Abstract
Background/Aim: Diabetic retinopathy (DR) is a common comorbidity of diabetes involving the formation of abnormal vascular structures in the retina. Tissue inhibitor of metalloproteinases 2 (TIMP2), initially identified as a key mediator of extracellular matrix turnover, is pivotal for inflammatory processes and tissue homeostasis. The current study examined the influence of TIMP2 gene variations on the risk for DR.
Materials and Methods: We investigated the association of TIMP2 gene variations with DR by analyzing four single-nucleotide polymorphisms (SNPs) of the TIMP2 gene (rs16971783, rs2889529, rs7220980, and rs8068674) in a cohort of 672 patients with DR and 919 diabetic controls with normal ophthalmoscopic findings.
Results: Our results showed that rs16971783 of TIMP2 gene was associated with a higher risk for DR (TA vs. TT, AOR=1.445, p=0.028; TA+AA vs. TT, AOR=1.179, p=0.046). We further demonstrated that the association of rs16971783 with DR was exclusively observed in diabetic individuals with proliferative DR (TA vs. TT, AOR=1.827, p=0.035; TA+AA vs. TT, AOR=1.351, p=0.027), whereas not detected among those who suffered from non-proliferative DR. In addition, preliminary exploration of gene expression data from public resources reveald that rs16971783 regulated TIMP2 gene expression in various human tissues.
Conclusion: Allele-specific expression of TIMP2 gene might contribute to the progression of DR.
Introduction
Diabetic retinopathy (DR), a major form of microvascular disease linked to prolonged diabetic conditions, is the primary culprit behind blindness among working-age individuals (1, 2). From a clinical perspective, DR can be classified into two main categories: non-proliferative and proliferative DR. The non-proliferative type represents the initial phase of disease, exhibiting clinical manifestations, such as hemorrhages, microaneurysms, and hard lipid exudates, and could deteriorate into the proliferative form, a more advanced condition typified by the growth of newly developed capillary structures, subsequently resulting in tractional retinal detachment (3). A number of etiologic factors have been identified as contributors to vascular abnormalities observed in DR, among which persistent inflammatory (4) and hyperglycemic conditions (5) are two well-established risks for DR. However, lately, attention has been drawn to the impact of pro-angiogenic factors and their cognate receptors on DR etiology (6). These parameters and the period of diabetics are thought to be intertwined and to some extent account for the observed heterogeneity in disease presentation.
A substantial body of evidence has established a clear link of genetic predisposition to the occurrence of both diabetes and its associated complications (7). In terms of the inheritance of DR, a multitude of genetic factors have been uncovered in association with its diverse phenotypes (8). These genetic components influence several molecular mechanisms that contribute to the pathogenesis of DR, such as promotion in polyol pathways, augmentation of non-enzymatic glycation, production of reactive oxygen species (ROS), and stimulation of protein kinase C (9).
Tissue inhibitors of metalloproteinases (TIMPs) constitute a conserved family of polypeptides that can form noncovalent complexes with matrix metalloproteinases (MMPs) and inhibit their activation or activity (10). As a key mediator of extracellular matrix (ECM) turnover, TIMPs have been recognized to play a crucial role in tissue homeostasis (11). Yet, these proteases demonstrate only a fraction of their cellular effects via metalloproteinase regulation. Even though commonly overlooked, noncanonical TIMP functions that are independent of MMP inhibition have been reported, particularly on the promotion of inflammatory processes and improvement of cognitive functions (12-14). Among four members of this protein family, TIMP2 is the most abundantly expressed. In addition to its eponymous function as a tumor suppressor through inhibition of metalloprotease activity, TIMP2 directly interacts with cellular receptors and matrisome elements to orchestrate cell signaling pathways, highlighting its therapeutic potential (11). Recently, elevation of TIMP2 levels was found to be associated with the severity of diabetic retinopathy (15).
To date, the relationship between TIMP2 gene polymorphisms and the development of retinopathy in patients with diabetes remains elusive, as an impact of TIMP2 variations on the predisposition to neovascular retinopathy in premature infants has been reported (16). Here, explored whether TIMP2 gene variants influence the risk for DR.
Materials and Methods
Subjects. In this targeted-gene survey, 672 DR cases were recruited in Chung Shan Medical University Hospital, Taichung, Taiwan. Disease diagnosis was determined as the presence of any of the following clinical manifestations: flame, dot-blot, or boat hemorrhages; microaneurysm; venous beading; cotton-wool spots; hard exudate; or intraretinal microvascular abnormalities. To assess the disease deterioration, patients with DR were classified into two groups, an initial phase (non-proliferative form, n=527) and an advanced stage (proliferative form, n=145), on the basis of any symptom of disease progression, such as vitreous hemorrhage, retinal neovascularization, neovascular glaucoma, and tractional retinal detachment. Besides, 919 patients with diabetes with normal ophthalmoscopic findings were recruited for comparisons. The study was approved by the institutional review board (CS1-20048), and informed written consent was obtained from all subjects while enrolled. Demographic and laboratory data concerning age, sex, kidney function, and indications of diabetes and hyperlipoproteinemia were collected, and 5 mL of whole blood was withdrawn from each participant.
Single nucleotide polymorphism analysis. Four single nucleotide polymorphisms (SNPs) of TIMP2 gene (rs16971783, rs2889529, rs7220980, and rs8068674) were selected on the basis of their potential for disease predisposition (16-19) and explored in this survey. Genotypic distribution was determined as described previously (20). In brief, DNA extraction was conducted by using QIAamp DNA Blood Mini kit (Qiagen, Valencia, CA, USA). Biallelic variants of four SNPs were solved by using the TaqMan assay (Applied Biosystems, Foster City, CA, USA).
Statistical analysis. Hardy-Weinberg equilibrium for four chosen TIMP2 gene variants was evaluated using the χ2 goodness-of-fit method. Comparison of demographic and laboratory results between cases and controls was conducted with the Mann-Whitney U test. Correlations of gene polymorphisms with the onset and progression of DR were assessed by multiple logistic regression analyses joined with the adjustment for putative confounding parameters. Differences in TIMP2 gene expression among genotypic groups from the Genotype-Tissue Expression (GTEx) database (21) were analyzed with one-way ANOVA. A p-value of <0.05 was considered statistically significant.
Results
Subject characteristics. To clarify a possible impact of TIMP2 gene polymorphisms on the development of DR, 672 patients with DR were enrolled and compared with 919 retinopathy-free controls with matched diabetic conditions. Clinical and demographic features of two cohorts were assessed (Table I). No difference in age and sex was observed between the two groups. In addition to enhanced signs of renal impairment (increased serum creatinine and reduced glomerular filtration rate), the DR group exhibited a higher level HbA1c and a longer duration of diabetic conditions in comparison with the control group. Moreover, matched status of hyperlipidemia, as indicated by levels of total cholesterol, LDL-cholesterol, and triglycerides, as well as ratio of total cholesterol to LDL-cholesterol, was detected between the cases and controls.
Clinical and laboratory characteristics of patients with diabetes with and without retinopathy.
Association between TIMP2 SNPs and DR. To examine whether TIMP2 gene polymorphisms are associated with the risk of developing DR, we genotyped four SNPs of the TIMP2 gene, including rs16971783, rs2889529, rs7220980, and rs8068674. The distributions of different genotypic groups for individual polymorphisms between DR cases and retinopathy-free subjects with diabetes were evaluated. For all TIMP2 gene variants tested, no deviation (p>0.05) from Hardy-Weinberg equilibrium was detected in any of the groups. Analysis of genotype distribution revealed that diabetic heterozygotes for rs16971783 of TIMP2 gene [TA; AOR=1.445; 95% confidence interval (CI)=1.040-2.009; p=0.028] were more commonly associated with the development of retinal complications (Table II), as compared to individuals with diabetes who were homozygous for the major allele (T) of rs16971783 (TT). Furthermore, patients with diabetes who carried at least one minor allele (A) of rs16971783 (TA and AA; AOR=1.179; 95%CI=1.003-1.386; p=0.046) had a higher risk for DR than did those who were homozygous for the major allele (TT). However, no association of the other three SNPs with the development of DR was detected (Table II). These findings suggest that TIMP2 rs16971783 alleles influence the susceptibility to retinal conditions in individuals with diabetes.
Odds ratio (OR) and 95% confidence interval (CI) of diabetic retinopathy associated with TIMP2 genotypic frequencies.
rs16971783 genotypes show a severity-specific effect on the development of DR. Since rs16971783 genotypes were associated with DR, we also tested whether TIMP2 gene polymorphisms affect the disease progression of DR. Our stratification analysis demonstrated that rs16971783 was associated with a higher risk for proliferative DR (TA vs. TT, AOR=1.827, p=0.035; TA+AA vs. TT, AOR=1.351, p=0.027) but not with non-proliferative DR (Table III, Table IV). These data indicate a correlation of TIMP2 gene variations with the severity of DR.
Odds ratio (OR) and 95% confidence interval (CI) of non-proliferative diabetic retinopathy associated with TIMP2 genotypic frequencies.
Odds ratio (OR) and 95% confidence interval (CI) of proliferative diabetic retinopathy associated with TIMP2 genotypic frequencies.
Functional insight of rs16971783 into DR. As rs16971783 is an intronic SNP in the TIMP2 gene, we also examined the putative function of this DR-associated allele through public resource data analysis. We observed alterations of TIMP2 gene expression in the lung, pancreas, and spleen tissues among donors with three rs16971783 genotypes in the GTEx database (Figure 1). These results suggest that changes of TIMP2 levels in an allele- and tissue-specific manner may contribute to the disease progression of DR.
Influence of rs16971783 genotypes on TIMP2 expression. Comparisons of TIMP2 expression among distinct genotypic groups in representative normal tissues based on data from the GTEx portal. p-Values were calculated among groups using one-way ANOVA.
Discussion
A considerable amount of research has highlighted that the development of DR is mediated by a synergy of heritable and acquired factors. In this investigation, we employed a targeted-gene strategy to show the correlation between TIMP2 variations (rs16971783) and the development of DR. Furthermore, rs16971783 was found to be associated with the advanced form of disease (proliferative DR) but not with the less severe early precursor type (non-proliferative DR), unveiling a role of TIMP2 gene polymorphisms in deteriorating the growth of abnormal vasculatures in the retina of subjects with diabetes.
In addition to being a key mediator of ECM turnover through the inhibition of MMP activity, TIMP2 exhibited an MMP-independent effect on inhibiting the mitogenic growth factors, thus blocking angiogenesis (22). It was demonstrated that binding of TIMP2 to α3β1 integrin on the cell surface of human microvascular endothelial cells can initiate a signaling cascade that results in enhanced Shp-1 (a protein tyrosine phosphatase) activity associated with angiogenic growth factors, FGF-2 and VEGF-A (23). The treatment of human microvascular endothelial cells with TIMP-2 decreased binding of Shp-1 with β1-integrin subunits and increased binding of Shp-1 with angiogenic growth factor receptors (VEGFR-2 and FGFR-1), leading to a concomitant decrease in the activation and phosphorylation of either VEGFR-2 or FGFR-1. Recently, elevation of TIMP2 levels was found to correlate with inner retinal vascularization (24) and the severity of diabetic retinopathy (15). These findings concerning TIMP-2-mediated mitogenic suppression of endothelial cell growth suggest possible new strategies for the development of antiangiogenic therapies against retinal complications in subjects with diabetes.
It has been demonstrated that many alleles of TIMP2 gene have configured intricate patterns of disease susceptibility. Across the relevant literature, rs8179090 (−418G/C), localized at the promoter region of TIMP2 gene, represents the most extensively studied polymorphism. rs8179090 was found to be associated with many diseases, such as chronic obstructive pulmonary disease (25), acne vulgaris (26), aortic aneurysms (27), breast cancer (28), oral cancer (29), and thoracic idiopathic scoliosis (18). Moreover, two intronic SNPs, rs796391657 and rs4789936, were found in association with primary open-angle glaucoma (30) and breast cancer (31), respectively. In addition to SNPs in non-coding regions, the polymorphic allele (T) of rs2277698, resulting in a synonymous amino acid change at codon position 101 (Ser101Ser), was linked to a lower risk for breast cancer (32). Here, we detected an association of another intronic SNP of TIMP2 gene, rs16971783, with proliferative DR. rs16971783 is located at the third intron of TIMP2 gene and, for the first time, reported to contribute to the susceptibility to retinal complications in patients with diabetes. These findings collectively indicate that the different disease-associated SNPs of TIMP2 have a cell type- or tissue-specific effect of on the pathogenesis of human diseases.
Besides the replacement of an amino acid residue that may alter the conformation, function, or binding affinity of the protein molecules, disease-associated gene polymorphisms can also affect gene expression via changes in splicing, non-coding RNA expression, and RNA stability (33-35). The presence of polymorphic alleles of rs8179090 (−418G/C), a SNP located in the promoter region of TIMP2 gene, has been shown to change the consensus sequence of Sp1 transcription factor binding site, resulting in lower TIMP2 gene expression and reduced cancer risk (28). Similarly, rs8179096, another SNP in the proximity of rs8179090 (−418G/C), exhibited distinct allele-specific effects on TIMP2 gene transcription through interfering with the binding capacity of NF-κB to the promoter (36). The functional roles of SNPs that do not result in protein-coding changes in disease susceptibility remain largely unknown. The risk allele of rs4789936 was associated with elevated TIMP2 expression (32), indicating that this polymorphism likely resides in regulatory regions of gene expression. In our study, alterations of TIMP2 gene expression in many human tissues were observed among different rs16971783 genotypic groups. Our results, together with the findings from others, suggest that changes in TIMP2 levels due to gene polymorphisms contribute to the susceptibility to DR in diabetic patients.
Study limitations. One potential issue is that a great degree of heterogeneity in disease comorbidities (e.g., diabetic neuropathy, diabetic cardiomyopathy, and diabetic nephropathy) among our cases and controls may produce different findings regarding the role of TIMP2 in the genetic predisposition to DR. Furthermore, we did not conduct functional experiments to dissect whether different rs16971783 genotypes alter the levels of TIMP2 in retinal vasculature. In addition, the genetic susceptibility detected here might be restricted to specific ethnic populations included in this study.
In conclusion, we demonstrated an association of TIMP2 gene variants with the progression of DR. This genetic predisposition likely links allele-specific expression of TIMP2 to the deterioration of retinal abnormalities in individuals with diabetes.
Acknowledgements
The Authors are grateful to the Human Biobank of Chung Shan Medical University Hospital, Taichung, Taiwan for sample preparation.
Footnotes
Authors’ Contributions
CTH: Conceptualization, data interpretation, writing and revision of the manuscript; HWC: data interpretation, writing and revision of the manuscript; KW: data interpretation, writing and revision of the manuscript; SCS: data interpretation, methodology, writing the manuscript; CHT: data interpretation, methodology; SFY: Conceptualization, data interpretation, statistical analysis, writing and revision of the manuscript. MYL: Conceptualization, data interpretation, statistical analysis, writing and revision of the manuscript. All Authors have approved the final version of the manuscript.
Conflicts of Interest
The Authors declare no competing interests in relation to this study.
Artificial Intelligence (AI) Disclosure
No artificial intelligence (AI) tools, including large language models or machine learning software, were used in the preparation, analysis, or presentation of this manuscript.
- Received February 13, 2026.
- Revision received March 14, 2026.
- Accepted March 16, 2026.
- Copyright © 2026 The Author(s). Published by the International Institute of Anticancer Research.
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
References
Jump to section
Related Articles
Cited By...
- No citing articles found.