Abstract

   Regulation of visual system function demands precise gene regulation.
   Dysregulation of miRNAs, as key regulators of gene expression in
   retinal cells, contributes to different eye disorders such as diabetic
   retinopathy (DR), macular edema, and glaucoma. MIR-96, a member of the
   MIR-183 cluster family, is widely expressed in the retina, and its
   alteration is associated with neovascular eye diseases. MIR-96
   regulates protein cascades in inflammatory and insulin signaling
   pathways, but further investigation is required to understand its
   potential effects on related genes. For this purpose, we identified a
   series of key target genes for MIR-96 based on gene and protein
   interaction networks and utilized text-mining resources. To examine the
   MIR-96 impact on candidate gene expression, we overexpressed MIR-96 via
   adeno-associated virus (AAV)-based plasmids in human retinal pigment
   epithelial (RPE) cells. Based on Real-Time PCR results, the relative
   expression of the selected genes responded differently to overexpressed
   MIR-96. While the expression levels of IRS2, FOXO1, and ERK2 (MAPK1)
   were significantly decreased, the SERPINF1 gene exhibited high
   expression simultaneously. pAAV-delivered MIR-96 had no adverse effect
   on the viability of human RPE cells. The data showed that changes in
   insulin receptor substrate-2 (IRS2) expression play a role in disrupted
   retinal insulin signaling and contribute to the development of diabetic
   complications. Considered collectively, our findings suggest that
   altered MIR-96 and its impact on IRS/PI3K/AKT/VEGF axis regulation
   contribute to DR progression. Therefore, further investigation of the
   IRS/PI3K/AKT/VEGF axis is recommended as a potential target for DR
   treatment.

Introduction

   microRNAs (miRNAs) are a class of small non-coding RNAs that influence
   gene expression patterns of cellular pathways and regulate diverse
   biological processes in this way. Their interaction with specific
   sequences in target mRNAs mediates post-transcriptional gene
   regulation. Recent studies have revealed that in addition to gene
   silencing, miRNAs can also activate gene expression through particular
   pathways [[34]1, [35]2]. They play a leading role in developmental,
   functional, and pathological conditions in the retina [[36]3]. The
   microRNA-183 cluster is a prominent family among the most frequently
   reported miRNAs in the retina, comprising MIR-183, MIR-96, and MIR-182.
   During retinal development, members of the MIR-183 cluster are
   overexpressed as a single transcript and regulate different target
   genes involved in cellular pathways [[37]4]. Given their pivotal roles,
   altered expression of the MIR-183 cluster has been shown to be linked
   to a variety of human diseases, including diabetes and diabetic
   vascular complications [[38]5, [39]6]. miRNA microarray analysis
   indicated that the MIR-96 level increased in the retinas of
   streptozotocin-induced diabetic rats [[40]7]. However, the MIR-96
   expression level was significantly downregulated in the serum of type 2
   diabetes patients compared with normal controls [[41]8]. A few studies
   have explored the potential involvement of miR-96 in the development
   and progression of diabetic retinopathy (DR). However, the exact
   mechanisms by which miR-96 influences the development of DR remain
   unclear and require further investigation. A recent study has
   investigated how the target genes regulated by miR-96 are connected to
   the signaling pathways associated with DR [[42]9]. Diabetic
   retinopathy, as one of the most common microvascular complications of
   diabetes, remains a dominant cause of visual impairments in diabetic
   patients. In the early stages of DR, the hyperglycemic condition
   contributes to vasodilation and increased retinal metabolism, pericyte
   apoptosis, microaneurysm, and eventually retinal microvascular damage.
   Advanced stages of DR, characterized mainly by neovascularization,
   culminate in severe vision impairments due to vitreous hemorrhage or
   detachment of the retina [[43]10, [44]11]. Altered retinal metabolism
   and elevated levels of inflammatory molecules affect multiple parts of
   the retina, including the vascular network, neuronal cells, choroid,
   and retinal pigment epithelium (RPE) [[45]12]. RPE cells secrete
   several growth factors and cytokines, including vascular endothelial
   growth factor (VEGF), pigment epithelium-derived factor (PEDF),
   monocyte chemoattractant protein-1 (MCP-1), interleukin-6 and 8 (IL-6,
   IL-8), and matrix metalloproteinase (MMPs). Studies have shown that
   high glucose exposure modifies the RPE proteome and secretome and can
   cause alterations in cell structure and function [[46]13].

   Recent studies have identified signaling pathways involved in DR
   development and progression. The most known pathways are insulin
   signaling, VEGF signaling, IL-6 signaling, and PI3K-AKT signaling
   pathways [[47]14]. According to the latest research, MIR-96, as well as
   other regulatory factors is involved in the control of signaling
   molecules. For instance, upregulation of MIR-96 leads to insulin
   signaling impairment through relevant repressing signaling molecules
   such as insulin receptor (INSR) and insulin receptor substrate-1 (IRS1)
   [[48]15, [49]16]. IRS proteins are cytoplasmic receptors involved in
   insulin action by activating the phosphoinositide 3-kinase (PI3K)/AKT
   pathway. Studies showed that IRS2 is implicated in type 2 diabetes, and
   its expression is upregulated in DR mouse models [[50]17, [51]18]. A
   previous study suggested activating the PI3K/AKT pathway upregulates
   VEGF expression in vascular cells [[52]18, [53]19]. Moreover, a new
   study determined that MIR-96 acts as a modulator of multiple angiogenic
   factors, including VEGF, ANG2, and VEGFR2, during pathological
   conditions in the retina [[54]20]. These suggest that the IRS/PI3K/AKT
   axis could be counted as a possible target in DR prevention and
   control.

   Considering the involvement of the MIR-96 target genes in specific
   signaling pathways, the present study was designed to achieve an
   understanding of MIR-96’s roles in molecular alterations in the retina.
   To that end, MIR-96 target genes were provided using four miRNA target
   prediction online databases, including TargetScanHuman 8.0, miRTarBase
   9.0, miRDB, and miRWalk 3. The results of three of these databases
   include predicted miRNA targets using bioinformatics tools. MiRTarBase
   9.0 provides miRNA targets with a higher confidence level validated
   experimentally by reporter assay, western blot, microarray, and
   next-generation sequencing experiments. The main signaling pathways
   linked to MIR-96 were identified using mirPath. By analyzing the data
   of the marked target genes in enriched KEGG pathway maps provided by
   DIANA-mirPath, the top co-expressed genes were determined utilizing
   GeneMANIA. In addition, the functional association network was acquired
   from the STRING online database. Taken together, the results provided a
   ranked list of genes to investigate the effects of MIR-96
   overexpression in human RPE cells.

Materials and methods

In silico target research

   To establish a comprehensive understanding of the target interactions
   for MIR-96, the in silico target research was conducted using multiple
   prediction tools and databases. In silico target prediction for MIR-96
   was done using TargetScan 8.0 ([55]http://www.targetscan.org/vert_80/)
   [[56]21], miRDB ([57]http://mirdb.org/) [[58]22], miRTarBase 9.0
   ([59]https://mirtarbase.cuhk.edu.cn/) [[60]23], and miRWalk 3
   ([61]http://mirwalk.umm.uni-heidelberg.de/) [[62]24]. Top-scored genes
   were identified and integrated with literature search and text mining
   results obtained from the EVEX database ([63]http://evexdb.org/)
   [[64]25]. Subsequently, the resulting gene list was introduced to
   STRING 11.5 ([65]https://string-db.org/) [[66]26], and the functional
   interactions among identified proteins were visualized using the
   Cytoscape stringApp (version 3.9.1) [[67]27].

   The miRNA-pathway interactions were predicted using DIANA mirPath v.3
   ([68]https://dianalab.e-ce.uth.gr/html/mirpathv3/index.php?r=mirpath)
   [[69]28], and the Kyoto Encyclopedia of Genes and Genomes (KEGG)
   pathway enriched terms were identified through Metascape website
   ([70]http://metascape.org/) [[71]29]. In addition, the DisGeNET
   database ([72]https://www.disgenet.org/) [[73]30–[74]32] was explored
   to retrieve DR-associated genes, and the gene-disease associations
   (GDA) network was visualized using the DisGeNET Cytoscape app [[75]33].
   This database covers a wide range of diseases and can be a valuable
   resource for studies on genes associated with human diseases. [76]Fig 1
   illustrates a schematic overview of the bioinformatic tools conducted
   in this study.

Fig 1. A schematic overview of the bioinformatic tools conducted in this
study.

   [77]Fig 1
   [78]Open in a new tab

   GDA = Gene-Disease Associations.

Vector constructs

   Following the procedure of the Molecular Cloning Laboratory Manual
   [[79]34], genomic DNA was isolated from hRPE cells. The MIR-96 sequence
   was amplified from genomic DNA using polymerase chain reaction (PCR)
   (primer sequences: 5′- CGGGGTACCACCGAAGGGCCATAAACAGA - 3ˊ (Forward),
   and 5ˊ- CTAGCTCGAGAGTGTAAGGCGATCTGGCT - 3ˊ (Reverse)). Extra base pairs
   were added to 5ˊ end of the primers to increase the cleavage efficiency
   of restriction enzymes. The purified PCR product was digested with XhoI
   and KpnI restriction enzymes and inserted into the pAAV-MCS (a human
   AAV-2 vector included in the AAV Helper-Free System designed by
   ©Agilent Technologies, USA) that had been previously manipulated in our
   lab to incorporate the enhanced green fluorescent protein (eGFP)-intron
   cassette [[80]35]. Thereafter, Escherichia coli XL10 bacteria (Agilent,
   USA) was transformed by the ligation product by the heat shock method.
   Bacterial clones comprising MIR-96 were identified using plasmid
   mini-preparation and digestion. Subsequently, the sequences of MIR-96
   and eGFP-intron fragments were verified using DNA sequencing analysis
   (Bioneer Corporation, South Korea). The pAAV-MCS-eGFP-int was also
   employed as a control vector.

Human RPE cell culture and transfection

   The human RPE (hRPE) cell line used in the present study was derived
   from the previously isolated, characterized, and established HRPE-2S
   cell line [[81]36]. The cells were cultured in DMEM/F-12 mixture
   (Gibco, Thermo Fisher Scientific, USA) supplemented with 10% fetal
   bovine serum (FBS, Biowest, France), penicillin (64 μg/ml), and
   streptomycin (100 μg/ml) in T25 culture flasks (SPL Life Science,
   Korea). Cells were incubated in a humidified 37 ˚C incubator with 5%
   CO[2] (Binder, USA). For transfections, hRPE cells were seeded in a
   six-well plate (SPL Life Science, Korea). After 24 hours (60–70%
   confluency), the cells were transfected with the MIR-96 expression
   vector by the calcium phosphate-mediated transfection method. For each
   well, 5 μg of the desired plasmid was added to sterile ddH[2]O and
   CaCl[2] 2M. The CaCl[2]/DNA mix was added dropwise to an equal amount
   of HBS 2X (HEPES Buffered Saline) while vortexing. Twenty minutes
   post-incubation at room temperature, the transfection mixture,
   containing calcium phosphate-DNA particles was added dropwise and
   evenly over the cell culture medium. After 6 hours, the medium was
   replaced with fresh DMEM/F-12 + 10% FBS, and the cells were harvested
   72 hours later for further experiments.

RNA extraction and RT-qPCR analysis

   Total RNA was isolated from both transfected and control hRPE cell
   cultures (which were transfected by pAAV-MIR-96-eGFP-int or
   pAAV-MCS-eGFP-int cassette as control) using TriPure Isolation Reagent
   (Roche, Germany). Mature MIR-96 expression was quantified following the
   stem-loop reverse transcription-quantitative polymerase chain reaction
   (RT-qPCR) procedure [[82]37]. Initially, RNA samples were reversely
   transcribed by a specific stem-loop primer using the Moloney murine
   leukemia virus (MMLV) enzyme (Qiagen Inc., USA), then the synthesized
   complementary DNA (cDNA) was used as a template to amplify MIR-96 with
   a specific forward primer and a universal reverse primer. The Real-Time
   PCR program was as follows: 15 min at 95 ˚C for activation, 15 s at 95
   ˚C, and 60 s at 60 ˚C for 45 cycles. The expression level of MIR-96
   transcripts was normalized to the expression level of RNU48 as
   endogenous control. To quantify gene expression in transfected hRPE
   cells, RNA samples were reversely transcribed to cDNAs, using a mixture
   of oligo (dT) and random hexamer primers and reverse transcriptase
   (Qiagen Inc., USA). Real-time PCR was performed with QuantiFast SYBR®
   Green PCR Kit (Qiagen, Germany, Catalog no. 204143) master mix and
   specific primer pairs of the genes, using Taq DNA polymerase (Qiagen
   Inc., USA). The utilized Real-Time PCR program was as follows: 10 min
   at 95 ˚C for activation, 15 s at 95 ˚C, and 60 s at 60 ˚C for 45
   cycles. The expression level of genes was normalized to the expression
   level of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as a verified
   housekeeping gene. The 2^- ΔΔCT method was used to calculate and
   analyze relative changes in gene expression for both the control and
   MIR-96-transfected groups. The relative expression of each gene in the
   transfected group was reported compared to the control group. The
   results are representative of at least three independent experiments
   with three replicates.

Cell viability assay (MTT)

   Cellular metabolic activity, an indicator of cell viability,
   proliferation, and cytotoxicity was assayed in transfected hRPE cells
   (pAAV-MIR-96-eGFP-int, and control pAAV-MCS-eGFP-int transfected hRPE
   cells) compared to non-transfected control. Both transfected and
   non-transfected cells were seeded at a density of 1 × 10^5 cells per
   well in 96-well plates (SPL Life Science, Korea). Following overnight
   incubation at 37 ˚C, the cells were treated with 10 μl
   3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT)
   (5 mg/ml, Sigma-Aldrich, USA). After 4 hours, the medium containing MTT
   was gently removed, and 100 μl dimethyl sulfoxide (DMSO)
   (Sigma-Aldrich, USA) was added to each well to dissolve formazan
   crystals. Thereafter, cell viability was determined by measuring
   optical density at 580 nm using a microplate reader. Non-transfected
   hRPE cells were considered for normalizing absorbance data.

Statistical analysis

   All transfections were independently done at least three times. RT-qPCR
   was performed in three independent experiments with three replicates in
   each experiment. Significant statistical differences were evaluated by
   one-way (to compare the relative expression levels among the candidate
   genes) and two-way (to assess the changes in gene expression at
   different time points) analysis of variance (ANOVA). Statistical
   analyses were carried out using GraphPad Prism version 8.0.1 (GraphPad
   Software, San Diego, California USA, [83]www.graphpad.com). The data
   were reported as the mean ± standard error of the mean, and P values
   less than 0.05 were considered statistically significant.

Results

Bioinformatics analysis

   We determined overlapping target genes from the four miRNA target
   prediction databases. The resulting genes were then integrated with
   text mining outcomes, extended, and visualized using stringApp ([84]Fig
   2). The functional enrichment analysis also showed that the target
   genes are mainly enriched in pathways including type II diabetes
   mellitus, VEGF signaling pathway, and insulin signaling pathway
   ([85]Table 1), indicating that these genes could be related to several
   biological processes underlying diabetic retinopathy.

Fig 2. KEGG pathway enrichment and PPI network of MIR-96 associated genes.

   [86]Fig 2
   [87]Open in a new tab

   The thickness of the connecting lines indicates the confidence level.
   Node borders with different colors present the specific biological
   pathway. KEGG = Kyoto Encyclopedia of Genes and Genomes; PPI =
   Protein-Protein Interaction; VEGF = Vascular Endothelial Growth Factor.

Table 1. Summary of KEGG enrichment analysis.

   Pathway  Description               Strength False Discovery Rate
   hsa04930 Type II diabetes mellitus 2.23     2.05e-11
   hsa04370 VEGF signaling pathway    2.14     4.92e-11
   hsa04910 Insulin signaling pathway 1.95     1.39e-14
   [88]Open in a new tab

   DIANA-mirPath analysis using the Tarbase algorithm with a threshold
   value of P < 0.05, and false discovery rate (FDR) correction implied
   that MIR-96 could be involved in the regulation of insulin signaling
   pathway. Metascape extracted all the protein-protein interactions among
   the input genes from the PPI data source and formed a PPI network. To
   identify densely connected proteins, the molecular complex detection
   (MCODE) algorithm was then applied to this network. Gene ontology (GO)
   enrichment analysis showed that the MCODE1 network components were
   significantly enriched in the insulin receptor signaling pathway and
   type II diabetes mellitus ([89]Fig 3).

Fig 3. Gene ontology enrichment analysis.

   [90]Fig 3
   [91]Open in a new tab

   (A) GO and KEGG pathway enriched terms recognized using Metascape
   website, sorted by -log10. (B) Network layout of representative terms
   visualized by Cytoscape. Each term was represented by a circle node.
   The colors describe the clusters’ identities. The thickness of the
   connecting line represents the similarity score. (C) The most
   significant MCODE model. GO enrichment analysis was applied to MCODE1,
   and the term “type II diabetes mellitus” was retained from the top
   three best p-value terms. GO = Gene ontology; KEGG = Kyoto Encyclopedia
   of Genes and Genomes; MCODE = Molecular Complex Detection.

   To explore related genes that are likely involved in the molecular
   mechanisms underlying diabetic retinopathy, DR-associated genes
   (altered expression) were retrieved from the DisGeNET database (v7.0)
   and gene-disease associations (GDAs) were visualized by the DisGeNET
   Cytoscape app ([92]Fig 4).

Fig 4. Network of differentially expressed genes associated with diabetic
retinopathy.

   [93]Fig 4
   [94]Open in a new tab

   (A) Association of common differentially expressed genes with diabetic
   retinopathy (DR) according to the DisGeNET database. (B) Top scored
   DR-associated genes. The enriched signaling pathway interrelated genes
   were highlighted in yellow, and the score ranges from 0.5 to 1.0.

Construction of the MIR-96 expressing vector, transfection and analysis of
GFP expression

   MIR-96 gene was amplified from human genomic DNA. The final PCR product
   (424 bp) was inserted into a pAAV-MCS-eGFP-int plasmid (5506 bp,
   control vector) to produce a 5.9-kb recombinant vector ([95]Fig 5A).
   The 1.3-kb DNA fragment containing MIR-96, eGFP-int, and cloning
   junction sequences was verified by DNA sequencing. Next, the
   recombinant plasmids were transfected to hRPE cells. 48 hours
   post-transfection, the cells were inspected for eGFP protein expression
   using fluorescence microscopy. As illustrated in [96]Fig 5B and 5C,
   both hRPE cultures transfected with either pAAV-MIR-96-eGFP-int or
   pAAV-MCS-eGFP-int display over 90% eGFP-expressing cells.

Fig 5. Schematic representation of pAAV-MIR-96-eGFP-int construct and hRPE
cells’ transfection.

   [97]Fig 5
   [98]Open in a new tab

   (A) MIR-96 fragment was inserted into pAAV-MCS-eGFP-int cassette (5.5
   kb) to construct pAAV-MIR-96-eGFP-int vector (5.9 kb). (B) GFP
   expressing hRPE cells, 48 h after transfection of the recombinant
   vector, and (C) GFP expressing hRPE cells, 48 h after transfection of
   the control vector. The fluorescence microscopy image demonstrates more
   than 90% of both cultures were GFP positive. hRPE = human Retinal
   Pigment Epithelium; AmpR = ampicillin resistance gene; CMV =
   cytomegalovirus; AAV2 = adeno-associated virus-2; ITR = inverted
   terminal repeats; GFP = green fluorescent protein.

Overexpression of MIR-96 in the transfected hRPE cells

   The result of stem-loop RT-qPCR, when compared with hRPE cells
   transfected with pAAV-MCS-eGFP-int as control, revealed that MIR-96 was
   upregulated by more than 82-fold in the transfected hRPE cells 48 and
   72 h post-transfection ([99]Fig 6A).

Fig 6. Effects of MIR-96 on cell viability.

   [100]Fig 6
   [101]Open in a new tab

   (A) The relative expression level of MIR-96 following hRPE cell
   transfection (at 48 and 72 hours). (B) Cell viability analysis using
   MTT assay. Data expressed as mean ± standard error of the mean (SEM). P
   < 0.05. MTT = 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium
   bromide; GFP = green fluorescent protein.

Impact of pAAV-MIR-96-eGFP-int overexpression on hRPE cell viability

   We further evaluated the effect of MIR-96 overexpression on the
   viability of the transfected hRPE cells in a time-dependent manner. MTT
   assay revealed that overexpression of MIR-96 had no statistically
   significant effect on cell viability when compared with the control
   group (P = 0.84, one-way ANOVA, [102]Fig 6B).

Effect of MIR-96 overexpression on target genes’ expression

   To explore the possible impacts of MIR-96 overexpression on insulin
   signaling pathway mediators and DR-related genes, the expression level
   of the nominated genes was assessed using RT-qPCR. As represented in
   [103]Fig 7, the expression levels of IRS2, FOXO1, and ERK2 genes were
   significantly downregulated (by 0.3, 0.8, and 0.4-fold, respectively)
   48 hours post-transfection in pAAV-MIR-96-eGFP-int transfected group
   compared with the control. Furthermore, SERPINF1 which displayed higher
   expression (2.7- fold) up until 48 h, decreased (0.4-fold) within 72 h
   after transfection compared with the control group.

Fig 7. Relative expression of selected genes in transfected hRPE cells.

   [104]Fig 7
   [105]Open in a new tab

   Error bars represent means ± standard error of the mean (SEM). *P <
   0.05, **P < 0.01. IRS1 = Insulin Receptor Substrate 1; IRS2 = Insulin
   Receptor Substrate 2; FOXO1 = Forkhead box protein O1; ERK2 =
   Extracellular signal-Regulated Kinase 2; SERPINF1 = Serpin Peptidase
   Inhibitor, Clade F, Member 1; CCL2 = Chemokine (C-C motif) Ligand 2;
   IL6 = Interleukin 6; PIK3R1 = Phosphoinositide-3-Kinase Regulatory
   Subunit 1. The p-values of IRS2, FOXO1, ERK2, and SERPPINF1 are 0.001,
   0.001, 0,002, and 0,014, respectively.

Discussion

   Recently, there has been considerable interest in comprehending the
   diverse molecular mechanisms behind DR to enhance the effectiveness of
   therapeutic strategies. Anti-VEGF therapy is a reliable and responsive
   treatment for proliferative diabetic retinopathy (PDR) and can also
   reduce the severity of DR in non-proliferative diabetic retinopathy
   (NPDR) patients. Despite its efficacy, reports show that DR lesions
   often recur quickly and underlying retinal ischemia also remains
   unchanged. Although anti-VEGF therapy is considered the main treatment
   option for DR, recent studies focus on the different pathways involved
   in the development of DR in the retina. Identifying the components of
   these pathways and designing effective drugs can provide new
   therapeutic options, whether used alone or in combination with
   anti-VEGF drugs, may enhance the effectiveness of treatments.
   Therefore, we need to target new pathophysiological pathways and
   stabilize drug effects [[106]38, [107]39]. Regarding the importance of
   DR-associated signaling pathways in developing new pharmaceutical
   agents targeting non-VEGF-driven pathways, the current study aimed to
   identify potent components of the related signaling pathways.

   Based on the results of the KEGG pathway enrichment analysis, MIR-96
   target genes are predominantly enriched in signaling pathways activated
   by insulin (phosphoinositide 3-kinase) and VEGF. Integrating the
   protein-protein interaction (PPI) network of MIR-96 associated genes
   with an in vitro study of MIR-96 overexpression in hRPE cell culture
   has shown that IRS2 and FOXO1 levels were significantly decreased after
   48 hours in pAAV-MIR-96-eGFP-int transfected hRPE cells.

   Insulin signaling requires IRS molecules to regulate glucose metabolism
   in the cells. Among IRS protein family members, IRS1, along with IRS2
   are essential molecules for the insulin signaling pathway. Recent
   evidence revealed that the IRS2 level has been elevated in mouse models
   of DR [[108]18]. Differently, IRS2 expression in macrophages was
   downregulated by hyperinsulinemia which is often associated with type 2
   diabetes [[109]40]. To regulate IRS2 expression, FOXO1 binds to the
   insulin-responsive elements (IRE) of the IRS2 promoter, upregulating
   IRS2 expression and activating the PI3K/AKT pathway in the process.
   There is, however, no clear understanding of how IRS2 dysregulation
   occurs [[110]41, [111]42].

   It has also been demonstrated that insulin regulates VEGF expression in
   several cell types, including epithelial cells, and thereby contributes
   to DR progression. VEGF by itself can damage the tight junctions and
   cause dysfunction in RPE cells [[112]43]. As far as we know, insulin
   upregulates the expression of VEGF mainly through the PI3K/AKT axis.
   Data support the idea that the IRS/PI3K/AKT/VEGF axis could be a
   potential target for the treatment of DR and accordingly, it should be
   further investigated to find new therapeutic targets [[113]44].

   Considering previous research, our findings suggest that the alteration
   of MIR-96 and its effects on IRS/PI3K/AKT/VEGF axis regulation and
   therefore the enhancement of VEGF, act as a facilitator in DR
   progression. According to Ji et al., suppression of IRS2 by MIR-7a
   inhibits PI3K/AKT cascade proteins in retinal pericytes and endothelial
   cells of DR mouse models. They suggested that the upregulation of
   MIR-7a and targeting IRS2 can inhibit VEGF and also the invasion
   capability of retinal cells [[114]18]. In the current study, we found
   that the overexpression of MIR-96, contrary to other in vitro
   experiments [[115]18, [116]45], can eventually result in the
   downregulation of PI3K/AKT cascade proteins, such as IRS2. Although the
   IRS2 expression differs from previous studies, it can nevertheless be
   argued that there is probably an interconnected system including
   epigenetic control, upstream genes, and miRNAs that can play a role in
   the regulation of IRS2 expression in diabetic patients [[117]41,
   [118]42]. In addition, our studies were performed on non-diabetic
   retinal cells in which the MIR-96 was overexpressed. Despite the
   decline of IRS2, the expression of IRS1 remained unaltered which was in
   line with the findings from the study reviewed by Pitale et al.
   [[119]45, [120]46].

   A more recent study by Zolfaghari et al. [[121]9] showed that despite
   the general similarity of retinal pigment epithelial cells in humans
   and mice, there are differences that should be considered for
   developing novel treatments [[122]47]. A comprehensive review of the
   cellular characteristics suggested significant differences associated
   with each species. For example, RPE cell proliferation was reduced in
   mice as a result of MIR-96 overexpression, while it remained unchanged
   in human retinal cells. Nonetheless, the decrease in the expression
   level of FOXO1 in human cells concurs well with the previous study on
   the retina of human donors with DR [[123]9].

   FOXO1, which activates through the insulin-stimulated PI3K pathway and
   contributes to diabetes hyperglycemia, is a direct target of MIR-96
   [[124]48, [125]49]. Moreover, FOXO1 has now been identified as a
   potential DR-specific diagnostic and therapeutic gene, and its aberrant
   expression is associated with the pathogenesis of DR. It has been
   reported that several signaling pathways including MAPK regulate the
   activity of FOXO proteins in response to hyperglycemic conditions in
   diabetes [[126]50, [127]51].

   Among autophagy-related genes identified by bioinformatics analysis of
   mRNA chip of DR samples, ERK2 (MAPK1) was found to be downregulated
   [[128]52]. In the present study, ERK2 gene expression level showed a
   significant reduction in pAAV-MIR-96-eGFP-int transfected hRPE cells.
   Several studies indicate that ERK1/2 plays a critical role in the
   development of DR [[129]53–[130]55]. It has been suggested that in
   addition to the PI3K/AKT signaling pathway, the ERK1/2 (p44/42 MAP
   Kinase) pathway may also contribute to long-term VEGF upregulation. As
   shown in previous studies, ERK1/2 can regulate VEGF expression by
   acting at the VEGF promoter. Furthermore, it has been shown that the
   ERK signaling pathway contributes to VEGF release in the retinas of
   diabetic rats [[131]56].

   Consistent with RT-qPCR results, Gene-disease associations (GDAs)
   provided by the DisGeNET (v7.0) database showed that among the
   DR-related genes, there are genes (such as IRS1, PIK3CA, AKT1, and
   MAPK1) related to the insulin signaling pathway and its downstream
   pathways whose expression has changed. In addition to these genes,
   MIR-183, according to a recent study, was also upregulated
   substantially in the retinas of DR rat models. The upregulation of
   MIR-183 activated the PI3K/AKT/VEGF signaling pathway promoting
   angiogenesis and endothelial cell proliferation [[132]57].

   Several studies have shown that PEDF (SERPINF1) is highly expressed in
   RPE cells, which affects retinal vasculature, and sustains retinal
   function. PEDF is a versatile protein that can inhibit the development
   of DR due to its antioxidant effect [[133]58–[134]61]. It is important
   to note that the balance between the expression of VEGF and PEDF in RPE
   cells has a profound impact on retinal vessels so an increase in VEGF
   and a decrease in PEDF can lead to the development of retinal
   neovascularization [[135]62]. Likewise, studies have linked DR
   pathogenesis to decreased levels of PEDF [[136]63]. In the present
   study, we observed that SERPINF1 expression started to decline over
   time after transfection, despite the elevation in the early hours. As
   the most potent neovascularization inhibitor in the eye, SERPINF1
   reduction can affect the protection of retinal cells, thus the RPE
   becomes defective and unable to digest photoreceptor outer segments,
   allowing the retina to degenerate [[137]64, [138]65].

   It has been shown that PEDF signals through both MAPK/ERK and AKT
   signaling pathways to regulate gene expression [[139]66]. It can also
   provide protection by activating the ERK pathway. Indeed, ERK
   activation has a protective function in the rescuing of RPE cells in an
   oxidative environment [[140]67, [141]68]. In addition, PEDF inhibits
   endothelial cell proliferation by regulating the MAPK/ERK pathway
   [[142]69]. Furthermore, there is evidence that PEDF can activate the
   PI3K/AKT pathway to reduce cytotoxicity in RPE cells that are being
   exposed to oxidative stress [[143]70]. SERPINF1 is also present in the
   GDAs network and is one of the top-scored DR-associated genes.
   Delivering SERPINF1 in order to repress angiogenesis in the retina
   revealed that this approach was effective at inhibiting intravitreal
   neovascularization [[144]71, [145]72].

   Studies have shown that the regulation of the MMP2 gene is closely
   associated with the activation of the PI3K/AKT pathway [[146]73]. Based
   on recent research [[147]74], the PI3KCA/IRS1/2/AKT1/IL6/MMP2 axis
   shown in the PPI network of MIR-96 associated genes ([148]Fig 2)
   contributes to retinal neovascularization by enhancing MMP2 expression
   through activation of PI3K/AKT signaling.

   In summary, this study demonstrated that overexpression of MIR-96
   influences the expression of IRS2, FOXO1, ERK2, and SERPINF1,
   consequently disrupting the PI3K/AKT pathway, which may contribute to
   retinal neovascularization and dysfunction. However, it is believed
   that further research is required to unravel the mechanism by which the
   PI3K/AKT pathway is dysregulated during DR pathogenesis, along with
   other associated mechanisms. The present study also confirms that the
   IRS/PI3K/AKT axis plays a prominent role in DR pathogenesis, which can
   be further investigated as a favorable therapeutic target.

Acknowledgments