Latest / Glaucoma, Vision & Longevity: Supplements & Science / AI Found 21 New Genetic Clues to Glaucoma — Could Some Become Future Drug Targets?

Transcript

AI Found 21 New Genetic Clues to Glaucoma — Could Some Become Future Drug Targets? · Glaucoma, Vision & Longevity: Supplements & Science

• 0:00
  SPEAKER_00: AI found 21 new genetic clues to glaucoma.
• 0:03
  SPEAKER_00: Could some become future drug targets?
• 0:06
  SPEAKER_00: A study published in NPJ Genomic Medicine on September 19, 2026, used artificial intelligence to look for glaucoma-related patterns in optical coherence tomography scans, then tested those patterns against genetic data.
• 0:20
  SPEAKER_00: The researchers reported 21 genetic loci not previously linked to glaucoma and prioritized 11 genes that might help explain why glaucoma damages the eye.
• 0:30
  SPEAKER_00: Five of those genes had not previously been reported in glaucoma research.
• 0:36
  SPEAKER_00: The work's main advance is not that artificial intelligence diagnosed glaucoma better.
• 0:41
  SPEAKER_00: It is that the researchers used measurable patterns of retinal damage, rather than a simple glaucoma or no glaucoma label, to search for genetic clues.
• 0:51
  SPEAKER_00: That approach produced a richer biological picture, including a particularly promising signal involving RSPO2 and WN signaling in cells that help drain fluid from the eye.
• 1:01
  SPEAKER_00: But these are genetic leads, not tested glaucoma treatments.
• 1:05
  SPEAKER_00: The study did not show that changing any of the 11 genes prevents vision loss.
• 1:10
  SPEAKER_00: The most exciting possibilities still require experiments to establish which genes are truly involved, whether changing them would help or harm, and whether a treatment can reach the right cells safely.
• 1:23
  SPEAKER_00: What the study found and what it did not PEX.
• 1:26
  SPEAKER_00: The peer-reviewed article is Deep Learning Derived Glaucoma-related endophenotypes enable novel genome-wide genetic and functional discovery.
• 1:35
  SPEAKER_00: It reports 36 genome-wide significant loci in the European Ancestry Analysis and 43 in a cross-ancestry analysis.
• 1:43
  SPEAKER_00: The authors identified 21 loci as previously unreported in glaucoma and used additional analyses to prioritize 11 genes, five of them novel to glaucoma.
• 1:53
  SPEAKER_00: A locus is a region of the genome associated with a trait.
• 1:57
  SPEAKER_00: It may contain several genes.
• 1:59
  SPEAKER_00: A gene prioritized at a locus is a candidate for explaining the association, not necessarily the gene that causes it.
• 2:07
  SPEAKER_00: The study therefore offers a set of hypotheses about glaucoma biology.
• 2:11
  SPEAKER_00: It does not establish that the 21 loci or any of the 11 genes cause glaucoma, nor that a drug aimed at them would work.
• 2:19
  SPEAKER_00: The methodological innovation in patient-friendly language, from a yes or no label to a detailed retinal fingerprint.
• 2:27
  SPEAKER_00: Many genetic studies classify each participant as having glaucoma or not having glaucoma.
• 2:32
  SPEAKER_00: That can be a noisy comparison.
• 2:34
  SPEAKER_00: Some people may have undiagnosed disease, and people with glaucoma can have very different patterns and amounts of retinal damage.
• 2:42
  SPEAKER_00: Instead, these researchers trained machine learning models on eye scans from people with glaucoma.
• 2:48
  SPEAKER_00: The models learned patterns in the ganglion cell complex and retinal nerve fiber layer, retinal structures that are affected as glaucoma damages the optic nerve.
• 2:58
  SPEAKER_00: They then represented those patterns as quantitative traits or measurable features of the scans.
• 3:05
  SPEAKER_00: A useful analogy is the difference between asking whether a house is damaged and measuring exactly where its walls are cracked, how deep the cracks are, and how much the structure has shifted.
• 3:17
  SPEAKER_00: The detailed measurements can reveal more about the process behind the damage.
• 3:21
  SPEAKER_00: These scan-derived traits, which the researchers call endophenotypes, were then linked to genetic differences.
• 3:28
  SPEAKER_00: They are glaucoma-informed measurements, not a replacement diagnosis or a measure of future vision loss.
• 3:34
  SPEAKER_00: The study's full-text manuscript describes two machine learning model families, labeled Otto and MOCO, used to produce traits from macular scans.
• 3:44
  SPEAKER_00: Reconstructing the research pipeline.
• 3:47
  SPEAKER_00: Train on glaucoma scans.
• 3:49
  SPEAKER_00: The models were developed using 18,985 clinical optical coherence tomography scans from 8,323 glaucoma patients at Mass I and Ear.
• 4:00
  SPEAKER_00: The earlier manuscript says most had open angle glaucoma codes, though the training group was not restricted to chart-confirmed primary open angle glaucoma cases.
• 4:09
  SPEAKER_00: Apply the models to UK Biobank.
• 4:11
  SPEAKER_00: The final peer-reviewed publication reports 39,046 UK Biobank participants with usable genetic and scan-derived data.
• 4:18
  SPEAKER_00: The ancestry groups were 35,739 European, 1,730 African, and 1,677 Asian participants.
• 4:27
  SPEAKER_00: The large difference in group sizes means the European results had much greater statistical power.
• 4:33
  SPEAKER_00: Run genome-wide association studies.
• 4:35
  SPEAKER_00: The researchers tested genetic variants against 21 scan-derived traits separately by ancestry.
• 4:41
  SPEAKER_00: A genome-wide association study looks across the genome for variants statistically associated with a measured trait.
• 4:48
  SPEAKER_00: Split samples for discovery and replication.
• 4:50
  SPEAKER_00: The manuscript describes a 75% discovery, 25% replication split within UK Biobank, followed by meta-analysis.
• 4:59
  SPEAKER_00: This is useful internal checking, but it is not the same as replication in a fully independent cohort.
• 5:05
  SPEAKER_00: Combine results across ancestry groups.
• 5:08
  SPEAKER_00: They conducted ancestry-specific analyses and then a cross-ancestry meta-analysis, finding 36 significant loci in the European analysis and 43 in the cross-ancestry analysis.
• 5:19
  SPEAKER_00: Smaller African and Asian sample sizes limit how confidently those group-specific effects can be estimated.
• 5:26
  SPEAKER_00: Prioritize possible genes and mechanisms.
• 5:28
  SPEAKER_00: Follow-up analyses included gene-based testing, expression and splicing quantitative trait lie, EQTLs and SQTLs, Bayesian collocalization, Mendelian randomization, and single cell expression enrichment in eye tissues.
• 5:43
  SPEAKER_00: These analyses help narrow a locus to possible genes and cell types, but each has its own assumptions and limitations.
• 5:49
  SPEAKER_00: Which UK biobank number should be used?
• 5:52
  SPEAKER_00: The peer-reviewed publications figure is 39,146 participants.
• 5:57
  SPEAKER_00: An earlier Metaarxive abstract reported 47,908, but the manuscript's detailed ancestry counts, 35,739 plus 1730 plus 1,677, sum to 39,146.
• 6:12
  SPEAKER_00: The final publication's participant number should be used for the peer-reviewed study.
• 6:16
  SPEAKER_00: The public versions do not clearly explain the origin of the earlier headline figure, so it is safest to treat it as an earlier or inconsistent version figure, rather than assume it represents the final analysis.
• 6:28
  SPEAKER_00: What 21 new genetic clues means.
• 6:31
  SPEAKER_00: The authors call 21 loci novel to glaucoma because they had not previously been reported as associated with glaucoma or established glaucoma-related traits.
• 6:41
  SPEAKER_00: Novel here does not mean the genes were newly discovered, that the DNA changes are unique to glaucoma, or that the genes' roles in the body are now understood.
• 6:51
  SPEAKER_00: The 11 prioritized genes were EIF3E, LIN-54, PBLD, RSPO2, SLC25A16, TSPAN10, 1I2, SLC25A12, A2H7, and C14 ORF39.
• 7:04
  SPEAKER_00: The paper identifies RSPO2, EIF3E, LIN-52, DO1I2, and SLC25A12 as the five genes not previously reported in glaucoma.
• 7:13
  SPEAKER_00: It also notes prior glaucoma-related or eye trait evidence for some of the other genes, including ATO7, C14 ORF-39, NPL4, TSPAN-10, PBLD, and SLC25A16.
• 7:26
  SPEAKER_00: The 11 prioritized genes, biology, eye cell evidence, and drug prospects.
• 7:31
  SPEAKER_00: The cell evidence below refers to the study's expression and enrichment analyses, not proof that a gene causes disease in that cell.
• 7:38
  SPEAKER_00: The single cell reference data came from non-diseased human eye tissues, and for several genes, the article does not identify a particular eye cell type as the key location.
• 7:47
  SPEAKER_00: For the full table, please open this article on VisualFieldtest.com.
• 7:51
  SPEAKER_00: Two shared regions make some gene assignments especially uncertain.
• 7:55
  SPEAKER_00: Several prioritized genes sit in the same genetic region.
• 7:58
  SPEAKER_00: For example, Dwine C1I2 and SLC25A12 share a locus in the study's prioritization table.
• 8:05
  SPEAKER_00: Both may be plausible candidates, but the association does not establish which gene or whether both accounts for the signal.
• 8:12
  SPEAKER_00: A similar caution applies to the chromosome 10 region involving 8OH7, PBLD, and nearby genes.
• 8:19
  SPEAKER_00: These are reasons for follow-up experiments, not reasons to discard the findings.
• 8:24
  SPEAKER_00: The main biological themes WN signaling and aqueous outflow, the clearest treatment-oriented lead.
• 8:30
  SPEAKER_00: Aqueous humor is the clear fluid inside the eye.
• 8:33
  SPEAKER_00: It normally drains through the trabecular meshwork.
• 8:36
  SPEAKER_00: If drainage meets too much resistance, eye pressure can rise.
• 8:40
  SPEAKER_00: Lowering pressure is an established way to slow glaucoma damage.
• 8:45
  SPEAKER_00: RSPO2 is notable because it can strengthen WNT signaling, and the study found it in trabecular meshwork and other anterior segment fibroblasts, as well as optic nerve head fibroblasts.
• 8:56
  SPEAKER_00: That puts a biologically plausible candidate in cells involved in tissue structure and outflow.
• 9:02
  SPEAKER_00: Importantly, earlier laboratory work supports a role for canonical WN signaling in the trabecular meshwork and pressure regulation.
• 9:10
  SPEAKER_00: In experimental models, blocking Went signaling raised pressure, and another study found that WN activation could reverse some abnormal features of trabecular meshwork cells.
• 9:20
  SPEAKER_00: Those results strengthen the general Dubine outflow hypothesis, but they do not prove RSPO2 causes glaucoma or establish whether RSPO2 itself should be increased or decreased.
• 9:33
  SPEAKER_00: That direction matters.
• 9:34
  SPEAKER_00: RSPO2 boosts WN signaling, and the prior I studies suggest that too little canonical W activity may be harmful in the trabecular meshwork.
• 9:44
  SPEAKER_00: A simple plan to block RSPO2 could therefore be wrong.
• 9:47
  SPEAKER_00: The best therapeutic idea, if the genetics is confirmed, may be to restore or tune local WN signaling rather than turn it off across the body.
• 9:56
  SPEAKER_00: Retinal ganglion cell vulnerability, axonal transport, and mitochondria.
• 10:01
  SPEAKER_00: The study's cross-ancestry gene expression enrichment linked Duang 1I2, NPLA-4, and SLC25A12 to a retinal ganglion cell cluster.
• 10:10
  SPEAKER_00: Their known functions suggest different ways a neuron might cope with glaucoma-related stress.
• 10:15
  SPEAKER_00: Duang 1i2, movement of cargo along cell structures, relevant to nerve cell transport.
• 10:20
  SPEAKER_00: NPLA4, protein quality control, relevant to handling damaged or unwanted proteins.
• 10:26
  SPEAKER_00: SLC25A12, mitochondrial metabolism, relevant to cellular energy supply.
• 10:32
  SPEAKER_00: These findings add genetic support to the broader idea that glaucoma is not only an eye pressure disease, it may also involve how resilient retinal ganglion cells are when under stress.
• 10:43
  SPEAKER_00: But the study did not show that any of these genes' functions were experimentally disrupted in glaucoma, or that changing one protects vision.
• 10:52
  SPEAKER_00: Lipids, inflammation, oxidized phospholipids, vascular stress, and calcium signaling.
• 10:58
  SPEAKER_00: The authors also reported gene set enrichment involving lipid and cholesterol regulation, innate inflammatory programs, responses to oxidized phospholipids, vascular and hypoxic signaling, tissue remodeling pathways, and calcium signaling.
• 11:13
  SPEAKER_00: These are useful clues that glaucoma-related scan patterns may sit within several intersecting biological systems, not just one pathway.
• 11:21
  SPEAKER_00: However, these are pathway enrichment results, not evidence that the study has identified a specific cholesterol drug, anti-inflammatory drug, blood flow treatment, or calcium channel medicine for glaucoma.
• 11:34
  SPEAKER_00: They support further testing of those hypotheses.
• 11:37
  SPEAKER_00: They do not justify choosing a medication based on the genetic findings.
• 11:41
  SPEAKER_00: Could existing medicines act on these pathways?
• 11:45
  SPEAKER_00: UNT modifying medicines.
• 11:47
  SPEAKER_00: The glaucoma-relevant UNT evidence remains mainly preclinical.
• 11:50
  SPEAKER_00: A small molecule WNT inhibitor reduced some dexamethasone-related changes in cultured human trabecular meshwork cells, while separate work found that WN activation could restore features of a normal cell phenotype.
• 12:03
  SPEAKER_00: The results illustrate why more dubent or less WNT cannot be assumed to be beneficial in every context.
• 12:11
  SPEAKER_00: Lithium is a prescribed medicine for other conditions that can influence the broad Duent GSK3 pathway, but it is not a selective RSPO2 drug, and this study did not test lithium, establish drug levels in the eye, or show glaucoma benefit.
• 12:26
  SPEAKER_00: Prior discussion of lithium and glaucoma is not evidence that patients should use lithium for glaucoma.
• 12:32
  SPEAKER_00: No patient should start or change lithium or another medication for glaucoma on the basis of this study.
• 12:38
  SPEAKER_00: There is some related drug development activity.
• 12:41
  SPEAKER_00: Experimental antibodies have been developed against R spondin family proteins in non-eye disease research, and an anti-RSPO3 antibody has been studied in cancer settings.
• 12:51
  SPEAKER_00: But RSPO3 is not RSPO2 and neither establishes a glaucoma treatment.
• 12:57
  SPEAKER_00: EIF3E, diene and mitochondrial pathways.
• 13:01
  SPEAKER_00: A patent application describes experimental compounds claimed to affect EIF3E.
• 13:05
  SPEAKER_00: That is a sign of drug discovery interest, not evidence of an approved drug or a glaucoma-ready candidate.
• 13:12
  SPEAKER_00: Research compounds such as ciliobrivine and dinerestin can interfere with dienine-dependent transport.
• 13:18
  SPEAKER_00: They are broad experimental tools, not selective medicines for Dwync1I2, and interfering with a transport system used by neurons may carry substantial biological risks.
• 13:28
  SPEAKER_00: For SLC25A12 and SLC25A16, the study points to mitochondrial biology, but there is no established selective glaucoma drug for either transporter.
• 13:40
  SPEAKER_00: An approach that supports mitochondrial health is not automatically a treatment that acts on either gene.
• 13:46
  SPEAKER_00: Metabolic and neuroprotective research, mitochondrial and metabolic approaches are already being investigated in glaucoma.
• 13:54
  SPEAKER_00: Small clinical studies of nicotinamide, alone or with pyruvate, have reported short-term signals in measures of retinal function.
• 14:02
  SPEAKER_00: Those findings do not yet establish that these supplements slow long-term visual field loss or treat the newly prioritized genes.
• 14:10
  SPEAKER_00: The American Glaucoma Society and American Academy of Ophthalmology have stated that nicotinamide is not approved for glaucoma and have warned about safety concerns, especially at high doses outside trials.
• 14:24
  SPEAKER_00: The practical point is that metabolic treatment is a plausible research direction, not an established add-on to glaucoma care.
• 14:31
  SPEAKER_00: Do not substitute supplements or experimental pathway modifiers for prescribed pressure-lowering treatment or an eye care professional's advice.
• 14:39
  SPEAKER_00: How strong is the evidence for causality?
• 14:42
  SPEAKER_00: It helps to separate four levels of evidence.
• 14:45
  SPEAKER_00: Association.
• 14:46
  SPEAKER_00: Genetic variants were associated with quantitative scan-derived traits.
• 14:51
  SPEAKER_00: This is the study's primary discovery.
• 14:53
  SPEAKER_00: Putative gene prioritization, gene-based tests, colocalization, expression patterns, and Mendelian randomization pointed to candidate genes.
• 15:02
  SPEAKER_00: These methods can strengthen a case, but they do not settle it.
• 15:07
  SPEAKER_00: Experimentally validated causality.
• 15:09
  SPEAKER_00: The study did not edit or manipulate these 11 genes in glaucoma models to show that changing the gene changes disease.
• 15:17
  SPEAKER_00: Therapeutic speculation.
• 15:19
  SPEAKER_00: Antibodies, small molecules, metabolic support, or other treatments are possibilities to investigate, not treatments demonstrated by this study.
• 15:27
  SPEAKER_00: Mendelian randomization can support a causal interpretation when its genetic instruments meet key assumptions.
• 15:33
  SPEAKER_00: It is not definitive proof that a particular gene causes glaucoma, that the effect applies to disease progression, or that a drug acting on that gene will be helpful.
• 15:44
  SPEAKER_00: Here, the genetic traits were scan-derived retinal structures, not a clinical trial endpoint, such as preserved vision.
• 15:52
  SPEAKER_00: The authors also note the lack of an independent cohort with both scan and genetic data and the European heavy sample as limitations.
• 16:00
  SPEAKER_00: Which genes look most promising as future drug targets?
• 16:04
  SPEAKER_00: The ranking below is a reasoned estimate, not a measured probability, from the paper.
• 16:09
  SPEAKER_00: It weighs the combination of genetic support, plausible biology, ocular cell relevance, pharmacology, and practical ability to develop a safe medicine.
• 16:18
  SPEAKER_00: For the full table, please open this article on VisualFieldTest.com.
• 16:24
  SPEAKER_00: Probability estimates.
• 16:26
  SPEAKER_00: These are judgment calls, not results calculated by the researchers.
• 16:30
  SPEAKER_00: Probability that at least one of the five novel genes becomes an experimentally validated glaucoma therapeutic target, approximately 30 to 40%.
• 16:39
  SPEAKER_00: Probability that a treatment based on one of these genes or its pathway reaches a human glaucoma trial within 10 years, approximately 10 to 20%.
• 16:48
  SPEAKER_00: Probability that a treatment based on one becomes an approved glaucoma medicine within 10 years, likely lower than the chance of a trial, because a trial is an early development milestone and an approved treatment requires much more evidence.
• 17:03
  SPEAKER_00: The estimates reflect the genuine convergence of genetic, cellular, and biological clues.
• 17:09
  SPEAKER_00: Balanced against the fact that the study did not experimentally test the genes, and the non-European UK biobank groups were comparatively small.
• 17:19
  SPEAKER_00: Which newly discovered gene is most interesting for glaucoma treatment?
• 17:23
  SPEAKER_00: RSPO2 is the most interesting overall candidate for a future pressure lowering treatment.
• 17:29
  SPEAKER_00: It combines a novel glaucoma association with a plausible mechanism in cells involved in aqueous fluid drainage and in optic nerve head fibroblasts.
• 17:38
  SPEAKER_00: Its secreted protein biology also makes local biologic approaches conceivable.
• 17:43
  SPEAKER_00: Earlier experimental work on UNT signaling in the trabecular mesh work gives the broader pathway a real biological foundation, not just a statistical association.
• 17:53
  SPEAKER_00: The largest caveat is that the study does not tell us whether RSPO2 activity should be raised, lowered, or finely tuned.
• 18:01
  SPEAKER_00: Because Duente signaling can have different effects depending on the tissue and context, a generic dub inhibitor or a generic dub activator could be the wrong treatment.
• 18:12
  SPEAKER_00: SLC25A12 is a strong runner-up for a different reason.
• 18:16
  SPEAKER_00: It connects a glaucoma-associated genetic signal to retinal ganglion cells and mitochondrial energy metabolism.
• 18:23
  SPEAKER_00: That makes it a valuable candidate for studying neuronal vulnerability, even though it is currently harder to translate into a drug.
• 18:30
  SPEAKER_00: The most realistic conclusion is therefore not that AI has found a glaucoma cure.
• 18:36
  SPEAKER_00: It is that AI-assisted retinal phenotyping has helped researchers identify testable biological leads that ordinary yes or no case control studies might miss.
• 18:46
  SPEAKER_00: RSPO2 offers the clearest near-term path for mechanistic experiments in eye fluid drainage.
• 18:52
  SPEAKER_00: SLC25A12 and Wine1I2 offer important but more technically challenging paths toward retinal ganglion cell protection.
• 19:01
  SPEAKER_00: All links to sources are available in the text version of this article.
• 19:05
  SPEAKER_00: You can find the full article at VisualFieldTest.com.
• 19:08
  SPEAKER_00: Thanks for listening.
• 19:09
  SPEAKER_00: To check your visual field, click the link at the bottom of this article or visit visualfieldtest.com.