Large Genetic Study Pinpoints Seven New DNA Loci and 23 Proteins Tied to Macular Degeneration

A sweeping integration of genome-wide association data and blood proteomics has mapped ten candidate genetic loci for age-related macular degeneration, seven of them newly identified, and flagged 23 circulating proteins whose genetically predicted levels are linked to the disease. Among these, two proteins, LAMC2 and PLTP, emerged as first-tier risk factors supported by convergent evidence across protein abundance, gene expression, and DNA methylation, offering some of the most detailed molecular leads yet for a condition that robs millions of central vision.

The meta-analysis surfaced 25 independent genome-wide significant loci in total. Fifteen corresponded to known signals, while ten prioritized candidate loci, located more than 500 kilobases from previously reported index variants, became the focus of downstream work. Seven of these ten were confirmed as independent new associations through conditional analysis that controlled for all known AMD variants.

The newly identified variants carry effect sizes comparable to established major risk loci. Their absolute effect estimates ranged from 0.060 to 0.300, with DENND1B at 0.300 and PTPRC at 0.202 matching or exceeding the replicated ARMS2/HTRA1 signal at 0.243. The authors argue this shows the expanded design captured both highly penetrant drivers and moderate-effect modulators of disease. Gene ontology analysis of the ten candidate genes pointed to lymphocyte-mediated immunity, transcriptionally active chromatin, and phosphatidylinositol phosphate binding, while disease enrichment linked them to neovascular AMD, geographic atrophy, and immune system diseases.

Using protein quantitative trait loci from deCODE Genetics as the discovery resource, covering 1,812 plasma proteins, and the UK Biobank Pharma Proteomics Project, covering 1,114 proteins, for replication, they found 89 proteins significantly associated with AMD in deCODE. Bayesian colocalization narrowed this to 12 proteins showing evidence of sharing a causal genetic variant with AMD, including LBP, MCEMP1, TNFSF14, LAMC2, C3d, CFHR3, C3b, BRD2, APOA5, and RBP1. After the two-stage discovery and replication process, 23 proteins carried suggestive evidence of association, 11 as potential risk factors and 12 as potentially protective.

Thirteen of the 23 proteins, including LAMC2, LBP, APOA5, PLTP, PILRA, NID2, AGER, WARS1, IGFBP6, GRN, HEBP1, C2, and MATN3, passed significance thresholds at the expression level, all with effects directionally consistent with the proteomic results. The methylation analysis identified 23 CpG sites mapped to seven genes, with risk-associated sites near LAMC2, NID2, and PLTP reinforcing the protein-level findings. Only LAMC2 and PLTP cleared every threshold at all three levels, earning first-tier status.

The data clustered into seven major categories, including monocytes and a stromal/progenitor-like population that was manually re-annotated after automated labeling flagged it as tissue stem cells. Eleven target genes showed cell-type-specific differential expression in AMD retinas, with HAVCR2, GRN, PLTP, and C3 enriched mainly in monocytes, and IGFBP6 and RBP1 in stromal/progenitor-like cells. Notably, LAMC2 and PLTP themselves did not show significant cell-type-specific enrichment, which the authors attribute to their roles as secreted proteins acting systemically rather than through localized transcription.

Screening 731 immunological traits identified 53 significantly associated with AMD, with regulatory T cell traits showing the most prominent links. Exploratory two-step mediation analysis suggested that LAMC2’s effect on AMD might be partially carried through several regulatory T cell subsets, with modest mediation proportions of 0.60 to 3.05 percent. C3d acted through myeloid and plasmacytoid dendritic cell populations, PILRA through myeloid, regulatory T cell, and natural killer traits with proportions up to 11.69 percent, and CFHR3 through a TBNK trait at just 0.25 percent. The authors stress these small proportions indicate partial, hypothesis-generating mechanisms rather than primary drivers.

Complementary evidence came from mouse knockout models and druggability databases. Of 11 queried genes, knockout models for RBP1, CFHR3, C3, VEGFA, GRN, and CSK showed ocular phenotypes including abnormal retinal pigment epithelium and vasculature morphology and abnormal electroretinography, though these loss-of-function models cannot distinguish specific mechanisms. The Drug-Gene Interaction Database flagged 15 target genes as druggable, including LAMC2 and PLTP. A computational enrichment analysis nominated cyclophosphamide, a cytotoxic immunosuppressant interacting with five prioritized genes, but the authors explicitly frame this as a pathway indicator only, given the drug’s well-documented systemic toxicity, not a therapeutic proposal.

The findings arrive against a substantial clinical backdrop. An estimated 196 million people worldwide have AMD, which primarily affects adults over 55 and progressively destroys central vision. Genetic factors are thought to contribute 40 to 60 percent of disease risk. While therapies for neovascular AMD, including anti-VEGF agents and the C3 inhibitor pegcetacoplan for geographic atrophy, have advanced, effective treatments for atrophic AMD remain limited, making primary prevention an unmet need that genetically supported targets could eventually address.

The study is explicitly exploratory, and its limitations are considerable. The cohorts were predominantly European ancestry, restricting global generalizability. Partial sample overlap exists within the GWAS cohorts and between proteomic exposure and outcome datasets, though sensitivity analyses, including a linkage disequilibrium score regression intercept near unity and consistent results in the independent FinnGen cohort, support robustness. The colocalization threshold of 0.5 was deliberately permissive, marking moderate rather than definitive evidence. Residual pleiotropy, limited instrumental variants for some proteins, and modest mediation effects all warrant caution. The authors conclude that the proposed therapeutic implications remain speculative and require validation in preclinical and clinical settings, positioning the work as a hypothesis-generating resource for future mechanistic studies rather than a route to immediate new treatments.

Subject of Research: Integrative genomic and proteomic analysis identifying genetic loci and genetically supported proteins associated with age-related macular degeneration.

Article Title: Integrative genomic and proteomic analyses identify candidate genetic loci and genetically supported proteins in age-related macular degeneration

Article References: Chen, N., Liu, Y., Chen, Z., Xiang, Y., & Yu, H. (2026). Integrative genomic and proteomic analyses identify candidate genetic loci and genetically supported proteins in age-related macular degeneration. Eye and Vision, 13(1), Article 35. https://doi.org/10.1186/s40662-026-00502-3

Image Credits: AI Generated

DOI: 10.1186/s40662-026-00502-3

Keywords: age-related macular degeneration, GWAS meta-analysis, Mendelian randomization, colocalization, LAMC2, PLTP, proteomics, single-cell RNA sequencing, immune mediation, drug targets, complement pathway, genetics