Functional variation of SLC52A3 rs13042395 predicts survival of Chinese gastric cancer patients

Abstract The solute carrier family 52 member 3 (SLC52A3) gene encodes riboflavin transporter protein which is essential to maintain mitochondrial function in cells. In our research, we found that SLC52A3 rs13042395 C > T variation was significantly associated with poor survival in a 926 Chinese gastric cancer (GCa) patients cohort (CC/CT genotype versus TT genotype, HR = 0.57, 95%CI (0.40‐0.82), log‐rank P = 0.015). The SLC52A3 rs13042395 C > T change led to its increased mRNA expression according to expression quantitative trait loci analysis (P = 0.0029). In vitro, it was revealed that rs13042395 C allele had higher binding affinity to inhibitory transcription factor Meis homeobox 1 (MEIS1) compared with T allele, knock‐down of MEIS1 could up‐regulate SLC52A3, and overexpression of SLC52A3 contributed to the increased ability of proliferation, colony formation, migration and invasion in GCa cells. Subsequently, the bioinformatics analysis combined with experiments in vitro suggested that Gap junction protein alpha 1 (GJA1) was the downstream effector of SLC52A3, SLC52A3 may promote the GCa cells aggressiveness by down‐regulating the GJA1 expression. Overall, SLC52A3 genetic variant rs13042395 C > T change was associated with poorer survival in Chinese GCa patients and increased SLC52A3 expression by interaction with MEIS1. SLC52A3 promoted the GCa cells aggressiveness by down‐regulating the GJA1 expression.

such as circulating non-coding RNA and some somatic genes mutations. 3,4 Several germline mutations have already been investigated for the prognosis in GCa. 5,6 However, few of them further explored the molecular mechanism underlying the prognosis effect.
The solute carrier family 52 member 3 (SLC52A3) gene encodes the riboflavin transporter protein, which was crucial for the riboflavin absorption. 7 Previous studies suggested that SLC52A3 gene was highly abnormally expressed in oesophageal tumour tissue, and the depletion of SLC52A3 gene would significantly inhibit the proliferation of oesophageal carcinoma cells, 8 which indicated that SLC52A3 gene played an important role in the progression of cancer cells. Germline mutation (rs13042395 C > T) on SLC52A3 gene was identified as a risk locus for gastric cancer incidence by genome-wide association studies, and it was found that rs13042395 C > T change significantly increased the risk of cancer incidence in various cancer types, notably in the GCa patients. 9,10 However, the molecular mechanism remained ambiguous and the research on the association of germline mutation in the SLC52A3 gene and the survival of GCa patients was hardly seen. In the present study, we investigated associations between genetic mutations of the SLC52A3 rs13042395 C > T change and survival of Chinese GCa patients and subsequently explored mechanistic basis of the observed associations.

| Genotyping, quality control and survival analysis
DNA of study patients was extracted from peripheral blood cells. Single nucleotide polymorphisms (SNPs) were genotyped via a matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometer using the MassARRAY Analyzer 4 platform (Sequenom, CA, USA). All the primers were designed by Assay Design Suite v2.0 from Mysequenom online software (www.myseq uenom.com). The standard polymerase chain reaction (PCR) reaction was conducted in a total volume of 5 μL system containing 10 ng of genomic DNA. One negative control and one duplicate sample were used for quality controls in every 96-well plates. Genotyping results of 5% patients were repeated, and the consistency was 100%.
A total of 926 patients were divided into two groups according to their rs13042395 genotype. Kaplan-Meier (KM) curve was used to illustrate the difference and log-rank test was performed to provide the statistics. Multivariate analysis was calculated by Cox proportional hazards regression which was adjusted by clinical variables.
Stratification analysis was conducted to explore the interaction of the genetic variant with clinical variables.

| Expression quantitative trait loci analysis and validation in vitro
Expression quantitative trait loci analysis was conducted using rs13042395 genotype and SLC52A3 mRNA expression data in normal gastric tissue from the Genotype-Tissue Expression (GTEx) database. The different expression violin plot was directly available from the GTEx website (https://www.gtexp ortal.org/).

| Target sequencing and electrophoresis mobility shift assay
We sequenced the rs13042395 locus in BGC823, MGC803 and AGS cell lines using Sanger sequencing method to make sure these cell lines used in our research were rs13042395 CC genotype.
Electrophoresis mobility shift assay (EMSA) analysis was performed at Viagene Co. Ltd (Shanghai, China) and was used to detect the different binding affinities of nucleoprotein with the rs13042395 C or T allele flanked by 58bp DNA up-and downstream. The DNA sequence of rs13042395 C allele flanked by 58 bp nucleotide was as follows: 5′-TGGG G T T C T G A C C A G G G C C A G T G C A C C G T C A T T G T G T G GGCTGGGCCATCTCCTCCAGG-3'. The rs13042395 T allele flanked by 58 bp nucleotide was as follows: 5'-TGGGGTTCTGA C C AG G G C C AG TG C AC C G T TAT TG TG TG G G C TG G G C C AT CTCCTCCAGG-3'. Bold characters in the sequence represent the rs13042395 locus. The probe representing the rs13042395 C allele (bio-CLwild) was as follow: 5'-CCAGTGCACCGTCATTGTGTGGGCT-3', and the probe representing the rs13042395 T allele (bio-CLmut) was as follow: 5'-CCAGTGCACCGTTATTGTGTGGGCT-3'. The positive control probe was as follow: 5'-CCAGTGCAAAGAGCTTGTGTGGGCT-3', and the negative control probe was as follow: 5'-CGTACGC GCTGTCATATTGACAGGT-3'. The nucleoprotein of BGC823 and AGS cell lines were extracted and incubated with different probes which were labelled with biotin. After that, nucleoprotein-probes complex were separated by electrophoresis and then biotin signals were detected to assess the quantities of nucleoprotein-probes complex.
In the competitive EMSA assay, the probe representing the rs13042395 C allele, which was not labelled with biotin was called In the super-shift EMSA assay, MEIS1 antibodies were added to ascertain whether the DNA-protein complex can bind the antibodies.

| Cell lines and culture
Human gastric cancer lines BGC823, MGC803 and AGS were obtained from the Cell Bank of Type Culture Collection of Chinese Academy of Sciences (Shanghai, China). All of the above cell lines have been authenticated by short tandem repeat (STR) DNA profiling analysis. Cells were cultured in RPMI 1640 medium supplemented with 10% FBS (Gibco, USA) and antibiotics at 37°C, 5% CO 2 .

| Gene sets enrichment analysis, lentiviruses and infection
The gene sets enrichment analysis (GSEA) of two Gene Expression Omnibus (GEO) data sets (GSE62254 and GSE15459) and TCGA gastric cancer data set were conducted to calculated the most likely downstream effector gene of SLC52A3.

Lentiviral vectors expressing SLC52A3 and Gap junction
protein alpha 1 (GJA1) were obtained from Shanghai Hanyin Biotechnology Co Ltd. MGC803 and AGS cells were infected with lentivirus carrying SLC52A3 or GJA1, and transfected cells were selected with indicated antibiotics to generate overexpressing stable cells. Western blotting was applied to detect the expression level of SLC52A3 and GJA1, and GAPDH was used as the internal reference protein.

| Western blotting and PCR
Firstly, the whole cell lysates were obtained as the supernatant, and protein concentration was determined using the Pierce BCA Assay Kit (Thermo Scientific). Then, a total of 30 μg protein was loaded into each lane and separated by a 10% SDS-PAGE gel. After that, proteins were transferred into PVDF membranes. The membranes were blocked with 5% of skim milk for one hour at room temperature and incubated in primary antibodies at 4°C overnight.

| Proliferation, migration and invasion assay
The EdU cell proliferation assay and plate colony formation assay were used to evaluate the ability of cell proliferation. For EdU cell proliferation assay, cells were seeded into 24-well plates with a density of 5 × 10 4 cells per well. After 24-hour culture, cells were incubated with EdU for 4 hours, then fixed with 4% paraformaldehyde and permeabilized by 0.5% Triton X-100. After that, EdU staining and nuclear staining with DAPI were applied to identify proliferative cells. Finally, images were captured using confocal microscopy and the proportion of EdU-positive cells was calculated. For plate colony formation assay, cells were seeded into 6-well plates with a density of 1 × 10 3 per well. After 10-day culture, colonies were fixed with 4% paraformaldehyde, stained with 0.1% crystal violet and counted. All experiments were performed in triplicate. The transwell assay was used to assess cell migration and invasion with the transwell system (24-well insert, pore size, 8 μm; Corning co. Ltd, Corning, New York, USA) For migration assay, 41 × 10 4 cells in 200 μL serum-free medium were plated into the upper chamber, and 600 μL medium containing 20% FBS was added into the lower chamber. 24 hours later, the cells that migrated to the lower side of chamber membrane were fixed with 4% paraformaldehyde, stained with 0.1% crystal violet and counted. For invasion assay, Matrigel (1:10) was polymerized in upper chamber for 45 minutes at 37°C, then 11 × 10 5 cells in serum-free medium were plated into the upper chamber, and 600 μL medium containing 20% FBS was added into the lower chamber.
48 hours later, the cells that invaded to the lower side of chamber membrane were fixed with 4% paraformaldehyde, stained with 0.1% crystal violet and counted.
All experiments were performed with mycoplasma-free cells.

| Statistic analysis
The survival analysis was assessed by KM method, log-rank test and Cox proportional hazards regression. The expressional differences of SLC52A3 gene were analysed by unpaired t test. The statistical analysis was performed by R language (version 3.5.1). P values were two-sided with a significance level of 0.05.

| SLC52A3 rs13042395 C allele predict better survival in GCa patients
In Figure 1, we demonstrated that SLC52A3 rs13042395 C > T change was significantly associated with OS in GCa patients with recessive model (ie TC/CC genotype versus TT genotype, log-rank P = 0.015). After adjusted by clinical variables, SLC52A3 rs13042395 was proved to be an independent prognosis factor by multivariate

| SLC52A3 rs13042395 T allele significantly increase the SLC52A3 expression by altering its binding affinities with MEIS1 compared to rs13042395 C allele
From GTEx database, we found that rs13042395 C > T change was significantly associated with higher SLC52A3 mRNA expression level in normal gastric tissue (P = 0.0029) (Figure 2A).
According to the Jaspar web server, among the transcription factors which could bind to the SLC52A3 rs13042395, the transcription factor MEIS1 was found to have the biggest binding  Targeting sequencing showed that MGC803, AGS and BGC823 cell all carried rs13042395 CC genotype ( Figure S1A). In EMSA analysis, it was revealed that the binding affinity of rs13042395 T allele probes with nucleoprotein was weaker than the rs13042395 C allele in BGC 823 cell line while this interaction of probes and nucleoprotein was not observed in AGS cell line ( Figure 3A). In the competitive EMSA assay ( Figure 3B), binding affinity of rs13042395 C allele probes (ie bio-CLwild probe) with MGC823 cell nucleoprotein was attenuated after adding the rs13042395 C allele probes without labelled with biotin (ie CLwild probe), and this effect was amplified when the concentration was elevated from 40 µmol/L to 80 µmol/L.
The binding affinity of bio-CLwild probe with nucleoprotein was also attenuated by adding the standard probes for detecting the MEIS1 protein (ie std-MEIS1 probe), and also, this effect was amplified when the concentration was elevated from 40 µmol/L to 80 µmol/L.

| SLC52A3 regulates malignant phenotype of gastric cancer cells through down-regulation of GJA1
The GSEA of two GEO data sets (GSE62254 and GSE15459) and TCGA gastric cancer data set revealed that GJA1 may be a downstream target gene for SLC52A3 ( Figure S2), we further investigated whether SLC52A3 regulated malignant phenotype of gastric cancer cells via regulating GJA1. We firstly detected protein ( Figure S1B).

| D ISCUSS I ON
It was well established that precise prognosis prediction played an important role for treatment decision-making for GCa, for the reason that high-risk patients may need more aggressive interventions and low-risk patients need avoid overtreatments. 11 In this relatively large survival analysis in Chinese GCa patients which contained over 900 patients, we identified that SLC52A3 rs13042395 C > T variation would independently significantly impact the OS of patients and demonstrated that SLC52A3 expression was increased due to the different binding affinities of rs13042395 C > T change with transcriptional factors. Specifically, rs13042395 C allele had a higher binding affinity with inhibitory transcription factors MEIS1 compared to the rs13042395 T allele so that rs13042395 T allele significantly increased the expression of SLC52A3 compared to C allele. Further mechanism investigation found that overexpression of SLC52A3 contributes to the increased ability of proliferation, colony formation, migration and invasion likely via down-regulation of GJA1 expression in vitro.
SLC52A3 was essential for absorption of riboflavin, which was a critical component of the mitochondrial electron transport chain, and loss functional mutation of SLC52A3 may lead to Brown-Vialetto-Van Laere syndrome, a rare neurological disorder characterized by bulbar palsies and sensorineural deafness. 12 As for the cancer risk, riboflavin supplement was thought to reduce various types of cancer risks, 13 and several cancer risk research found that SLC52A3 SNPs was associated with oesophageal cancer risk, [14][15][16] probably activated by NF-κB p65/Rel-B. 17 Previous studies also found that SLC52A3 rs13042395 C > T change could promote glioma cancer cell progression and migration in vivo and in vitro. 18 Interestingly, researchers found that SLC52A3 rs13042395 TT genotype favoured reduced lymph node metastasis rate and longer relapse-free survival in oesophageal squamous cell carcinoma. 19  GJA1, which was enriched in the downstream of SLC52A3 according to our research, was a member of the connexin gene family and encoded cell gap junction protein, and it was crucial for low molecular weight materials transportation from the cell to cell. 20 Researchers already confirmed that the connexin gene family played a complicated role in tumorigenesis and progression. 21 Recently researcher found that GJA1 was associated in glioblastoma cancer cells apoptosis and promote the progression and invasion in breast cancer cells. 22,23 Few research about GJA1 was conducted in GCa, one of them suggested that overexpression of GJA1 leads to decreased ability of colony forming and invasive ability in BGC-823 cells, 24 which was consistent with our research results.
In conclusion, we found that SLC52A3 rs13042395 C > T change independently predicted the survival in Chines eastern GCa patients. SLC52A3 rs13042395 T allele had a lower binding affinity with inhibitory transcription factor, MEIS1, leading the up-regulation of SLC52A3 gene, SLC52A3 overexpression is associated with aggressive phenotype in GCa cells likely via down-regulation of GJA1 gene. However, because our research was conducted just in one medical centre and only include Han ethnicity patients, additional multi-centres and multi-ethnicities studies or clinical trials as well as mechanism investigations in vivo are needed to confirm these results.

ACK N OWLED G EM ENTS
We gratefully acknowledge the financial supports by the National Science Foundation of China, and we also gratefully acknowledge the tissue bank of Fudan University Shanghai Cancer Center (FUSCC) for their assistance and kindly advise. This work was supported by the Natural Science Foundation of China (grant numbers. 81101808 and 81902875).

F I G U R E 5
The interaction of SLC52A3 and GJA1 in GCa cells. Overexpression SLC52A3 down-regulated GJA1 expression (A), GJA1 overexpression by lentiviral vectors was validated by Western blotting assay (B). Overexpression of SLC52A3 promoted cell proliferation, colony formation, migration and invasion, while GJA1 overexpression inhibited malignant phenotype (C-G). Co-transfection of SLC52A3 and GJA1 restored proliferation, colony formation, migration and invasion of GCa cells (C-G). GCa, gastric carcinoma. *P < 0.05, **P < 0.05

CO N FLI C T S O F I NTE R E S T
The authors have no conflicts of interest to declare. Writing-original draft (supporting); Writing-review & editing (lead).

DATA AVA I L A B I L I T Y S TAT E M E N T
The deidentification data that support the findings of this study are available on request from the corresponding author. The raw data are not publicly available due to privacy or ethical restrictions, any researcher who are interested in getting raw data are welcome to contact Fudan University institutional review board to obtain permission.