Integrated analysis of Solute carrier family-2 members reveals SLC2A4 as an independent favorable prognostic biomarker for breast cancer

ABSTRACT Most of Solute carrier family-2 (SLC2) members play a key role of facilitative transporters, and glucose transporter (GLUT) proteins encoded by SLC2s can transport hexoses or polyols. However, the function and mechanism of SLC2s remain unclear in human cancers. Here, we explored the dysregulated expression, prognostic values, epigenetic, genetic alterations, and biomolecular network of SLC2s in human cancers. According to the data from public-omicsrepository, SLC2A4 (GLUT4) was found to be significantly downregulated in most cancers, and higher messenger RNA (mRNA) expression of SLC2A4 significantly associated with better prognosis of breast cancer (BRCA) patients. Moreover, DNA hypermethylation in the promoter of SLC2A4 may affect the regulation of its mRNA expression, and SLC2A4 was strongly correlated with pathways, including the translocation of SLC2A4 to the plasma membrane and PID INSULIN PATHWAY. In conclusion, these results provide insight into SLC2s in human cancers and suggest that SLC2A4 could be an unfavorable prognostic biomarker for the survival of BRCA patients.

Several studies have found aberrant expressions and unique mechanisms in some members of GLUTs family. For example, GLUT1 overexpression promoted the glycolysis process in many human cancers, such as gastric adenocarcinoma and breast cancer carcinoma and adenocarcinoma [7]. Besides, GLUT1 expression was regulated by long non-codingRNA HOX transcript antisense RNA and microRNA miR-150 [8,9]. GLUT2 transporters showed the activity in hepatocellular carcinoma cell, and the knockdown of GLUT2 can induce the apoptosis in HepG2 cells [10]. Moreover, GLUT3 may play an important role in proliferation and apoptosis in human cancers [11].
GLUT4 was found to be correlated with 18 F-flurodeoxyglucose uptake in gastrointestinal stromal tumor [12]. Furthermore, GLUT5 can act as fructose transporter in vivo in human breast cancer [13]. However, the prognostic significance and molecular mechanisms of SLC2s remain unclear.
Nowadays, with the development of high throughput technologies, as well as public attention to cancer genomes, researchers can use The Cancer Genome Atlas (TCGA) to explore the molecular mechanisms and genomic changes of a variety of human tumors. Thus, cancer prevention, diagnosis, and precision therapy were greatly speeded up. In this study, we analyzed the messenger RNA (mRNA) expression, gene mutation, and methylation modification of SLC2s in human cancers based on TCGA datasets and discussed their prognostic value and gene regulatory network in breast cancer.

SLC2s-mRNA expression analysis
The oncomine (www.oncomine.org) database was used to analyze the expression of SLC2s-mRNA between different cancer tissues [14]. The cut-off of p-value and fold change were as follows: p-value: 0.01, fold change: 1.5, gene rank: 10%. In addition, UALCAN (http://ualcan.path.uab.edu) is an interactive web resource to perform in-depth analyses of gene expression between tumor and normal samples based on individual clinicopathologic features from the TCGA data. The mRNA expression of SLC2s in different cancer subtypes and grades/stages was also analyzed by UALCAN [15].

Kaplan-Meier survival analysis
The relevance between the SLC2s' expression and prognosis was analyzed by Kaplan-Meier plotter (http://kmplot.com/analysis) [16]. "Overall survival (OS)," "auto-select best cut-off," and "only JetSet best probe set" was chosen for calculating and draw Kaplan-Meier survival curve. Best cut-off values were calculated by all possible cut-off values between the lower and upper quartiles, and the best generated execution threshold was used as a cut-off (Supplementary Table 1 216236_s_at. The Affymetrix probeset ID of SLC2A7 was not found in KM plotter. Furthermore, GSE62254 dataset was excluded when KM plots were generated for SLC2s in stomach cancer because of GSE62254 having markedly different characteristics (longer survivals, shifted expression) than the other datasets.

Mutations and copy-numberalterations analysis
cBioPortal (www.cbioportal.org) is an online open access website resource that can be used to interactively explore multidimensional cancer genomics datasets [17]. Gene mutations and copy number alternation of SLC2A4 was analyzed from Invasive Breast Cancer (TCGA, firehorse legacy, 1101 patients/1108 samples) in cBioPortal.

Methylation modification analysis
MEXPRESS (https://mexpress.be) was used to analyze the correlation between DNA methylation and SLC2A4 mRNA expression in 1268 breast invasive carcinoma samples [18]. Besides, MethHC (http:// MethHC.mbc.nctu.edu.tw) was used to study the relationship of DNA methylation in the promoter and SLC2A4-mRNA expression for 839 breast invasive carcinoma samples [19].

Gene regulatory networks analysis
The Search Tool for the Retrieval of Interacting Genes/Proteins (STRING,https://www.string-db. org) was used to construct the interaction network of SLC2A4 with a confidence score of 0.4 [20]. Additionally, Metascape (https://metascape.org) was chosen for gene ontology and pathway enrichment analysis of the genes related to SLC2A4 [21].

SLC2s-mRNA expression in breast cancer
Using the Oncomine database, the mRNA expression of SLC2s in 20 cancer types was analyzed ( Figure 1). Results revealed that SLC2A1, 3,4,5,6,10,13 and 14 were abnormally expressed in most cancer types. Among them SLC2A1, 3, 4 and 14 has the most data available in breast cancer. Compared with normal tissues, SLC2A3, 4 and 14 were significantly downregulated in breast cancer. In the Curtis dataset, SLC2A3 mRNA expression was observed 2.665-fold decrease in invasive lobular breast carcinoma samples, while SLC2A4 downregulation was found in invasive breast carcinoma samples with a fold change of 5.226 from TCGA Breast dataset [22]. TCGA Breast datasets also showed 2.617-fold decrease of SLC2A14 mRNA expression in mucinous breast carcinoma samples. Concurrently, SLC2A1 was significantly up-regulatedin breast cancer. In Zhao Breast dataset, SLC2A1 was overexpressed in invasive ductal breast carcinoma compared with normal tissues with a fold change of 2.800 [23]. However, in Finak's datasets, SLC2A1 was significantly downregulated in invasive breast carcinoma stroma sample with a 3.780-fold change [24].

Prognostic value of SLC2s expression in breast cancer patients
Next, KM plotter was used to investigate the prognostic values of the mRNA expression of SLC2s in breast cancer patients ( Figure 3 and Table 2). Results showed that higher mRNA expression of SLC2A4 and SLC2A11 significantly associated with better prognosis of breast cancer patients (HR: 0.7 [0.55-0.88];p =0.0024 and HR: 0.45 [0.33-0.62]; p =5E-07, respectively). Moreover, higher mRNA expression of SLC2A5 and SLC2A12 significantly   Expression of SLC2A4-mRNA in different major subclasses of TCGA breast cancer samples are shown as Figure 4. Compared with normal tissues, SLC2A4 were significantly reduced in Overall, this result showed that SLC2A4 is a potential prognostic biomarker for breast cancer patients.

Relationship between DNA hypermethylation in the promoter of SLC2A4 and SLC2A4-mRNA expression in breast cancer
We use MEXPRESS to assess whether DNA hypermethylation in the promoter of SLC2A4 was related to mRNA expression of SLC2A4. As shown in Figure 6, according to the Pearson correlation coefficient, mRNA expression of SLC2A4 was positively correlated with its DNA methylation in the promoter (probe ID: cg07287120, r = 0.122; probe ID: cg03670302, r = 0.093; all p < 0.01) and can be negatively correlated with its DNA methylation in the promoter (probe ID: cg17663577, r = −0.104; probe ID: cg21994579, r = −0.222; probe ID: cg06891043, r = −0.113; probe ID: cg27067158,r = −0.391; all p < 0.01). The expression level of SLC2A4 in breast cancer tissues was lower than that in normal samples. Besides, DNA methylation in SLC2A4 promoter region was significantly related to histological type, sample type, and subtype (p = 2.087E-6,p = 1.692E-129 andp = 0.003, respectively). We further compared DNA hypermethylation in the promoter of SLC2A4 in breast cancer patients to normal samples; data from MethHC showed that there were significant differences between the two groups (p < 0.05). All in all, these results demonstrated that DNA hypermethylation in the promoter of SLC2A4 can affect the regulation of its mRNA expression.
To explore the relationship between low mRNA expression of SLC2A4 and genetic alterations, we used cBioPortal to analyze the mutations, copy-number alterations, and mRNA expression transformations of SLC2A4. As shown in Figures 6(c), 6% (68) of 1108 breast invasive carcinoma samples have alternations (p < 0.05). Among them, 1 sample was missense mutation (unknown significance), 1 sample was amplification (unknown significance), 11 samples were deep deletion (unknown significance), 15 samples were mRNA low, and 39 samples were mRNA high. Interestingly, 1 sample was both amplification and mRNA high. These outcomes indicated that alternations of SLC2A4 were most mutually independent.

Biomolecular network regulated by SLC2A4
We further searched potential regulated genes to seek the regulation function of SLC2A4 in    Figures 7(b), 20 gene sets were significantly enriched. The most significantly enriched five gene sets were translocation of SLC2A4 (GLUT4) to the plasma membrane and protein localization to plasma membrane. These results demonstrated that SLC2A4 was related with cellular material transportation and insulin in breast cancer. In addition, SLC2A4 may play an important role in breast cancer, including AMPK pathway, glucose homeostasis, and protein localization.

Discussion
In this study, we systematically explore the mRNA expression, prognostic value, epigenetic and genetic alterations, and gene regulatory network of SLC2A4 in human cancers. Results from our  study showed that the mRNA expression of SLC2A4 was significantly decreased in BRCA tissues, and high level of SLC2A4-mRNA was significantly relevant to better prognosis in breast cancer patients. Compared with normal tissues, SLC2A4 was significantly downregulated in patients with all subtypes and stages of BRCA, and high level of SCL2A4 in luminal A breast cancer and stage III breast cancer can significantly predict better prognosis. Additionally, SLC2A4 can also abnormally express in other types of cancers. According to the latest cancer epidemic trend from China National Cancer Center, we also investigated the relevance between the expression and prognosis of SLC2A4 in lung cancer, liver cancer, and stomach cancer [25]. Results are shown in Supplementary Table 3. It suggested that SLC2A4 can be a common prognostic biomarker in specific stages of multiple cancers. The latest progress in tumor epigenetics shows that DNA methylation, histone modification, and non-codingRNA play an important role in regulating gene expression and chromatin structure, which may lead to the occurrence and development of cancer [26]. In addition, the increase of DNA copy number will also lead to the increase of gene mRNA expression. Above results indicated that SLC2A4 may be an important biomarker to promote the development of accurate diagnosis and prognosis. As one of the sugar transporter proteins in human genome, GLUT4 regulates glucose transport and is encoded by SLC2A4 [27,28].GLUT4 plays an essential role in maintaining body glucose homeostasis and is regulated by insulin. GLUT4 expression was downregulated in adipocytes in obesity and was upregulated in adipocytes and muscle cells in response to exercise [29]. GLUT4 was expressed differently among normal humans  and those with obesity and diabetes, and the overexpression of GLUT4 in skeletal muscle can alter substrate utilization and improve the benefits of insulin [30,31]. Besides, three studies have shown that GLUT4 may play a potential role in tumorigenesis and progression [32][33][34]. GLUT4 downregulation may inhibit glucose uptake and induce metabolic reprogramming; research also found that GLUT4 downregulation can suppress cell proliferation and critically decrease cell viability under hypoxic conditions, especially in MCF7 and MDA-MB-231 breast cancer cells [33]. Besides, 7 antimicrobial peptides (RAB1-7) became anti-cancerdrugs by inhibiting SLC2A4 to impair the energy gained by cancer cells during angiogenesis [32]. Another study demonstrated that high glucose increased SLC2A4 and VEGF/ VEGFR expression by upregulating estrogen receptor and further promoted epithelial-mesenchymal transition process and accelerated the development of uterus endometrial cancer [34]. Moreover, by targeting GLUT4, the silence of krüppel-like transcription factor 8 (KLF8) expression decreased the glycolysis rate of gastric cancer cells in vitro [35]. However, the physiological and pathophysiological mechanisms of GLUT4 protein are still unclear. Multiple studies revealed that GLUT4 was associated with the regulation adipogenesis and blood glucose regulation [36][37][38][39][40]. Interestingly, exercise may be a key role of affecting the translocation and expression of GLUT4; exercise can activate AMPK, PPAR β positive feedback loop and PGC-1 α, and further upregulate GLUT4 expression to enhance glucose uptake by tissue cells [41][42][43]. Some extracts from plants, including quercetin and Moringa concanensis nimmo extracts, can upregulate the expression of GLUT4 [44,45].
We performed functional enrichment analysis of 35 SLC2A4-related genes and discovered some important pathways that may play a key role in tumorigenesis and progression. Pathways that SLC2A4related genes most significantly enrichedinclude translocation of SLC2A4 (GLUT4) to the plasma membrane, PID INSULIN PATHWAY, PID INSULIN GLUCOSE PATHWAY, response to insulin and protein localization to plasma membrane. These pathways may contribute to the discovery of novel mechanisms of SLC2A4.
Insulin is a kind of protein hormone secreted by islet β cells, and it is also the only hormone that can reduce blood glucose in human body. Increasing evidence suggested that the upregulated insulin was associated with tumorigenesis and cancer growth [46][47][48]. Although increased insulin production was a common phenomenon during cancer development, the insulin resistance also occurred in the normal tissues and lead to alterations in carbohydrate and lipid metabolism [48]. Insulin and insulin-like growth factor receptors may play a pivotal role in cell fate determination, and they can regulate cell proliferation, differentiation, apoptosis, glucose transport, and energy metabolism [47]. When exposed to hyperinsulinemia, cancer cells gain a selective growth advantage compared to normal tissues [46]. It has been found that the higher breast cancer incidence and higher all-cause mortality after breast cancer were significantly correlated with higher levels of insulin resistance in postmenopausal women [49,50]. Analyzing protein-protein interaction network in STRING database, SLC2A4 also participates in the translocation of SLC2A4 (GLUT4) to the plasma membrane and protein localization to plasma membrane through enriched proteins such as EXOC5 and C2CD5, which play an important role in human cancers. However, most of them were rarely studied in the pathways we found. Another important pathway related to SLC2A4 was AMP-activated proteinkinase (AMPK) pathway, which was a highly conserved and widely expressed energy balance regulator in eukaryotic cells and play a key role in carcinogenesis and cancer drug resistance [51][52][53]. Seventeen related genes we found were closely associated with AMPK signaling pathway, including AKT1, RAB8A, and IRS1. Studies demonstrated that Insulin receptor substrate 1 (IRS1) promoted tumor growth in colorectal cancer targeted by miR-30a-5p and was stabilized by RNA-binding protein lin-28 homolog B (LIN28B) [54]. It is important to correlate the data with the insulin levels of each cancer patient. If the cancer patient had Type 2 diabetes characterized by insulin resistance of course GLUT4 levels will be reduced. In that case, SLC2A4 expression level has no prognostic level for cancer outcome but diabetes and increased insulin secretion have. Reduced GLUT4 expression is only a byproduct of high chronic plasma insulin levels.
The most common genes enriched in 23 pathways were ARF6, AKT1, VAMP2, MYO1C, MYO5A, and RAB8A, which were found in 14, 13, 13, 12, 12, and 12 gene sets, respectively. Other genes such as INS, RNB10, PRKCZ, RAB13, and RAB10 can also participate in most pathways. Li et al. [55] indicated that ARF6 can regulate the functions of membrane traffic, and overexpression of ARF6 was correlated with poor prognosis in multiple invasive cancers, such as triple-negative breast cancer and invasive clear cell renal cell carcinoma. Riggio et al. [56] demonstrated that AKT1 promoted cell proliferation via upregulating cyclin-dependent kinase 1 and S6, and C-X-C motif chemokine receptor 2 (CXCR2) promoted breast cancer metastasis and chemoresistance by inhibiting AKT1 and activating COX2 (Cyclooxygenase 2). Moreover, Wang et al. [57] showed that the expression of vesicle-associated membrane protein 2 (VAMP2) was negatively regulated by miR-493-5p and further suppressed the proliferation and migration in liver cancer. Several studies had showed that Myosin 1 C (MYO1C) may play a key role in regulatingautophagosome-lysosome fusion through F-actin remodeling, and miR-137 overexpression inhibited the cell migration, proliferation by targeting Krűppel-likefactor 12 (KLF12) and MYO1C in gastric cancer cell lines [58,59]. Besides, myosin VA (MYO5A) may play a crucial role of diagnosis and prognosis in glioblastoma multiforme and gastric cancer [60,61]. Li et al. [62] found that Rab8a can regulate GLUT4 trafficking in muscle and adipose cells, and the suppression of Rab8A inhibited insulin-stimulated GLUT4 translocation. These results revealed that SCL2A4 had undeniable potential functions in the development and progression of human cancers.
There were some limitations in our research. First, all of our analysis were conducted based on publicly available datasets and further experimental studies consisting of larger sample sizes may contribute to validate our results. Meanwhile, performing abundant investigations will validate whether the bioinformatics results from public datasets is uniform with immunohistochemistry staining or western blot. Second, the study required more comprehensive and detailed analysis; researching all the gene sets enriched by Metascape will help to explore more potential functions and mechanisms of SLC2A4 in breast cancer. The potential diagnostic and therapeutic role of SLC2A4 can be assessed in great detail. Finally, we only investigated the most likely prognostic biomarker of SLC2s in breast cancer. Biomolecular network of other SLC2 family members will help us to deeply understand the role of SLC2s in breast cancer.

Conclusion
In conclusion, we systematically analyzed the potential function and molecular mechanism of SLC2s in human cancer. High expression of SLC2A4 was significantly correlated with better prognosis in breast cancer patients. Furthermore, the expression of SLC2A4 mRNA was found to be regulated by DNA hypermethylation. To our knowledge, this is the first study that revealed that SLC2A4 could be a prognostic biomarker for survivals of breast cancer patients.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This work was supported by the Science and Technology Department of Henan Province [192102310379] and Science and Technology Department of Henan Province [192102310350].