Glucose withdrawal triggers disulfidptosis-like cell death in diabetic retinopathy vessels

A new study published in Eye and Vision reports that the cystine/glutamate antiporter SLC7A11 is markedly upregulated in the pathological blood vessels of proliferative diabetic retinopathy (PDR), and that exploiting this expression with localized glucose deprivation can trigger a form of cell death with features consistent with disulfidptosis, attenuating disease progression in a mouse model. The work, led by Bowei Yuan and colleagues with Jing Zhuang as senior author, combines bioinformatic screening, human patient tissue analysis, cell culture experiments, and a preclinical animal model to link a recently identified regulated cell death pathway to the abnormal neovascularization that characterizes advanced diabetic eye disease.

Disulfidptosis is a mode of cell death initiated when intracellular cystine and other disulfide compounds accumulate to excess, promoting disulfide stress, anomalous cross-linking of actin cytoskeleton proteins, and ultimately structural collapse of the cell. The process depends critically on SLC7A11, the transporter that imports cystine into cells. Under normal conditions, this import supports glutathione production and antioxidant defense, and glucose fuels the pentose phosphate pathway, generating the NADPH needed to reduce incoming cystine. When glucose is withdrawn, however, NADPH supply plummets, cystine import continues unchecked, and the resulting disulfide overload becomes lethal. The researchers hypothesized that this same vulnerability might operate in the diabetic retina, where chronic hyperglycemia drives both metabolic inflexibility and pathological vessel growth.

To investigate, the team first mined public sequencing data. Analysis of a bulk RNA sequencing dataset comparing fibrovascular membranes from 20 patients with PDR against three healthy controls identified 9,464 differentially expressed genes, of which 14 were disulfidptosis-related genes with significantly altered expression. Single-cell RNA sequencing of a PDR dataset covering 7,156 retinal cells from five patients revealed nine distinct cell types, and among the candidate genes, SLC7A11 showed the highest and most endothelial-specific expression pattern, pointing to a primary role in vascular pathology rather than in neuronal or immune compartments.

The bioinformatic findings were then validated in human clinical samples. Immunofluorescence staining of PDR-associated fibrovascular membranes showed strong SLC7A11 signal, while the protein was virtually undetectable in epiretinal membranes from nondiabetic controls. Western blotting of vitreous humor samples from 12 patients with PDR and 12 nondiabetic controls confirmed significantly higher levels of soluble SLC7A11 in the PDR group. Double immunofluorescence staining further demonstrated that intense SLC7A11 expression colocalized with CD31-positive vascular endothelial cells within the microvascular structures of the pathological membranes, confirming that the upregulation is concentrated in the endothelium of the abnormal vessels themselves.

In laboratory experiments, human retinal microvascular endothelial cells cultured in high-glucose medium, 30 millimolar glucose for 72 hours, showed significantly elevated protein levels of both SLC7A11 and vascular endothelial growth factor A, alongside enhanced proliferation, migration in wound healing assays, and capillary-like tube formation on Matrigel. When the researchers knocked down SLC7A11 using small interfering RNA, these pro-angiogenic effects were significantly inhibited, with reduced proliferation, slower migration, and fewer nodes, junctions, and meshes in tube formation assays. Conversely, stable lentiviral overexpression of SLC7A11 exacerbated all of these behaviors and upregulated VEGFA. Transcriptomic analysis of the overexpressing cells revealed gene programs enriched in angiogenesis, cell migration, wound healing, and blood vessel morphogenesis, indicating that SLC7A11 drives a coordinated pro-angiogenic program under hyperglycemic conditions.

The pivotal experiment tested whether chronic high-glucose adaptation creates a lethal dependency. When high-glucose-adapted cells, including SLC7A11-overexpressing cells, were switched to a glucose-free medium, they underwent rapid cell death exhibiting classic disulfidptosis features: cell shrinkage and collapse of the F-actin cytoskeleton. The cytotoxic effect was dramatically accelerated in SLC7A11-overexpressing cells, which lost profound viability within 8 hours of glucose withdrawal, significantly faster than control cells. Critically, the ferroptosis inhibitor ferrostatin-1 failed to rescue the dying cells, conferring no survival benefit and leaving the same severe F-actin collapse and rounding morphology. As a positive control, ferrostatin-1 completely reversed cell death induced by the canonical ferroptosis inducer erastin, validating the inhibitor and confirming that the glucose-deprivation-induced death was ferroptosis-independent and consistent with disulfidptosis instead.

Biochemical measurements supported the proposed mechanism. High-glucose-cultured endothelial cells showed significant elevations in intracellular free sulfhydryl content, and both the cultured cells and vitreous humor from patients with diabetic retinopathy exhibited significantly higher NADP+/NADPH ratios than their respective controls. When high-glucose-adapted cells were subjected to glucose deprivation, their redox homeostasis collapsed: the free sulfhydryl pool rapidly depleted, and the NADP+/NADPH ratio surged to levels far exceeding those of controls. Immunofluorescence analysis of patient-derived fibrovascular membranes also correlated SLC7A11 expression with clinical glycemic control, as measured by HbA1c. Membranes from patients with high HbA1c displayed high SLC7A11 expression and an F-actin cytoskeleton organized into peripheral cortical rings forming a stable monolayer, with a highly complex cytoarchitecture quantified by fractal dimension analysis, whereas membranes from patients with normalized HbA1c showed markedly reduced SLC7A11 and a broken-down cytoskeleton of loose skeletal fibers.

To test the therapeutic principle in vivo, the team established a streptozotocin-induced diabetic mouse model in C57BL/6J mice and confirmed that SLC7A11 was markedly upregulated in the pathological retinal vasculature, predominantly localized to the retinal vascular endothelium with a filamentous pattern along the pathological blood vessels. The authors note an important limitation: this model successfully recapitulated early-stage diabetic retinopathy features such as vascular leakage, nonperfusion, and edema, but did not consistently develop the structured fibrovascular membranes characteristic of human PDR, so the findings primarily demonstrate SLC7A11 upregulation and the principle of glucose-deprivation therapy rather than modeling advanced fibrovascular proliferation.

Nevertheless, the therapeutic experiment produced striking results. Diabetic mice received weekly vitreous lavage for four consecutive weeks, in which the vitreous cavity was perfused for 90 seconds with either glucose-free balanced salt solution or glucose-replete solution containing 30 millimolar glucose. Optical coherence tomography quantification showed that the glucose-replete control group exhibited progressive edema, with an average increase in retinal thickness of 172.8 plus or minus 42.0 micrometers, while the glucose-free treatment group showed an average decrease of 24.2 plus or minus 16.2 micrometers. The glucose-deprived group also showed a reduction in proliferative membrane thickness, whereas the control group showed continued thickening. The authors interpret these results as evidence that inducing localized glucose deprivation can exploit the SLC7A11-mediated metabolic vulnerability to attenuate progression of pathological proliferative membranes.

The study’s authors are careful to acknowledge that definitive confirmation of canonical disulfidptosis requires identification of actin-specific interprotein disulfide bonds via specialized biochemical assays, which they recommend for future work; their current morphological, biochemical, and pharmacological rescue data demonstrate features highly consistent with SLC7A11-mediated disulfidptosis. They also note that prior work reported decreased SLC7A11 in the retinas of diabetic mice, a discrepancy they attribute to the cell-type-specific nature of the upregulation, which whole-retina measurements may mask, to differences between early and established disease stages, and to interspecies divergence between human tissue and mouse models. The findings suggest that targeting the metabolic dependency of pathological neovessels, rather than only downstream angiogenic signaling, may offer a new therapeutic paradigm for proliferative diabetic retinopathy, though clinical translation in humans remains to be established.

Subject of Research: The role of SLC7A11-mediated disulfidptosis in proliferative diabetic retinopathy

Article Title: SLC7A11-mediated disulfidptosis underlies the pathogenesis of proliferative diabetic retinopathy

Article References: Yuan, B., Chen, S., Liu, J., Yang, T., Zeng, F., Zhang, R., Xiao, Z., Chen, X., Qiu, J., Wei, C., Zhou, Z., Pan, Q., Lu, Y., & Zhuang, J. (2026). SLC7A11-mediated disulfidptosis underlies the pathogenesis of proliferative diabetic retinopathy. Eye and Vision, 13(1), Article 45. https://doi.org/10.1186/s40662-026-00511-2

Image Credits: AI Generated

DOI: 10.1186/s40662-026-00511-2

Keywords: SLC7A11, disulfidptosis, proliferative diabetic retinopathy, diabetic retinopathy, neovascularization, glucose deprivation, retinal endothelial cells, fibrovascular membranes, NADPH, redox metabolism, vitreous lavage, cell death