RNA methyltransferase NSUN2 promotes gastric cancer cell proliferation by repressing p57Kip2 by an m5C-dependent manner.

The RNA methyltransferase NSUN2 has been involved in the cell proliferation and senescence, and is upregulated in various types of cancers. However, the role and potential mechanism of NSUN2 in gastric cancer remains to be determined. Our study showed that NSUN2 was significantly upregulated in gastric cancers, compared to adjacent normal gastric tissues. Moreover, NSUN2 could promote gastric cancer cell proliferation both in vitro and in vivo. Further study demonstrated that CDKN1C (p57Kip2) was the potential downstream gene of regulated by NSUN2 in gastric cancer. NSUN2 could promote gastric cancer cell proliferation through repressing p57Kip2 in an m5C-dependent manner. Our findings suggested that NSUN2 acted as an oncogene through promoting gastric cancer development by repressing p57Kip2 in an m5C-dependent manner, which may provide a novel therapeutic target against gastric cancer.


Introduction
During the past a few years, RNA modifications have been found to play an important role in the occurrence and development of many tumors. More than 100 types of chemical modifications have been identified in various types of RNAs, with methylation being the most common 1 . Methylation is a prevalent post-transcriptional modification that occurs in almost all RNA species. N 6methyladenosine (m 6 A) is the most abundant internal modification in mammalian messenger RNA (mRNAs) and widely involved in various biological processes of mRNAs [2][3][4] . Recently, many studies revealed that aberrant m 6 A modification is closely related to tumorigenesis, including acute myeloid leukemia 5 , hepatocellular carcinoma 6,7 , breast cancer 8,9 , bladder cancer 10,11 , cervical cancer 12 , and lung cancer 13 .
Another important RNA modification, 5-methylcytosine(m 5 C), was first identified in stable and highly abundant transfer RNAs (tRNAs) and ribosome RNAs (rRNAs) 14 . Recently, m 5 C modification and related m 5 C sites have been found in mRNA by advanced highthroughput techniques combined with next-generation sequencing in mRNAs. Yang et al. 15 found that NSUN2 (NOP2/Sun domain family, member 2; MYC-induced SUN domain-containing protein, Misu) was the main enzyme catalyzing m 5 C formation, while Aly/REF export factor (ALYREF, an mRNA transport adaptor, also named THOC4) functioned as a specific mRNA m 5 C-binding protein regulating mRNA export. It was found that m 5 C could promote the pathogenesis of bladder cancer through stabilizing mRNAs 16 . Recent studies showed that NSUN2 was linked to cell proliferation, stem cell differentiation and testis differentiation 17,18 . Wang and colleagues 19 found that NSUN2 could delay the replicative senescence by repressing Cyclin-dependent kinase inhibitor 1B (CDKN1B, p27 Kip1 ) translation and promote cell proliferation by elevating Cyclin-dependent kinase 1 (CDK1) translation 20 . Moreover, elevated protein expression of NSUN2 was found in various types cancers, including the esophageal, stomach, liver, pancreas, uterine cervix, prostate, kidney, bladder, thyroid, and breast cancers by immunohistochemistry (IHC) analysis 21 . Indeed, Wang and colleagues 22 found that NSUN2 was associated with metastatic progression by affecting DNA hypomethylation in human breast cancer. Gao et al. 23 also found NSUN2 could promote tumor progression via its interacting partner RPL6 in gallbladder carcinoma. However, the role and related mechanisms of NSUN2 in gastric cancer has not been investigated.
In the present study, we firstly showed that NSUN2 was significantly upregulated in gastric cancers, compared to adjacent normal gastric tissues. Moreover, NSUN2 could promote the gastric cancer cells proliferation both in vitro and in vivo. Further study demonstrated that p57 Kip2 was the potential downstream gene regulated by NSUN2 in gastric cancer. Furthermore, NSUN2 could promote gastric cancer cell proliferation by repressing p57 Kip2 in an m 5 C-dependent manner. This study suggested that NSUN2-mediated m 5 C methylation of p57 Kip2 mRNA may serve as novel mechanism for gastric cancer development and progression.

Results
NSUN2 was upregulated in human gastric cancers compared to adjacent normal gastric tissues Firstly, TCGA database analysis showed that NSUN2 was upregulated in tumors compared to adjacent normal gastric tissues (Fig. 1a, b). Meanwhile, to determine NSUN2 expression in gastric cancer tissues, we examined expression of NSUN2 in gastric cancer patients' tissues by performing quantitative real-time PCR (qRT-PCR) and western blot assay. As shown in Fig. 1c-e, both the mRNA and protein expressions of NSUN2 was significantly upregulated in gastric tissues, compared to corresponding adjacent normal gastric tissues. These findings implied that NSUN2 was upregulated in human gastric cancer, compared to adjacent normal gastric tissues.

NSUN2 promoted the human gastric cancer proliferation in vitro
Firstly, qRT-PCR and western blot assay revealed that NSUN2 was stably knockdown or overexpression in SGC 7901 and MGC 803 cells (Fig. 2a). Subsequently, the CCK-8 assay and colony formation assay showed that NSUN2 knockdown significantly inhibited cell proliferation and colony formation, whereas NSUN2 overexpression promoted cell proliferation and colony formation (Fig. 2b, c). Flow cytometry analysis further revealed that NSUN2 knockdown induced at the G1/G0 cell cycle arrest and NSUN2 overexpression decreased the percentage of G1/G0 phase, compared with wild-type cells, both in SGC 7901 and MGC 803 cells (Additional file: Supplementary Fig. S1a, b). Taken together, our results suggested that NSUN2 could promote the gastric cancer cell proliferation in vitro.
NSUN2 promoted human gastric cancer tumorigenesis in vivo MGC 803 cell line was used in this experiment. Firstly, stable NSUN2 overexpress cells, NSUN2 knockdown cells and their corresponding wild-type cells were injected into female nude mice. Up to 3 weeks after injection, we found that stable knockdown of NSUN2 suppressed tumor growth in nude mice effectively. The average tumor weight and volume of NSUN2 knockdown cells group were also significantly lower than those in the wild-type group (Fig. 3a, b). Moreover, IHC analysis confirmed that the tumors formed from the wild-type cells group displayed stronger Ki-67 staining than those from NSUN2 knockdown cells group (Fig. 3c). whereas, overexpression of NSUN2 significantly promoted tumor growth in nude mice ( Fig. 3d-f). Our results indicated that NSUN2 could promote human gastric cancer tumorigenesis in vivo.
p57 KIP2 was identified as potential target regulated by NSUN2 by RNA sequencing (RNA-seq) To identify the downstream targets regulated by NSUN2 in gastric cancer, we performed RNA-Seq assay to determine the mRNA expression changes in NSUN2 knockdown and corresponding wild-type cells. In this study, Kyoto Encyclopedia of Genes and Genome (KEGG) analysis revealed that differentially expressed genes were significantly changed in ribosome, endocytosis, cell cycle, protein processing in endoplasmic reticulum and apoptosis, suggesting that NSUN2 may play important roles in protein modification and cell proliferation ( Table 1). As cell proliferation plays an important role in tumor development, we selected cell cycle related pathways as a candidate targets of NSUN2 for further study. In total, we found 86 differential genes expression from cell cycle related pathways in NSUN2 stable knockdown cells, compare with corresponding wild-type cells ( Table 2). We then found that p57 Kip2 was obviously upregulated among these genes and was identified as potential target regulated by NSUN2.
NSUN2 destabilized the p57 Kip2 transcript and it may be involved in the oncogenic function of NSUN2 in gastric cancer The level of p57 Kip2 was remarkably upregulated in NSUN2 knockdown cells identified by qRT-PCR and Western blot (Fig. 4a, b). In addition, we found the relative half-life of p57 Kip2 mRNA increased from 3.44 to 6.26 h in MGC 803 cells, and from 4.93 to 7.14 h in SGC 7901 cells, following the NSUN2 knockdown (Fig. 4c). We silenced p57 Kip2 via transfecting with siRNA in NSUN2 knockdown cells and wild-type cells (Fig. 4d, e). Upon silencing of p57 Kip2 expression by siRNAs, the ability of cell proliferation in both NSUN2 knockdown cells and wild-type cells was enhanced (Fig. 4f, g). These results suggested that NSUN2 might exert its oncogenic effects in gastric cancer cells by repressing p57 Kip2 expression.
NSUN2 destabilized the p57 Kip2 mRNA relies on its methyltransferase activity and m 5 C modifications in the 3′untranslated region (UTR) of p57 Kip2 mRNA The Dot blot assay showed that NSUN2 knockdown significantly decreased the m5C levels, whereas NSUN2 overexpression increased the m 5 C levels (Fig. 5a). To assess the role of the m 5 C modifications in p57 Kip2 mRNA regulated by NSUN2, we conducted wild-type and mutant 3′-UTR of p57 Kip2 reporter plasmids for the luciferase reporter assays (Fig. 5b). Relative luciferase activity of the wild-type and mutant 3′-UTR of p57 Kip2 reporter genes Fig. 1 NSUN2 was upregulated in human gastric cancer tissues, compared to adjacent gastric normal tissues. a, b NSUN2 was significantly upregulated in gastric cancer tissues, compared with adjacent gastric normal tissues from the TCGA database, *p < 0.05; c Relative expression of NSUN2 mRNA in gastric cancer tissues and compared with corresponding adjacent normal gastric tissues. NSUN2 expression was examined using qRT-PCR and normalized to β-actin expression. The horizontal lines and numbers represent the median values of the distribution. *p < 0.05. d Expression of NSUN2 at protein level in eight paired gastric cancer tissues and adjacent normal gastric tissues by western blot. T: Gastric tumor tissues, N: Adjacent normal tissues. e The protein levels of NSUN2 were quantified by densitometry and the relative gray value of NSUN2 protein (normalized to β-actin) in gastric cancer tissues and adjacent gastric normal tissues was statistically significant. *p < 0.05. Ctrl, pLOV-control vector shNSUN2#2, pLKD-shNSUN2-1; shNSUN2#3, pLKD-shNSUN2-2; shCtrl, pLKD-control vector. Data were presented as the mean ± SD; *p < 0.05. b, c CCK-8 and colony formation assays were performed to determine the growth ability of overexpress and knockdown of NSUN2. Data were presented as the mean ± SD; *p < 0.05.
was measured in MGC 803 and SGC 7901 cells. As expected, the luciferase activity of the wild-type 3′-UTR of p57 Kip2 reporter gene was significantly enhanced after NSUN2 silencing. However, knockdown of NSUN2 had    no effect on the expression of the mutated 3′-UTR of p57 Kip2 reporter gene (Fig. 5c). More importantly, by m 5 C RNA immunoprecipitation (RIP) assay and qPCR assay, we found that m 5 C antibody significantly enriched 3′-UTR of p57 Kip2 mRNA and knockdown of NSUN2 reduced the m 5 C levels on 3′-UTR of p57 Kip2 mRNA (see figure on previous page) Fig. 4 p57 KIP2 was identified as potential target regulated by NSUN2 by RNA sequencing. a RNA was isolated from SGC 7901 and MGC 803 cells with stable knockdown of NSUN2 and p57 Kip2 mRNAs expression was determined by qRT-PCR. Data were showed as the mean ± SD; *p < 0.05. b The protein levels of p57 Kip2 and β-actin was assessed by western blot. c The mRNA half-life (t1/2) of p57Kip2 in SGC 7901 and MGC 803 cells with stable knockdown of NSUN2 or corresponding wild-type cells. d, e The relative mRNA and protein expression of p57 Kip2 in NSUN2 knockdown cells or corresponding wild-type cells, transfected with small-interfering RNAs (siRNAs), was tested using qRT-PCR and western blot. siNC: negative control siRNA; sip57 Kip2 : siRNA against p57 Kip2 . Data were showed as the mean ± SD; *p < 0.05. f, g CCK-8 and colony formation assays were performed to determine the growth ability of NSUN2 knockdown cells after transfection of siRNA against p57 Kip2 . Data were showed as the mean ± SD; *p < 0.05.  ). b, c Pezx-FR02-p57 Kip2 -3′-UTR plasmid with either wild-type or mutant (CCT mutation) m 5 C sites were constructed. The pattern diagram was shown. Above constructed plasmid was transfected into stable knockdown of NSUN2 or corresponding wild-type cells. Firefly luciferase activity was measured and normalized to Renilla luciferase activity. Data were presented as the mean ± SD; *p < 0.05. d m 5 C RIP and qRT-PCR analysis of m 5 C level in mRNA of p57 Kip2 in MGC 803 cells transduced with stable knockdown of NSUN2 or corresponding wild-type cells. Data were showed as the mean ± SD; *p < 0.05. (Fig. 5d). Altogether, our data indicated that NSUN2 destabilized the p57 Kip2 mRNA relies on its m 5 C methyltransferase activity in its 3′-UTR.

Discussion
Gastric cancer is the most common gastrointestinal tumors, representing one of the leading causes of cancerrelated deaths worldwide 24,25 . Despite the improvement in surgical techniques and patient management, there has been unsatisfactory improvement in the 5-year overall survival rate. Many patients were diagnosed with advanced stages that limited the successful therapeutic strategies. Furthermore, the molecular mechanisms underlying gastric cancer progression is still poorly understood. Therefore, better understanding of the tumor formation and diagnostic markers will improve the diagnosis and treatment of gastric cancer. m 5 C modification is another important posttranscriptional RNAs modification beside m 6 A modification. As a main m 5 C methyltransferase, NSUN2 was reported to promote cell proliferation, mobility, invasion in breast cancer 22 , gallbladder carcinoma 23 , and associated with poor prognosis in head and neck squamous carcinoma by bioinformatics analysis 26 . However, NSUN2 was few studied in tumors formation related its m 5 C modification activity, especially gastrointestinal cancers. In this study, we found NSUN2 was upregulated in gastric cancer tissues, compared to adjacent normal gastric tissues in mRNA and protein levels. Interestingly, there is much more heterogeneity at protein than mRNA levels both in gastric cancer tissues and normal adjacent tissues. There may be multiple reasons for inconsistent changes in proteins and mRNAs 27,28 . Post-translational regulation of NSUN2 may be one of the important reasons, which may be different in various gastric cancer cases. Its detailed mechanism needs further investigation. Subsequently. We also found NSUN2 could increase gastric cancer cells proliferation both in vitro and in vivo significantly.
Subsequent RNA-seq and KEGG analysis found that cell cycle was the main pathway regulated by NSUN2 in gastric cancer. Cell cycle dysregulation is a hallmark of cancer due to uncontrolled proliferative signaling 29 . The affected transitions in the cell cycle are regulated by the balanced activities of cyclin-dependent kinases (CDKs) and CDK inhibitors. Although p57 Kip2 might not be the only targeted gene of NSUN2, our results confirmed that p57 Kip2 was an important downstream gene regulated by NSUN2 in gastric cancer. p57 Kip2 is the recently found CDK inhibitors of the Cip/Kip family, and has been involved in many biological processes, including cell cycle control, differentiation, apoptosis, tumorigenesis and development 30,31 . Recent studies indicated that p57 Kip2 was frequently downregulated in multiple types of human cancers such as breast cancer, hepatocellular carcinoma, colorectal cancer, and ovarian cancer 32,33 . Importantly, De and colleagues 34 found that p57 Kip2 could serve as a tumor suppressor in gastric cancer. In our study, we found that the expression level of NSUN2 was negatively correlated with p57 Kip2 and the ability of NSUN2 knockdown cells proliferation was enhanced after p57 Kip2 silencing in gastric cancer. It revealed another regulatory mechanism that NSUN2 play an oncogenic role by repressing p57 Kip2 expression in gastric cancer.
Previous studies showed that RNA methylase play potential role in regulating mRNA decay, translation, and processing 35 . In our study, we demonstrated that relative half-life of p57 Kip2 mRNA increased in NSUN2 knockdown cells. These results indicated that RNA methyltransferase NSUN2 could affect p57 Kip2 mRNA stability. Wang et al. found that different functional mechanisms of RNA methyltransferase NSUN2 depended on the location of methylation by m 5 C modification 19,20,[36][37][38] . Our dualluciferase reporter assay also found that the luciferase activity of the wild-type 3′-UTR of p57 Kip2 reporter gene was significantly enhanced, compared with the mutated 3′-UTR of p57 Kip2 reporter gene. Subsequently, our m 5 C RIP and qRT-PCR assays found that m 5 C antibody significantly enriched 3′-UTR of p57 Kip2 mRNA and knockdown of NSUN2 reduced the m 5 C levels on 3′-UTR of p57 Kip2 mRNA, indicating that RNA methyltransferase NSUN2 regulated p57 Kip2 expression by m 5 C modification in 3′-UTR of its mRNA. Based on the above results, we elaborated a novel mechanism indicating that NSUN2 mainly methylated the 3′-UTR of p57 Kip2 mRNA, which led to the downregulation of p57 Kip2 at the RNA level and related protein levels. In the present study, we explored the effects of NSUN2 on gastric cancer cell cycle progression and demonstrated that NSUN2 could repress p57 Kip2 by an m 5 C-dependent manner.
In summary, we found that NSUN2 acted as an oncogene through promoting gastric cancer development by repressing p57 Kip2 in an m 5 C-dependent manner, which may provide a novel therapeutic target against gastric cancer.

Bioinformatics analysis
Clinical data for Bioinformatics analysis were downloaded from The Cancer Genome Atlas (TCGA) database (https://cancergenome.nih.gov/), and applied for analyzing the expression of NSUN2 in 415 gastric cancer tissues and 34 normal tissues.

Clinical sample
Twenty pairs of gastric tumor and adjacent normal tissues were obtained from the First Affiliated Yijishan Hospital with Wannan Medical College from 2017 to 2018. All samples were obtained with written informed consent from patients and the ethics committee of Wannan Medical College approved these tissues for research use.

Cell culture and transfection
Human gastric cancer cell lines MGC 803 and SGC 7901 were obtained from the Chinese Academy of Sciences Committee on Type Culture Collection Cell Bank (shanghai, China), and has recently been tested for mycoplasma contamination. The cells were maintained in RPMI 1640 medium (GIBCO, USA) supplemented with 10% fetal bovine serum (GIBCO, USA), 100 U/ml penicillin-streptomycin (GIBCO, USA). All the cell lines were maintained at 37°C, 5% CO 2 .
Lentivirus constructs for NSUN2 overexpression and knockdown were obtained from Obio Technology (Shanghai, China), and generated as described previously 39 . Briefly, the gastric cancer cells were stably transfected with NSUN2 overexpression lentivirus (termed as NSUN2) and negative control (termed as Ctrl) using polybrene (Obio Technology, China). Similarly, cells were stably transfected with negative control (termed as shCtrl) and NSUN2 knockdown lentivirus (termed as shNSUN2#2 and #3). Subsequently, stably transfected gastric cancer cells were used for further studies by selection using puromycin (5 μg/ml) for 1-2 weeks.
To explore the further relationship between NSUN2 and p57 Kip2 in gastric cancer cells, small-interfering RNAs (siRNAs) against p57 Kip2 or negative control RNAs were purchased from GenePharma (Shanghai, China), and transfected into the stably transfected gastric cancer cells using EndoFectin TM -Max (GeneCopoeia, China) according to the manufacturer's instructions.

RNA m 5 C dot blot assay
Total RNA was first isolated from stable NSUN2 overexpression and knockdown cells and their corresponding negative control cells, and then treated with deoxyribonuclease I (DNase) according to the manufacturer's protocol. RNA quality was analyzed by NanoDrop 2000 (Thermo Scientific, USA). Different amounts of RNA (400,600,800 ng) were loaded onto the Amersham Hybond N + membrane (GE Healthcare) fixed on the Bio-Dot apparatus (Bio-Rad). After ultraviolet rays crosslinking for 5 min at 254 nm, the membrane was blocked with 5% non-fat dried milk in phosphate buffer solution with Tween-20 (PBST) followed by incubation with the primary mouse anti-m 5 C antibody (ab10805, Abcam, USA) and corresponding anti-mouse secondary antibody. After, the membranes were washed three times with 0.1% PBST, the intensity of the dot blot was determined by autoradiographed and analyzed by image J.

Cell proliferation assay
The transfected gastric cells were monitored in 96 wells at~2000 cells per well incubated at 37°C in an atmosphere of humidified air 5% CO 2 incubator, and the cell proliferation assay was carried out with cell counting kit-8 (CCK-8, Obio Technology, China) following the manufacturer's protocol after 5 days of culture. Optical densities (ODs) were measured at 490 nm with a microplate reader (BioTek, USA). For colony formation assay, a certain number of cells were placed into each well of 6well plates and cultured for 14 days before stained 0.5% crystal violet (Beyotime, China). The colonies with more than 50 cells was manually counted.

Flow cytometry analysis
The transfected gastric cells were seeded in 96-well plates and harvested by trypsinization with 0.25% EDTA (Sigma, USA), when the cells were grown to 80% confluence. Cells for cell cycle were stained with propidium iodide (BD Biosciences, USA) and the percentages of cells in G0/G1, S, and G2/M phase were analyzed by flow cytometry (Cty-toFLEX, Beckman,USA) and ModFit LT 5.0software.

Tumor xenograft
Four-to eight-week-old female BALB/c nude mice were purchased from model animal research center of Nanjing University, and randomly divided into four groups (five per group). Five mice/group were subjected to the experiment to have statistical importance. Sable NSUN2 overexpression cells, NSUN2 knockdown cells and their corresponding wild-type cells (6 × 10 6 cells in 100 ml PBS) were injected into the upper left flank region of each nude mice. Mice were sacrificed and tumor tissues were collected after 3 weeks. Tumor volumes (1/2 × length × width 2 ) and weight were measured in mice. Hematoxylineosin (HE) staining and IHC were performed on processed and sectioned tissues from Serviebio company (Wuhan, China). The investigators were not blinded to the mice group during experiments. All procedures were approved by the animal care and use committee of Nanjing Medical University (acceptance no.: IACUC1804027).

RNA-seq
RNA samples were isolated from stable NSUN2 knockdown cells and corresponding negative control cells in MGC 803 cells. The RNA samples were sequenced by the Allwegene Technology Inc (Beijing, China). Three biological replicates for each sample were included in this experiment. The threshold for screening differential genes (DEG) is generally as follows: |log2(Foldchange)| > 1, qvalue < 0.05 and read count ≥100. DEG function analysis was by Gene Ontology (GO) and KEGG.

RNA stability assay
To analyze RNA stability, stable NSUN2 knockdown cells and corresponding wild-type cells were treated with actinomycin D (1 μg/ml). Cells were collected at different time points (0, 2, 4, 6, 8 h), and RNA was extracted using Trizol reagent. Reverse transcription was performed using oligo primers and mRNA levels were measured using qRT-PCR.

Dual-luciferase reporter assay
Dual-luciferase reporter assay was performed using Luc-Pair TM Duo-luciferase HS assay kit (GeneCopoeia, USA) according to the manufacturer's instructions. Wild-type and mutant 3′-UTR of p57 Kip2 reporter plasmid were constructed from GenePharma (USA). Briefly, wild-type 3′-UTR of p57 Kip2 reporter plasmid was made by inserting the 3′-UTR of p57 Kip2 transcript after the Fluc coding sequence, and mutant 3′-UTR of p57 Kip2 reporter plasmid was made by inserting the 3′-UTR of p57 Kip2 transcript without "CCG" sequence according to Yang et al. 15 . Stable NSUN2 knockdown and corresponding negative control cells were plated in 96 wells dishes and infected with 100 ng of wild-type and mutated 3′-UTR of p57 Kip2 transcript reporter plasmid using EndoFectin TM -Max (GeneCopoeia, China). After 24 h, cells were collected and assayed with NanoDrop 2000 (Thermo Scientific, USA). Firefly luciferase (F-luc) was used to assess the effect of m 5 C modification on p57 Kip2 expression. Renilla Luciferase (R-luc) was used to standardize the transfection efficiency of the reporter plasmid.

m5C RNA RIP assay
For m 5 C RIP, the standard procedure was described as previous study with some modifications 37,38 . Briefly, total RNAs were firstly isolated and treated with DNase. Then, total RNAs were chemically fragmented (~100 nucleotide) with 1×fragmentation buffer (100 mM Tris-HCl, 100 mM ZnCl 2 ) and incubated with the m 5 C antibody (Abcam, USA) in IP buffer (50 mM Tris-HCl, 750 mM NaCl and 0.5% (vol/vol) Igepal CA-630). The IP samples were washed with elution buffer [1×IP buffer, 6.7 mM 5methylcytosine hydrochloride (Sigma, USA)]. Enrichment of m 5 C containing mRNA was analyzed by RT-PCR.

Statistical analysis
All experiments in this study were repeated in triplicate, unless otherwise specified. All dates were presented as the mean ± SD, and student's t-tests (unpaired, two-tailed) were performed using the SPSS 19.0 software (SPSS, Chicago, IL, USA) and graphical presentations were conducted with GraphPad Prism 7.0 software (San Diego, CA). p-value < 0.05 was statistically significant.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.