PRPH2, peripherin-2, and the retinal diseases they cause
PRPH2 is a gene. It carries the instructions for peripherin-2, a protein that holds the light-catching parts of the retina in shape. A change in that gene can cause anything from a mild pattern dystrophy noticed late in life to a more serious cone-rod dystrophy that begins early; most people with a PRPH2 variant first notice something in adulthood. This site explains how that works in plain language, and lets you look up any PRPH2 variant against what is publicly known about it.
Written by a parent, for families, patients, clinicians and the curious. Not medical advice. Words with a dotted underline and a ? can be tapped for a definition; the full glossary is one page. Large type, dark, light and high-contrast modes are in the Display menu.
Genes, proteins, variants, and what a genetic report says
If you have never needed these words before, this section is for you. Everything else on the site builds on it.
A gene is a recipe; a protein is the dish
Your DNA holds about 20,000 gene: A stretch of DNA that carries the instructions for making one protein. PRPH2 is a gene; peripherin-2 is the protein it makes. Glossary. Each one is the instructions for making one protein: A molecule built from a chain of amino acids, folded into a shape that lets it do a job in the cell. Peripherin-2 is a structural protein: its job is to hold a membrane in shape. Glossary, and proteins are the working parts of cells. PRPH2 is a gene. The protein it makes is called peripherin-2. Everyone has two copies of PRPH2, one inherited from each parent.
Proteins are chains of amino acids
A protein is a long chain of building blocks called amino acid: One of the 20 building blocks that proteins are made of. Peripherin-2 is a chain of 346 of them. Each has a three-letter name (Gly, Arg) and a one-letter code (G, R). Glossary, folded into a shape. Peripherin-2 is a chain of 346 of them. Each position in the chain has a number, and each amino acid has a name with a three-letter and a one-letter short form: glycine is Gly or G, arginine is Arg or R.
A variant is a change in the recipe
A variant: A change in the DNA sequence compared with the usual version. "Mutation" means the same thing; "variant" is the neutral word laboratories use because not every change causes disease. Glossary (older word: mutation) is a change in the DNA of a gene. Some change one amino acid for another (missense: A variant that swaps one amino acid for another, so the protein is made in full but with one wrong building block. Written like p.Gly208Asp or G208D: glycine at position 208 replaced by aspartate. Glossary, written like p.Gly208Asp: glycine at position 208 became aspartate). Some stop the protein from being made at all (truncating (loss-of-function): A variant that stops the protein being made properly, either by putting a premature stop signal in (nonsense) or shifting the reading frame (frameshift). The cell usually destroys the faulty message, so that copy of the gene makes nothing. Glossary). Some disturb how the gene's message is cut and joined (splice variant: A change at the boundary of an exon that disturbs how the gene's message is cut and joined. Often behaves like a loss-of-function variant, but not always. Glossary). Most variants in any gene are harmless; the ones on this site are the ones that are not, or are still being worked out.
Reading your report
A report usually gives the same change two or three ways: c. notation: How a variant is written at the DNA level: c.623G>A means position 623 of the gene's coding sequence changed from G to A. c.828+1 means one base into the intron after position 828. Glossary describes the DNA letter that changed, p. notation: How a variant is written at the protein level: p.Gly208Asp means the amino acid glycine at position 208 became aspartate. Often shortened to G208D. Glossary (or G208D) describes the amino acid that changed as a result, and sometimes a chromosome position. They are one change with three addresses. "heterozygous: Having one altered copy and one usual copy of the gene. This is the normal situation in PRPH2 disease. Glossary" means one of your two copies carries it. The classification (pathogenic, likely pathogenic, variant of uncertain significance (VUS): A variant that has been seen but for which there is not enough evidence to say whether it causes disease. It is not a verdict either way. Many VUS are later reclassified as more evidence arrives. Glossary, likely benign, benign) is the laboratory's judgment of whether it causes disease.
One altered copy is enough
PRPH2 disease is autosomal dominant: One altered copy of the gene, inherited from one parent, is enough to cause disease. Each child of an affected person has a 1 in 2 chance of inheriting it. PRPH2 disease is dominant in about 95% of families. Glossary in about 19 families out of 20. One altered copy causes it. Each child of someone who carries the variant has a 1 in 2 chance of inheriting it, whether or not the parent has symptoms. It does not skip generations, and people who did not inherit it cannot pass it on.
What "uncertain significance" means
Not enough is known yet to say whether the change causes disease. It is not a verdict in either direction. Variants move out of this category as more families are reported or as laboratory tests are done; a relative who has the disease and carries the same variant is one of the strongest pieces of evidence. If your report says this, the practical answer is to be examined by a retinal specialist and to ask whether affected relatives can be tested.
Where peripherin-2 works
The retina's light-sensing cells catch light with a stack of tiny membrane discs that is rebuilt about every ten days. Peripherin-2 holds those discs in shape.
The eye
Light enters through the cornea and lens and lands on the retina, the thin sheet of nerve tissue lining the back of the eye. The macula is the small central patch that gives you reading and face-recognition vision.
Two features of this architecture matter for PRPH2 disease. First, the disc stack is continuously renewed: new discs form at the base every day and old ones are shed from the tip. A protein that is slightly wrong does not cause a single failure; the cell keeps rebuilding with it, thousands of times over a lifetime. Second, the retina has two kinds of photoreceptor: A light-sensing cell in the retina. Rods work in dim light and cover the periphery; cones give sharp daytime and color vision and are packed into the macula. Glossary with the same protein doing subtly different jobs: rods stack flat closed discs and give night and side vision; cones stack open discs and give daytime, color and reading vision, and are concentrated in the macula: The small central area of the retina responsible for reading, faces and fine detail. Most PRPH2 disease affects it. Glossary. That difference in geometry is one reason the same variant can spare cones and damage rods, or the reverse.
The peripherin-2 protein
It crosses the disc membrane four times, and the part that matters most is a large loop inside the disc.
Look at the drawing at the top of the page. Peripherin-2 threads through the disc membrane four times (the four pillars, labeled TM1 to TM4, TM for "transmembrane"). Between the pillars it forms loops. Two small loops and the two ends hang inside the cell. The big loop, called the D2 loop: The large part of peripherin-2 that sits inside the disc (residues 124 to 264). About seven in ten disease-causing variants are here, because this is the part that links molecules together. Glossary, sits inside the disc. Of all the disease-causing changes reported in PRPH2, about seven in ten are in this loop.
The D2 loop matters because it is where peripherin-2 molecules hold on to each other. Seven of its amino acids are cysteine: An amino acid whose sulfur atom can form a covalent bond (a disulfide bond) with another cysteine. Peripherin-2 uses seven of them in the D2 loop as staples to hold its shape and to link to neighbors. Glossary, which can form a strong chemical bond with another cysteine, like a staple. Three of those staples hold the loop's own shape. One reaches out to a neighboring molecule. Losing a cysteine, gaining one in the wrong place, or bending the loop so the staples cannot close, all end in the same result: peripherin-2 that cannot link into the chains the disc rim needs.
Peripherin-2 rarely works alone. It pairs with a closely related protein called ROM1. A 2022 study froze the two together and photographed them with an electron microscope (a cryo-EM structure: A 3D picture of a protein made by freezing it and imaging it with an electron microscope. The 2022 structure of peripherin-2 with ROM1 shows where each amino acid sits. Glossary), which is how we know where each amino acid sits; that picture is the basis of the 3D view on the Explore page. ROM1 changes on their own rarely cause disease, but a ROM1 change alongside a PRPH2 change can make things worse, which is why a thorough genetic test looks at both.
Three independent patient studies found the same thing:
How peripherin-2 builds a disc rim
Single molecules join into groups of four, groups of four join into groups of eight, and the groups link into chains that bend the membrane.
Monomer
One peripherin-2 molecule: four passes through the membrane, a small loop (D1) and a large loop (D2) inside the disc. The D2 loop carries seven cysteines.
Assembly is also a checkpoint. Only groups of four (tetramer / oligomer: A group of protein molecules locked together: four is a tetramer, eight an octamer. Peripherin-2 only works in these groups, which is why one faulty molecule can spoil several good ones. Glossary) that are correctly built are allowed out of the cell body and into new discs. A faulty molecule that cannot join a group is held back and destroyed; the cell is short of protein, but what it has is good. A faulty molecule that joins a group well enough to pass the checkpoint, but then cannot bond properly, gets built into the disc rim and does its damage there. This distinction, worked out in frog photoreceptors twenty years ago, is the reason different kinds of variant cause different diseases, as the next two sections explain.
Two ideas describe what the chains then do, and both are probably right. In the classic view they are the scaffold of the rim, forcing the membrane into its tight curve. A newer view adds that peripherin-2 is what stops a new disc from pinching off as a loose bubble and being lost. Either way, less working peripherin-2 means fewer, shorter, more disorganized discs.
Why one altered copy can do more damage than a missing one
In PRPH2 disease one altered copy of the gene is enough. What that copy does, though, is not one thing, and the difference matters for prognosis and for treatment.
Dominant-negative
The mutant copy makes a protein that folds well enough to join tetramers but cannot bond correctly. Any tetramer with even one mutant subunit is spoiled. If half the subunits are mutant, only one tetramer in sixteen (0.5 to the fourth power) is clean. That is why a D2-loop missense variant can be far worse than having no protein at all.
The picture above is the core of it. If the altered gene makes no protein (a premature stop signal, a shifted reading frame, most splice changes), the cell has half the normal amount of a normal protein. That is called haploinsufficiency: When one copy of the gene makes nothing and half the normal amount of protein is not quite enough. Milder than dominant-negative, because the protein that is made is normal. Glossary, and photoreceptors tolerate it surprisingly well. If instead the altered gene makes a faulty protein that folds well enough to be assembled but not well enough to work, it spoils every group it joins. That is called a dominant-negative: When the faulty protein does not just fail to work but actively interferes with the normal protein made by the other copy of the gene. Function can fall far below half. Glossary effect. Groups of four are built from whichever molecules are nearby, so if half the molecules are faulty, only about one group in sixteen (6%) is made entirely of good ones. Function does not fall to half; it falls much further. That sounds absolute, and in a living retina it is not: the cell keeps making both kinds of protein, some discs still form, and useful vision usually remains for years. The point is that these variants are the more serious kind, not that vision is gone.
There is a third documented pattern that complicates this. In a study of eleven D2-loop changes in living mouse retina, the ones associated with cone disease were made in excess, reached the right place and yet did not work properly. The authors suggest that too much of a non-working protein is its own kind of problem. It is one study, but it is a warning that where a change sits in the loop is not the same as knowing what it does.
Why does this matter to a family? Because the kind of treatment follows the kind of damage. Adding a healthy copy of the gene helps when the problem is too little protein; it does nothing for a faulty protein that has to be silenced, corrected, or helped to fold. So the first question about any PRPH2 variant is not "is it bad" but "what kind of bad".
The diseases PRPH2 causes
PRPH2 variants are diagnosed as pattern dystrophy, adult vitelliform macular dystrophy, central areolar choroidal dystrophy, cone-rod dystrophy or retinitis pigmentosa. Underneath, it is one problem: discs that are not built properly. The name describes which part of the retina is failing, and how fast. That is why the outlook differs from name to name even though the cause is shared.
Pattern dystrophy
Yellow-gray deposits in the macula arranged in butterfly, net-like or "pseudo-Stargardt" shapes. Often found on a routine exam with few symptoms. Central vision typically stays useful into the fifties or beyond, with a risk of late thinning (atrophy) or of a leaking blood vessel (see CNV (choroidal neovascularization): Abnormal new blood vessels growing under the retina, which leak and can damage vision quickly. It is the one complication of PRPH2 disease that can be treated with injections. Glossary). The commonest PRPH2 label in most studies.
Adult-onset vitelliform macular dystrophy (AVMD)
A single yellow "egg-yolk" spot under the center of the macula, usually appearing between 40 and 60. It can stay stable for years or gradually break down. Overlaps with pattern dystrophy; the same variant can produce either.
Central areolar choroidal dystrophy (CACD)
A sharply bounded patch of thinning in the center of the macula that grows slowly over decades, taking reading vision with it. Strongly linked to missense changes in the D2 loop such as R142W, R172W and R195L; in the largest study no truncating variant produced it. This is the structural name for the cone-dominant end of the spectrum.
Cone-rod dystrophy (CRD)
A label from the ERG (electroretinogram): A test that measures the retina's electrical response to flashes of light, with a small electrode on or near the eye. It tells whether rods, cones, or both are underperforming, before that shows on a scan. Glossary, the test that measures the retina's electrical response to light: cones fail first and more than rods. Loss of sharpness, color vision and light tolerance, later night vision. R172W is the classic PRPH2 cone-rod variant. A minority label overall, but the one most often applied to children.
Retinitis pigmentosa (RP)
Rods fail first: night blindness, then a shrinking side field, with central vision preserved longest. More common with truncating variants and in East Asian families (64% of Chinese PRPH2 families). Can occur alongside macular disease in the same family.
Early-onset severe disease
Rare, and not the situation of anyone whose report shows one altered copy. Two altered copies, one from each parent, produce severe disease from infancy. About one PRPH2 family in twenty in a national registry was of this recessive kind rather than dominant.
| Pseudo-Stargardt pattern dystrophy | 41% |
|---|---|
| Central areolar choroidal dystrophy (CACD) | 28% |
| Retinitis pigmentosa | 15% |
| Butterfly / vitelliform pattern dystrophy | 11% |
| Normal autofluorescence | 5% |
Sorted by what the retina looks like on fundus autofluorescence, most patients fall into pattern dystrophy or central areolar choroidal dystrophy (CACD). "Cone-rod dystrophy" is not a category on this axis at all. Median symptom onset in this cohort was 40, range 4 to 78; 6% of carriers had no symptoms.
The chart above shows why the labels vary. Sort the same 241 patients by what their retina looks like on an autofluorescence photograph and most are "pattern dystrophy" or "CACD". Sort them by what their photoreceptors are doing on an ERG and the groups change completely, and "cone-rod dystrophy" is a minority. Neither is wrong. When a clinician says "cone-rod dystrophy" about a child with a reduced cone ERG and a normal-looking macula, they are describing function; the description of structure is still to come.
What a variant can and cannot tell you
Studies of hundreds of patients show a tendency: the kind of variant leans toward one end of the spectrum or the other. It is a real pattern, and it is only a pattern.
Missense changes in the D2 loop lean toward the more serious, cone-involving, macular end (CACD on imaging, cone-rod dystrophy on ERG), with earlier and larger loss of central vision.
Truncating changes (premature stop, frameshift, most splice) lean toward the pattern-dystrophy end, with useful vision kept into adulthood.
Seen in the international, Spanish, Japanese and Chinese studies.
- Some carriers never get symptoms. About 6% in the largest study had none when examined, a few into their seventies (penetrance: How often people who carry a variant actually develop the disease. If penetrance is incomplete, some carriers never show symptoms. In PRPH2 about 6% of carriers in the largest study had none. Glossary).
- Severity varies inside families. The classic R172W family had members with severe cone-rod dystrophy and members with normal vision. One relative can have pattern dystrophy and another RP from the same variant.
- Position is not mechanism. Some D2-loop changes (R220Q among them) do not show the usual signs of a harmful protein when tested in the laboratory, so where a change sits does not settle what it does.
- Other genes move the outcome. Changes in modifier gene: A second gene whose variants change how severe a disease is. For PRPH2 the known ones are ROM1 and ABCA4. Glossary such as ROM1 and ABCA4, and factors nobody has named, shift severity in both directions.
So what is a variant good for?
It tells you which end of the range is more likely, not where on it you will land. Three things narrow it down, and all three are things a family can obtain: a careful baseline (an OCT (optical coherence tomography): A painless scan that takes cross-section pictures of the retina's layers, like an ultrasound with light. The main tool for tracking structural change. Glossary scan, an FAF (fundus autofluorescence): A photograph of the retina taken with blue light that makes waste pigment glow. Patterns of bright and dark tell where the retinal pigment epithelium is stressed or lost. Glossary photograph, an ERG, ideally microperimetry: A test that maps how sensitive each small spot of the central retina is, by showing tiny lights while tracking the eye. Glossary) so that change over time can be measured rather than guessed; a test of the modifier genes; and, when two PRPH2 variants turn up in one person, whether they are on the same copy of the gene or on opposite copies (cis / trans: When someone has two variants in the same gene: cis means both are on the same copy (the other copy is fine), trans means one on each copy (no fully normal copy). It matters for prognosis and for some treatments. Glossary), which decides whether this is one altered copy or two.
A variant is also what makes you findable for research. Almost every treatment in development is matched to a kind of damage, and some need the exact DNA sequence. A resolved variant with a mechanism attached is what lets a trial find you when one opens.
What to expect, in broad strokes
Nearly all published numbers come from adults, because that is who has been studied; a child with an early diagnosis is at the rare end of the range and the numbers below apply to them only loosely. In the largest study the median age when symptoms began was 40, with a range from 4 to 78. Pattern dystrophy found late in life usually stays mild for years, with the leaking-vessel complication as the main thing to watch. CACD and cone-rod dystrophy tend to begin earlier and to take central vision over a decade or more, while side vision and mobility are usually kept. RP takes side and night vision first. A carrier with no symptoms has a real chance of staying that way, and should still be examined so that a baseline exists. None of this is a schedule for any one person; it is what the population of PRPH2 patients has looked like.
The Explore page applies the pattern to any variant you give it, and says out loud where it is inferring from the kind of change rather than reporting from data about that exact variant.
Relatives, testing and children
Who might carry it
If the disease is dominant, each parent, sibling and child of a carrier has up to a 1 in 2 chance of carrying the same variant. Testing relatives for a known family variant is simpler and cheaper than the original search: the laboratory looks only at that one spot. It is done on a blood sample or a cheek swab, arranged through a genetics clinic or an eye-genetics service. Some programs, listed on the Resources page, offer it free.
Why testing relatives helps everyone
Finding the variant in relatives who have the disease, and not in those who do not, is called segregation evidence. It is one of the strongest ways a "variant of uncertain significance" gets reclassified. So testing an affected parent or sibling can settle a question about your own report.
Adult children
An adult can decide for themselves whether to be tested. Reasons to test: a baseline eye exam can be arranged, avoidable risks (high-dose vitamin A, for instance) can be avoided, and a positive result makes them findable for research. Reasons some choose to wait: there is no treatment yet to start, and a result cannot be un-known. A genetic counselor exists to help with exactly this decision, and the consultation is usually covered.
Where it came from
Usually from a parent, who may have mild or no symptoms and not know they carry it. Sometimes the change arose new in the child (a "de novo" variant), in which case neither parent carries it and the chance of it happening again in a sibling is very small. Testing both parents is how you tell. Either way, nobody chose it and nothing anyone did caused it; a DNA copying error is not a fault.
Young children
Children of a carrier are commonly examined by an eye doctor from an early age even without a genetic test, because the ERG and imaging can show changes before the child notices anything. Genetic testing of a child who has no symptoms is a decision to make with a counselor; it becomes clearly useful once a child has symptoms, when a trial would need it, or when a treatment exists.
These are the general rules for a dominant condition. The one-in-twenty families where PRPH2 disease is recessive follow different rules, and a genetic counselor can tell you which applies to yours.
Diagnosis, monitoring and daily life
There is no approved treatment yet. Monitoring, symptom management and being ready for a trial still matter a great deal.
Abnormal blood vessels can grow under the macula (CNV (choroidal neovascularization): Abnormal new blood vessels growing under the retina, which leak and can damage vision quickly. It is the one complication of PRPH2 disease that can be treated with injections. Glossary) in roughly one PRPH2 patient in six. It is the one complication that a drug treats well, and it is treated best in the first days. Call your retina specialist the same day if you notice:
- a sudden worsening of central vision in one eye
- straight lines that look wavy or bent
- a new dark or blank spot in the center of vision
- familiar objects or faces looking distorted
- OCT (optical coherence tomography): a painless scan that shows the retina's layers in cross-section, like an ultrasound made with light. Tracks structure.
- FAF (fundus autofluorescence): a photograph taken with blue light that makes stressed areas of the retina glow. Tracks the spread of damage.
- ERG (electroretinogram): measures the retina's electrical response to flashes, with a small sensor on or near the eye. Tells whether rods, cones or both are struggling, often before a scan shows it.
- Microperimetry: maps how sensitive each small spot of the central retina is. Tracks function.
What exists today
The honest version: as of September 2026 there is no approved treatment that changes the course of PRPH2 retinal disease. Care is monitoring, symptom management, low-vision rehabilitation, and being ready for a trial. That is not second best. For most people with this disease it is the care.
Confirm the diagnosis genetically
PRPH2 is the most common look-alike of (a "phenocopy" of) ABCA4 Stargardt disease and is regularly mislabeled as Stargardt or age-related macular degeneration on a chart. Genetic confirmation wins over the imaging label. If first-line testing came back empty, ask for a broader panel: in a 2026 Irish registry of macular and cone dystrophy patients, 59% were unresolved after first-line testing, and PRPH2 was the single most common gene found on the second pass.
Get a proper baseline
A high-density OCT volume scan with total macular volume (the volume of retinal tissue in the central area, measured by OCT) computed and the raw data stored; fundus autofluorescence with a reduced-illumination protocol; a full-field ERG; and best-corrected acuity with the chart recorded. Ask explicitly for ellipsoid zone thickness in microns (the ellipsoid zone is the band on the OCT scan that corresponds to the photoreceptors' energy-producing region), since a thickened ellipsoid zone was present in 34 of 36 patients in a 2025 Oxford series and may be an early structural sign. Progression can only be measured against a baseline, and progression data is what every trial and every individualized-therapy program asks for.
Refraction, every time
Uncorrected refractive error compounds retinal vision loss. Best-corrected acuity should be confirmed at every visit, and high myopia deserves its own attention: cone-dominant inherited retinal disease phenotypes averaged about -8 diopters in a 2025 series, plausibly because early cone dysfunction degrades the foveal image and drives eye growth.
Photophobia management
The highest-yield daily intervention in cone-involving disease. Damaged cones plus rods saturated by bright light make bright environments actively disabling, not just uncomfortable. Outdoors: 100% UV-A and UV-B blocking lenses (polycarbonate is inherently UV-blocking) plus a wide-brimmed hat. Indoors: warm lighting, glare control, and permission to wear tinted lenses in school or at work written into an accommodation plan.
Tinted lenses, staged to disease state
Red, orange, amber and plum filters (NoIR 16, 65 and 14 are common starting points) cut the short wavelengths that overstimulate surviving rods and improve contrast. The trade-off is color discrimination, so do not start color-filtering tints prophylactically in someone whose color vision is still normal; photochromic lenses are the right early choice. Get a filter trial at a low-vision clinic rather than buying blind, because responses differ from person to person.
Low-vision rehabilitation, started early
Magnifiers, video magnifiers, accessible technology, orientation and mobility training, and for children a teacher of the visually impaired and an IEP or 504 plan (the written plans a school uses to provide support and accommodations). The literature is consistent that rehabilitation should run in parallel with medical care from the start, not begin after medical options are exhausted.
The one treatable complication
Choroidal neovascularization, abnormal blood vessels growing under the retina, occurs in roughly 17% of PRPH2 cohorts. It is the one acute event in this disease that a drug fixes, and catching it in days rather than weeks changes the outcome.
Anti-VEGF injections work
The same drugs used for wet macular degeneration (bevacizumab, ranibizumab, aflibercept, faricimab) injected into the eye under topical anesthesia stop the leakage and often reverse the vision loss when caught early. This is standard of care for CNV regardless of the underlying gene.
Know the warning signs
Sudden worsening of central vision, straight lines that look wavy or bent, a new dark spot in the center of vision, or familiar objects looking distorted. Any of these means calling the retina specialist that day, not waiting for the next scheduled visit.
Surveillance
OCT angiography (an OCT scan that also maps blood flow) at annual visits is the non-invasive tool for spotting neovascularization before it causes symptoms. Ask for it as part of the standard imaging set.
Monitoring
Track structure and function in parallel, and lean on the metrics that have been shown to move reliably. Default cadence is a comprehensive reassessment every 12 months, tightened to 6 if imaging shows rapid change and loosened to 18 to 24 if everything is stable.
OCT: total macular volume is the best progression marker
In the Perth longitudinal cohort (12 patients over about 5 years), total macular volume (the volume of retinal tissue in the central area, measured by OCT) fell 0.071 cubic millimeters per year with tight agreement between the two eyes, and was the most reliable objective metric. Not every clinic computes it by default; ask for the volume scan, the calculation, and the raw data.
Fundus autofluorescence: the earliest detector
FAF images lipofuscin in the retinal pigment epithelium and catches PRPH2 changes before OCT or acuity move. Decreased-autofluorescence area grew about 0.10 millimeters per year in the Perth cohort. Ask for a reduced-illumination protocol (about 25%) with longer excitation wavelengths to limit phototoxicity, green-channel FAF in addition to blue if the center has it, and ultrawide-field FAF for tracking lesion area over time.
ERG: the functional axis
Electroretinography tells you whether cones, rods, or both are failing and in which order, which is a different question from what the retina looks like. One FAF appearance can correspond to several ERG phenotypes. Repeat every 2 to 3 years unless something specific prompts it sooner.
Microperimetry: useful, but noisy
Macular sensitivity fell about 0.10 dB per year in the Perth cohort, but learning effects made it unreliable for tracking change in individuals. Treat it as supporting evidence, not the anchor.
Document for trial eligibility
Genotype in standard genetic notation (the c. and p. forms on your report), serial OCT volumes as raw data, serial FAF, ERG amplitudes and implicit times (how quickly the retina responds), acuity with the chart used, refraction, and every medication or supplement. Registry enrollment (My Retina Tracker, Uni-Rare) is the structured place this data lives so a sponsor can find you.
What to avoid
A short list, and every item on it is something a well-meaning person will suggest.
High-dose vitamin A palmitate
The 1993 Berson recommendation of 15,000 IU per day was for adults with typical RP, was re-analyzed in 2018 and found not to slow vision loss, and was rolled back by the American Academy of Ophthalmology's 2024 Ophthalmic Technology Assessment, which found no clear evidence that any dietary supplement helps RP. For cone-rod, Stargardt-like and PRPH2 phenotypes specifically, vitamin A is not endorsed and may be actively harmful: in ABCA4 disease it speeds up the build-up of a toxic vitamin A by-product (bisretinoid). The September 2026 gildeuretinol result in Stargardt does not change this; that drug works by producing less of the toxic vitamin A dimer, and no retail supplement approximates it.
Vitamin E
The 2024 AAO assessment actively recommends avoiding vitamin E supplementation in inherited retinal disease.
Repeated low-level red-light therapy for myopia
Independent radiometry published in JAMA Ophthalmology (March 2026) found two commercial devices reached the retinal exposure safety limit within seconds of a 180-second protocol, and the documented harms are cone and macular harms. That safety margin was measured in normal retinas; it has no reason to transfer to a retina with a structural outer-segment defect. China has reclassified these devices as Class III. The finding was publicly contested in August 2026 and survived, with an undisclosed industry conflict on the challengers' side.
Valproic acid and off-label NAC
Valproic acid failed its Phase 2 trial in dominant RP and carries liver, pancreas and pregnancy risks. N-acetylcysteine is in a Phase 3 trial for adults with RP that has not read out; the trial dose (1800 mg twice daily) causes GI side effects, there is no pediatric data, and supplement-strength doses are not the trial.
Unregulated stem cell clinics and off-label gene therapy
Real cell therapy happens in trials approved by an ethics board (an "IRB") at academic centers. Clinics selling stem cell injections for retinal disease are predatory and have blinded people. Luxturna is for biallelic RPE65 disease; any clinic offering it for PRPH2 is committing fraud.
Smoking and secondhand smoke
Depletes macular lutein and zeaxanthin and thins the choroid, including in children. A smoke-free home is one of the few environmental variables fully in a family's control.
Supplements with a story but no data
Lutein and zeaxanthin, omega-3 and DHA, and AREDS2 formulations are safe and plausible from food but none is proven to slow PRPH2 disease; the DHAX trial in X-linked RP was negative on its primary endpoint and AREDS was designed for intermediate age-related macular degeneration. A Mediterranean-pattern diet rich in leafy greens, egg yolks and fatty fish is the reasonable dietary strategy, with essentially no downside.
A note on myopia (mainly relevant to children and young adults who are short-sighted)
Cone-involving retinal disease and high myopia travel together more often than chance, and myopia is the one partially modifiable retinal risk in the picture.
The risk is the eye length, not the prescription
High myopia at -6 diopters or worse (diopters measure the strength of a glasses prescription; -6 is strongly short-sighted) carries roughly 20 times the lifetime risk of retinal detachment, plus higher risk of glaucoma and early cataract, and those risks come from the eyeball growing too long from front to back (axial elongation) rather than the refractive number. Ask whether anyone is measuring the eyeball's front-to-back length (axial length), not just refraction.
What the evidence supports
At least two hours a day outdoors is the one free, no-risk recommendation in the April 2026 AAPOS statement, and it is compatible with photophobia management: outdoor time with photochromic lenses and a hat, not less outdoor time. Low-dose atropine is standard for children in general, but the only study in children with inherited retinal disease (April 2026, n=27) found no significant effect once age was accounted for. Orthokeratology carries corneal infection risk that stacks a second source of vision loss on top of the first, and red-light therapy is a hard no.
Current as of 2026-09-16. Sources are listed on the About page.
Where treatment stands
The plain state of things, as of 2026-09-16: no treatment for PRPH2 disease is approved, and no PRPH2-specific trial is enrolling patients. One PRPH2-specific treatment has been given to a single person. Everything below is either at the laboratory stage or is a treatment for the retina in general that PRPH2 patients could one day qualify for. That is hard to read, and it is also a field with more money, more groups and more approaches aimed at this gene than at any point before.
PRPH2 is not an ultra-rare disease: estimates run from several thousand to over twenty thousand patients in the United States. What has held programs back is not the number of patients but the slow, uneven pace of the disease, which makes a treatment effect hard to measure inside a trial. That is a solvable problem, and solving it is what natural history study: A study that follows people with a disease over time without treating them, to measure how it progresses. Trials need this data to prove a treatment works. Glossary are for.
| Modality | Preclinical | Phase 1 | Phase 2 | Phase 3 | Filed | Approved |
|---|---|---|---|---|---|---|
| antisense oligonucleotide | ||||||
| gene-agnostic gene therapy | ||||||
| optogenetics | ||||||
| neuroprotection | ||||||
| oral small molecule | ||||||
| RNA exon editing | ||||||
| gene augmentation | ||||||
| knockdown-and-replace | ||||||
| genome editing | ||||||
| small-molecule corrector | ||||||
| cell therapy | ||||||
| retinal prosthesis |
Recruiting PRPH2-specific trials: 0. Verified against the ClinicalTrials.gov API monthly from April through September 2026: no PRPH2-specific trial is recruiting anywhere. Exactly one PRPH2 interventional trial exists in any status, the n-Lorem single-patient ASO for G208D (NCT07177196), which is active and not recruiting with no results posted. The realistic enrollments for a PRPH2 patient today are observational: the Uni-Rare natural history study (NCT05589714) and the NEI genotyping network (NCT06491615).
The same programs as a list
- n-Lorem personalized ASO for PRPH2 G208D (n-Lorem Foundation with UCSD (Borooah)) · antisense oligonucleotide · Phase 1 trial · PRPH2-specific. The first and only PRPH2-targeted therapy ever given to a human. It is built for one patient's exact variant, so it is not a drug other people can take, but it proves the regulatory and manufacturing path for a PRPH2 ASO exists, and n-Lorem's model is repeatable for other tractable variants.
- OCU400 (AAV5-hNR2E3) (Ocugen) · gene-agnostic gene therapy · Phase 3 trial. The most plausible near-term gene therapy that a PRPH2 patient could in principle receive, because it does not depend on which gene is broken. Phase 1/2 showed about a two-line low-luminance acuity gain sustained at three years across several genotypes. The Phase 3 enrolled children from age 3.
- MCO-010 (sonpiretigene isteparvovec) (Nanoscope Therapeutics) · optogenetics · filed for approval. Gene-agnostic vision restoration for people who have already lost most of their photoreceptors; cone-rod and macular dystrophies are inside its orphan designation. Phase 2b RESTORE met its primary endpoint and gains of about three lines held through three years.
- RTx-015 (Ray Therapeutics) · optogenetics · Phase 1 trial. A second, independent optogenetic shot at gene-agnostic vision restoration, one cell layer downstream of MCO-010. First efficacy signal disclosed June 2026: vision gain across all dose levels in advanced RP.
- SPVN06 (RdCVF / RdCVFL) (SparingVision) · neuroprotection · Phase 2 trial. When rods die they stop making the factor that keeps cones alive. SPVN06 replaces it directly, which fits PRPH2 biology well: D2-loop mutations hit rods hardest at the cellular level, and the cones that carry reading vision die second. Mutation-independent, so PRPH2 is a plausible expansion target.
- SPVN20 (GIRK cone reactivation) (SparingVision) · gene-agnostic gene therapy · Phase 1 trial. The only clinical program trying to restore light signalling inside the cone itself rather than protecting it or wiring an opsin into a downstream cell. It only works if cone cell bodies are still present, which is exactly the situation in most PRPH2 disease for a long time.
- NAC Attack (oral N-acetylcysteine) (Wilmer Eye Institute / Johns Hopkins, NEI-funded) · oral small molecule · Phase 3 trial. A cheap oral antioxidant aimed at the secondary cone death that follows rod loss. Gene-agnostic in principle, so it could apply to PRPH2 if it works, but the trial population is adults with RP.
- ACDN-01 (Ascidian Therapeutics) · RNA exon editing · Phase 1 trial. The first human test of a new modality: rewriting a block of exons at the RNA level without touching the genome. Conceptually well suited to a dose-sensitive, dominant-negative gene like PRPH2, but this product is for ABCA4 and its exclusions bar pathogenic variants in other cone-rod genes, so PRPH2 patients are not eligible.
- Non-viral DNA nanoparticle PRPH2 augmentation (Naash lab, University of Houston) · gene augmentation · laboratory stage · PRPH2-specific. The longest-duration rescue data for any PRPH2 therapy: structural rescue held to 15 months and ERG and acuity improved at 10 months in Prph2+/- mice. But that model is haploinsufficiency. When the same lab tried augmentation against the dominant-negative C213Y variant, structure improved and function did not, and they concluded that eliminating the mutant protein is a prerequisite for a cure in that class.
- Mutation-independent knockdown-and-replace (Al-Ubaidi and Naash, University of Houston (PARD award $481,269 to June 2027)) · knockdown-and-replace · laboratory stage · PRPH2-specific. The modality that works regardless of which variant a patient carries and does not need a spare healthy allele, so it survives the dominant-negative problem and the cis-versus-trans question. The lab's 2023 mouse comparison showed cones tolerate reduced peripherin-2 far better than rods, which suggests knockdown could sustain cone function longer.
- Rhodopsin downregulation (Rutan Woods, Naash and Al-Ubaidi, University of Houston) · antisense oligonucleotide · laboratory stage · PRPH2-specific. Does not touch PRPH2 at all. Less rhodopsin means less demand on broken disc-assembly machinery. In PRPH2 mutant mice, an intravitreal rhodopsin ASO improved disc ultrastructure and photoreceptor function. Mutation-agnostic, and an ASO avoids subretinal surgery.
- ROM1 antisense oligonucleotide (Tebbe lab, University of Houston (Knights Templar Eye Foundation, 2025)) · antisense oligonucleotide · laboratory stage · PRPH2-specific. In a PRPH2 mutant mouse, reducing ROM1 switched the phenotype from severe progressive vision loss to night blindness only. If an eye-drop ASO works, it changes the delivery paradigm for children entirely.
- Prime editing of PRPH2 variants (Quinn (PrimeSight Lab, Penn), Ryals (OHSU Casey), Palczewski (UC Irvine)) · genome editing · laboratory stage · PRPH2-specific. Prime editing repairs the DNA itself without double-strand breaks, so it is mutation-specific but phase-agnostic. The c.828 founder splice variants were the first correctable PRPH2 targets in cell culture (2025); as of September 2026 the field has moved into D2-loop missense variants.
- HITI knock-in of a healthy PRPH2 copy (Qin Liu, Harvard / Mass Eye and Ear (2026 PARD award)) · genome editing · laboratory stage · PRPH2-specific. The first PRPH2 grant aimed at directly correcting the gene in a variant-agnostic way: one reagent for any patient rather than one per mutation.
- Small-molecule folding correctors for PRPH2 (Imanishi (Indiana, PARD screen to March 2027); Comander (Mass Eye and Ear, variant map to March 2028); Octant Bio (rhodopsin precedent, OCT-980)) · small-molecule corrector · laboratory stage · PRPH2-specific. The only modality that would work the same way regardless of variant or phase, with no surgery and no vector. The cryo-EM structure shows nearly all high-risk PRPH2 mutations sit on the PRPH2-ROM1 dimer interface, which fits the mechanism. Both halves of the discovery method (a saturation variant map and a compound screen) are PARD-funded and running now.
- Luxturna (voretigene neparvovec) (Spark Therapeutics) · gene augmentation · approved. The first FDA-approved gene therapy for an inherited disease (2017) and the regulatory template any PRPH2 therapy would follow. Durable to 8 to 9 years in most patients. It will not work for PRPH2: it supplements a missing protein in a recessive disease, and most PRPH2 variants are dominant.
- Bota-vec (botaretigene sparoparvovec) (MeiraGTx) · gene augmentation · filed for approval. A regulatory precedent, not a PRPH2 option. Phase 3 LUMEOS missed its mobility primary endpoint but hit a broad slate of retinal-sensitivity and patient-reported secondaries; the sponsor reacquired the asset from J&J in April 2026, raised $400 million, and is filing anyway. How FDA responds sets the bar for how much functional-vision evidence is enough in a slowly progressive IRD, which is exactly the problem PRPH2 faces.
- Laru-zova (laruparetigene zovaparvovec) (Beacon Therapeutics) · gene augmentation · Phase 3 trial. The other RPGR program. Twelve-month Phase 2 DAWN data held (50 to 67% gained two or more lines of low-luminance acuity). A clean XLRP approval would make X-linked RP the second IRD with an approved gene therapy and further de-risk the FDA and manufacturing path.
- ASP2020 (Astellas) · cell therapy · Phase 1 trial. The first industry protocol whose genotype criterion PRPH2 patients meet: it enrolls STGD-like macular dystrophy with one or more pathogenic variants in any gene known to cause macular dystrophy, not just ABCA4, and doses from age 6. That is a new and durable enrollment category, even though the trial itself is built for eyes with established macular atrophy and acuity between 20/40 and 20/500.
- Gildeuretinol (ALK-001) (Alkeus Pharmaceuticals) · oral small molecule · Phase 3 trial. The first oral drug to slow Stargardt lesion growth (21.6% versus placebo over 24 months, JAMA Ophthalmology, September 3, 2026). It works by making less of the toxic vitamin A dimer, which confirms the mechanism that makes ordinary vitamin A supplementation dangerous in this disease class. It is not a reason to take vitamin A; it is the opposite.
- EDIT-101 (BRILLIANCE) (Editas Medicine) · genome editing · laboratory stage. The first in-body CRISPR therapy for an inherited retinal disease: safe in 14 patients including two children, measurable improvement in 11 (NEJM 2024). Then the company shelved it because the target population was too small to be commercial. That tension, a therapy that works and a company that walks away, is the central risk for any variant-specific PRPH2 therapy.
- iPSC-derived photoreceptor cell therapy (OpCT-001, CLARICO) (BlueRock Therapeutics and others) · cell therapy · Phase 1 trial. Gene-agnostic replacement of photoreceptors that have already died, explicitly aimed at RP and cone-rod dystrophy. Theoretically the most direct therapy for end-stage PRPH2 disease, and useless for photoreceptors that are still alive.
- PRIMA photovoltaic subretinal implant (Science Corporation (formerly Pixium Vision)) · retinal prosthesis · Phase 3 trial. The backstop. In a 38-participant trial for geographic atrophy (October 2024), patients gained an average of 4.6 lines on the eye chart. Only relevant for end-stage disease with no remaining photoreceptor function, and worth knowing about precisely because it exists.
What the approaches are
- gene-agnostic: A treatment that works regardless of which gene is faulty, because it targets the retina's response to damage rather than the specific cause. Glossary treatments do not care which gene is faulty. OCU400, a gene therapy that resets how photoreceptors respond to stress, has enrolled children from age three in its final-stage trial and could be filed for approval in 2027. Optogenetic treatments give some light sensitivity back to eyes that have already lost most photoreceptors, so they are for late disease. Neuroprotection trials try to slow loss whatever the cause.
- Silencing the faulty copy (antisense oligonucleotide (ASO): A short synthetic piece of genetic code that binds to a gene's message and can switch it down. Given by injection into the eye. The first PRPH2-specific treatment tried in a person is an ASO. Glossary, or "knockdown-and-replace") is the strategy for variants that make a harmful protein. The only PRPH2-specific treatment ever given to a person is one of these: a made-to-order ASO for a single G208D patient, built by the n-Lorem Foundation and given at UC San Diego from August 2025.
- Adding a healthy copy (gene therapy: Delivering a working copy of a gene into cells, usually with a harmless virus (AAV) injected under the retina. It helps when the problem is a missing protein; it does not remove a harmful one. Glossary) suits variants that make too little protein. It has worked in mice with one PRPH2 copy missing.
- Correcting the DNA (prime editing: A precise form of gene editing that rewrites one variant back to the usual sequence inside the cell. Tested in cells and animals; not yet in people for PRPH2. Glossary) has fixed the c.828 splice variants in patient-derived cells, and is being developed further.
- A pill that helps the protein fold (a small-molecule corrector) is being searched for with foundation money, alongside a laboratory map of which of the roughly 6,000 possible PRPH2 changes actually break the protein.
None of this is available to patients yet. It is a real pipeline with a timeline measured in years, and it is the reason that being enrolled in a registry, imaged at baseline and findable now is the most useful thing a family can do today.
Organizations and support
Foundation Fighting Blindness (FFB)
Founded in 1971, FFB is the largest non-governmental funder of inherited retinal disease research in the world, and the connective tissue of the field. For PRPH2 specifically it runs the dedicated PRPH2 and Associated Retinal Diseases program with the Nixon Visions Foundation: eight research awards of up to $500,000 each since 2023, and the only PRPH2-specific workshops ever held (March 2023 and September 2026, both in San Diego). It also runs the My Retina Tracker registry, a Peer Guide program that pairs newly diagnosed people with a trained mentor who has the same condition, roughly 50 local chapters, free webinars, and a phone line staffed by information specialists.
My Retina Tracker Registry and free genetic testing
FFB's patient registry is the de facto national database that trial sponsors search when a study opens, and enrollment also connects you to a free 285-gene inherited retinal disease panel (run with PreventionGenetics) that includes genetic counseling. As of FFB's March 2026 forum the registry held about 44,000 participants, more than 31,000 of them genotyped. Being in it does not put you in a trial automatically. It makes you findable, which is the step most families miss.
Nixon Visions Foundation
The patient-family foundation that has done more than any other organization to fund PRPH2-specific research. Janine Nixon noticed changes in her vision for more than 15 years before genetic testing in 2020 found a PRPH2 variant; she and Brandon Nixon then partnered with FFB to build the PRPH2 and Associated Retinal Degenerations program. The foundation set up the Macular Dystrophy-PRPH2 Research Fund at UC San Diego's Shiley Eye Institute in 2021 (the Borooah and Ayyagari labs), has funded six of the eight PARD Science Awards since 2023, and co-organized both PRPH2 workshops with FFB and Shiley. Those workshops are where many PRPH2 families have met each other in person.
n-Lorem Foundation
A nonprofit that designs a custom antisense oligonucleotide (ASO) for one patient's specific mutation and provides it free for life. It sponsors the only PRPH2-targeted therapy ever given to a human: an ASO for the G208D variant, dosed at UCSD since August 2025 (NCT07177196). Its 2026 Nucleic Acids Research report describes more than 35 patients treated with no ASO-related serious adverse events. Two things to know before you get your hopes up: only a research physician at a major referral hospital can submit an application, and the mutation has to be ASO-tractable, which for a dominant gene usually means an allele-selective design that spares the healthy copy.
Nome
A company that sells the coordination of an individualized genetic medicine: matching a mutation to a modality, then managing the labs, manufacturers, and regulators. Its founder, Stevie Ringel, is an inherited retinal disease patient himself (a KIZ variant with fewer than 200 patients worldwide) who previously ran precision-medicine units at Tempus and GeneDx. The relevant difference from n-Lorem is that Nome takes intake directly from families and the eligibility assessment is free. The honest caveat: as of September 2026 it is a very early-stage company with no dosed patient, no published pricing, and no PRPH2 program.
Uni-Rare natural history study (NCT05589714)
FFB's largest natural history study, run by the Jaeb Center through about 40 FFB Clinical Consortium sites worldwide, recruiting since May 2023 with primary completion estimated December 2029. It is two studies in one: a one-visit registry tier (one visit, then annual phone calls, capped at 100 participants per gene) and a repeat-visit natural-history tier that opens gene by gene based on funding and registry enrollment. For a dominant gene like PRPH2, a single disease-causing variant qualifies. Why it matters: the missing piece for any PRPH2 trial is progression data, and the FDA's 2026 rare-disease pathways require natural history data in untreated patients. Registry enrollment is the lever families can pull for free.