Selenoprotein S: a therapeutic target for diabetes and macroangiopathy?

When Walder et al. [14] first discovered SELENOS, they confirmed that it was a receptor for the acute inflammatory response protein, serum amyloid A (SAA); in addition, inhibition of SELENOS expression could up-regulate the expression of SAA in lipopolysaccharide (LPS)-induced HepG2 human liver cancer cells [43]. These results suggested a potential relationship between SELENOS and inflammatory reactions. Subsequently, Fradejas et al. [17] showed that SELENOS expression increased after induction of inflammatory injury in brain tissues of C57BL/6 mice, indicating that SELENOS was associated with inflammation. Next, in vitro induction of inflammatory injury in human and mouse astrocytes using LPS and interleukin-1β (IL-1β) also showed the up-regulation of SELENOS expression in corresponding cells. Furthermore, induction of SELENOS overexpression showed that it could reduce the expression of the inflammatory factors IL-1β and interleukin-6 (IL-6) in astrocytes that was stimulated by LPS. In contrast, inhibition of SELENOS expression further increased the expression of IL-1β and IL-6 stimulated by LPS. These results partially explained the molecular mechanism of the anti-inflammatory function of SELENOS. Liu et al. [44] established a cerebral ischemia/reperfusion injury model in SD rats to induce inflammation, and they found that neuronal cells in ischemic penumbra swelled, the cell density decreased, and SELENOS expression increased, which also confirmed the association between SELENOS and inflammation. In addition to inhibition of the expression of the inflammatory factors IL-1β and IL-6 by SELENOS, supplementation of selenium could also reduce the levels of tumor necrosis factor α (TNFα) and monocyte chemoattractant protein 1 (MCP 1) and reduce the secretion of interleukin-2 (IL-2) in primary porcine splenocytes [39, 45]. Selenium is the raw material for selenoprotein synthesis. Proper supplementation of selenium could increase SELENOS biosynthesis [45, 46], while inhibition of SELENOS expression could weaken the function of selenium supplementation on the reduction of IL-2 secretion in primary porcine splenocytes [39]. These results suggested that supplementation of selenium reduced the levels of TNFα, MCP-1, and IL-2 by increasing SELENOS biosynthesis, which further corroborated the molecular mechanism of the anti-inflammatory function of SELENOS.

Zhang et al. [35] found that the SELENOS gene in pigs had high homology with the human SELENOS gene (88%) and that the gene structures were very similar. In addition, the −601 to 398 region of the SELENOS gene promoter had an NF-κB binding site, a positive regulatory element for regulation of SELENOS gene expression. This binding site could increase SELENOS gene expression, and its sequence was confirmed to be the same as that in humans. In addition, a negative element that was suspected to regulate SELENOS gene expression in the −161 to 601 region of the SELENOS gene promoter was discovered. The above conclusion of Zhang et al. may be able to explain the study results of Su et al. [47], who injected LPS into the abdominal cavities of BALB/c mice to establish a septic mouse model. Their results showed that SELENOS expression in the livers of the mice increased, reached a peak value after 24 h of injury, gradually decreased afterward, and was reduced to normal levels after 72 h. Therefore, it was speculated that the increase of SELENOS expression at the early stage of injury may be associated with an up-regulation induced by NF-κB to protect tissues and cells from inflammatory injury. For the subsequent down-regulation of SELENOS expression, it was speculated that during the inflammatory injury process induced by LPS, some factors that interacted with the SELENOS negative regulatory region were produced to inhibit SELENOS expression, attenuate the protective function of SELENOS, and finally cause tissue and cell damage. Whether these regulatory transcription factors are specifically present in inflammation still awaits further in-depth studies.

Furthermore, studies in recent years on the association between SELENOS gene polymorphisms and inflammation have also become a hot research topic. Previously, Curran et al. [48] showed that the −105G→A single nucleotide variation (rs28665122) in the SELENOS gene promoter and the +3705G→A variation in the SELENOS gene were associated with serum TNFα, IL-6, and IL-1β and that the +5227C→T variation (rs4965373) in the SELENOS gene was associated with serum IL-1β, of which the association of the −105G→A variation was the most evident. Further studies showed that the −105G→A variation decreased the activity of the SELENOS promoter and reduced SELENOS expression in HepG2 liver cancer cells under the stimulation of ER stress, whereas inhibition of SELENOS expression increased TNFα and IL-6 production in RAW264.7 macrophages under ER stress. It is known that the −105G→A variation is located in the ER stress-response element (ERSE) of the SELENOS gene; therefore, it was speculated that this variation changed the ERSE sequence to affect SELENOS expression under the stimulation of ER stress and further influenced the levels of inflammatory factors [48]. However, the subsequent studies of Seiderer et al. [49] and Bos et al. [50] confirmed that the above three SELENOS single nucleotide polymorphisms (SNPs) were not associated with inflammatory factors. This controversial phenomenon may be explained by the study of Stoedter et al. [45], which confirmed that the selenium level in the body and gender could affect SELENOS expression and further affect cytokine levels. Therefore, these two confounding factors should be controlled in a study of SELENOS SNPs, and a stratified analysis should be performed based on the selenium levels in the body and gender. Bos et al. [50] also performed haplotype association analysis on three SNPs: SELENOS SNP rs28665122, rs4965814, and +6218 A→G variation (rs9874). The haplotype GAG was associated with serum IL-2, IL-6, interleukin 10 (IL-10), and granulocyte colony-stimulating factor (G-CSF). The above studies suggested that SELENOS SNPs play important roles in mediating inflammatory reactions. However, except for the study by Curran et al. [48], there is no report on the mechanism underlying the regulation of inflammatory reactions by SELENOS gene variations.

Christensen et al. [18] showed that SELENOS also had reductase activity. This reduction function was exerted through ¹⁸⁸Sec and was maintained through the restoration of the selenosulfide bond between ¹⁷⁴Cys and ¹⁸⁸Sec. In other words, in the reduction process, ¹⁸⁸Sec of SELENOS and substrates formed a mixed selenosulfide bond to cause substrate reduction. Next, through the restoration of the selenosulfide bond between ¹⁷⁴Cys and ¹⁸⁸Sec, SELENOS broke the mixed selenosulfide bond between SELENOS and substrates to maintain the activity of the SELENOS reductase. Liu et al. [51] showed that the reduction of the oxidized SELENOS to restore the selenosulfide bond between ¹⁷⁴Cys and ¹⁸⁸Sec depended on the function of thioredoxin (Trx), indicating that SELENOS was a Trx-dependent reductase. In addition, they also showed that SELENOS had peroxidase activity and could break down its substrate, hydrogen peroxide (H₂O₂), into H₂O [51]. The above studies indicated that SELENOS was closely associated with oxidative stress. This viewpoint was confirmed in the studies of Zahia et al. [52] and Zeng et al. [43]. They found that H₂O₂ could up-regulate SELENOS expression in HEK293 human embryonic kidney cells, while inhibition of SELENOS expression could further aggravate the LPS-induced increase of reactive oxygen species levels in HepG2 human liver cancer cells, down-regulation of GPx expression, and inhibition of cell viability. These results suggested that SELENOS may have the function of protecting tissues and cells from oxidative stress-induced damage. Previously, Gao et al. [19] showed that in vitro induction of SELENOS overexpression in Min6 islet β cells could increase its resistance to H₂O₂ damage and enhance cell viability. Our study group induced SELENOS overexpression in human umbilical vein endothelial cells (HUVECs) and showed that SELENOS overexpression could significantly increase HUVEC viability after H₂O₂ stimulation, enhance the activity of superoxide dismutase (SOD), and reduce the production of methane dicarboxylic aldehyde (MDA) [41]. Gan et al. [53] showed that SELENOS-overexpressing porcine kidney PK15 cells could exert an antioxidant protection function through the reduction of reactive oxygen species (ROS) and the increase of glutathione (GSH) expression. Next, the study by Ye et al. [42] indicated that inhibition of SELENOS expression in vascular smooth muscle cells (VSMCs) further aggravated the reduction of the cell viability induced by H₂O₂, which was accompanied by the further increase of ROS and MDA levels and further reduction of the GPx activity. They also confirmed that the antioxidant protection function of SELENOS was associated with mitogen-activated protein kinase (MAPK) and c-JUN N-terminal kinase (JNK).

As an ER membrane protein, SELENOS plays an important role in the maintenance of the morphology and distribution of ER in cells [54]. In addition, it forms a complex with degradation in endoplasmic reticulum protein 1 (Derlin1)-ubiquitin ligase E3-p97ATPase to degrade unfolded or misfolded proteins in the ER, a process known as ER-associated protein degradation (ERAD) [20, 55‐58]. Most recent studies have shown that selenoprotein K also participated in the composition of this complex [59]. Studies showed that the ER stress inducers tunicamycin (TM), thapsigargin (TG), and β-mercaptoethanol could up-regulate SELENOS expression in RAW264.7 macrophages, HepG2 liver cancer cells, HEK293T human embryonic kidney cells, and mouse astrocytes [21, 60‐62], suggesting that SELENOS is involved in the regulation of ER stress. Overexpression of SELENOS attenuated the reduction in RAW264.7 macrophage viability and the increase in cell apoptosis induced by TG and TC [21], inhibition of SELENOS expression in mouse astrocytes further aggravated the reduction in cell viability induced by TG and TC [60], and inhibition of SELENOS expression in HepG2 liver cancer cells further aggravated cell apoptosis and further reduced the cell viability induced by β-mercaptoethanol [61]. These results indicated that SELENOS could protect cells from damage induced by ER stress. The above conclusion was explained in the studies of Kelly et al. [63] and Kim et al. [64]. They found that SELENOS overexpression could reduce the activity of the glucose-regulated protein 78 (GRP78) promoter, an ER stress marker protein, in HepG2 liver cancer cells and decrease the GRP78 protein expression induced by TC in HEK293T human embryonic kidney cells. Therefore, it was speculated that the up-regulation of SELENOS expression in the process of ER stress was to defend the further increase of GRP78 expression to exert its protective function in ER stress.

In addition to investigating the relationship between SELENOS and inflammation, Fradejas et al. [17] also analyzed the relationship between SELENOS and ER stress. They found that after human and mouse astrocytes were treated with the ER stress inducers TM and TG, SELENOS expression levels in these two cells were up-regulated and were significantly higher than that in the group with inflammatory injury induced by LPS and IL-1β. These results indicated that SELENOS was associated with ER stress and that the function of ER stress in the up-regulation of SELENOS expression was stronger than that of inflammatory stimulation. Fradejas et al. [17] further induced high SELENOS expression in astrocytes and showed that SELENOS could reduce the expression of the ER stress markers X-box binding protein 1 (XBP-1) and CCAAT/enhancer-binding protein homologous protein (CHOP), stimulated by TM and TG, which further corroborated the molecular mechanism underlying the protection of cells from ER stress by SELENOS. This conclusion was contradictory to the study of Speckmann et al. [38]. They inhibited SELENOS expression in LS174T colon cancer cells and showed that in the presence of the ER stress inducer TM, the expression of GRP78, XBP-1, and CHOP between the low SELENOS expression group and the normal control group were not different and the expression levels of the above ER stress markers in the low SELENOS expression group did not further increase. Subsequently, they repeated the above experimental processes in another two colon cancer cell lines, HT29 cells and Caco-2 cells, and also obtained the same results [38]. The difference in the types of studied cells may be the major reason for the inconsistent conclusions between these two studies. It was speculated that the participation of SELENOS in the regulation of ER stress was cell-type dependent; however, further studies are required for confirmation and explanation.

Martínez et al. [65] compared the genotype frequencies of six types of SELENOS SNPs between type 1 DM (T1DM) patients (n = 310) and non-T1DM healthy controls (n = 550) in a Spanish population. The genotype frequencies of SELENOS SNPs rs11327127, rs28665122, rs4965814, rs12917258, rs4965373, and rs2101171 between these two groups were not different; the comparison after sex-stratification still did not show a difference. These results indicated that the above six types of SELENOS SNPs were not risk factors of T1DM. Olsson et al. [66] randomly selected 1236 subjects from the INTERGENE program, which analyzed the interplay between genetic susceptibility and environmental factors for the risk of chronic diseases in a Swedish population to analyze the relationship between three types of SELENOS SNPs and metabolic risk factors. The serum insulin levels and homeostasis model assessment of insulin resistance (HOMA-IR) between the group that carried SELENOS SNP rs4965373 and the group that did not carry this SNP were significantly different, while the above indicators between the groups that carried or did not carry rs28665122 and groups that carried or did not carry rs4965814 were not significantly different. The multiple linear regression analysis results after age, sex, and DM status were adjusted showed that SELENOS SNP rs4965373 correlated with the serum insulin level, SELENOS SNP rs4965814 correlated with the blood glucose level and HOMA-IR, and SELENOS SNP rs28665122 did not correlate with the above metabolic risk factors. Currently, there is still no study focusing on the association between SELENOS SNPs and DM. Case–controlled studies that involve more SELENOS SNPs, different races, and larger sample sizes still need to be performed (Table 1).

Hyrenbach et al. [71] used the Italian and German populations as study subjects and reported that the SELENOS SNP rs28665122 was not associated with ischemic stroke caused by spontaneous cervical artery dissection. In addition, Alanne et al. [72] selected 2222 study subjects from the FINRISK Study in Finland to perform a case-controlled study. They also confirmed that SELENOS polymorphism at that locus was not associated with CVD and ischemic stroke. However, they found that the risk ratio of the development of CVD in females carrying SELENOS SNP rs8025174 was 2.95; the risk ratio of the development of ischemic stroke in the population carrying SNP rs7178239 was 1.75, and the risk ratio in females reached 3.35. Li et al. [73] used the Chinese population as subjects to perform a case-controlled study (ischemic stroke group n = 239; non-ischemic stroke control group n = 240) and showed that SELENOS SNP rs4965814 could increase the risk of ischemic stroke by 1.54-fold; this risk ratio in females reached 2.43. These results were similar to those of Silander et al. [74]. They found that carrying SELENOS SNP rs4965814 increased the risk of ischemic stroke by 2.89-fold in Finnish women; in addition, they also confirmed that carrying SELENOS SNP rs9874 increased the risk of ischemic stroke in Finnish women by 3.32-fold [74]. Furthermore, Cox et al. [75] used European American T2DM patients (n = 1220) as study subjects in the Diabetes Heart Disease Study to analyze the relationship between 10 types of SELENOS gene polymorphism and the risk of atherosclerosis in T2DM patients. The SELENOS SNPs rs28665122, rs4965814, rs28628459, rs7178239, and rs12917258 were associated with subclinical atherosclerosis, and SELENOS SNPs rs4965814, rs28628459, and rs9806366 were associated with clinical atherosclerosis. The above studies suggest that SELENOS gene polymorphism is expected to become one of the indicators for the evaluation of the risk of macroangiopathy in non-DM and T2DM patients and one of the theoretical bases for the adoption of the early enhanced primary prevention measure for macroangiopathy complications in the population carrying relevant SNPs (Table 1).

Vascular endothelial cells play an important role in the maintenance of cardiovascular homeostasis, while endothelial dysfunction is the initiating step in the development of atherosclerosis [76‐78]. Our study group induced high/low SELENOS expression in HUVECs in vitro and showed that high SELENOS expression could increase the viability of HUVECs and SOD activity and reduce MDA production stimulated by H₂O₂; the changes after inhibition of SELENOS expression had the opposite results [41]. These results indicated that SELENOS was a protective factor in vascular endothelial cells and could increase the resistance of HUVECs to oxidative stress damage.

VSMCs are important components of the tunica media of vascular walls. Their abnormal proliferation is involved in the formation of early atherosclerosis plagues, and the apoptosis of VSMCs in advanced atherosclerosis plaques will induce the reduction of collagen production to cause fibrous cap thinning and reduce plague stability [79], which is a risk factor for cardiovascular events. Recent studies discovered that inhibition of SELENOS expression in VSMCs could aggravate cell damage caused by H₂O₂ or TM and increase VSMC apoptosis. Therefore, it was speculated that SELENOS could increase the resistance of VSMCs to oxidative stress and ER stress [42].

Our study group further showed that high SELENOS expression could inhibit the increase of caveolin-1 (Cav-1) induced by H₂O₂. When SELENOS expression was inhibited, Cav-1 and protein kinase Cα (PKCα) expression levels increased [41], suggesting that the protection of HUVECs from oxidative stress damage by SELENOS was associated with the regulation of Cav-1 and PKCα expression. Ye et al. [42] confirmed that inhibition of SELENOS expression in VSMCs further increased the phosphorylation levels of MAPK and JNK stimulated by H₂O₂, indicating that SELENOS increased the resistance of VSMCs to oxidative stress damage through the MAPK/JNK pathway. In addition, the antioxidant function of SELENOS in blood vessels may also be associated with the aforementioned ¹⁸⁸Sec in its molecular structure. Furthermore, the study of Liu et al. [51] also showed that SELENOS had peroxidase activity and could break down H₂O₂ into H₂O, which could partially explain the study results of our group [41] and those of Ye et al. [42]. These results elucidated the possible mechanism underlying the macrovascular protective function of SELENOS (Fig. 2).

Shibata et al. [80] confirmed that carrying SELENOS SNP rs28665122 was associated with a 1.99-fold increased risk of intestinal type gastric cancer in Japanese individuals. Mao et al. [81] also showed that the risk ratio of gastric cancer in the population carrying SELENOS SNP rs34713741 was 1.62. The studies of Méplan et al. [82] and Sutherland et al. [83] confirmed that this variation could increase the risk of colorectal cancer by 1.68- and 1.57-fold, respectively, and the risk ratios in women increased to 1.83 and 2.25, respectively. In addition, Hart et al. [84] showed that a combination of gene variations in SELENOS SNP rs28665122, caspase 8, matrix metalloproteinase 1 (MMP 1), and IL-10 increased the risk of the development of non-small cell lung cancer in the Norwegian population by 4.62-fold. The above studies suggest that the SELENOS gene polymorphism is expected to become one of the indicators for the risk of cancer. The association between SELENOS gene variation and the risk of cancer and the associations between SELENOS gene variation combined with other gene variations and the risk of cancer still require more studies for investigation and validation. In addition, there is currently no in vivo and in vitro functional study of SELENOS in the occurrence and development of cancer (Table 1).

It has been confirmed that various SELENOS SNPs are associated with the risk of the development of autoimmune inflammatory diseases. Marinou et al. [85] showed that the co-presence of SELENOS SNP rs28665122 and IL-1β SNP rs16944 increased the risk of development of rheumatoid arthritis in Caucasians by 2.3-fold. Santos et al. [86] showed that carrying SELENOS SNP rs28665122 could increase the risk of Hashimoto’s thyroiditis in Portuguese by 2.22-fold. This conclusion was subsequently confirmed by Li et al. [87] in a case-controlled study using the Chinese population as their study subjects. They discovered that the risk ratio of the development of Hashimoto’s thyroiditis in the population carrying SELENOS SNP rs28665122 was 1.28. In addition, SELENOS SNP rs28665122 has also been confirmed to be associated with preeclampsia, spontaneous preterm birth, and Kashin–Beck disease (KBD) [88‐90]. The study by Du et al. [90] showed that the expression of PI3K and Akt in the whole blood of carriers with that variation was higher than that of non-carriers and that the expression levels of PI3K and Akt were higher in the whole blood of KBD patients compared with the control group. Therefore, it was speculated that SELENOS SNP rs28665122 participated in the development of KBD through the regulation of PI3K/Akt pathway (Table 1).