SNORA42 Promote Oesophageal Squamous Cell Carcinoma Development Through Interacting With Dhx9

Background:Oesophageal squamous cell carcinoma (ESCC) is a high malignant cancer, which is the most common subtype of oesophageal cancer. Small nucleolar RNAs (snoRNAs) are a group of novel non-coding RNAs that have been found play a key role in various cancers. Methods:The expression of SNORA42 in ESCC samples and cell lines was measured by using real-time PCR and a series of in vitro and in vivo assays were performed to determine the function of SNORA42 in ESCC. Furthermore, RNA pull-down assay combined with Mass Spectrometry was applied to identify the protein that is associated with SNORA42. Silencing SNORA42 in ESCC cell followed by Next-Generation mRNA Sequencing was used to investigate the signaling pathway regulated by SNORA42. Results: We identied H/ACA box snoRNA42 (SNORA42) wasupregulated in ESCC and had the potential to be applied as a prognostic marker. Specically, overexpression of SNORA42 promotedESCCprogressionin vitro and in vivo, whereas knockdown of SNORA42 had opposite effects. We identied SNORA42 interacted with DHX9, which was up-regulated in ESCC and had a positive correlation with the expression of SNORA42. Furthermore, the promotion of SNORA42 on ESCC phenotypes can be reversed by knockdown of DHX9. Mechanically, SNORA42 promoted DHX9 expression by attenuating DHX9 transports into cytoplasm, to protect DHX9 from being ubiquitinated and degraded. From KEGG analysis of Next-Generation Sequencing, the NF-κB pathway was one of the most regulated pathways by SNORA42. SNORA42 enhanced phosphorylation of p65 and this effect could be reversed by NF-κB inhibitor, BAY11-7082. Moreover, SNORA42 activated NF-κB signaling through promotingthe transcriptional co-activator DHX9 interacted with p-p65, inducingNF-κB downstream genes expression. Conclusions:These ndings suggest that SNORA42 is up-regulated in ESCC and promotes ESCC developments,partly via interacting with DHX9 and triggering SNORA42/DHX9/p65 axis. The noncoding RNAs SNORD50A and SNORD50B bind K-Ras and are recurrently deleted in immunoblotting assay was performed to measure the expression DHX9. conducted on overexpressed SNORA42 ESCC cells. (b) NF-κB downstream genes were tested by real-time PCR after co-transfected siDHX9 and SNORA42 in ESCC cells.Data present the mean ± SEM (n=3). Different letters above bars denote statistical signicance. (c) The schematic of proposed role of SNORA42 in ESCC. SNORA42 is over-expressed in ESCC clinical tissues and cell lines, and promotes ESCC metastasis and proliferation in vitro and in vivo. Mechanically, SNORA42 associates with DHX9 and stabilizes DHX9 protein expression. On the other hand, SNORA42 promotes DHX9 and p-p65 association to further induce p65 downstream genes expression, resulting ESCC metastasis and proliferation.


Introduction
Oesophageal cancer is one of the most human malignant cancers with the 7th incidence rate and the 6th mortality rate over the world [1]. Oesophageal squamous cell carcinomas (ESCC) is the most common histologic subtype of oesophageal cancer [2]- [4]. However, there is no speci c symptom for ESCC in early stages; thus, in most cases, it cannot be detected until advanced stages accompany organ metastasis leading to miss best therapy opportunity. Identi cation of early diagnosis biomarkers of ESCC is an ideal screening strategy while there are fewer studies about it [5].
Small nucleolar RNAs (snoRNAs) is a group of conserved noncoding RNAs with 60-300nt in length [6]. It has been classi ed into 2 groups: C/D box snoRNAs (SNORDs) and H/ACA box snoRNAs (SNORAs). C/D box snoRNAs contain a closed loop which is consisted with C and D motif whereas H/ACA box snoRNAs contain 2 loops linked by H box and ACA motif located at 3' end [7]. snoRNAs have been well known to exert its function through binding to ribonucleoprotein (RNP) to form snoRNP[8]- [10]. H/ACA box snoRNAs are reported as housekeeping genes to guide pseudouridylation of ribosomal RNAs (rRNAs) and Loading [MathJax]/jax/output/CommonHTML/jax.js small nuclear RNAs (snRNAs) [7], [11]. However, recent studies have revealed novel roles of snoRNAs in human disease [12], included cancer [13], [14]. Moreover, more evidence indicated that snoRNAs may be treated as potential biomarkers in clinical application [15], [16].
SNORA42 is an H/ACA box snoRNA and encoded by KIAA0907. The previous study indicated that SNORA42 was overexpressed in non-small cell lung cancer (NSCLC) and promoted tumorigenicity in a p53-dependent manner [17]. In addition, overexpressed SNORA42 was positively correlated with poor prognosis in NSCLC [15], [17]. Another study revealed that SNORA42 expression was negative related with overall survival, disease-free survival and a risk factor for distant metastasis in colorectal cancer. In this study, SNORA42 was also identi ed the patients with high risk for recurrence and poor prognosis in stage II colorectal cancer [16]. Most recently, SNORA42 has been reported to enhance prostate cancer tumorgenesis by promoting EMT progression [18]. However, there is no research about the role of SNORA42 in ESCC. DEAH box protein 9 (DHX9), also named ATP dependent RNA helicase A (RHA) is a member of the DEAH family of human RNA helicases [19]. DHX9 has been reported upregulated in different human cancers and promoted tumor process through interacting with a number of transcription factors [20]- [25]. The previous study reported that DHX9 interacted with nuclear factor-κB (NF-κB) dimer p65 and enhanced NF-κB transcriptional ability, leading to NF-κB downstream genes expression [22].
The aim of this study was to investigate the function of SNORA42 in ESCC. Here, we found that SNORA42 is upregulated in ESCC patient's tissues and serum, and has a positive relationship with poor survival and metastasis. Furthermore, SNORA42 promoted ESCC development ability by interacting with DHX9 to enhance ESCC proliferation, migration and invasion ability. On one hand, SNORA42 blocks ubiquitination of DHX9 to prolong DHX9 lifespan. On the other hand, SNORA42 enhances the interaction between DHX9 and p-p65 to trigger NF-κB activity, and then promotes transcription of NF-κB target genes.

Materials And Methods
Cell Culture and transient transfection All  Two sets of samples were used in this study. The rst set, which contained ESCC tissues and paired paracancer tissues, was collected from the Fourth Hospital of Hebei Medical University during 2014-2015. No patients received chemotherapy or radiation therapy before surgery. Informed consent for the use of samples was obtained from patients and approval was obtained from the hospital. All samples information was shown in Table 1. The other set, which contained serum from ESCC patients before surgery and control serum from healthy volunteers with age in 25-55. Informed consent for the use of samples was obtained from patients and volunteer, and approval was obtained from the hospital.

Fluorescence In Situ Hybridization ( sh) Assay
The FISH assay was performed on Eca-109 cell to determine the cellular location of SNORA42. Cells were seeded on slides and xed with 4% paraformaldehyde for 20 min at room temperature. Quantum dotuorescent in situ hybridization (QD-FISH) was performed to detect the presence of SNORA42 using a digoxin-labeled oligonucleotide probe (RiboBio) indirectly labeled with streptavidin-conjugated quantum dots. 4′, 6-diamidino-2-phenylindole (DAPI) was used as a marker for nucleus and 18S rRNA was used to indicate cytoplasm.

Immuno uorescence Assay
Immuno uorescence was performed on Eca-109 and KYSE30 cells to determine the subcellular localization of DHX9. The cell was seeded and xed on a slice, followed blocked by 2% BSA and incubated with DHX9 antibody overnight. In the second day, slides were incubated with uorescencelabeled (IF488, Abcam) secondary antibody, and DAPI to mark the location of the nucleus. The slide was observed and taken by the confocal microscope (LSM 700). All male BALB/c nude mice (4-6 weeks old) were obtained from the Vital River Laboratory Animal Technology Co. Ltd., Beijing. Animal experiments contain two parts. The rst one, nude mice (4-6 weeks old) were injected subcutaneously under left armpit with SNORA42 knockdown cells and control cells (1 × 10 6 /mouse, n = 6/group). Each group contains 6 nude mice. Tumor size was measured every 5 days until its volume up to 1500 mm 3 . 27 days later, the tumor was removed to weigh and used to be performed IHC assay. The other part, nude mice were separated into 2 groups (n = 6) randomly and injected knockdown SNORA42 cells or control cells (5 × 10 6 /mouse). After 90 days, organs (involved liver, lung, brain) were removed and tested through H&E staining assay to determine if organ metastasis occurred.
H&e Staining H&E staining was used to determine the situation of tissue suspected cancer. Slices were dewaxed and dehydrated independently, followed by being stained with hematoxylin for 5-15 min and chromatic separated with alcohol hydrochloride. The nucleus was alkalized with lithium carbonate until the color of it became blue. Then slides were stained with eosin for 1-5 min, followed by dehydrated and covered with neutral balsam. Each slide was assessed by two pathologists independently.

Protein Extraction And Immunoblotting
Cell lysates and immunoblotting were prepared as described previously [26].

Biotin-rna Pull-down Assay
The pGEM-3zf(+)-SNORA42 and pGEM-3zf(+)-SNORA42-antisense plasmid were linearized to be used as templates for the transcription of SNORA42 and anti-SNORA42, followed by ampli ed through RT-PCR. Biotin-labeled RNAs were transcribed using Biotin-RNA Labeling Mix (Roche, Indianapolis, IN, USA), treated with RNase-free DNase I (TaKaRa, Kyoto, Japan) and puri ed with an RNeasy Mini Kit (Roche). Biotinylated RNAs were heated at 98 °C for 2 min, put on ice for 5 min and left at RM 25 min to form secondary structure. Total protein lysates were mixed with the Biotinylated RNA incubated at 30 °C and added Streptavidin agrose beads (GE Healthcare, Little Chalfont, UK) incubated at RM with rotation. The RNA-protein-beads complex washed with binding buffer, included RNA nuclease and proteinase, followed by boiled with loading buffer for 10 min. RNA a nity was tested by 12% sodium dodecyl sulfatepolyacrylamide gel electrophoresis and visualized by silver staining. The speci c bands were excised and identi ed by mass spectrometry.

Rip Assay
RIP assay was used to con rm the interaction between protein and RNA. The assay was performed according to the manufacturer protocol of Magna RIP kit (Millipore, USA). The antibody against DHX9 was used to form the DHX9-SNORA42-magnetic beads complex, followed detected by RT-PCR and agrose Loading [MathJax]/jax/output/CommonHTML/jax.js gel. Total RNA was used as input and anti-rabbit IgG antibody was used as negative control to determine the RNA was speci c binding to protein DHX9.

Co-immunoprecipitation (co-ip) Assay
Co-immunoprecipitation was performed to evaluate the effect of SNORA42 on the interaction between DHX9 and p-p65, as well as DHX9 and ubiquitin. Brie y, cell lysate was incubated with DHX9 overnight in the same quanti cation for each group. In the second day, the cell lysate and antibody complex was added magnetic beads to obtain a protein that speci cally interacted with DHX9, followed boiled for 10 min and performed immunoblotting with antibodies against p-p65 (1:500, CST)and Ubiquitin (1:1000, CST). Reverse veri cation was performed for further con rmation. The total cell lysate was used as input and anti-rabbit IgG antibody was used as negative control.

Statistical analysis
All in vitro experiments were done at least 3 biological replicates in each group with repeated at least 3 times. The results are evaluated as mean ± SEM in bar graphs and line graphs. Statistical analyses such as, paired or unpaired student's t-test, χ 2 test or Mann-Whitney U test, Pearson Correlation test were used through SPSS version 13.0 and GraphPad Prism 8.

Results
Clinical signi cance of SNORA42 in ESCC Using ESCC tissue samples from The Fourth Hospital of Hebei Medical University, we found that SNORA42 expression in ESCC tissues was signi cantly higher than that in para-cancer tissues (Fig. 1a). From the analysis of TCGA database, SNORA42 level in oesophageal carcinoma was also signi cantly higher than that in normal oesophageal tissues (Fig. 1b) [27]. To further con rm the impact of SNORA42 in ESCC development, the analysis about SNORA42 and clinic pathological features of ESCC was performed. Patients with ESCC were divided into low-expression group and high-expression group based on median value of SNORA42 expression level in all patients. As shown in Table 1, high SNORA42 level was associated with tumor invasion, region lymph nodes invasion, distant metastasis and oesophageal cancer stage. We also tested the level of SNORA42 in patient's peripheral serum with early stage ESCC. To note, the serum was measured before patients got surgery, chemotherapy and radiotherapy. As shown in Fig. 1c, SNORA42 was signi cantly upregulated in ESCC patient's serum than in healthy serum. Moreover, SNORA42 expression in ESCC cell lines was higher than oesophageal immortalized epithelial cell (NE 2 -hTERT), especially, in Eca-109, KYSE30, and KYSE180 (Fig. 1d). So we will use these cell lines to investigate the function of SNORA42 in our following experiment.In addition, Kaplan-Meier survival analysis determined that patients with higher SNORA42 level were negatively correlated with 5-year survival of ESCC patients (Fig. 1e) [27].

SNORA42 exerts tumor-promoting effects in vitro and in vivo
The up-regulation of SNORA42 in ESCC tissue, serum and cell lines has been determined, however the role of SNORA42 in ESCC is still unknown. The expression of SNORA42 was signi cantly increased in KYSE30, Eca-109, and TE-1 cells after transfected with SNORA42 plasmid (Additional Fig. 1a, see Additional le 2). Due to the nuclear location and structure of SNORA42 constrained the knockdown e ciency of RNA interference (Additional Fig. 1b, see Additional le 2), antisense oligonucleotides (ASOs) was used to suppress endogenous SNORA42 level. The level of SNORA42 was signi cantly suppressed in KYSE30 and Eca-109 cells by ASOs (Additional Fig. 1c, see Additional le 2). Cell proliferation, migration and invasion were tested by MTS assay, Transwell migration assay, Transwell matrigel assay and wound-healing assay respectively. Overexpressing SNORA42 promoted cell proliferation, migration and invasion (Fig. 2a, b, Additional Fig. 1c, d, see Additional le 2) whereas knockdown of SNORA42 signi cantly inhibited these phenotypes (Fig. 2c, d, Additional Fig. 1f, see Additional le 2).
To further investigate the function of SNORA42 on ESCC in vivo, SNORA42 stably down-regulated cell line was generated using Lentivirus infected KYSE30 (Additional Fig. 1g, see Additional le 2). Both subcutaneous xenograft and tail vein injection nude mice models were utilized with the SNORA42 knockdown cell. In the subcutaneous xenograft mice model, the size and weight of tumor from SNORA42downregulated group were signi cantly inhibited compared with the control group (Fig. 2e, f). Moreover, the expression of cell proliferation marker Ki-67, metastasis markers MMP2 and MMP9 in tumor tissues was signi cantly suppressed (Additional Fig. 1h, see Additional le 2). In the tail vein injection metastasis model, nude mice in SNORA42-downregulated group had less distant organ metastasis con rmed by H&E staining assay with 1 nude mouse had lung metastasis, 1 nude mouse had brain metastasis and none of them had liver metastasis. However, nude mice in the control group all had lung metastasis, 5 nude mice had brain metastasis and no liver metastasis (Fig. 2g, h). In summary, SNORA42 promoted ESCC proliferation, migration and invasion in vivo and in vitro.

Snora42 Interacts With Dhx9 Protein
To investigate the mechanism that how SNORA42 performs its promotion function in ESCC, a biotinlabeled RNA pull-down assay combined with mass spectrometric analysis was performed to identify the nucleus proteins that might interact with SNORA42. Three bands were found speci cally interacted with biotinylated SNORA42 oligo (Fig. 3a). The liquid chromatography-mass spectrometric analysis was heterogeneous nuclear ribonucleoprotein and showed that 75 proteins potentially bound to SNORA42 (Additional Table 2 biotinylated SNORA42 oligos whereas it cannot be pulled down by antisense of SNORA42 or nonbiotinylated SNORA42 oligos. Furthermore, the result of RNA immunoprecipitation (RIP) assay showed that SNORA42 was speci cally enriched by anti-DHX9 antibody in ESCC cells (Fig. 3c). As shown in Fig. 3d and e, the cellular location of SNORA42 was in the nucleus which was con rmed by Fluorescence in situ hybridization (FISH) assay, and it was the same location with DHX9. These results indicated that SNORA42 was associated with DHX9 in ESCC cells.

Snora42 Promotes Escc Progress Through Interacting With Dhx9
Though DHX9 has been reported as an oncogene in multiple cancers, there is no report in ESCC. We tested the level of DHX9 gene in ESCC clinical tissues and found that DHX9 level was upregulated in ESCC (Fig. 3f). Knockdown of DHX9 signi cantly suppressed ESCC cells proliferation, migration, and invasion (Additional Fig. 2b, c, see Additional le 2). To test if SNORA42 exerts its function through associating with DHX9 in ESCC, We rst tested whether the expression of DHX9 is associated with SNORA42 level in ESCC cells. The result showed that SNORA42 can suppress DHX9 protein expression in ESCC cells but had no effects on DHX9 gene level ( Fig. 3g and Additional Fig. 2a, see Additional le 2). This result was coincident with ESCC clinical tissues, SNORA42 expression had a positive relationship with DHX9 protein in the same cohort clinical tissues in immunohistochemistry (IHC) assay, and DHX9 was more accumulated in nucleus in ESCC compared with para-cancer which was consistent with FISH assay (Fig. 3i). Furthermore, the IHC assay about mice subcutaneous xenograft tumour showed that DHX9 expression was signi cantly suppressed after knocking down SNORA42 (Fig. 3h). These results suggested that SNORA42 is associated with DHX9 in ESCC and has a positive correlation with DHX9 protein expression.
Whether the association of SNORA42 and DHX9 has an effect in ESCC development was evaluated by co-transfecting SNORA42 and siDHX9 in ESCC cells. As shown in Fig. 4a, the enhanced proliferation by SNORA42 was reversed by knocking down of DHX9. Moreover, the promotion of migration and invasion were partly rescued by knocking down of DHX9 ( Fig. 4b and Additional Fig. 2d, see Additional le 2). Taken together, these results indicated that SNORA42 exerts its promotion function partly through interacting with DHX9 in ESCC.

Snora42 Attenuates Dhx9 Ubiquitination
Above studies veri ed that SNORA42 enhanced DHX9 protein level, but little is known about molecular mechanism. Given the results that SNORA42 played no effects on DHX9 gene level (Additional Fig. 2a protein generation, and half-life period of DHX9 protein was measured. Speci cally, ESCC cells were transfected with SNORA42 plasmid or vector, and treated with CHX. Cell lysate was collected at different times after treatment. As shown in Fig. 4c, DHX9 protein level decreased 50% with CHX treatment 4 hours in the control group, whereas in SNORA42 overexpression group half-life of DHX9 was dramatically delayed until later than 6 hours. Furthermore, we used MG132, a proteasome inhibitor, to protect DHX9 protein from being degraded. From Fig. 4d, knockdown of SNORA42 reduced DHX9 protein level whereas treated with MG132 blocked this effect. These results suggested that SNORA42 promoted DHX9 expression in ESCC cells by attenuated its protein degradation.
Ubiquitin-proteasome pathway is known involved in protein degradation of eukaryotic cells. The degradation process of DHX9 was still unknown so we rstly used immunoprecipitation assays with anti-DHX9 antibody in ESCC cells to determine if DHX9 degradation through the ubiquitin-proteasome pathway. As shown in Fig. 4e, the ubiquitinated DHX9 was detected by the anti-ubiquitin antibody. Next, we tested our hypothesis that SNORA42 might involve in DHX9 ubiquitin degradation pathway. ESCC cells were transfected with SNORA42 plasmid followed by MG132 treatment, and Immunoprecipitation assay was performed. As shown in Fig. 4f, overexpression of SNORA42 reduced the DHX9 ubiquitination level, indicating that SNORA42 promoted DHX9 protein expression by attenuating DHX9 ubiquitination.
To further investigate how SNORA42 promoted DHX9 protein level, we measured the DHX9 in ESCC cell nucleus and cytoplasm. As shown in Additional Fig. 2e (see Additional le 2), SNORA42 could protect DHX9 accumulated in the nucleus from being translated into cytoplasm. Taken together, SNORA42 associated with DHX9 to accumulate DHX9 in nucleus, resulting to protect DHX9 protein from being degraded.

Snora42 Regulates Nf-κb Pathway In Escc Cells
To determine the molecular mechanism of SNORA42 regulating ESCC development, we performed the Next Generation Sequencing on KYSE30 after knocking down SNORA42 (GSE146195) and the sequencing result was veri ed by real-time PCR (Fig. 5a). It was found that in the genes that changed greater than 5 times between control and SNORA42 low-expressed groups, 70 genes were down-regulated whereas 8 genes were up-regulated (Additional Table 3, see Additional le 1). From KEGG analysis of the sequencing, we found that NF-κB was one of most regulated and cancer-related pathways by SNORA42 (Additional Fig. 3a, see Additional le 2). To investigate if SNORA42 could regulate the NF-κB pathway, we measured downstream genes of NF-κB after overexpressing and silencing SNORA42. It was shown that MMP2, MMP9, VEGFC, and BCL2 levels were changed consistent with SNORA42 expression (Fig. 5b). In addition, we tested two metastasis markers MMP2 and MMP9, which are also NF-κB downstream genes. As shown in Fig. 5c  inhibitor of NF-κB pathway, BAY11-7082 could reverse this promoted effect (Fig. 5e). These results suggested that SNORA42 activated NF-κB pathway in ESCC cells.
SNORA42 promotes DHX9 interacted with p-p65 to regulate the NF-κB pathway Studies have revealed that DHX9 interacts with p65 and plays as an transcriptional co-activator to promote p65 downstream gene transcription [28], [29], we hypothesized that SNORA42 might regulate the interaction of p65 and DHX9 to modulate NF-κB signalling. We performed Co-Immunoprecipitation assay to test whether SNORA42 affected the association of p-p65 and DHX9. It was found that overexpression of SNORA42 increased the interaction of p-p65 and DHX9 in both Eca-109 and KYSE30 cells (Fig. 6a). In addition, SNORA42-induced NF-κB pathway downstream genes (MMP2, MMP9, VEGFC, and BCL2) by was reversed by knocking down of DHX9 (Fig. 6b). These results indicated that SNORA42 promoted p-p65 and DHX9 interaction to enhance p65 transcription ability.
As the schematic graph shown in Fig. 6d, SNORA42 is over-expressed in ESCC to promote migration, invasion and proliferation in vitro and in vivo, in part through associating with DHX9 and promoting its protein level, furthermore activating p-p65 to promote target gene expression and nally regulating the biological activity.
Discussion snoRNA has been considered as the house-keeping gene for a long time to guide rRNA modi cations. The snoRNAs are conserved non-coding RNAs existing in different species and encoded from the introns of genomic genes. However, recent evidence reported their abnormal expressions in various human cancers, indicated that these snoRNAs might have other functions in human disease. It has been shown that snoRNAs, especially C/D box snoRNAs, are highly expressed in human acute myelogenous leukemia (AML) [30]. Studies also showed that both two classes of snoRNAs were signi cantly overexpressed in human colorectal cancer (CSC) [31]. Therefore, we selected several snoRNAs that overexpressed in ESCC from TCGA database (data not shown) [27]. We measured these snoRNAs expression in human ESCC tissues and found SNORA42 was one of the most highly expressed snoRNA. Furthermore, studies revealed that snoRNAs were as variable as miRNA expression and can be used potential biomarkers for early detection and prognostication of non-small cell lung cancers (NSCLC) [32]. Note, SNORA42 expression has been found negatively correlated with overall survival in an additional independent cohort and identi ed the patients with high risk for recurrence and poor prognosis in stage II CRC [31]. In this study, it analyzed the correlation between SNORA42 expression and clinical pathology characteristics to test the possibility of SNORA42 as a prognostic biomarker in ESCC. behaviour. It has been reported SNORA42 promoted cell growth and colony formation partially in a p53dependent manner [33]. In addition, Studies showed that SNORA42 enhanced prostate cancer cell viability, migration and epithelial-to-mesenchymal transition (EMT) [34]. In this study, we used two ways to knockdown and overexpress SNORA42 respectively to test the function of SNORA42 in ESCC cells.
Moreover, we utilized two in vivo models further con rmed SNORA42 as an oncogene in ESCC.
DExH-Box Helicase 9 (DHX9) is a member of the DEAH-containing family of RNA helicases which has the ability to bind RNAs and catalyze the ATP-dependent unwinding of double-stranded RNAs (dsRNAs) [35]. It has been reported it is dysregulated in multiple human cancers [20], [21]. In this study, we found that the overexpression of SNORA42 suppressed the degradation of DHX9 and ubiquitination of DHX9, whereas had no effects on DHX9 mRNA level, indicating that SNORA42 promotes DHX9 expression by inhibiting its degradation, not the generation. Consistently, the expressions of DHX9 and SNORA42 correlated closely with ESCC clinical tissues and in vivo. We thought the reason might be the interaction of SNORA42 and DHX9 in nucleus attenuates DHX9 being transported to cytoplasm and degradation while further studies still needed. The recent study showed that SNORA18L5 reduces p53 expression by retaining RPL5 and RPL11 in the nucleolus that less RPL5 and RPL11 binding with MDM2 and inhibits its ubiquitin ligase activity toward p53 resulting p53 accumulation in hepatocellular carcinoma [36]. It also might be other reasons outside of the nucleus, as the study reported that the intracellular location of snoRNA changed from the nucleolus and other parts of the cell[8].
p65 (also named RELA) plays a key role in canonical NF-κB pathway inducing target gene transcription. p65 translocates into the nucleus and being phosphorylated by various kinases to trigger its interaction with CBP/p300 and thereby increases p65 transcription activity [37]. Numerous studies reported that p-p65 involved in different tumorgenesis processes by promoting growth, proliferation, angiogenesis, and invasion [38]. Moreover, nuclear NF-κB p65 phosphorylation frequently existed in head and neck squamous cell cancer and NF-κB downstream genes are signi cantly upregulated [39].
A study revealed that the loss of amino-terminal enhancer of split (AES) interacts with multiple snoRNAs resulting in decreased leukemia self-renewal potential in AML [40]. Here, we demonstrated that SNORA42 promoted ESCC growth, migration and invasion partly through interacting with DHX9. It has been reported that DHX9 directly bind to nucleolar p65 and mediate the transcription activity of NF-κB [22]. Consistently, the KEGG analysis of next-generation sequencing showed the NF-κB pathway was one of the most modi ed pathways after overexpression of SNORA42. A study showed that the loss of SNORD50A and SNORD50B direct binding to K-Ras in human cancers leading to activate the K-Ras and synergize the effects of KRAS mutation [8]. Therefore, we hypothesized that SNORA42 might work as a scaffold to associate with both p65 and DHX9, but from RNA pull-down assay, neither p-p65 nor p65 associated with SNORA42. Then we hypothesis it might due to SNORA42 can regulate the interaction between DHX9 and p-p65. We found that overexpression of SNORA42 can promote their interaction and activate downstream genes of NF-κB expression, including metastasis-related genes in vitro and in vivo. performed using paired two-tailed Student's t-test (a) and unpaired two-tailed Student's t-test (b, c ). **p<0.01, ***p<0.001, ****p<0.0001vs. controls.       (e) p-p65 level was measured in SNORA42 overexpressed cells and treated with NF-κB inhibitor, BAY11-7082. Data present the mean ± SEM (n=3). Different letters above bars denote statistical signi cance.

Figure 6
SNORA42 promotes DHX9 interacting with p-p65 to enhance the NF-κB signaling. (a) Coimmunoprecipitation (Co-IP) assay was performed using an anti-DHX9 antibody or anti-p-p65 antibody