Fetuin-B Overexpression Promotes Inflammation in Diabetic Retinopathy Through Activating Microglia and the NF-κB Signaling Pathway

Abstract Purpose To investigate the expression, source, role, and mechanism of Fetuin-B (FETUB) in diabetic retinopathy (DR). Methods ELISA and immunofluorescence were used to analyze the concentration of FETUB in plasma, aqueous fluid, and tissue specimens of patients with DR and healthy controls. Immunofluorescence, q-PCR, and western blotting were used to examine the expression of FETUB in DR mice and cells cultured with different concentrations of glucose. BV2 microglia cell line and DR mice were treated using FETUB recombination protein and FETUB shRNA to explore the function of FETUB in DR by q-PCR, western blotting, and immunofluorescence. Results FETUB concentrations in plasma, aqueous fluid, and tissue specimens were significantly increased in DR patients. The mice in DR group had a higher concentration of FETUB in the retina and liver tissues than those in the control group, and the expression of FETUB was increased in both ARPE19 and BV2 cells under a high-glucose environment. The ratio of p-P65 (Phospho-P65)/P65 and the expression levels of TNF-α, VEGF, and ionized calcium binding adaptor molecule (IBA)-1 were increased in BV2 cells cultured with FETUB recombinant protein, while they were decreased in BV2 cells transfected with FETUB shRNA. Immunofluorescence staining showed that there were more IBA-1+ activated microglia in the retinas of the FETUB recombination protein group than in the retinas of the DR group, and there were fewer IBA-1+ activated microglia in the retinas of the FETUB shRNA group than in the retinas of the DR group. Conclusions FETUB sourced from endocrine, autocrine, and paracrine pathways could promote inflammation in DR by activating the NF-κB pathway and microglia.


Introduction
Diabetic retinopathy (DR) is one of the most common complications of diabetes and the main cause of vision loss in working-age population.The main therapeutic methods of DR include intravitreal injection of anti-vascular endothelial growth factor (VEGF), pan retinal photocoagulation, and vitreoretinal surgery.Because of unclear pathogenesis, these treatments do not serve a radical cure for DR. 1,2Thus, further research on the pathogenesis and the new therapeutic targets of DR is warranted.
Fetuins belong to the cystatin superfamily, which includes Fetuin-A (FETUA) and Fetuin-B (FETUB) proteins. 3ETUB, a serum protease inhibitor, is mainly secreted by the liver as an important liver cytokine.The FETUB expression level is closely related to the metabolism of glucose and lipid.5][6][7] The concentration of FETUB in plasma was significantly higher in patients with T2D than that in plasma of pre-diabetes patients with impaired glucose tolerance. 6TUB was also linked to insulin resistance. 8Besides, previous studies reported that high level of FETUB predicted the frequency and probability of acute episodes of chronic obstructive pulmonary disease. 9The concentration of serum FETUB closely reflected the FETUB level in the follicular fluid to predict the outcome of fertilization to a certain extent. 10,11FETUB was also strongly associated with the acute myocardial infarction 12 and involved in modulating the rupture of atherosclerotic plaques. 13Therefore, it is logical and necessary to investigate the role of FETUB in DR.In this study, we aimed to study the location, expression, potential role, and pathological effect of FETUB in the development and progress of DR.

Subjects and exclusive criteria
Overall, 17 patients without diabetes mellitus (DM), 21 patients with DM (without DR), 21 patients with non-proliferative diabetic retinopathy (NPDR) and 18 proliferative diabetic retinopathy (PDR) subjects participated in the study.The patients with DM were diagnosed based on the fasting plasma glucose (FPG) (�7.0 mmol/L) and oral glucose tolerance tests (2-h postprandial blood glucose � 11.1 mmol/L).The patients with DR were diagnosed by clinical examination.Based on the fluorescein fundus angiography, the patients were divided into NPDR and PDR groups according to whether the retinal neovascularization or fibrovascular membranes (FVMs) existed.Patients with malignant tumor, hepatic or renal diseases, blood diseases, acute and chronic inflammatory diseases, cognitive disorders, or women who were menstruating, pregnant, or lactating currently were excluded.All patients signed the informed consent.All experimental protocols were approved by the Ethics Committee of the Second Affiliated Hospital of Xi'an Jiaotong University (Approval number: 2021-771).

Assessment of plasma and aqueous fluid FETUB levels
Enzyme-linked immunosorbent assay (ELISA) kits (Cloud-Clone Crop, Wuhan, China) were used to determine the concentration of FETUB and VEGF levels.Use a vacuum tube to collect venous blood and set the samples for 2 h before centrifugation for 15 min at approximately 1,000�g at 4 � C. The samples were kept in freezing tubes at -80 � C. Aqueous fluid samples were collected using 25G needles and stored at -80 � C freezing tubes.Each sample underwent three duplicate readings.The detection range was 125-8000 pg/mL.There was no detectable cross-reactivity to other molecules or interference by other proteins.

Human tissue specimens
Idiopathic macular epiretinal membranes (IMEM) of three control patients without DM and FVMs of three PDR patients were obtained from patients who needed vitreoretinal surgery and fixed in 4% paraformaldehyde (Biosharp, Hefei, China).
BV2 mouse microglia cell line (Procell Life Science and Technology) were grow in DMEM medium (Procell Life Science and Technology Co., Ltd.) containing 10% FBS and 1% penicillin-streptomycin solution at 37 � C in 5% CO 2 .The cell lines were identified by short tandem repeats (STR) and free of mycoplasma contamination.ARPE19 and BV2 cells were grown to 70% before the experiments.
Cells were stimulated with D-glucose (LEAGENE, Beijing, China) at non-(D-) glucose, low (5.5 mmol/L), medium (15 mmol/L) and high (30 mmol/L) concentrations for 48 h.In order to eliminate the effect of osmotic pressure, mannitol was added to non-glucose medium as the osmotic control groups of low-, medium-, and high-glucose groups.In addition, BV2 cells were stimulated with FETUB recombinant protein (Cloud-Clone Crop) at non-FETUB, 100 ng/ mL, 200 ng/mL and 300 ng/mL concentrations for 48 h.Lentivirus containing a shRNA (Genecarer, Xi'an, China) against FETUB were added to culture medium to transfect and block FETUB expression in the BV2 cells.

Animal studies
All procedures with animals were approved by the Xi'an Jiaotong University of Animal Experimentation Ethics Committee (Approval number: XJTU2022-1641).Male C57BL/6 mice (Chengdu Gembio, Chengdu, China) were used in this study.Mice were bred at 26 � C (12/12 h, light/ dark cycle) with normal chow diet and drinking water, with regular changes of beddings.Subsequently, 50 five-week-old mice were divided into DR and control groups randomly.Blood glucose levels were measured before medications were administered.The DR mice were made by 100 mg/kg streptozotocin (STZ) (Solarbio, Beijing, China) (10 mg/ml in 0.1 M citric acid-sodium citrate buffer) treatment after 12 h fasting via intraperitoneal injection.The control group received the same volume of citric acid-sodium citrate buffer.The tail vein blood glucose was measured to confirm hyperglycemia (>16.7 mmol/L) in the DR group.The mice were executed by cervical dislocation at 8, 16, 24 weeks after establishment of diabetes and the retina and hepatic tissues were collected for further experiments.
Five mice in control group and 20 mice in DR group were randomly selected at 6 weeks after establishing diabetes.20 DR mice were divided into 4 groups as DR group, FETUB group, FETUB shRNA group, and NC shRNA group randomly.After 12 h fasting, the mice were intraperitoneally injected with FETUB recombination protein (1 lg/g body mass) (twice a week) in FETUB group, 50 ll adeno-associated virus containing a FETUB shRNA (10^12PUF/ml) in FETUB shRNA group, 50 ll adeno-associated virus containing a NC shRNA (10^12 PUF/ml) in NC shRNA group, or 50 ll phosphate buffer solution (pH 7.4) in DR group.The mice were executed by cervical dislocation at 24 weeks after diabetes.
Overall, 20 lg protein was resolved by 12.5% SDS gel electrophoresis, transferred to a poly-vinylidene fluoride membranes (Epizyme Biotech) and incubated overnight at 4 � C with one of the primary antibodies (Supplementary Table 1).Next, incubate the membranes with corresponding secondary antibodies (Supplementary Table 1) for 2 h at room temperature.The antibodies were diluted in a universal antibody dilution buffer (Epizyme Biotech).All antibodies were validated by the manufacturer.
Electrochemiluminescence (BioRad, Hercules, CA, US) was used to capture the blots.Enhanced chemiluminescence (Epizyme Biotech) reagents were used to visualize protein blots.ImageJ software (National Institutes of Health, MD, USA) was used to measure the blot intensity.

Immunofluorescence
The BV2 cell slices, human epiretinal membranes, and mice eyeballs were fixed overnight in 4% paraformaldehyde.Following various established protocols, 4-lm tissue sections were incubated with 1% bovine serum albumin for 30 min.Next, the sections were incubated overnight at 4 � C with the primary antibodies (Supplementary Table 1).After washing with PBS (pH 7.4) for 15 min, sections were incubated with corresponding secondary antibodies (Supplementary Table 1) and then incubated with DAPI (Beyotime Biotechnology).Panoramic MIDI (3DHISTECH, Budapest, Hungary) was used for image capture.

Statistical analysis
SPSS 24.0 software (IBM, NY, USA) and GraphPad Prism 8.0 software (GraphPad, CA, USA) were used to perform all analyses.At least 3 data were used for statistical analysis in each group.Data were expressed as mean ± standard deviation (SD).Student's t test (unpaired) or one-way ANOVA were used to determine statistical significance.The p value <0.05 was defined as statistical significance.

The concentration of FETUB in plasma and aqueous fluid increased in DR patients
Table 1 shows the clinical characteristics of all patients (n ¼ 77).The duration of diabetes was significantly shorter in patients with DR than in those with NDR (p < 0.05).Systolic blood pressure was significantly higher in DR patients than in NDR patients (p < 0.05).FPG was significantly higher in DR and NDR patients than in controls (p < 0.001).
The concentrations of plasma VEGF and FETUB were significantly higher in DR patients than in control and NDR groups (both p < 0.05) (Figure 1A).Compared with NDR groups, plasma FETUB concentrations was significantly higher in PDR and NPDR patients (p < 0.05) (Figure 1B).Compared with the control and NDR groups, aqueous fluid FETUB concentrations were significantly higher in DR group (both p < 0.001) (Figure 1C).What's more, aqueous fluid FETUB concentrations was significantly higher in PDR and NPDR patients compared with NDR groups (p < 0.001) (Figure 1D).However, there was no significant correlation between FETUB and VEGF concentrations.The patients in PDR group were younger than NPDR group (p < 0.01).Otherwise, there were no significant differences between the two groups in other characteristics (Table 2).

The expression of FETUB increased in FVMs-PDR
The immunofluorescence staining showed that the FETUB was strongly positive in PDR-FVMs while nearly undetectable in IMEM (Figure 1E) (Supplementary Figure 1), confirming that the expression of FETUB increased in the FVMs of PDR patients.Retinaldehyde binding protein 1 (RLBP1) were used to provide colocalization, which is a maker of RPE cells.The characteristics of two representative patients (Figure 1F,G) were as following.One, IMEM patient (without DM) was collected IMEM from left eye: 70 years old; female; FPG: 5.8 mmol/L; diagnosis: IMEM (left eye).The vision of left eye was 0.4, while the right eye vision was 1.0.There was not any pathological changes detected in the right retina.Two, PDR patient was collected FVMs from right eye: age: 42; male; FPG: 6.8 mmol/L; diagnosis: PDR (both eyes), traction retinal detachment (right eye), vitreous hemorrhage (right eye), diabetic macular edema (left eye), and type 2 diabetes (10 years).There were several obvious large areas of retinal neovascularization in left retina.The vision of left eye was hand motion/10 cm, and the vision of right eye was 0.3.

Expression level of FETUB increased in retina and liver tissue of DR mice
The expression level of FETUB increased in tissue of DR mice.After STZ injection, the FPG of DR group significantly increased and met the diagnostic criteria (>16.7 mmol/L) for DM (Supplementary Figure 2A).The weight of DR group mice was less than that in the control group (Supplementary Figure 2B).The hematoxylin-eosin staining of retina showed that DR appeared in the retinal tissue from DR group (Supplementary Figure 2C).The mice were executed at weeks 8, 16, and 24 after diabetes was established.Detached retina and hepatic tissue were collected for western blotting and q-PCR analysis.
In retina, q-PCR analysis showed that the expression level of FETUB increased at 16 (p < 0.001) and 24 weeks (p < 0.001) after STZ injection (Figure 2A).Western blotting of retina demonstrated similar protein level changes in 16 (p < 0.001) and 24 weeks (p < 0.001) after STZ injection (Figure 2B,C).Western blotting of liver tissue showed that the protein level of FETUB increased at 24 weeks (p < 0.001) (Figure 2E,F) after STZ injection, but q-PCR indicated that the FETUB level increased by 8 weeks (p < 0.05), 16 (p < 0.01) and 24 weeks (p < 0.001) after STZ injection (Figure 2D).The eyeballs were detached at 24 weeks after STZ injection to perform immunofluorescence staining of the retina.The images showed that compared with control mice, the FETUB concentration in the retina of DR mice was higher (Figures 2G).Besides, ionized calcium binding adaptor molecule-1 (IBA-1) were used to provide colocalization, which is a maker of activated microglia.

Expression level of FETUB increased in ARPE19 and BV2 cells after glucose stimulation
Since FETUB has been co-stained with RLBP1 and IBA-1 in previous immunofluorescence staining, adult retinal pigment epithelial cell line19 (ARPE19) and BV2 mouse microglia cells were used in in vitro experiments.
The FETUB expression in ARPE19 cells was examined with or without 48 h glucose stimulation.In order to eliminate the effect of osmotic pressure, mannitol was added to non-glucose medium as the osmotic control groups of low-, medium-, and high-glucose groups.q-PCR indicated that the expression level of FETUB was higher in high-glucose group than in the medium-(p < 0.01), low-(p < 0.01), no-(p < 0.001), and osmotic control of high-glucose (p < 0.001) groups (Figure 3A).Western blotting confirmed similar FETUB protein level changes among these groups (Figure 3B,C).
The FETUB expression in BV2 mouse microglia cells after 48 h glucose stimulation was examined by q-PCR and western blotting.q-PCR indicated that compared with noglucose and low-glucose groups, FETUB expression was higher in the medium-(p < 0.01) and high-glucose (p < 0.01) conditions.Compared with the medium-glucose condition, the expression level of FETUB was higher in the high-glucose (p < 0.05) group.Compared with osmotic control of medium-and high-glucose groups, the expression level of FETUB was higher in the medium-(p < 0.01) and high-glucose (p < 0.01) groups respectively (Figure 3D).Western blotting confirmed that FETUB protein level was higher in the medium-(p < 0.01) and high-glucose (p < 0.01) conditions than that in the no-glucose group.Compared with the low-glucose condition, FETUB protein level was higher in the medium-(p < 0.01) and high-glucose (p < 0.001) groups (Figure 3E,F).Compared with osmotic control of medium-and high-glucose groups, the expression level of FETUB was higher in the medium-(p < 0.01) and high-glucose (p < 0.001) groups respectively.The immunofluorescence staining demonstrated that FETUB expression was significantly increased under the high-glucose condition (Figure 3G).
In summary, hyperglycemia increases the expression of FETUB in ARPE19 and BV2 cells in vitro.

FETUB activated NF-jB signaling pathway in BV2 cells and increased the expression of inflammatory factors
To explore the function of FETUB in DR, BV2 cells were treated by the high-glucose medium with FETUB recombinant protein for 48 h.The concentrations of FETUB recombinant protein were 0, 100, 200 and 300 ng/ml.The TNF-a  expression by qPCR was higher dose-dependently in all four groups.Western blotting confirmed that TNF-a protein expression was also higher in the 300 ng/ml (p < 0.001), 200 ng/ml (p < 0.01), and 100 ng/ml (p < 0.01) groups than that in the 0 ng/ml group.It was also dose-dependent among different doses (Figure 4A-C).
In addition, VEGF expression by qPCR was increased dosedependently among the four groups.Western blotting confirmed that VEGF protein expression was also increased in groups which cultured with FETUB recombinant protein (Figure 4A-C).
IBA-1 is a marker of the activation of BV2 cells.The q-PCR results show that IBA-1 expression in BV2 cells was significantly higher in the 100 ng/ml (p < 0.001), 200 ng/ml (p < 0.01), and 300 ng/ml (p < 0.01) groups than in the 0 ng/ ml group.Western blotting indicated that IBA-1 expression was higher in the 100 ng/ml (p < 0.05), 200 ng/ml (p < 0.01), and 300 ng/ml (p < 0.01) groups than in the 0 ng/ml group.These results demonstrated that BV2 cells were activated by all tested concentrations of FETUB (Figure 4A-C).

The activation of BV2 cells was inhibited and the expression of inflammatory factors was decreased by FETUB shRNA transfection
BV2 cells were cultured with medium containing 30 mmol/L glucose and transfected with shRNA (Supplementary Figure 3).All groups were cultured under high glucose condition for 48 h to perform q-PCR and western blotting studies.
The transcriptional levels of TNF-a, VEGF, and IBA-1 were higher in both high glucose and NC shRNA groups than those in the control group (p < 0.01) assessed by qPCR.There was no significant difference between the high-glucose and NC shRNA group (p > 0.05), demonstrating that the transfection of negative control shRNA had no effect on the transcriptional levels of TNF-a, VEGF, and IBA-1.In the FETUB shRNA group, however, the transcriptional levels of TNF-a, VEGF, and IBA-1 were all significantly lower than those in the highglucose and NC shRNA groups (p < 0.01) (Figure 5A).
Western blotting confirmed the result with protein levels.The expression levels of TNF-a, VEGF, and IBA-1 were higher in both high-glucose and NC shRNA groups than those in the control group (p < 0.05).There was no significant difference between the high-glucose group and NC shRNA group (p > 0.05).The expression levels of TNF-a, VEGF and IBA-1 in FETUB shRNA group were decreased significantly (p < 0.05) (Figure 5B,C).
The ratio of p-P65/P65 NF-jB was higher in both high-glucose and NC shRNA groups than that in the control group (p < 0.01), and it was lower in FETUB shRNA group compared with the high glucose and NC shRNA groups (p < 0.01) (Figure 5B,D).

FETUB promoted activation of microglia in the retina of DR mice
The mice eyeballs were detached at week 24 after confirming diabetes to perform immunofluorescence staining.The image showed that the IBA-1 þ activated microglia in the retina of DR mouse were more pronounced than those in the control mouse.The IBA-1 þ activated microglia in FETUB recombinant protein group were more than that in DR group.The IBA-1 þ activated microglia in DR group were more than those in FETUB shRNA group (Figure 5E).These results demonstrated that FETUB promoted the activation of microglia in the retina of DR mouse.

Discussion
Fetuins are serum proteins with multiple functions. 14They belong to the cystain superfamily, which includes Fetuin-A (FETUA) and FETUB proteins. 3Harold 15 confirmed that fetuins were expressed in the eye including retina of postnatal and adult mice.This report was the first time demonstrating the derived location of Fetuins at the retina.The immunofluorescence of fetuins in the ganglion cells and optic nerves of adult rat retina was reported, suggesting that retina-derived fetuins performed important functions in adult eyes. 16Subsequently, Danda 17 confirmed FETUA expression was increased in retinoblastoma tissues.
Recently, FETUB has been found to have a potentially important regulatory role in T2D. 8 Qu et al. found that patients with T2D had significantly higher concentrations of plasma FETUB than that with normal glucose tolerance (NGT). 6ralisch et al. 18 found higher circulating FETUB levels in pregnant women with gestational diabetes than that in NGT group.In diabetic kidney disease, circulating FETUB was negatively and independently associated with renal function. 19Serum FETUB was independently associated with nonalcoholic fatty liver disease in patients with T2D. 20These results demonstrated that FETUB played an important role in various diseases, including T2D.In addition, the relationship between the concentrations of FETUA in serum and DR has been demonstrated.The higher FETUA value was associated to the increased risk of DR.Serum FETUA concentrations are independently correlated with VEGF and CRP concentrations in T2D patients with DR.][23] Based on the previous researches, we tried to research the link between Fetuin-B and DR.In this study, we found that the concentration of FETUB in plasma and aqueous fluid were higher in patients with DR than that in controls and that in DM patients without DR.FETUB, therefore, may play an important role in DR.Moreover, there was no significant difference of FETUB levels between NPDR and PDR patients, suggesting that FETUB plays a role in both the early and terminal stages of DR.
In in vivo experiments, C57BL/6 mice were used for a DR model.The expression of FETUB was increased in both retina and liver tissues of DR mice.In addition to hepatocyte secreting FETUB into the blood, retina also secreted FETUB in this experiment.ARPE19 and BV2 cells were chosen to explore the change of FETUB levels under a hyperglycemia or high glucose environment.As well as transporting nutrients from the vascular choroid, RPE cells form the blood-retinal barrier and absorb scattered light.RPE plays a significant role in retina physiology. 24In addition, microglia cells are the resident monocytes of the retina.Internal environment imbalance, triggered by ischemia, hyperglycemia, and hypoxia in DM duration, activates microglia. 25,26The activated microglia cells release various proinflammatory mediators. 27The FETUB expression was increased in ARPE19 and BV2 cells under hyperglycemia or high glucose culture condition in this experiment.In retina, both ARPE19 and BV2 cells secrete FETUB.It is therefore speculated that FETUB may trigger inflammatory neuro-vasculopathy of DR through endocrine, autocrine, and paracrine effects.
Currently, DR is considered as an inflammatory response, manifesting as leukocyte adhesion, microglia activation, and increased secretion of inflammatory factors. 28,29Microglia cells play a central role in the immunomodulatory process in retina.Activated microglia migrate to specific areas and perform unique functions.1][32] IBA-1 is a maker of the activated microglia.In this study, the up-regulation of IBA-1 demonstrated the activity of microglia cultured in the medium with addition of FETUB recombinant protein.It indicated that FETUB promoted inflammation in DR through activating microglia.
VEGF increases intercellular cell adhesion molecule (ICAM)-1 expression through NF-jB and stimulates vascular growth directly during DR lesions.VEGF plays an important role in neovascularization. 33In mouse retinal M€ uller cells, repression of VEGF reduced the expression of NF-jB, TNF-a, and ICAM-1 in the retina of DR mice. 34In this study, FETUB increased TNF-a and VEGF expression, affecting the pathological progress of DR.
Activation of NF-jB signaling pathway is involved in the development of DR. 28,[35][36][37] The expression of NF-jB was elevated in both NPDR and PDR. 38Several factors, such as hyperglycemia, oxygen radicals, cytokines and protein kinase C induced the activation of NF-jB. 39The activated NF-jB stimulates the expression of various adhesion factors and cytokines, such as ICAM-1, VEGF, bFGF and TNF-a, thereby enhances the adhesion between leukocytes and retinal vascular endothelial cells, and promotes the proliferation of vascular endothelial cells and neovascularization.In this report, FETUB increased the ratio of p-P65/P65 NF-jB, indicating the activation of NF-jB pathway.
In conclusion, the present study demonstrated that the expression of FETUB increased over the course of DR.FETUB contributed to the inflammatory neuro-vasculopathy of DR through endocrine, autocrine, and paracrine effects at retina.FETUB activated the NF-jB pathway and microglia cells in retina.Therefore, FETUB may serve a therapeutic target and a diagnostic biomarker of DR.However, patient number in this study is limited, and the molecular mechanism needs further investigations.

Figure 2 .
Figure 2. The expression level of FETUB increased in retina and hepatic tissues of DR mice.(A) the FETUB expression levels were increased in retina of DR mice measured by qPCR.(B,C) Western blot analysis showed that the FETUB protein levels were increased in retina of DR mice.(D) q-PCR analysis showed that the FETUB expression levels were increased in hepatic tissue of DR mice.(E,F) Western blot analysis showed that the FETUB expression levels were increased in hepatic tissue of DR mice.b-actin was used as an internal reference control and loading control.(G) Representative immunofluorescence images of ionized calcium binding adaptor molecule-1 (IBA-1) (green) and FETUB (red) of mice retinas (24 weeks after diabetes).IBA-1 is a maker of activated microglia cells.Scale bars: 50 lm.GCL: ganglion cell layer; INL: inner nuclear layer; ONL: outer nuclear layer; RPE: retinal pigment epithelium; C: control group; DR: diabetic retinopathy group.Data are mean ± SD of four independent experiments.� p < 0.05 versus control group; �� p < 0.01 versus control group; ��� p < 0.001 versus control group.

Figure 3 .
Figure 3. Expression of FETUB was increased in ARPE19 and BV2 cells under the stimulation of high glucose.(A) q-PCR analysis showed that the FETUB expression levels were increased in ARPE19 under the stimulation of high glucose.b-actin was used as the internal reference control.(B,C) Western blot analysis showed that the FETUB expression levels were increased in ARPE19 under the stimulation of glucose.b-actin was used as the loading control.� p < 0.05.�� p < 0.01.��� p < 0.001.(D) q-PCR analysis showed that the FETUB expression levels were increased in BV2 cells under the stimulation of glucose.b-actin was used as the internal reference control.(E,F) Western blot analysis showed that the FETUB expression levels were increased in BV2 cells under the stimulation of glucose.b-actin was used as the loading control.�� p < 0.01 versus NG group; ��� p < 0.001 versus NG group; † † p < 0.01 versus LG group; † † † p < 0.001 versus LG group.' p < 0.05 versus MG group.‡ ‡ p < 0.01 versus MG-C group or HG-C group.‡ ‡ ‡ p < 0.001 versus MG-C group or HG-C group.(G) Representative immunofluorescence images of FETUB (red) in BV2 cells after stimulation of high glucose.Scale bars: 50 lm.Data are mean ± SD of four independent experiments.NG: no glucose group; LG: low glucose group; MG: medium glucose group; HG: high glucose group; LG-C: mannitol osmotic control group of low glucose group; MG-C: mannitol osmotic control group of medium glucose group; HG-C: mannitol osmotic control group of medium glucose group.

Figure 5 .
Figure 5.Effect of FETUB shRNA transfection on BV2 cells (A-D) and DR mice (E).(A) q-PCR analysis showed that the TNF-a, VEGF, and IBA-1 expression levels were decreased in FETUB shRNA transfection group.b-actin was used as the internal reference control.(B,C) Western blot analysis showed that the TNF-a, VEGF and IBA-1 expression levels were decreased in FETUB shRNA transfection group.b-actin was used as the internal reference control.(B,D) FETUB shRNA transfection inhibited the P65 phosphorylation.Data are mean ± SD of four independent experiments.� p < 0.05 versus FETUB shRNA group.�� p < 0.01 versus FETUB shRNA group.† p < 0.05 versus control group.† † p < 0.01 versus control group.† † † p < 0.001 versus control group.C: control group; HG: high glucose group; FETUB shRNA: FETUB shRNA transfection group; NC shRNA: negative control shRNA transfection group.(E) Representative immunofluorescence images of IBA-1 (green) of mice retina.Scale bars: 50 lm.GCL, ganglion cell layer; INL: inner nuclear layer; ONL: outer nuclear layer; RPE: retinal pigment epithelium; DR: diabetic retinopathy group; FETUB: DR mice treated with FETUB recombinant protein group; NC shRNA: DR mice treated with negative control shRNA transfection group; FETUB shRNA: DR mice treated with FETUB shRNA transfection group.

Table 1 .
Clinical characteristics of the study participants.

Table 2 .
Clinical characteristics of the patients in NPDR and PDR groups.