorigenesis and progression to facilitate the development of effective treatments.
The Human Genome Project has demonstrated that less than 2% of the human genome consists of protein-coding genes, while greater than 90% produces noncoding RNAs (ncRNAs). 4 The ncRNA family can be divided into many categories-according to size, structure, and function -including microRNAs (miRNAs) and long noncoding RNAs (lncRNAs).
Many studies have focused on ncRNA-mediated protein-coding gene regulation, overall indicating that ncRNAs form an interaction network which fine-tunes regulation of gene expression. More recently, the competing endogenous RNA (CeRNA) hypothesis was proposed. This suggests that a large number of ncRNA types may interact with and sequester miRNAs, thereby derepressing the function of alternate and unsequestered transcripts of these miRNAs, providing a novel mechanism contributing to posttranscriptional regulation of target genes. 5,6 Moreover, aberrant expression of ncRNAs has been documented in several human tumors, suggesting that ncRNAs may significantly contribute to their pathogenesis. 7,8 For example, lncRNA-HULC overexpression may suppress miR-186 expression, resulting in increased expression of its target protein, HMGA2, which serves as an oncogene in hepatocellular carcinoma. 9 Such mRNA/miR-lncRNA coregulatory pairs (eg, TCF7L2/miR-217-CRNDE and TUSC7/miR-211-CDK6) have also been demonstrated in CRC. 10,11 SNHG1, located on chromosome 11q12.3, is 1134 base pairs (bps) long and its intronic sequence encodes eight small nucleolar (sno) RNAs: SNORD22 and SNORD25-31. 12 Previous studies suggested that abnormal upregulation of SNHG1 was observed in lung cancer, liver cancer, and neuroblastoma, which was negatively correlated with prognosis. [13][14][15] It has been documented that SNHG1 promoted the growth of primary esophageal cancer, hepatocellular carcinoma, and non-small-cell lung cancer via sponging several tumor-suppressive miRNAs including miR-338, miR-195, and miR-101-3p. 12,16,17 Based on the above results, SNHG1 possesses the characteristics of an oncogene in many tumor types; however, the role of SNHG1 in CRC remains unclear.  Island, New York) supplemented with 10% fetal bovine serum (Gibco BRL) and 1% penicillin/streptomycin (Invitrogen, Carlsbad, CA), at 37°C in a humidified incubator (5% CO 2 atmosphere). All cell lines were cultured for at least 6 months.

| Quantitative real-time polymerase chain reaction
The expression levels of SNHG1, miR-211, miR-137 and other potential target miRNAs of SNHG1 such as miR-101-3p, miR-4735-3p were examined by qRT-PCR. Total RNA was extracted from cultured cells

| Flow cytometry
Target plasmids or negative controls were transfected into cell lines in six-well plates (as described above). After 48 hours, cells were incubated with propidium iodide (PI; Sigma-Aldrich) in the dark for 30 minutes, then harvested and subjected to flow cytometry. Cell cycling is expressed as the percentage of cells in G0/G1, S, and G2/M.

| MDC staining
LoVo cells were transfected with siR-SNHG1, miR-137 mimics, or miR-137 inhibitor for 48 hours. After that, the cells were incubated with 50 μM monodansylcadaverine (MDC) at 37°C for 10 minutes in the dark. Later on, cells were washed with PBS for three times, and images were obtained with a fluorescent microscope.

| Animal study
To establish an in vivo metastatic cancer model, 1 ×

| Immunohistochemistry
To evaluate the expression of RICOTR in tumor tissues, formalin-fixed and paraffin-embedded tissues were subjected to immunohistochemical staining using anti-RICTOR antibodies. After deparaffinization and rehydration, tissue sections were incubated with 3% hydrogen peroxide, and endogenous peroxidase was quenched in methanol.
Tissue sections were blocked with 1% bovine serum albumin for 30 minutes before incubation with antibodies at 4℃ overnight. As a negative control, staining was also performed in the absence of primary antibodies for each tissue section.

| SNHG1 expression is upregulated in CRC
Relative SNHG1 expression level in 80 pairs of CRC tissues and corresponding normal tissues was examined by qRT-PCR.
Compared with the normal group, SNHG1 expression was significantly upregulated in CRC tissues group ( Figure 1A). As indicated in Figure 1B, the expression of SNHG1 was higher in patients with CRC in T3 and T4, compared with that in T1 and T2. In addition, the level of SNHG1 was markedly upregulated in patients with lymph node metastasis (N1/2) group, compared with patients without lymph node metastasis (N0) group ( Figure 1C). Detailed correlations between SNHG1 expression and clinicopathological features are summarized in Table 1

| miR-137 is a biological target of SNHG1
Accumulating evidence indicates that lncRNAs act as "sponges" which regulate the expression and activity of miRNA. By exploring the bioinformatic database, Starbase, we found that miR-137 was a hypothetic miRNA targeting SHNG1 (Figure 2A). To confirm whether miR-137 is a direct binding target of SNHG1, a luciferase reporter gene assay was performed. As shown in Figure 2B and 2C, overexpression of miR-137 significantly reduced SNHG1-WT reporter activity in LoVo and HT-29 cells, respectively. However, it failed to repress the mutated SNHG1-3′UTR. To avoid off-target effects, we designed two siRNAs targeting different SNHG1 regions. As indicated in Figure 2D, the level of SNHG1 was significantly decreased following transfection with si-SNHG1s. In addition, downregulation of SHNG1 markedly increased the level of miR-137 ( Figure 2).
Ectopic SNHG1 overexpression (pcDNA3.1-SNHG1) increased the transcription level of SNHG1 in HT-29 cells ( Figure 2E). In addition, the level of miR-137 was notably decreased following transfection with pcDNA3.1-SNHG1-Wt ( Figure 2F). Expression of SNHG1 was unaffected following transfection with miR-137 ( Figure 2G). Meanwhile, the expression level miR-137 was significantly downregulated in CRC tissues, compared with that in normal tissues ( Figure 2H). These results suggested that MiR-137 is a biological target of SNHG1.

| miR-137 is negatively mediated by SNHG1
Previous studies have demonstrated that miRNAs exist in the cytoplasm in the form of miR-nucleoprotein complexes, including the protein Ago2, which is a key RISC component with an important role in siRNA or miRinduced gene silencing. 20 Ago2 coimmunoprecipitation may thus facilitate identification of potential miRNA targets. An RNA pull-down assay was carried out using SNHG1 probes. Meanwhile, Ago2 and miR-137 were investigated to determine whether SNHG1 and miR-137 exist in the same RISC complex. RNA pull-down assays were performed to determine the physical relationship between SNHG1 and Ago2 ( Figure 3A). To confirm that miR-137 and SNHG1 are in the same Ago2 complex, we synthesized biotin-labeled SNHG1 RNA probe and mixed with the cellular extract. After pull-down experiment with beads, we detected miR-137 by Western blot analysis, suggesting SNHG1 directly interacted with miR-137 ( Figure 3B). To further investigate the relationship between SNHG1 and miR-137, we examined miR-211, which was negatively regulated by loc285194. The qRT-PCR assay results demonstrated that a significant amount of miRNA-211 in the loc285194 pulled down pellet, while the amount of miR-211 in the SNHG1 pulled down pellet was only slightly increased compared with control ( Figure 3C). These data indicated that the activity of SNHG1 is mediated through negative regulation of miR-137.

| RICTOR is a target gene of miR-137
TargetScan, PITA, and miRanda software were used to predict the downstream targets of miR-137. The results showed that RICTOR is a potential target of miR-137 ( Figure 4A). To validate whether RICTOR is a downstream target of miR-137, luciferase reporter plasmids containing the RICTOR miR-137-binding sites (WT), or a mutant RICTOR 3′UTR were constructed. As indicated in Figure 4B, overexpression of miR-137 significantly reduced luciferase activity of RICTOR-WT but not the activity of the RICTOR-MUT in HEK293T cells, demonstrating that miR-137 could specifically target the RICTOR 3'UTR. In addition, Western blot analysis was used to confirm the results. Upregulation of miR-137 markedly decreased the level of RICTOR in LoVo and HT-29 cells, while downregulation of miR-137 had no effect on RICTOR levels ( Figure 4C).

To confirm whether SNHG1 regulates RICTOR expression in
LoVo and HT-29 cells, we firstly analyze the expression of RICTOR following transfection with siR-SNHG1. As shown in Figure 4D, the downregulation of SNHG1 markedly decreased the level of RICTOR.
However, the level of RICTOR was significantly increased after cotransfection with si-SNHG1 and miR-137 inhibitor in LoVo and HT-29 cells, respectively ( Figure 4E and 4F). These data suggest that RICTOR is a direct target of miR-137, and that SNHG1 can regulate RICTOR expression through interaction with miR-137.

| SNHG1 exerts carcinogenic activity through regulation of the mTORC2 pathway in vitro
For further study, we investigated the effects of RICTOR on CRC cells. As shown in Figure 5A, the expression of RICTOR was decreased the most following transfection with si-RICTOR-2. In addition, the percentage of the S phase was markedly decreased in the si-RICTOR-2 group, compared with the si-NC group ( Figure 5B).
Meanwhile, downregulation of RICTOR significantly inhibited the invasion ability of LoVo cell ( Figure 5C).
As shown in Figure 5D, the downregulation of SNHG1 significantly inhibited the percentage of the S phase. However, downregulation of miR-137 notably increased the percentage of the S phase, which were markedly reversed after cotransfection with si-SNHG1 ( Figure 5D). In addition, the invasion ability of LoVo cell was significantly increased after  transfection with miR-137 inhibitor, while the pro-invasive effect was notably decreased after cotransfection with si-SNHG1 ( Figure 5E).
Moreover, cell proliferation study indicated miR-137 inhibitor notable increased cell proliferation, while there effects were completely blocked by si-RICTOR2 ( Figure 6A and 6B). All these data suggest that RICTOR may influence CRC cell proliferative and invasive capabilities.
To further validate that SNHG1 and miR-137 exert biological activities mediated through the regulation of RICTOR, the expressions of RICTOR-associated proteins were detected. As shown in Figure 7A and

| SNHG1 exerts carcinogenic activity in vivo through regulation of miR-137
To further demonstrate that the carcinogenic activity of SNHG1 is mediated through negative regulation of miR-137, LoVo cells were subcutaneously injected into nude mice. Four weeks later, tumor volume and tumor weight were significantly decreased in the LoVo-si-SNHG1 group ( Figure 8A and 8B). Downregulation of miR-137 markedly increased the tumor weight and volume, which were significantly decreased following transfection with sh-SNHG1 ( Figure 8A and 8B).
Additionally, immunohistochemistry (IHC) results indicated that the level of RICTOR was markedly decreased after transfection with sh-SNHG1, which was significantly increased after transfection with miR-137 inhibitor. However, the upregulated RICTOR level in miR-137 inhibitor group was notably downregulated following cotransfection with sh-SNHG1 ( Figure 8C and 8D). These data indicate that SNHG1 exerts carcinogenic activity in vivo through regulation of miR-137. miR-137 is located on human chromosome 1p22 and is involved in many biological processes and diseases. 29 To date, research

| DISCUSSION
indicates that miR-137 may play a dual role during tumorigenesis, and that the nature of this role may be dependent on tumor type and target messenger RNA (mRNA) identities. 30,31 As a tumor suppressor in melanoma, miR-137 may inhibit cell migration by targeting the TBX3 transcription factor. 32 As an oncogene in breast cancer, miR-137 may enhance the epithelial-to-mesenchymal transition (EMT) The expression of miR-137 is frequently downregulated during oncogenesis, but its regulation and possible mechanisms remain to be clarified. It has been documented that methyl-CpG-binding protein 2 (MeCP2) and DNA methyltransferases (DNMTs) may work cooperatively to enhance methylation of the miR-137 promoter, leading to transcriptional silencing. 34 The current study found that SNHG1 may-instead of inhibiting expression of miR-137-sequester mature miR-137 (a novel posttranscriptional mechanism of miR-137 regulation and miR-137 mimic appears to decrease phosphorylation of AKT, SGK1, and p70S6K1, as well as increasing the LC3II/ LC3I ratio.
In context, p-AKT, p-SGK1, and p-p70S6K1 are known to promote proliferation, while an increased LC3II/I ratio may suggest activation of autophagy. 41,42 It is therefore reasonable to infer

| CONCLUSION
The current study identifies a novel mechanism of lncRNA and miRNA interaction which appears to promote CRC tumorigenesis and progression. However, miR-137 is not the only miRNA bound by SNHG1, as 27 miRNAs were identified that are expected to exhibit sufficient basepair complementarity for this purpose.
Additionally, miR-137 biological function is not RICTOR-specific, as it is also able to bind the 3′UTR of mRNA for other proteincoding genes, such as TBX3 and BMP7. Taken together, SNHG1 and miR-137 likely engage in significantly more comprehensive biological functions than those observed in the present study.
Future studies should identify additional oncogenic activities of SNHG1 in clinical samples from a larger number of patients. Such information will contribute to the development of novel targeted anticancer therapies.