Studies identified by database searches
Our systematic search identified a total number of 311 studies. After removing duplicates, the title and abstract of 304 studies were reviewed. Of these, 282 were excluded due to irrelevance to our study or inclusion of non-diabetic subjects. Consequently, 23 articles were selected for full-text screening. Out of these, 13 articles met the eligibility criteria and were included in this systematic review. The remaining 9 articles were excluded due to reasons such as incorrect population (T1D), absence of exosome-related content or unavailability of full-text. The PRISMA flow chart elucidating the study selection process is represented in Fig. 1.
Flowchart depicting the study selection process using the PRISMA guidelines. This figure visually outlines the inclusion and exclusion of articles at different stages of the review
All the studies included were case-control studies with sample sizes ranging from 18 to 226 participants. The sample size for the study by Yu et al. (2021) could not be obtained despite efforts to contact the author. Nine studies reported differentially expressed exosomal miRNAs [24,25,26,27,28,29,30,31,32], three studies focused on altered lncRNAs [15, 33, 34], and one study examined dysregulated circRNAs [35]. RNA extraction samples varied across studies: vitreous humour was used in four studies [15, 26, 27, 33], blood plasma in three studies [7, 28, 32], blood serum in four studies [25, 30, 31, 35] and tear samples in two studies [24, 29].
All studies compared the DR patients with T2DM patients, except for two studies. Kot and Kaczmarek, 2022 compared the PDR with Macular hole [26] and Li et al. 2021 compared the PDR group with the senile cataract group [35]. Table 1 summarises the characteristics of the 13 included studies. Methodological aspects of the study including exosome isolation, RNA isolation and quantification techniques are summarised in Table 2.
Finally, we have also comprehensively synthesised the data from these studies and tabulated the key findings from these studies providing its clinical relevance (Table 3).
Studies characterising exosomal microRNA modulations and their impact on DR progression
Among the included studies, nine focussed on miRNA expression profiles in DR. Wan et al. initiated this exploration by identifying increased serum miR-7 in T2DM and its microvascular complications including DR [30]. The study revealed that circulating miR-7 predominantly existed in an exosome-free form rather than within membrane-bound exosomes which targets multiple components of the mTOR signaling pathway.
Jiang et al. elucidated the angiogenic role of miR-377-3p by showing its negative regulation of VEGF expression [25]. Yu et al. observed that miR-431-5p among the top four differentially expressed miRNAs (miR-431-5p, miR-142-5p, miR-361-3p and miR-181 d-5p) as a significant regulator associated with MAPK, RAP1, and RAS signalling pathways [36]. Their KEGG pathway analysis linked miR-431-5p to processes such as cell adhesion, cAMP signalling, and glucose metabolism, demonstrating the broad impact of miRNA on cellular functions relevant to DR. Overall, the study revealed that MiR-431-5p exhibited higher expression levels in small extracellular vesicles (sEVs) derived from patients with diabetic retinopathy (DR) as opposed to circulating miRNAs in DR patients.
Liu et al. further demonstrated that miR-9-3p among the top five upregulated miRNAs (miR-9-3p, miR-6511b-5p, miR-1285-3p, miR-505-5p, miR-4685-3p) activates VEGFR2 signalling pathway through the S1P1 axis, both in the presence and absence of exogenous VEGF-A [27]. Similar findings of aberrated miRNA expressions were described by Kot and Kaczmarek [26]. They identified 26 differentially expressed miRNAs, including 16 downregulated (miR-125a-5p, miR-125b-5p, miR-204-5p, miR-412-3p, miR-137, miR-361-3p, miR-211-3p, miR-9-3p, miR-30e-3p, miR-375, miR-9-5p, miR-30a-5p, miR-328-3p, miR-345-5p, miR-100-5p, and miR-543-3p) and 10 upregulated ones (miR-21-5p, let-7 g-5p, miR-660-5p, miR-142-3p, miR-19a-3p, miR-142-5p, miR-15a-5p, miR-103a-3p, miR-92a-3p, and miR-16-5p) between PDR and controls. Through experimental validation, they provided strong evidence linking miR-125 family, miR-204-5p, miR-21-5p, miR-41-3p, and let-7 g-5p 2 to mechanisms driving fibrovascular membrane development in PDR, affecting processes like angiogenesis, inflammation, and tissue remodelling. Notably, miR-21-5p upregulation, and the miR-125 family downregulation played a significant role in altering the TGF-β and VEGF signalling pathways. These studies emphasise the role of miRNAs in modulating angiogenic factors via VEGF signalling pathway which is pivotal in DR progression.
Another study conducted by Santovito et al., identified 12 upregulated miRNAs (let-7a-5p, miR-16-5p, miR-23a-3p, miR-25a-3p, miR-27a-3p, miR-92a-3p, miR-150-5p, miR-197-3p, miR-223-3p, miR-320a-3p, miR-320b, miR-486-5p) and 2 down regulated miRNAs (miR-346 and miR-495-3p) between DR and T2DM [28]. Further experimental validation downstream highlighted upregulation of miR-23a-3p, miR-25-3p and miR-320b, alongside downregulation of miR-495-3p leading to angiogenesis through NOTCH1 and β-Catenin pathways. Hu et al. identified several dysregulated miRNAs, including upregulated miR-145-5p, miR-214-3p, miR-9-5p, and miR-218-5p, as well as downregulated miR-146a-5p, miR-31-5p, and miR-96-5p [24]. KEGG pathway analysis connected these miRNAs to energy metabolism, insulin secretion, and resistance, including ErbB signaling, highlighting their widespread impact on metabolic and signaling pathways in DR.
In a study by Yang et al. among 18 differentially expressed miRNAs, four miRNAs (miRNA-3976, PC-5p-39533, and PC-3p-37421) showed a larger fold change [31]. The study highlighted in particular that, miRNA-3976 overexpression inversely correlates with NFκB, suggesting a protective role in regulating retinal ganglial cells (RGC) proliferation and apoptosis, adding another layer to the miRNA-mediated regulation of cell survival in DR. Torimura et al. in a pilot study identified higher expression of miR-151-5p in AMD and miR-422a in DME, although their specific pathways remain unspecified, indicating potential new areas for further research [29].
Though observations from the above investigated studies signifies the potential role of exosomal miRNAs in DR pathogenesis, there is no consistent pattern of dysregulation of specific miRNAs across all the studies.
Studies characterising exosomal lncRNA modulations and their impact on DR progression
Three of the included studies investigated the role of exosomal lncRNA in DR. Ye et al. identified exosomal Distal-less homeobox 6 antisense 1 (DLX6-AS1), and Aminoacylase-1 (ACY1), and Rho GTPase activating protein (ARHGAP), which was overexpressed and regulating the p38–MAPK pathway [34]. They also found Psoriasis-susceptibility-related RNA gene induced by stress (PRINS) and Family with Sequence similarity 190, Member A3 (FAM190A-3), which was modulating the TGF-β signalling pathway to be underexpressed in DR.
In another study, Hu et al. were the first to report that lncRNA, LOC100132249 in vitreous samples from PDR patients, mediated endothelial dysfunction through the Wnt/β-Catenin signaling pathway [15]. The study revealed the crucial role of LOC100132249/miR-199a-5p/SNAI1 axis in endothelial-to-mesenchymal transition causing endothelial dysfunction. Lastly, Li et al. focused on the role of exosomal lncRNA-MIAT in DR, demonstrating that it modulates the MMP-X1 expression pathway by sponging miR-133a-3p [33]. This regulation influences tube formation, migration, and proliferation, thereby affecting retinal neovascularization. Their findings highlight how lncRNA-MIAT contributes to angiogenesis and the formation of new blood vessels in the retina, a critical aspect of DR progression. Together, these studies illustrate a complex interplay of ncRNAs-DLX6-AS1, PRINS, LOC100132249, miR-199a-5p, and lncRNA-MIAT that regulate various signalling pathways such as TGF-β, Wnt/β-Catenin, and VEGF.
Studies exhibiting variations in exosomal circRNA profiles: insights into mechanisms and applications
Only one among the included studies examined circRNAs expression pattern in DR. circRNAs are widely expressed in mammalian serum and serve as a sponge for miRNAs to regulate gene expression [36, 37]. Li et al. 2021 identified 26 exosomal circRNAs in serum specimens [35]. Further they constructed a Competing endogenous RNA (ceRNA) network involving circRNA–miRNA–mRNA exosomal circRNA from the patients with DR. They observed that circFndc3b and circFAM13B are upregulated in DR. These circRNAs regulate PI3K-AKT, RAS, and MAPK signalling pathways, promoting angiogenesis and reducing cardiomyocyte apoptosis and fibrosis.
Abnormal angiogenesis, a benchmark in the development of DR, can disrupt the blood-retinal barrier and retinal microenvironment [38]. From all the above investigated studies, we have strong evidence that exosomal ncRNAs play an important role in angiogenesis and are expressed through the VEGF signaling pathway. Its effects are executed through a cascade of multiple signaling pathways such as mTOR, MAPK, Wnt/β-catenin, NF-κB, TGF-β signaling pathways, which are involved in the angiogenesis, cell growth and proliferation.
Together, these studies depict a comprehensive network of exosomal ncRNAs intricately involved in various signaling pathways, elucidating their critical roles in the pathogenesis of diabetic retinopathy.
Studies with differentially expressed miRNAs across the studies
To identify potential common ncRNAs among the studies, we conducted a systematic comparison of the differentially expressed ncRNAs reported in each study. Across the 13 included studies, a total of 35 differentially expressed ncRNAs were identified. These were categorized based on their upregulation or downregulation compared to control groups, as summarized in Table 3.
Despite this comprehensive comparison, no single ncRNA was found to be consistently differentially expressed across multiple studies. This lack of commonly expressed ncRNAs across studies highlights the variability due to the potential differences in experimental conditions, methodologies, or samples used among the studies. These inconsistencies underscore the need for further research as there are no specific ncRNAs that are universally differentially expressed to identify strong biomarkers for diabetic retinopathy.
Risk of bias assessment for the included studies
The evaluation of potential bias was carried out for the thirteen included case-control studies using the NOS tool (Table 4). The overall quality of the included studies could be described as having a high risk of bias based on NOS assessment. All studies were showing high risk of bias owing to poor study design and lack of sample size justification and ascertainment of exposure, with the scores ranging from 2 to 6. A relatively low risk was observed for the questions ‘Representativeness of the cases’ (10/13) and ‘Comparability between cases and controls’ (10/13). While, a high risk of bias was observed for most of the studies in terms of inadequate case definition (7/13), Selection bias (3/13), Definition of Controls (4/13), Ascertainment of exposure (2/13), Same method of ascertainment for cases and controls (1/13). The results of the ROB assessment is presented in the Table 4.
Leave a Reply