Stmn1 up-regulates Cdx2 expression and participates in gastric intestinal metaplasia in vitro and in vivo: a randomized controlled trial

BACKGROUND AND PURPOSE: Stmn1 is over-expressed in almost all pathological stages during gastric cancer development, such as chronic atrophic gastritis, dysplasia, and gastric cancer. IM is an important precancerous lesion of gastric cancer, however, whether Stmn1 was up-regulated or down-regulated in this stage is still unknown. We aimed to evaluate the expression level of Stmn1 in IM in vivo and its relationship with important gene of IM named Cdx2 in vitro. EXPERIMENTAL APPROACH: Wistar rats (n=12, sex in half) were gavaged with MNNG (167μg/ml) to induce IM model in stomach. After pathological examination with AB staining to conrm that the model was successful, relative expression level of Stmn1 was detected between normal group and model group using RT-qPCR. Human gastric cell line GES-1 was used to investigate whether Stmn1 inuence expression level of IM essential gene Cdx2 by over-expressing or down-expressing experiments, RT-qPCR and western blot. KEY RESULTS: We have demonstrated that Stmn1 was up-regulated in IM model induced by MNNG in rats in vivo, and it could signicantly up-regulate Cdx2 expression level in human gastric cell line GES-1 in vitro. CONCLUSIONS AND IMPLICATIONS: We demonstrated that Stmn1 was involved in IM in this model and it could up-regulating Cdx2 in human gastric cell line GES-1 in vitro. These results suggested that Stmn1 might be a potential biomarker or candidate treatment target of IM in stomach.


Background
There were 1,033, 701 new gastric cancer cases worldwide in 2018 (WHO). IM (IM), which refers to the transformation of gastric epithelium into intestinal phenotype, is the most common precancerous lesion of gastric cancer [1,2]. Control and reversion of precancerous lesions is the key point to prevent deteriorating into gastric cancer. Cognition of the mechanisms of precancerous lesions is a prerequisite.
Stathmin 1 (Stmn1) belongs to the stathmin gene family, and the protein it encoded prevents assembly and promotes disassembly of microtubules. In gastric cancer, Stmn1 expression level is up-regulated [16], and it positively correlated with lymph node metastasis, TNM stages and vascular invasion, and negatively with recurrence-free survival [17]. Stmn1 gene silencing could suppress proliferation, migration and invasion of gastric cancer cells via AKT/sCLU and STAT3 signaling [17][18][19][20]. Over-expression of Stmn1 is associated with poor prognosis of patients with gastric cancer [21]. However, its expression in IM is unknown. The gastric cancer model [22] induced by 1-methyl-3-nitro-1-nitrosoguanidine (MNNG) in rats can simulate the whole process from normal human gastric to chronic gastritis, chronic atrophic gastritis, IM, intraepithelial neoplasia, gastric adenocarcinoma, the most common type of gastric cancer in human [23], which provides possibility and su cient window period for drug intervention to prevention or treatment of gastric cancer.
Previously, we found that Stmn1 is signi cantly higher expressed in chronic atrophic gastritis in rats [24], here we investigate whether Stmn1 and IM in stomach of IM model rat, and detected the effect of Stmn1 on the expression of Cdx2 in human gastric cells GES-1.

Rat experiments
Twenty-four 6-week-old wistar rats, weighing around 200g, were bought from Shanghai SLAC Laboratory Animal Co. Ltd (Shanghai, China) and randomly divided into normal control group (n=12) and model group (n=12) using strati ed random sampling, sex in half. All rats were kept on a 12-hour light-dark cycle in a pathogen-free animal feeding facility at Zhejiang Academy of Traditional Chinese Medicine.
The maximum caging density was six mice. The animal study protocols were approved by the Medical Rats in the model group were provided with ample fresh MNNG solution (167μg/ml) stored in light-free bottles for 20 weeks. On the last day, all rats were anesthetized by intraperitoneal injection of pentobarbital sodium (50mg/kg, 2%), then stomachs were removed along with 0.5 cm of esophagus above the stomach and 1.5 cm of duodenum below the stomach. After these, liver and blood of each rats were collected for further experiments, as a result every rat was sacri ced on test-bed. Each stomach was vertically divided into 2 parts; one part was frozen in liquid nitrogen for total RNA isolation, and the other part was xed in 10% formalin. AB staining was used to observe the IM of gastric mucosa (goblet cells).
For each rat, ve different investigators were handled as follows: a rst investigator (RZ) administered the treatment following strati ed random sampling and the randomization table. A second investigator (CY) was responsible for the anaesthetic procedure, whereas a third investigator (CX, TY, CS, DG) performed the surgical and x procedure. A fourth investigator (ZF) (also unaware of treatment) performed pathological section and evaluation of pathological results. Finally, a fth investigator (WR) performed data analysis.

Histological section assay
The rat stomach samples in formalin were embedded in para n and sliced at 4-μm thickness. The slices were then separately stained with AB-PAS following routine procedures. Sections were examined under light microscope. The investigator was blind to group assignment and outcome assessment. IM was determined according to intestinal type cells in gastric gland cells.

RT-qPCR
The RT reaction was performed using 500 ng of total RNA according to the manufacturer's instructions (RR036A, Takara). The qPCR reactions were performed on the StepOnePlus system (ABI) with the SYBR® Premix ExTaqTM kit (RR820A, Takara). The endogenous control gene was RPL13A for cell experiments and EF-1a for rat experiments. The relative expression levels were calculated using the 2 -△△Ct method [25]. Primers were as follows: Stmn1 (rat, homo), Forward: Western Analysis (ProteinSimple ® ) was used to detect protein abundance according to the manufacturer's protocol. In brief, 0.6 µg whole protein sample was loaded into each well, and Size Range of the separation matrix was 12-230 kDa. Target proteins were identi ed with following primary antibodies: Cdx2 (12306; Cell Signaling Technology ® ), Stmn1 (3352; Cell Signaling Technology ® ), SOX2 (23064; Cell Signaling Technology ® ). GAPDH (60004-1-lg; Proteintech ® ) was used as a reference control. According to primary antibodies, mouse or rabbit secondary antibodies provided by the manufacturer were used. Chemiluminescent signals were detected and quanti ed using Compass Software (ProteinSimple), the area value was used as the protein expression level.
Statistical analysis SPSS Statistics 17.0 was used to analyze the data. Body weight and gene expression level among 5 groups were analyzed using one way ANOVA, and between 2 groups were analyzed using independent sample t-tests. Differences were considered statistically signi cant at P<0.05.

Stmn1 was signi cantly higher expressed in IM in stomach tissues in rats
To investigate whether Stmn1 is over-expressed in IM, we detected its expression level in MNNG induced IM in stomach of rats. In typical pictures of pathological sections (Fig 1a.), there was a very small amount of blue staining in the deep part of gastric mucosa in the antrum of normal stomach; there was blue staining in the whole layer from the deep part to the shallow part of the antrum mucosa, and there were goblet cells, which are the sign of IM. In model group, there were 10 IM rats, the other 2 rats failed to be induced into IM model Then we divided all 24 rats in the normal group and model group into two groups: the group without IM (n=14) and the group with IM (n=10). The t-test showed that the expression level of Stmn1 in IM group was signi cantly higher than that in normal group (P =0.003; Fig 1b). The relationship between Stmn1 and important genes of intestinal meplasia, such as Cdx2 needed to be established. However, we could not detect the expression of Cdx2, a key regulatory gene of IM, in rat IM gastric tissue, and no other literature has reported about Cdx2 in IM model rats induced by MNNG.

Stmn1 signi cantly up-regulates Cdx2 expression level in GES-1 cells
To explore the effect of Stmn1 on Cdx2, we over-expressed or knock-down its expression along with overexpression of Cdx2 in GES-1 cells (Fig 2).
Relative mRNA expression level of Cdx2 and Stmn1 in each group was detected by RT-qPCR. The results showed that Cdx2 was not expressed in normal group (Fig 3). Among other 4 groups, relative expression level of Cdx2 was tested by Levene variance homogeneity test and the variance of each group was not homogeneous (P = 0.011), then one-way ANOVA of grouping design was used to evaluate the difference, and the results showed that the expression level of Cdx2 among 4 groups was signi cantly different (F= 65.179, P =0.000; Fig 3). Compared with Cdx2 over-expression group, Cdx2 expression level in Cdx2 overexpression + Stmn1 over-expression group was signi cantly higher (P = 0.000 ; Fig 3.), indicating that Stmn1 could signi cantly up-regulate Cdx2 expression level. Compared with Cdx2 over-expression + Stmn1 over-expression group, Cdx2 expression level in Cdx2 over-expression + Stmn1 over-expression + Stmn1 siRNA group decreased signi cantly (P = 0.000 ; Fig 3.), which was lower than that in Cdx2 group (P = 0.046 ; Fig 3.), indicating that blocking the over-expression of Stmn1, Cdx2 was signi cantly down regulated. However, expression level of Cdx2 was no difference between Cdx2 over-expression group and Cdx2 over-expression+Stmn1 siRNA group.
Relative protein expression level of Cdx2 and Stmn1 in each group was detected by western blot on Wes instrument (n=3 ; Fig 4a, b). Cdx2 was not expressed on protein level in normal group, which consistent with mRNA level results. In the other 4 groups, Cdx2 protein abundant among groups was signi cantly different (F = 56.622, P = 0.000). Compared with Cdx2 over-expression group, the protein level of Cdx2 in Cdx2 over-expression + Stmn1 over-expression group was signi cantly higher (P = 0.006), indicating that Stmn1 could signi cantly up regulate Cdx2 on protein level, which support the mRNA level results. Compared with Cdx2 over-expression + Stmn1 over-expression group, Cdx2 protein abundance in Cdx2 over-expression + Stmn1 over-expression + Stmn1 siRNA group was signi cantly lower (P = 0.000), indicating that blocking Stmn1 over-expression, Cdx2 protein expression was signi cantly inhibited. Compared with Cdx2 over-expression group, Cdx2 protein expression level in Cdx2 over-expression + Stmn1 siRNA group did not change (P = 0.540), indicating that blocking Stmn1 expression, Cdx2 protein level expression did not change, consistent with mRNA level change.

Discussion
Gastric cancer is one of the most common malignant tumors worldwide. The chains of process from normal gastric mucosa to intestinal type gastric carcinoma include super cial gastritis, atrophic gastritis, IM, dysplasia, carcinoma, metastasis [23]. We found it over-expressed in chronic atrophic gastritis in rats [24]. Here we found that Stmn1 was up-regulated in IM in stomach in rats. Moreover, it could induce signi cant higher expression of Cdx2, a key gene in IM, in human gastric GES-1 cells, which suggested that Stmn1 might involve in the development of IM by up-regulating Cdx2. These results initially ll the IM gap of expression level of Stmn1 in development of gastric cancer. However, clinical evidences about whether Stmn1 was over-expressed in IM and/or chronic atrophic gastritis in patients need to be studied further.
Over-expression of Stmn1 alone did not induce Cdx2 expression, indicating that Stmn1 could not trigger Cdx2 expression, but promoted Cdx2 expression.
Down-regulation of Stmn1 in Cdx2 over-expressed group did not inhibit the expression of Cdx2, indicating that low expression level Stmn1 could not regulate Cdx2 and also knock-down low expression Stmn1 could not effect Cdx2. Only over-expressed Stmn1 could promote Cdx2 expression, and knock-down high expression level Stmn1 could inhibit over-expressed Cdx2.
Over-expression of Cdx2 alone had no effect on Stmn1 expression, indicating that Cdx2 could not affect the expression level of Stmn1.

Conclusions
Stmn1 was up-regulated in IM in rats in vivo and it could signi cantly up-regulate an IM essential gene Cdx2 in human gastric GES-1 cells in vitro. All these results suggested that Stmn1 might involve in the development of IM by up-regulating Cdx2. Availability of data and materials All data generated or analysed during this study are included in this published article.

Competing interests
The authors declare that they have no competing interests. Authors' contributions XC, LC and RW generated the conception; XC and GD designed the work; XC, ZR, YT, SC completed the experiments and analyzed the data. XC was a manor contributor in writing the manuscript. All authors read and approved the nal manuscript.