Aberrant expression of TRIM44, transcriptionally regulated by KLF9, contributes to the process of diabetic retinopathy

Abstract

Background

Diabetic retinopathy (DR) is the common cause of diabetic vascular complications and it causes blindness. Until now, there are still some patients with DR who lack effective treatment. Tripartite motif containing 44 (TRIM44) has been shown to play a significant role in endothelial cells. However, the role of TRIM44 in DR remains unknown.

Methods

Diabetes was induced in rats through the administration of an intraperitoneal injection of 65 mg/kg of streptozotocin (STZ). Rat retinal microvascular endothelial cells (RMECs) were subjected to stimulation under high glucose (HG) conditions. A thorough proteomic investigation and bioinformatic analysis were performed to identify the differentially expressed proteins (DEPs) in rat RMECs after blocking TRIM44. A dual luciferase reporter assay was employed to assess the luciferase activity of TRIM44.

Results

TRIM44 was highly expressed in the retinal tissues of rats with diabetes and HG-induced RMECs. In vivo assays suggested that TRIM44 silencing improved the pathological alterations of DR rats as demonstrated by the downregulated expression of isolectin-B4 and VEGFA, along with a decrease in acellular capillaries within the retinal tissues. Knockdown of TRIM44 markedly reduced cell viability, proliferation, migration, invasion, and angiogenesis in HG-evoked RMECs. Mechanistically, TRIM44 was demonstrated to be activated transcriptionally by KLF transcription factor 9 (KLF9), a known facilitator of angiogenesis in DR. In HG-induced cells, the loss of TRIM44 resulted in the reverse of the endothelial cell function caused by KLF9 overexpression. After the comprehensive analysis, 64 upregulated and 38 downregulated DEPs were screened out for a series of functional enrichment analyses.

Conclusions

Collectively, this study demonstrates that TRIM44 knockdown suppressed diabetes-induced retinal vascular dysfunction in DR.

Introduction

Diabetic retinopathy (DR) is the most common complication of diabetes mellitus and is the leading cause of blindness and visual impairment, especially in the working-age adult population [1, 2]. The main pathologic features of DR are loss of pericytes, disruption of the blood-retina barrier (BRB), formation of abnormal neovascularization, etc [3]. Anti-vascular endothelial growth factor A (VEGFA) therapy is currently the mainstay of treatment for patients with advanced DR, but this intervention necessitates frequent intravitreal injections and may not be appropriate for all patients [4]. Furthermore, a significant number of patients with DR exhibit insufficient responses to anti-VEGFA agents, even following prolonged treatment [5]. Therefore, developing novel and mechanism-based therapeutic approaches is extremely important for the effective clinical management of patients suffering from DR.

Tripartite motif containing 44 (TRIM44) is a member of the TRIM family, and its dysregulation is related to a variety of biological processes, including mitosis, apoptosis and proliferation, cell cycle progression, migration, and invasion [6]. Research has shown that TRIM44 expression is evoked in rats with pathological cardiac hypertrophy and that the knockout of TRIM44 improves cardiac remodeling and dysfunction and reduces myocardial fibrosis [7]. TRIM44 is highly expressed in rats with traumatic brain injury, and downregulation of TRIM44 suppressed the inflammatory response and ameliorated brain injury [8]. TRIM44 has been shown to enhance the proliferation, migration, and invasion of multiple cancer cell types, including renal cell carcinoma [9], breast cancer [10], and testicular germ cell tumors [11]. However, the function of TRIM44 in DR remains unclear.

Angiogenesis is one of the main pathological features of DR, and neovascularization is a dynamic and balanced process that is regulated by angiogenic factors (e.g., VEGFA) and antiangiogenic factors (e.g., pigment epithelium-derived factor, PEDF) [12,13,14]. In response to angiogenic factors, vascular endothelial cells proliferate, migrate, and form lumens, which subsequently form new blood vessels. TRIM44 has been documented to influence the proliferation, migration, and angiogenic processes of human umbilical vein endothelial cells [15]. Hypoxia-inducible factor-1α (HIF-1α), an upstream regulator of VEGFA, is a promising target for DR therapy [16, 17]. Unlike most TRIM family members, TRIM44 contains a zinc-finger ubiquitin protease structural domain and acts as a deubiquitinating enzyme. A prior study has indicated that TRIM44 diminishes the ubiquitination of HIF-1α in myeloma cells, thereby enhancing the stability of HIF-1α [18]. Consequently, we speculated that TRIM44 might be involved in the DR process by influencing angiogenesis in vascular endothelial cells.

Our prior research has demonstrated that KLF transcription factor 9 (KLF9) exhibits elevated expression levels in DR rats as well as human umbilical vein endothelial cells treated with HG. Moreover, it reveals that suppressing KLF9 impeded cell proliferation, migration, invasion, and the capacity for angiogenesis [19]. KLF9 as a transcription factor has been described to transcriptionally promote the expression of downstream genes, such as TRIM33 [20]. JASPAR website analysis predicted the multiple binding sites between KLF9 and the promoter sequence of TRIM44. However, the underlying mechanisms connecting KLF9 and TRIM44 remained unclear.

The present study was designed to determine (1) whether TRIM44 was functional in DR; (2) whether TRIM44 was transcriptionally regulated by KLF9, and (3) the potential molecular targets downstream of TRIM44.

Materials and methods

Animals and diabetic models

All animal experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health, and the protocols were approved by the Animal Care and Use Committee of the Second Hospital of Jilin University. Healthy male Sprague-Dawley rats (6–8 weeks old) were housed in 12 h white and 12 h black with a temperature of 22 ± 1℃ and humidity of 45–55%. They were fasted overnight after one week of adaptive feeding. Subsequently, the diabetic rat models were established by a single intraperitoneal injection of 65 mg/kg streptozotocin (STZ, Macklin, Shanghai, China), which was dissolved in 10 mM citrate buffer (pH 4.5). The control group of rats received an equal volume of citrate buffer only. Measurements of fasting blood glucose levels and body weight were conducted at intervals of 1, 4, 8, and 12 weeks following the administration of STZ. After 6 days of STZ injection, rats exhibiting blood glucose concentrations greater than 16.7 mmol/L were regarded as diabetic and utilized further experimental procedures.

Lentiviral production

The short hairpin RNA (shRNA) target rat TRIM44 (LV-shTRIM44) was generated by Generalbio (Chuzhou, China) and ligated into the lentiviral vector pLVX-shRNA1 (Fenghui Biotechnology, Changsha, China). The pLVX-shRNA1 vector carrying the non-target shRNA (LV-shNC) was used as the control vector. The sequences of shRNAs were as follows: shTRIM44: 5’-CCGGGAAGAAGAGAGCGAGACAGATTCAAGAGATCTGTCTCGCTCTCTTCTTCCTTTTT-3’; shNC: 5’-CCGTTCTCCGAACGTGTCACGTTTCAAGAGAACGTGACACGTTCGGAGAATTTTT-3’. The constructed vector was co-transfected into the HEK-293T cells (iCell Bioscience, Shanghai, China) with an envelope plasmid, and a packaging plasmid using Lipofectamine 3000 reagent (Invitrogen, Carlsbad, CA, USA). Following a 48-h transfection period, the viral supernatants were harvested and subsequently subjected to filtration through a 0.45-µm cellulose acetate filter to achieve virus packaging.

To upregulate KLF9 expression in rat retinal microvascular endothelial cells (RMECs), lentiviral plasmid pLVX-IRES-puro (Fenghui Biotechnology) contained the coding sequence (CDS) of rat KLF9 (NM_057211) was generated (LV-KLF9), and the empty vector was used as the control (LV-NC). The corresponding lentivirus was then packaged using the same methods described above.

Intravitreal lentivirus injection

The diabetic rats were randomly divided into three groups: DR, DR + LV-shNC, and DR + LV-TRIM44. Four weeks post-STZ injection, the rats in the DR + LV-TRIM44 group received an injection of lentivirus silencing TRIM44 into the vitreous body of both eyes, with a volume of 1 µL per eye at a titer of 1 × 10⁹ TU/mL. As a control, the rats in the DR + LV-shNC group were injected with the same dose of control lentivirus. Following a 12-week period of STZ administration, all rats were killed, and the retinal tissues were subsequently fixed and embedded in paraffin for further analysis.

Hematoxylin-eosin (H&E) staining

H&E staining was carried out to observe the histological changes as described previously [21]. In brief, the embedded retinal tissues were cut into 5 μm slices using an RM2235 microtome (Leica Microsystems, Wetzlar, Germany). The sections were then dewaxed with xylene (Aladdin, Shanghai, China) and rehydrated in graded alcohol solutions (Sinopharm, Shanghai, China). Following a 5-min staining with hematoxylin (Solarbio, Beijing, China), the sections were subsequently stained with eosin (Sangon, Shanghai, China) for 3 min, followed by dehydration. Stained tissue sections were observed using a BX53 microscope (Olympus, Tokyo, Japan).

Retinal trypsin digestion

Periodic acid Schiff (PAS) staining was performed to assess acellular capillaries by the retinal trypsin digestion procedure [22]. The retina was washed in distilled water several times and then digested with 3% trypsin at 37 °C for 3 h. After digestion, retinal vasculature was stained with PAS staining solution (Leagene, Beijing, China) for 15 min and counterstained with hematoxylin for 2 min. The retinal capillaries were observed by a BX53 microscope (Olympus).

Immunohistochemistry (IHC) and Immunofluorescence staining

The embedded retinal sections were used for the IHC staining of VEGFA. Endogenous peroxidase activity was eliminated by the addition of 3% H₂O₂ (Sinopharm) for 15 min at room temperature. After blocking, the sections were incubated with VEGFA antibody (1:100, Affinity, Changzhou, China) overnight at 4˚C and then incubated with an HRP-labeled secondary antibody (1:500; ThermoFisher Scientific, Pittsburgh, PA, USA) for 1 h at room temperature. Subsequently, sections were counterstained with 3,3’ diaminobenzidine tetrachloride (DAB, Maixin Biotechnology, Fuzhou, China) and counterstained with hematoxylin. Finally, the images were visualized using a BX53 microscope (Olympus).

For immunofluorescence staining, sections were incubated with 1% bovine serum albumin (BSA, Sangon Biotech, Shanghai, China) to block the background and incubated with the primary antibody against TRIM44 (1:100, Proteintech, Wuhan, China) and cluster of differentiation 31 (CD31, 1:100, Novus Biologicals, Littleton, CO, USA) overnight at 4˚C. The sections were then incubated with secondary antibodies including FITC-labeled IgG (anti-rabbit, 1:200, Abcam, Cambridge, UK) or Cy3-labeled IgG goat (anti-mouse, 1:200, Invitrogen, Carlsbad, CA, USA) for 1 h at room temperature. Nuclei were counterstained with diamidino-2-phenyl-indole (DAPI, Aladdin, Shanghai, China).

Cell culture and treatments

Rat RMECs purchased from iCell Bioscience were grown in PriMed-iCell-002 medium at 37 °C in 5% CO₂. In this study, the experiments were divided into 3 groups: normal glucose (NG): RMECs were subjected to 5.5 mmol/L glucose (Macklin); mannitol osmotic control (OS): RMECs were subjected to 5.5 mmol/L glucose and 19.5 mmol/L mannitol (Macklin); high glucose (HG): RMECs were treated with 25 mmol/L glucose. Cell viability was assessed using the MTT method following 0, 12, 24, and 48 h treatment. After 48 h-culture, the cells were utilized for the detection of other experiments. To explore the function of TRIM44 in RMECs induced by HG, we infected the lentivirus silencing TRIM44 or control lentivirus into RMECs, and the infection efficiency was detected by western blot (WB) assay 72 h post-infection.

Quantitative real-time PCR (qRT-PCR)

RNA was isolated from rat retinal tissues or whole cells using the TRIpure (BioTeke, Corporation, Beijing, China) protocol, and cDNA was synthesized using the BeyoRT II M-MLV reverse transcriptase (Beyotime, Shanghai, China). After that, the expression of TRIM44 was measured by qRT-PCR using the SYBR Green (Solarbio), and the expression level was normalized to the β-actin gene. The 2^−ΔΔCT method was used to calculate the fold change of TRIM44 expression. The primers used were as follows:

TRIM44-forward: 5’-GCTGGATTTGAGCACCTAT-3’;

TRIM44-reverse: 5’- CCTTTCCACCAACTCTATCA-3’;

β-actin-forward: 5’-GGAGATTACTGCCCTGGCTCCTAGC-3’;

β-actin-reverse: 5’-GGCCGGACTCATCGTACTCCTGCTT.

WB assay

The expression of indicated molecules in the rat retinal tissues or cells was estimated by WB assay. In brief, the protein was extracted utilizing the RIPA protein lysis buffer supplemented with 1 mM PMSF (Solarbio). Equal amounts of proteins were separated by SDS-PAGE and transferred to the PVDF membrane (Millipore, Billerica, MA, USA). The membranes were then incubated with the following primary antibodies: TRIM44 antibody (Proteintech, Rosemont, IL, USA, dilution ratio: 1:1000), VEGFA antibody (Affinity, Changzhou, China, dilution ratio: 1:1000), KLF9 antibody (Abclonal, Wuhan, China, dilution ratio: 1:1000), or β-actin (Proteintech, dilution ratio: 1:10000) overnight at 4 °C. After blocking, the membranes were incubated with the secondary antibody conjugated with horseradish peroxidase (Solarbio, dilution ratio: 1:3000). The signal was visualized by adding the ECL substrate (Solarbio).

In vitro immunofluorescence staining.

For the immunofluorescence staining of TRIM44, the cells were fixed by 4% paraformaldehyde and permeabilized with 0.1% tritonX-100 (Beyotime). After that, the cells were incubated with the TRIM44 antibody (Proteintech, dilution ratio: 1:100) overnight at 4 °C and subsequently incubated the goat anti-rabbit IgG (Abcam, dilution ratio: 1:200) at 37 °C for 1 h. Finally, slices were subjected to incubation with DAPI (Aladdin) to facilitate the visualization of cell nuclei through fluorescent staining, and the images were captured and analyzed using microscopy (Olympus).

Enzyme-linked immunosorbent assay (ELISA)

The VEGFA level in the cell supernatants was determined by ELISA. Rat VEGFA ELISA Kit (LIANKE Biotech, Hangzhou, China) was utilized following the instructions of the manufacturer. The sample absorbance at 450 nm and 570 nm was determined using a microplate reader (BioTek, Winooski, VT, USA).

MTT assay

Cell viability was determined by MTT assay as described previously [23]. RMECs treated with NG, OS, or HG were seeded to a 96-well plate with 5 × 10³ cells per well. MTT solution (50 µL, KeyGen Biotech, Nanjing, China) were added to the cells and incubated at 37 °C for 4 h. After incubation, DMSO (KeyGen Biotech, 150 µL) was added to dissolve the purple crystals formed by the cells. The absorbance at 490 nm was measured using a microplate reader (BioTek).

Flow cytometry analysis

Cell proliferation was detected using a kFluor488-EdU kit (KeyGen Biotech) according to the atomic covalent reaction between azide compound and alkyne catalyzed by copper. In brief, cells were seeded to a 6-well plate with 1 × 10⁶ cells per well. The cells were incubated with 10 µM EdU for 4 h, and EdU was chemically conjugated to kFlour647-azide using the Click-reaction. EdU incorporation was detected by flow cytometry at 488 nm (Agilent, Santa Clara, CA, USA).

Wound healing assay

A wound healing assay was carried out to assess cell migration. First, cells were pretreated with mitomycin C (Sigma-Aldrich, St. Louis, MO, USA) for 1 h to completely inhibit cell proliferation. Cells were then scratched with the pipette tip, and the cell surface was washed with serum-free medium. Finally, the cells were photographed at 0 h and 24 h by using a microscope (Olympus), and the wound closed area was quantified at 24 h.

Transwell assay

The cells were planted into the transwell upper chamber pre-coated with Matrigel (Corning, NY, USA). The lower chamber was supplemented medium with 10% fetal calf serum. After 24 h-incubation, the invasive cells were fixed with 4% paraformaldehyde (Aladdin, Shanghai, China) for 20 min, stained with 0.5% crystal violet (Amresco, Solon, OH, USA) for 5 min, and counted under the microscope (Olympus).

Angiogenesis assay

RMECs (2 × 10⁴ cells/well) were seeded into a 96-well plate coated with 30 µL Matrigel (Corning) for 2 h at 37 °C. Following 6-h incubation, tubular structures were imaged with a microscope (Olympus). The number of tube formations was counted to evaluate the angiogenesis capacity.

Luciferase reporter assay

The CDS of rat KLF9 was inserted into the eukaryotic pcDNA3.1 vector (Generalbio, Chuzhou, China), which was named OE-KLF9, and the empty vector (EV) was used as the control. The pGL3 vector containing TRIM44 promoter (from − 7000 to + 50 bp, -470 to + 50 bp, and − 270 to + 50 bp) or not was transfected into the HEK-293T cells with the OE-KLF9 or EV vector, respectively. Luciferase activity was assessed after 48 h-transfection using the Luciferase Assay Kit according to the manufacturer’s protocol (KeyGen Biotech).

Proteomic analysis

Rat RMECs infected with TRIM44 shRNA-mediated lentivirus or control lentivirus were subjected to the HG treatment. Label-free proteomic analysis was performed to select the downstream target proteins of TRIM44. The extracted lysates were digested and subsequently analyzed by LC-MS/MS. Proteome Discoverer 2.4 was used for peptide identification, and a 1% false-discovery rate (FDR) was used to reduce false positives.

Differentially expression proteins (DEPs) were defined as those for which|log2FC| ≥ 0.8 and a P value < 0.05. Volcano and clusters of Eukaryotic Orthologous Groups (KOG) plots were produced by the R software (version 4.3.2) with the ggplot2 (version 3.5.0) and ggrepel packages (version 0.9.5). The heatmap was generated in R software using the pheatmap package (version 1.0.12). Gene Ontology (GO) enrichment analysis was performed to identify the functional enrichment of biological processes (BP), cellular components (CC), and molecular function (MF), and it was made using R software with the circlize (version 0.4.16) and ComplexHeatmap (version 2.18.0) packages. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was carried out using the ClueGO (version 2.5.10) via Cytoscape software (version 3.9.1) [24], and the networks were configured as “medium” with a 0.4 kappa score.

Statistical analysis

Statistical significance was calculated using GraphPad Prism 8.0 software. Differences between groups were analyzed using an unpaired t-test or one-way analysis of variance (ANOVA). Comparisons of weight and blood glucose levels across time points were analyzed by two-way repeated measures ANOVA. Data were expressed as mean ± standard deviation (SD). Asterisks above data points indicate significance, with 1, 2, or 3 asterisks indicating P < 0.05, P < 0.01, or P < 0.001, respectively. Non-significant changes are denoted with ns.

Results

TRIM44 is widely expressed in the retina of diabetic rats

As illustrated by Fig. 1A, the diabetic SD rats with STZ injection were established. Compared with the control rats, diabetic rats displayed high blood glucose concentration and decreased body weight (Fig. 1B). Specifically, the blood glucose level rose significantly to 26 ± 2.1 mmol/L in STZ-induced rats after 12 weeks. The higher mRNA and protein levels of TRIM44 were found in the tissues of rat retinas (Fig. 1C-D). As shown in Fig. 1E, we observed that TRIM44-positive retinal ganglion cells (RGCs) in the ganglion cell layer (GCL) were increased in diabetic rats.

Fig. 1

TRIM44 is upregulated in the retina of diabetic rats. (A) A schematic illustration of the experiments was presented. (B) The blood glucose (left) and body weight (right) of rats were measured at indicated time points. (C-D) After 12 weeks of induction, the mRNA and protein expression levels of TRIM44 in the rat retinal tissues were measured. (E) Representative images of immunofluorescence staining for TRIM44 (red) in the rat retina (magnification × 200). DR, diabetic retinopathy; STZ, streptozotocin; GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer. ^***P < 0.001. n = 6 per group

Knocking down TRIM44 attenuates retinal lesions in diabetic rats

To systemically evaluate the role of TRIM44 in DR, we injected the TRIM44 shRNA-mediated lentivirus into the vitreous cavity of rats. The enhanced expression of isolectin-B4 was observed in the rat retinal tissues after STZ induction, and the loss of TRIM44 impaired the angiogenesis (Fig. 2A). In diabetic rats, we found the swelling of the GCL and the thinning of the inner nuclear layer (INL) and outer nuclear layer (ONL). The morphological changes were significantly improved after the knockdown of TRIM44 (Fig. 2B). As shown in Fig. 2C, we found that TRIM44 deficiency decreased the number of retinal acellular capillaries (indicated by black arrows) induced by STZ. VEGFA, as one of the main cytokines contributing to DR development is responsible for retinal vascular leakage [25]. The IHC analysis revealed a significant elevation in VEGFA expression, indicated by a pronounced dark brown signal, in rats subjected to STZ administration. Furthermore, the knockdown of TRIM44 was found to partially restore the effects (Fig. 2D). Immunostaining images displayed the colocalization of TRIM44 and CD31⁺ endothelial cells in the GCL of retinal tissues (Fig. 2E).

Fig. 2

TRIM44 knockdown alleviates retinal lesions in diabetic rats. Representative images of (A) immunofluorescence staining for isolectin-B4 (magnification × 200) and (B) H&E staining (magnification × 200 & × 400) in the rat retina. (C) PAS staining was used to identify the abnormal lesions of retinal blood vessels (magnification × 400). Black arrows indicated vascular leakage. (D) IHC staining for VEGFA expression in rat retinas (magnification × 400). (E) Immunofluorescence staining of TRIM44 (green) and CD31 (red) in retinal sections (magnification × 200). H&E, hematoxylin and eosin; PAS, periodic acid-Schiff; IHC, immunohistochemistry; VEGFA, vascular endothelial growth factor A; IF, immunofluorescence; CD31, cluster of differentiation 31. n = 6 per group

TRIM44 is widely expressed in HG-induced RMECs

To assess the significance of TRIM44 in DR, we established cell models utilizing rat RMECs subjected to HG treatment. The viability of the cells exposed to HG was assessed, revealing a progressive increase in cell viability corresponding to the duration of the treatment. We also found that HG induction caused a modest increase in cell viability at 24 h and 48 h (Fig. 3A). Consistent with in vivo assay, TRIM44 expression was significantly expressed in RMECs under HG conditions (Fig. 3B-D). Immunofluorescence for TRIM44 suggested that TRIM44 was primarily located in the cytoplasm (Fig. 3D).

Fig. 3

TRIM44 is upregulated in rat RMECs induced by high glucose. Rat RMECs were subjected to three distinct treatments: NG (5.5 mmol/L glucose), OS (5.5 mmol/L glucose plus 19.5 mmol/L mannitol), or HG (25 mmol/L glucose). (A) Cell viability in RMECs was assessed at indicated time points. (B-C) The mRNA and protein expression levels of TRIM44 were detected in RMECs. (D) Representative images of immunofluorescence staining for TRIM44 in RMECs (magnification × 400). NG, normal glucose; OS, osmotic glucose; HG, high glucose; RMECs, retinal microvascular endothelial cells. ^nsP > 0.05; ^*P < 0.05; ^***P < 0.001. n = 3 per group

Knocking down TRIM44 attenuates HG-induced cell migration and invasion

To further investigate TRIM44’s role, we transfected the lentivirus silencing TRIM44 into the RMECs, and the transfected efficiency was verified by WB assay (Fig. 4A). The decreased protein amount of TRIM44 in HG-administrated cells was also measured after blocking of TRIM44 (Fig. 4B). Following MTT assay and EdU incorporation assay indicated that the knockdown of TRIM44 markedly inhibited cell viability (Fig. 4C) and cell proliferation (Fig. 4D and G). Additionally, the silencing of TRIM44 mediated by shRNA significantly inhibited the migration and invasion of endothelial cells (Fig. 4E-F and H-I).

Fig. 4

TRIM44 knockdown alleviates the migration and invasion of RMECs induced by high glucose. (A-B) Western blot assay was performed to verify the efficiency of lentiviral infection in untreated or treated RMECs cells. (C) Cell viability was assessed in RMECs after 72 h-infection. (D) EdU incorporation was detected via flow cytometry, and (G) the number of EdU-positive cells was quantified. (E-F) Cell migration (magnification × 100) and invasion (magnification × 200) ability were detected by wound healing assay and Transwell assay, respectively. The quantification of (H) the percentage of wound-healing rate and (I) the number of invasive cells were exhibited, respectively. EdU, ethynyl-2-deoxyuridine, ^*P < 0.05; ^**P < 0.01; ^***P < 0.001. n = 3 per group

Knocking down TRIM44 attenuates HG-induced vascular formation

Given the possibility of its inhibitory effects on epithelial cell morphologies, we then wondered whether TRIM44 contributed to angiogenesis. Compared to the NG group, the RMECs in the HG group exerted enhanced tube formation both at 12 h and 24 h (Fig. 5A-B). Knockdown of TRIM44 markedly inhibited HG-induced angiogenesis only at 24 h of HG treatment. Consistently, we discovered that the expression of VEGFA (the angiogenic marker) also declined in HG-triggered RMECs after 24 h-treatments (Fig. 5C-D).

Fig. 5

TRIM44 knockdown inhibits vascular formation of RMECs induced by high glucose. (A) Angiogenesis assay was carried out to assess the tube formation in RMECs after 12 h-treatment or 24 h-treatment (magnification ×100). (B) The number of tube formation was quantified. (C-D) VEGFA levels in RMECs were detected by ELISA and western blot assays. ELISA, enzyme-linked immunosorbent assay. ^nsP > 0.05; ^*P < 0.05; ^***P < 0.001. n = 3 per group

TRIM44 is transcriptionally upregulated by KLF9

To further explore the mechanism for the regulation of TRIM44 by KLF9, we measured the protein amount of KLF9 in NG, OS, and HG-treated RMECs. A high expression level of KLF9 in HG condition was demonstrated (Fig. 6A). Lentivirus harboring KLF9 overexpression was infected into RMECs with or without HG stimulation, and the infection largely induced KLF9 upregulation (Fig. 6B-C). The results of qRT-PCR confirmed that KLF9 indeed upregulated TRIM44 expression in endothelial cells induced by HG (Fig. 6D). The expression of TRIM44 in rat RMECs without any administration was further detected to determine the regulation between KLF9 and TRIM44, and the results showed that the upregulation of KLF9 led to an increase in the protein levels of TRIM44 (Fig. 6E). However, the degree of TRIM44 upregulation was less than that in cells stimulated by HG (Fig. 6F). As indicated in Fig. 6G, the promoter region of TRIM44 from − 700 to + 50, -470 to + 50, and − 270 to + 50 bp were inserted in the pGL3 vector, respectively. We found that KLF9 upregulation significantly increased luciferase activity driven by TRIM44 promoter, and the region from − 700 to + 50 bp had strong promoter activity in HEK-293T cells. These results suggested that the promoter fragments spanning positions − 700 bp to -270 bp encompassed the potential positive regulatory elements and TRIM44 was transcriptionally activated by KLF9.

Fig. 6

TRIM44 is transcriptionally regulated by KLF9. (A) Western blot assay was performed in RMECs treated with different conditions to assess the protein expression of KLF9. (B-C) The efficiency of lentiviral infection was verified by detecting KLF9 expression in untreated or treated RMECs. (D) The mRNA level of TRIM44 were detected in KLF9-overexpressed RMECs with HG treatment. (E-F) The protein expression levels of TRIM44 were detected in KLF9-overexpressed RMECs without or with HG stimulation. (G) Luciferase activity of TRIM44 promoter was measured in 293T cells co-transfected with KLF9 overexpression vector or emptor vector. ^*P < 0.05; ^**P < 0.01; ^***P < 0.001. n = 3 per group

TRIM44 knockdown affects KLF9-mediated cell function in HG-induced RMECs

We carried out rescue assays in HG-driven RMECs co-infected with LV-KLF9 and LV-shTRIM44 to elucidate the regulatory mechanism between KLF9 and TRIM44. MTT assay showed that the significantly enhanced cell viability in KLF9-overexpressed cells driven HG was suppressed by TRIM44 knockdown (Fig. 7A). In the HG conditions, endothelial cells with silenced TRIM44 exhibited reduced cell migration and invasion upon the overexpression of KLF9 (Fig. 7B-E).

Fig. 7

TRIM44 mediates the effects of KLF9 on DR pathology. (A) Cell viability was assessed in RMECs by MTT assay. (B) Cell migration ability in RMECs was detected by wound healing assay (magnification ×100). (C) Angiogenesis assay was carried out to assess the tube formation in RMECs (magnification ×100). The quantification of (B) the percentage of wound-healing rate and (E) the number of tube formations were exhibited. ^*P < 0.05; ^**P < 0.01. n = 3 per group

Screening for downstream target proteins of TRIM44

After exploring the function and upstream mechanisms of TRIM44 in DR, we also desired to have a comprehensive overview of TRIM44 downstream target proteins using proteomics. As shown in Fig. 8A, TRIM44-silenced RMECs or control cells after HG administration were used for proteomic analysis, and bioinformatics analysis were carried out to screen out the DEPs. Proteins with P value < 0.05 and|log₂FC| ≥ 0.8 were considered to have significant changes. We found that there were 38 downregulated and 64 upregulated proteins (Fig. 8B), and the top 10 DEPs in proteomics were listed in Table 1. The clustering heatmap demonstrated that the protein expression patterns across the three samples within each group exhibited a degree of similarity (Fig. 8C).

Fig. 8

Identification of differentially expressed proteins by label-free proteomic analysis. (A) Flow diagram of proteomic analysis. (B) A volcano plot was utilized to depict the differential expressed proteins in HG-induced RMECs between TRIM44 knockdown and its negative control. The accession of the top 10 differentially expressed proteins was labeled in the diagram. (C) A heatmap with hierarchical clustering was utilized to depict the differential expressed proteins. A single row represented each differentially expressed protein, whereas a single column represented each cell sample. (D) KOG annotation of differentially expressed proteins. These proteins were classified into 19 functional clusters. LC-MS/MS, liquid chromatography-mass spectrometry; DEPs, differentially expression proteins; KOG, Eukaryotic Orthologous Groups; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes

Table 1 List of top 10 differentially expressed proteins in proteomics

According to KOG annotation, we found that the functional distributions of the DEPs were grouped into 19 categories, and posttranslational modification, protein turnover, and chaperones (indicated by blue font) was the functional category with the largest number of proteins (Fig. 8D). The detailed information of the 11 DEPs in KOG category O was listed in Table 2.

Table 2 List of 11 differentially expressed proteins in the most enriched class of KOG annotation

GO and KEGG enrichment analysis of deps

To obtain further insight into the functional mechanism, we performed GO and KEGG enrichment analysis on the 64 upregulated and 38 downregulated proteins. DEPs were enriched and clustered according to GO terms. As indicated in Fig. 9A, upregulated proteins were mainly enriched in 3 GO-MFs, including ganglioside GM1 binding, kininogen binding, and beta-N-acetylhexosaminidase activity. Additionally, the downregulated proteins were significantly enriched in 2 GO-MFs, including protein-glutamate O-methyltransferase activity, and fucokinase activity (Fig. 9B). Following KEGG results showed that there were abundant proteins enriched in drug metabolism, collecting duct acid secretion, MAPK signaling pathway, nitrogen metabolism, glycosphingolipid biosynthesis, and sphingolipid signaling pathway (Fig. 10A). The results presented in Fig. 10B showed that the knockdown of TRIM44 was associated with a downregulation of a greater number of proteins that are implicated in various processes, including microRNAs in cancer, ECM-receptor interaction, human papillomavirus infection, focal adhesion, herpes simplex virus 1 infection, endocytosis, glycosaminoglycan biosynthesis, synthesis and degradation of ketone bodies.

Fig. 9

GO enrichment analysis of differentially expressed proteins. GO enrichment analysis of the (A) upregulated and (B) downregulated proteins was performed. The results were presented via circle diagrams, which included 4 circles from outside to inside: the Term ID (green, orange, and blue gray), BgRatio (purple), the number of upregulated (light green) or downregulated (yellow) proteins, and Richfactor (grey)

Fig. 10

KEGG enrichment analysis of differentially expressed proteins. ClueGo in Cytoscape was used to conduct KEGG pathway enrichment analysis to find pathways that contain the (A) upregulated or (B) downregulated proteins

Discussion

DR is a vision-threatening diabetic complication, and it typically results in functional and structural damage to the retina [26]. DR has been extensively acknowledged as a microvascular complication linked to endothelial dysfunction. This condition is characterized by several pathological changes, including retinal cell apoptosis, BRB destruction, retinal capillary basement membrane thickening, loss of pericytes and endothelial cells, as well as the formation of abnormal neovascularization [27, 28]. Although researchers have made great efforts to unravel the underlying mechanisms of DR pathology, the exact causes of DR mostly remain unknown. Due to the multifactorial nature of the etiology of DR, there is currently no effective treatment for DR. The current investigation revealed several new findings regarding the function of TRIM44 in retinal lesions associated with diabetes. First, TRIM44 was demonstrated to be upregulated in the retinas of rats with diabetes induced by STZ as well as in rat RMECs subjected to HG conditions. Second, the loss of TRIM44 remitted diabetes-triggered structural damage to the retinal vasculature, indicating that TRIM44 might contribute to the pathogenesis in DR. Third, in vitro assays demonstrated that TRIM44 is implicated in the processes of cell migration, invasion, and vascular formation in RMECs exposed to HG, and the transcriptional activity of TRIM44 was markedly activated by the KLF9, a transcription factor that positively regulates angiogenesis.

VEGFA, a major contributor to retinal neovascularization, plays a central role in the pathogenesis of DR-related hyperpermeability and angiogenesis [29]. In many patients with DR, overproduction of VEGFA in the retina leads to hyperproliferation of epithelial cells and neovascularization, resulting in microaneurysms and tissue damage [30]. Therefore, the development of anti-VEGFA therapies has been one of the most important advances in the field of DR. It has been proven that the mRNA level of VEGFA is upregulated within 1 to 2 weeks after induction of diabetes in rats, and elevated VEGFA is observed in a rat model even 6 months after diabetes [31]. In the present investigation, TRIM44 deficiency was found to impede angiogenesis through the downregulation of VEGFA expression, subsequently delaying the progression of DR. Furthermore, the in vitro results corroborated the findings observed in vivo. Therefore, studies dedicated to TRIM44 inhibition may provide us with new directions for the treatment of DR.

The significant roles of TRIM44 in various diseases have been reported in recent studies. For example, elevated TRIM44 expression is related to the T-cell infiltration and modulates tumor immunity in gastric cancer. It suggests that TRIM44 may serve as a novel immunotherapy indicator of gastric cancer [32]. Kim et al. indicated that TRIM44 may be a valuable target for evaluating the sensitivity of PARP inhibitors, and focusing on TRIM44 presents opportunities to expand its application to tumors that exhibit homologous recombination proficiency [33]. In addition to its recognition as a prognostic biomarker for multiple types of cancer, TRIM44 is instrumental in modulating the sensitivity of autophagy to oxidative stress [34]. The multiple functions of TRIM44 have continued to expand our understanding, and future studies could also focus on the development of biomarkers based on TRIM44 expression or activity to improve disease diagnosis or prognosis prediction.

Accumulating evidence for the involvement of inflammatory processes, including cytokine upregulation and leukocyte infiltration, has led to diabetic retinopathy being recognized as an inflammatory disease [35,36,37]. There are many cells involved in the progression of DR, including microglia, Müller cells, astrocytes, endothelial cells, and pericytes [38]. With the damage of RGCs, there is accompanied by the possible cascade reaction under diabetic conditions. In other words, the integrity of the BRB depends on the cells fulfilling their duties in DR progress [38]. For instance, Müller cells exhibit a significant production of cytokines when exposed to hyperglycemia and may also exert direct effects on microglia. This indicates that Müller cells serve as a primary source of inflammatory mediators in DR [39]. Here, considering the background of TRIM44 in endothelial cells, we only explored the expression and function of TRIM44 in HG-induced rat RMECs, as for the function in other cells affecting BRB function remains to be further explored. The above results provide compelling evidence for the hypothesis that TRIM44 enhanced vascular permeability in DR progress. Therefore, focusing on TRIM44 has the potential to develop novel therapeutic interventions to improve DR.

By analyzing the proteomics data, we found that upregulated NASP (log₂FC = 1.58, P < 0.05) and downregulated THBS1 (log₂FC = -1.92, P < 0.05) expression in TRIM44-silenced RMECs after HG administration. The role of NASP in the DR process has not been reported. The evidence suggests that THBS1 is a potent endogenous inhibitor of angiogenesis, and it is found to be overexpressed in DR [40]. Another study indicates that THBS1 expression is suppressed in retinal tissues of DR rats and THBS1 inhibition promotes tube formation in rat endothelial cells [41]. However, the detailed mechanism of THBS1 overexpression in DR remains unknown, which warrants further investigation. Furthermore, KOG enrichment analysis showed that the DEGs were enriched in posttranslational modification, protein turnover, and chaperones. Here, we do not intend to put all our findings within a single article. Instead, subsequent research endeavors will persist in the identification of the downstream targets of TRIM44, with the aim of formulating efficacious therapeutic approaches for DR.

There is an abundance of evidence to support the function of TRIM44 which exerts an important role in many cancers [10, 11, 42]. Even though dysregulation of TRIM44 has been discovered in the development of different diseases, the regulation mechanisms are still largely unclear. A previous study claimed that ELFN1-AS1 regulated TRIM44 expression at post-transcription level, but it did not influence the activity of TRIM44 promoter [43]. In this study, we demonstrated that a novel transcription factor, KLF9, enhanced TRIM44 transcription in HG-triggered RMECs, and TRIM44 knockdown affects KLF9-mediated cell function in HG-induced cells. Additionally, evidence has been provided indicating that KLF9 can enhance the expression levels of another TRIM family component, TRIM33 in rheumatoid arthritis [20].

In this study, the experiments were repeated at least 3 times in both cell and animal models. Although the sample size utilized in this study was adequate to meet the initial research objectives and yield statistically significant outcomes [44, 45], we acknowledge that a larger sample size could potentially enhance the generalizability and robustness of our findings. Notably, the morphology and structural characteristics of retinal microvascular networks in the rat retina show some variances compared to that in humans [46], which could potentially impact the role of TRIM44 in the context of DR. As described by Brubaker et al., the generalization of results obtained from animal models to human patients constitutes a significant challenge within the biomedical field [47]. Additionally, while cell lines offer a simplified model for studying cellular mechanisms, they lack the complexity of the in-vivo human environment [48]. In patients with DR, various factors, including comorbidities, diverse genetic backgrounds, and lifestyle, can affect the progression of the disease [49,50,51]. Therefore, the inability to directly translate these findings to human subjects represents a limitation of our study. However, our study offers significant insights and establishes a basis for subsequent investigations in human-related fields. We will explore more in-depth studies to bridge this gap in the future.

Conclusion

In summary, we found that the upregulation of TRIM44 in the retinas of diabetic rats and HG-induced endothelial cells. Knocking down TRIM44 ameliorated pathological structural changes and decreased the level of VEGFA in the retina, suggesting TRIM44 was tightly associated with neovascularization. We discovered that TRIM44 was transcriptionally activated by KLF9 under HG conditions. Considering the protective role of TRIM44 knockdown in DR, future studies should aim to determine whether TRIM44 inhibition is also an effective treatment for other ocular neovascular diseases.