Peptide Therapies for Diabetic Eye Disease

Peptide Therapies for Diabetic Eye Disease

173,846 patients

In the largest comparative study of tirzepatide versus lifestyle intervention, tirzepatide reduced the risk of diabetic macular edema by 37.6% over 12 months.

Shah et al., Ophthalmology, 2026

Shah et al., Ophthalmology, 2026

View as image

Diabetic retinopathy affects roughly one-third of people with diabetes and remains the leading cause of vision loss in working-age adults globally. Current treatments center on anti-VEGF injections and laser photocoagulation, both targeting vascular complications after they have already developed. The neuronal component of diabetic eye disease, where retinal ganglion cells and photoreceptors degenerate before visible vascular changes appear, receives no approved therapy. Peptides are emerging across multiple pathways in this space: GLP-1 receptor agonists already prescribed to hundreds of millions of diabetes patients may indirectly affect retinal health; neuroprotective peptides like PACAP directly shield retinal neurons in animal models; PEDF-derived peptide eye drops reduce vascular leakage by roughly 60% in diabetic mice; somatostatin analogs target retinal neurovasculature; integrin peptide therapy blocks the upstream production of VEGF; and peptide-drug conjugates are enabling eye drop delivery of anti-VEGF molecules that currently require intravitreal injection.^[1] This article maps the full peptide pipeline for diabetic eye disease, from approved drugs with emerging retinal data to preclinical candidates. For deeper coverage of specific subtopics, see our articles on GLP-1 agonists and retinal health and somatostatin analogs for retinal neuroprotection.

The Diabetic Retinopathy Problem

Diabetic retinopathy progresses through stages defined by vascular changes: microaneurysms, hemorrhages, hard exudates (nonproliferative), then new blood vessel growth (proliferative). Diabetic macular edema (DME), involving fluid leakage into the central retina, can occur at any stage and is the most common cause of vision loss. Anti-VEGF injections (ranibizumab, aflibercept, bevacizumab) are the standard of care for DME and proliferative disease, requiring monthly or bimonthly intravitreal injections that many patients find burdensome.

What this framework misses is the neurodegeneration that begins before vascular disease becomes clinically apparent. Retinal ganglion cells, dopaminergic amacrine cells, and photoreceptors deteriorate in the diabetic retina through mechanisms including glutamate excitotoxicity, oxidative stress, and inflammatory signaling. By the time a patient presents with visible retinopathy, substantial neuronal loss may have already occurred. No approved therapy addresses this neuronal component.

Peptides are relevant to diabetic eye disease across this entire spectrum: metabolic peptides (GLP-1 agonists, tirzepatide) that improve the systemic environment driving retinopathy; neuroprotective peptides (PACAP, somatostatin, PEDF) that directly shield retinal neurons; anti-angiogenic peptides (integrin peptide therapy) that target vascular proliferation; endogenous peptides (C-peptide, GHRH agonists) with direct retinal effects; and peptide-drug conjugates that solve the delivery problem for existing anti-VEGF therapies.

GLP-1 Receptor Agonists: The Largest Dataset

GLP-1 receptor agonists (semaglutide, liraglutide, dulaglutide, exenatide) are now among the most prescribed peptide drugs worldwide, used for type 2 diabetes and obesity. Their potential effects on diabetic retinopathy represent the largest real-world dataset for any peptide-eye disease interaction.

The SUSTAIN-6 Signal

The controversy began with SUSTAIN-6, a cardiovascular outcomes trial of semaglutide in type 2 diabetes. The trial reported a hazard ratio of 1.76 for retinopathy complications (vitreous hemorrhage, blindness, need for intravitreal treatment or photocoagulation) in the semaglutide group compared to placebo. This finding generated alarm and prompted a wave of follow-up investigations.

Subsequent analysis attributed the signal primarily to rapid glycemic improvement rather than direct retinal toxicity. Rapid reduction in HbA1c is a well-established risk factor for early worsening of diabetic retinopathy, first documented with insulin initiation decades ago. Patients in SUSTAIN-6 with the most severe baseline retinopathy and the largest HbA1c drops showed the highest risk. Varughese and colleagues noted in 2024 that this phenomenon demands enhanced screening protocols for patients starting any glucose-lowering therapy that produces rapid improvements.^[2]

Large-Scale Safety Data

Barkmeier and colleagues published the largest comparative safety analysis in 2026, examining over 150,000 patients prescribed GLP-1 receptor agonists during routine clinical practice. They found no difference in the risk of sight-threatening diabetic retinopathy between semaglutide, dulaglutide, liraglutide, and exenatide: semaglutide versus dulaglutide showed a hazard ratio of 0.88 (95% CI 0.70-1.11) for composite treatment of DME or proliferative disease.^[3]

A systematic review and meta-analysis by Alwafi and colleagues in 2025, covering 39 studies (24 RCTs and 15 observational studies), found that GLP-1 receptor agonists were not significantly associated with diabetic retinopathy risk (pooled RR 1.00, 95% CI 0.71-1.43). After excluding studies with high risk of bias, the estimate remained nonsignificant (RR 1.06, 95% CI 0.67-1.67).^[4]

Nadeem and colleagues provided context in 2025, noting that while the retinopathy signal has been extensively debated, clinical recommendations have converged on obtaining baseline eye exams and closer follow-up in the first 12-18 months of GLP-1 therapy, particularly for patients with severe baseline retinopathy.^[5] For our full analysis of this evidence, see GLP-1 agonists and retinal health: help or harm?

Direct Retinal Protective Effects

Beyond the safety question, emerging preclinical evidence suggests GLP-1 agonists may actively protect retinal tissue. Liu and colleagues demonstrated in 2025 that exendin-4, a GLP-1 receptor agonist, suppressed diabetic retinopathy progression in both in vivo and in vitro models by regulating TGFB2 expression levels, reducing vascular leakage and retinal cell death.^[6]

Cheng and colleagues showed in 2025 that semaglutide (100 micrograms/kg/week for 5 weeks) increased retinal inner nuclear layer thickness, alleviated neuronal apoptosis, restored ZO-1 expression in retinal microvasculature, and downregulated VEGFA expression in high-glucose-treated human retinal microvascular endothelial cells. These findings suggest the peptide has direct retinal protective effects independent of systemic glycemic control.^[7]

Gong and colleagues published a comprehensive review in 2026 covering the mechanisms by which GLP-1 receptor agonists may affect ocular disease, including anti-inflammatory effects, anti-oxidative stress pathways, and modulation of retinal vascular permeability. The GLP-1 receptor is expressed in retinal tissue, meaning these drugs could exert local effects beyond their systemic metabolic benefits.^[1]

Tirzepatide and Multi-Agonists

Tirzepatide, a dual GIP/GLP-1 receptor agonist, represents the next generation of incretin-based therapy. Shah and colleagues examined 173,846 patients (86,923 per cohort) and found tirzepatide was associated with reduced 12-month risk of incident mild NPDR (RR 0.864), proliferative DR (RR 0.705), DR with macular edema (RR 0.624, 95% CI 0.536-0.727), vitreous hemorrhage (RR 0.607), and tractional retinal detachment (RR 0.370) compared to lifestyle intervention alone.^[8] Whether the additional GIP receptor component contributes to retinal protection or simply reflects better metabolic control is unknown. For more on how GLP-1 drugs affect other organ systems, see GLP-1 drugs and heart disease and autoimmune diabetes and GLP-1.

Neuroprotective Peptides

PACAP

Pituitary adenylate cyclase-activating polypeptide (PACAP) is the most extensively studied neuroprotective peptide for diabetic retinopathy in preclinical models. Szabadfi and colleagues demonstrated that three intravitreal injections of 100 pmol PACAP over one week preserved cone photoreceptors, dopaminergic amacrine cells, and ganglion cell layer neurons in streptozotocin-induced diabetic rats. PACAP treatment also increased PAC1 receptor expression and tyrosine hydroxylase levels, suggesting active upregulation of protective signaling.^[9]

PACAP addresses the neuronal component of diabetic retinopathy that GLP-1 agonists and anti-VEGF therapies miss. It inhibits apoptosis through PAC1 receptor activation of cAMP/PKA signaling, counteracts glutamate excitotoxicity, modulates microglial inflammation, and protects retinal pigment epithelial cells from oxidative stress. The limitation is that all evidence comes from animal models with no human retinal trials conducted. For the full evidence review, see PACAP: the neuropeptide protecting against retinal damage.

PEDF-Derived Peptide Eye Drops

Pigment epithelium-derived factor (PEDF) is a multifunctional protein with neuroprotective, antiangiogenic, and anti-inflammatory properties. Liu and colleagues tested two PEDF-derived peptides as eye drops in Ins2(Akita) diabetic mice: the antiangiogenic PEDF60-77 (P60) and neuroprotective PEDF78-121 (P78), applied once weekly for 15 weeks. Both peptides penetrated the cornea within 1-4 hours and reached the ciliary body, retinal pigment epithelium-choroid complex, retinal microvasculature, and vitreous. Both reduced vascular leakage by approximately 60% and restored tight junction proteins ZO-1 and occludin to nondiabetic levels. P78 also prevented diabetes-induced microglia activation by approximately 60%, retinal ganglion cell death by approximately 22%, and inner plexiform layer thinning by approximately 13%. P78 reduced levels of 9 of 20 cytokines in diabetic vitreous including IFN-gamma, IL-6, IL-3, and TNF-alpha.^[10]

The PEDF eye drop data are striking because they demonstrate topical delivery of a neuroprotective peptide to the retina, a route that would eliminate injection burden. Peak vitreous levels reached 0.2 micrograms/mL for P60 and 0.9 micrograms/mL for P78. The dual neuroprotective and antiangiogenic activity of the PEDF peptide fragments makes them relevant to both early (neuronal) and later (vascular) stages of diabetic retinopathy. No human trials have been conducted.

SOCS1-Derived Peptide Eye Drops

Suppressors of cytokine signaling (SOCS) proteins regulate inflammatory responses, and SOCS1 is downregulated in diabetic retinas. Hernandez and colleagues tested a SOCS1-derived peptide administered as eye drops twice daily for 2 weeks in db/db diabetic mice. The peptide significantly reduced glial activation, neural apoptosis, and retinal levels of proinflammatory cytokines. Electroretinogram parameters improved, confirming the histological findings translated to functional retinal improvement. The SOCS1 peptide also prevented disruption of the blood-retinal barrier, addressing vascular leakage through an anti-inflammatory rather than anti-VEGF mechanism.^[11]

C-Peptide Supplementation

C-peptide, the 31-amino-acid byproduct of insulin processing that is absent in type 1 diabetes and reduced in late type 2 diabetes, has direct retinal protective effects. Lee and colleagues demonstrated in 2023 that systemic C-peptide supplementation via osmotic pumps ameliorated retinal neurodegeneration in diabetic mice by inhibiting VEGF-induced pathological events. C-peptide reduced hyperglycemia-induced activation of macroglial and microglial cells, prevented downregulation of glutamate aspartate transporter 1, and inhibited neuronal apoptosis through a mechanism involving reactive oxygen species generation.^[12]

The C-peptide findings are notable because they demonstrate a systemic peptide replacement strategy for retinal neuroprotection, distinct from the topical approaches of PEDF and SOCS1. C-peptide supplementation would be most relevant to type 1 diabetes patients who completely lack the peptide and to late-stage type 2 diabetes patients with severely diminished beta-cell function.

GHRH Agonists

Growth hormone-releasing hormone (GHRH) agonists represent an unexpected entry in the retinal protection pipeline. Thounaojam and colleagues published in PNAS in 2017 that GHRH and its receptor are significantly downregulated in both diabetic rat retinas and human diabetic retinas (postmortem). Treatment with the GHRH agonist MR-409 preserved retinal ganglion cell survival, exerted antioxidant effects by upregulating NRF-2-dependent gene expression, reduced proinflammatory cytokines and adhesion molecules, downregulated VEGF, and increased PEDF expression. The GHRH antagonist MIA-602 worsened retinal morphology, confirming the pathway specificity of the effect.^[13]

Somatostatin Analogs

Somatostatin, a 14- or 28-amino-acid polypeptide produced by neuroendocrine cells, has anti-angiogenic and neuroprotective properties relevant to diabetic retinopathy. Its receptors (five subtypes, SST1-5) are expressed in the retina, where they contribute to neuromodulation and vascular regulation. Diabetic retinas show decreased somatostatin expression, suggesting a loss of endogenous protection.

The EUROCONDOR trial, a European multicenter phase II-III study, randomized 449 patients with early diabetic retinopathy to topical somatostatin 0.1%, brimonidine tartrate 0.2%, or placebo eye drops twice daily for 96 weeks. In the whole population, no neuroprotective effect was found. In the 34.7% of patients with preexisting retinal neurodysfunction, however, implicit time worsened in the placebo group but remained unchanged in both treatment groups. This finding points to screening retinal neurodysfunction as critical for identifying patients who may benefit from neuroprotective peptide therapy.

The delivery challenge remains: systemic octreotide affects too many hormonal pathways for chronic use; intravitreal delivery adds injection burden. Research into sustained-release nanoparticle formulations is ongoing. For detailed coverage, see somatostatin analogs for retinal neuroprotection in diabetes.

Anti-Angiogenic Peptide Approaches

Integrin Peptide Therapy

ALG-1001 (risuteganib/Luminate) is the first of a class called integrin peptide therapy. Unlike anti-VEGF agents that neutralize VEGF after release, integrin peptide therapy targets integrin heterodimers (alphavbeta3, alphavbeta5, alpha5beta1, alphaMbeta2) on endothelial cells, forcing them to stop producing VEGF and halt the entire neovascular cascade. This upstream mechanism addresses the cause of pathological vessel growth rather than its downstream effects.^[14]

In a phase I study of diabetic macular edema, 55% of patients treated with ALG-1001 monotherapy improved 3 to 5 lines of visual acuity, with at least 30% reduction in central macular thickness. These results held for at least 3 months after the final of three monthly loading doses. The patients enrolled were end-stage DME cases refractory to existing anti-VEGF therapy, making any improvement clinically meaningful. Preclinical studies also showed improved Mueller cell viability and mitochondrial health, suggesting effects beyond pure anti-angiogenesis.

Peptide-Drug Conjugates for Eye Drop Anti-VEGF

Perhaps the most transformative peptide approach is using cell-penetrating and melanin-binding peptides to deliver existing anti-VEGF drugs as eye drops rather than intravitreal injections. Fan and colleagues demonstrated in 2025 that peptide-bound aflibercept eye drops achieved sustained therapeutic concentrations in the retina and vitreous of nonhuman primates. The peptide component binds to melanin in ocular tissues, creating a depot effect that prolongs drug residence time.^[15]

If this approach translates to human use, it would address the single largest barrier to diabetic eye disease treatment: patient adherence. Roughly 40% of patients with diabetic macular edema receive fewer anti-VEGF injections than recommended, primarily due to the burden of repeated eye injections. An eye drop formulation maintaining therapeutic retinal drug levels could fundamentally change treatment compliance.

The Pipeline at a Glance

Limitations and Open Questions

GLP-1 early worsening risk. The rapid glycemic correction issue is real, even if it does not represent direct retinal toxicity. Clinicians initiating GLP-1 therapy in patients with moderate-to-severe baseline retinopathy need clear screening and monitoring protocols. The current evidence base, while reassuring at the population level, does not eliminate risk for individual high-risk patients.

Preclinical translation barriers. PEDF, SOCS1, PACAP, C-peptide, and GHRH agonist data all come from rodent models. Rodent retinas lack a true macula, the region most affected by diabetic macular edema. Peptide neuroprotection demonstrated in rat models may not translate to the anatomically distinct human macula. The EUROCONDOR trial with somatostatin showed this gap: whole-population results were negative despite strong preclinical signals.

Delivery challenges. Systemic peptide delivery (C-peptide, GHRH agonists, octreotide) achieves retinal concentrations but affects multiple organ systems. Topical delivery (PEDF, SOCS1, PACAP eye drops) has demonstrated corneal penetration in animals but has not been validated in human eyes. The peptide-bound aflibercept eye drop approach is the most advanced topical concept but remains in primate models.

Integrin peptide replication. ALG-1001 results come from small, early-phase trials. Whether the 55% response rate in refractory DME holds in larger, controlled studies is unknown.

Combination therapy rationale. The most promising future likely involves multiple peptide classes addressing different mechanisms simultaneously: metabolic control (GLP-1/tirzepatide), neuroprotection (PACAP or PEDF), and anti-angiogenesis (integrin peptide or anti-VEGF). No combination study has been designed.

Where These Approaches Converge

The clinical question is evolving from "do GLP-1 drugs harm the retina?" to "how can we harness peptide biology to protect it?" Samanta and colleagues reviewed the management landscape in 2025, noting that GLP-1 receptor agonists are increasingly recognized not just as glucose-lowering agents but as potential retinal protectors.^[16]

A patient in the near future might receive tirzepatide for systemic metabolic control (improving the environment driving retinopathy), PEDF or SOCS1 peptide eye drops (protecting retinal neurons from early degeneration), and a peptide-delivered anti-VEGF formulation (treating vascular complications without injection burden). Each peptide class addresses a different mechanism in the disease cascade. No single peptide will treat diabetic eye disease. The pipeline is building toward a multi-target strategy that matches the multi-mechanism nature of the disease itself.

For related peptide approaches to eye disease, see our articles on endothelin and glaucoma and PACAP and retinal neuroprotection.

The Bottom Line

Peptide therapies for diabetic eye disease span seven distinct classes with verified preclinical or clinical data: GLP-1 agonists (large safety dataset, emerging direct neuroprotective evidence), tirzepatide (reduced DR risk in 173,846-patient cohort), PEDF-derived eye drops (~60% vascular leakage reduction), SOCS1 peptide eye drops (anti-neuroinflammatory with functional improvement), PACAP (strongest preclinical neuroprotection), C-peptide (systemic retinal neuroprotection via ROS inhibition), GHRH agonists (VEGF downregulation and PEDF upregulation in PNAS data), somatostatin analogs (EUROCONDOR clinical trial, effective in neurodysfunction subset), integrin peptide therapy (upstream VEGF suppression, early clinical data), and peptide-drug conjugates (enabling eye drop delivery of anti-VEGF). The field is moving from single-target approaches toward multi-peptide strategies addressing metabolic, neuronal, and vascular components simultaneously.

Frequently Asked Questions

Do GLP-1 drugs cause diabetic retinopathy?

GLP-1 receptor agonists do not appear to cause diabetic retinopathy directly. The SUSTAIN-6 trial found more retinopathy events with semaglutide, but subsequent analysis attributed this to rapid glycemic correction, not drug toxicity. A 2025 meta-analysis of 39 studies found no significant association between GLP-1 receptor agonists and diabetic retinopathy risk (pooled RR 1.00, 95% CI 0.71-1.43). A 2026 study of over 150,000 patients found no difference in sight-threatening DR risk between four common GLP-1 agents (Barkmeier et al., Ophthalmology Retina, 2026).

Can peptide eye drops treat diabetic retinopathy?

Several peptide eye drop approaches are in preclinical development. PEDF-derived peptide eye drops reduced vascular leakage by approximately 60% and prevented microglia activation by approximately 60% in diabetic mice over 15 weeks of weekly application (Liu et al., 2012). SOCS1-derived peptide eye drops reduced neuroinflammation and improved retinal function in diabetic mice after just 2 weeks (Hernandez et al., 2019). Peptide-bound aflibercept eye drops achieved sustained anti-VEGF concentrations in nonhuman primate retinas (Fan et al., 2025). None has reached human clinical trials.

What is integrin peptide therapy for eyes?

Integrin peptide therapy (ALG-1001/risuteganib) is a drug class that blocks integrin receptors on blood vessel cells, stopping them from producing VEGF rather than neutralizing VEGF after release. In a phase I trial, 55% of end-stage diabetic macular edema patients refractory to standard anti-VEGF treatment improved 3-5 lines of visual acuity with ALG-1001 monotherapy, sustained for 3 months after the final dose. The peptide targets four integrin heterodimers involved in neovascularization.

Does tirzepatide protect eyes from diabetic retinopathy?

A 2026 cohort study of 173,846 patients found tirzepatide was associated with a 37.6% lower risk of diabetic macular edema (RR 0.624, 95% CI 0.536-0.727) and 29.5% lower risk of proliferative diabetic retinopathy (RR 0.705) compared to lifestyle intervention alone over 12 months (Shah et al., Ophthalmology, 2026). Whether this reflects the additional GIP receptor component, superior metabolic control, or patient selection differences remains unknown.

What is PEDF and how does it protect the retina?

Pigment epithelium-derived factor (PEDF) is a protein with neuroprotective, antiangiogenic, and anti-inflammatory properties. Two PEDF-derived peptide fragments, P60 (antiangiogenic) and P78 (neuroprotective), penetrated the cornea within 1-4 hours when applied as eye drops in diabetic mice and reached the retina. P78 reduced 9 of 20 inflammatory cytokines, prevented ganglion cell death by 22%, and reduced inner plexiform layer thinning by 13% (Liu et al., Molecular Medicine, 2012).

Does semaglutide have direct effects on retinal cells?

Preclinical evidence suggests it may. Cheng and colleagues showed in 2025 that semaglutide increased retinal inner nuclear layer thickness, reduced neuronal apoptosis, restored tight junction protein ZO-1 in retinal microvasculature, and downregulated VEGFA expression in diabetic rat retinas and human retinal endothelial cells. The GLP-1 receptor is expressed in retinal tissue, meaning semaglutide could have local effects beyond its systemic metabolic benefits. Human retinal neuroprotection data do not yet exist.