Executive Summary

ER-100 has reached an important translational milestone. On October 8, 2026, Life Biosciences reported preliminary Phase 1 findings from three participants with open-angle glaucoma who received the first dose level of its investigational epigenetic reprogramming gene therapy. The therapy was reportedly well tolerated through Day 56, with no dose-limiting toxicities, serious adverse events or adverse events of special interest. The Data Safety Monitoring Board recommended escalation to the next dose level.

Two of the three participants demonstrated improvements in Humphrey visual-field sensitivity. In each participant, at least 28% of tested locations improved by more than 4 decibels relative to baseline. These observations are potentially relevant because ER-100 is intended to restore function in damaged but surviving retinal ganglion cells, rather than simply lower intraocular pressure.

The preclinical program demonstrated large improvements in retinal electrophysiology after experimentally induced optic nerve injury. The human observations are directionally consistent with recovery of neuronal function, but the comparison is indirect: the nonhuman primate results were measured primarily by pattern electroretinography, whereas the human efficacy observations disclosed so far concern psychophysical visual-field testing. Human electrophysiologic data, global visual-field changes and structural outcomes have not been disclosed.

A pointwise visual-field improvement exceeding 4 decibels could be clinically meaningful if reproducible, but such changes must be assessed against substantial test-retest variability and learning effects associated with automated perimetry in glaucoma. The current study is open-label, includes only three participants at the first dose level, and was designed principally to characterize safety and tolerability.

The critical next question is whether the early visual-field findings persist and are corroborated by global perimetric indices, retinal electrophysiology, independent visual-function measures and structural assessments. If so, ER-100 could provide human evidence supporting the proposition that some apparently irreversible vision loss reflects recoverable dysfunction in surviving neurons. The present findings do not yet establish neurorestoration or clinical efficacy.

ER-100 Moves Beyond Safety Into Preliminary Human Efficacy

ER-100 is an investigational adeno-associated virus serotype 2 (AAV2) gene therapy designed to produce controlled expression of OCT4, SOX2 and KLF4, collectively designated OSK. Its proposed mechanism is to alter the biological state of injured retinal ganglion cells (RGCs), thereby preserving or potentially restoring their function. The therapy is administered through a single intravitreal injection and followed by 56 days of systemic doxycycline to induce OSK expression. 12

The program is being evaluated in two optic neuropathies: open-angle glaucoma (OAG), a chronic neurodegenerative disease, and non-arteritic anterior ischemic optic neuropathy (NAION), an acute ischemic injury. The registered first-in-human Phase 1 study, NCT07290244, employs an open-label, sequential-cohort design, with dose evaluation in OAG preceding assessment in NAION. 1

Life Biosciences' October 8, 2026, disclosure concerned the first three OAG participants at the initial dose. The company reported no dose-limiting toxicities, serious adverse events or adverse events of special interest through Day 56. It characterized reported adverse events as associated with the procedure, steroid treatment or pre-existing conditions, rather than directly with ER-100. A prednisolone taper is used to mitigate inflammatory risk associated with ocular AAV2 administration. 2

Progression to a dose three times higher is an important development milestone for a first-in-human therapy based on inducible expression of reprogramming-associated transcription factors. However, an absence of major adverse events in three individuals over 56 days cannot establish longer-term safety of a potentially persistent gene therapy vector. 23

The exploratory visual-field findings represent a change in the evidence available for the program. Until this disclosure, the most compelling efficacy observations were derived from animal models. The new human findings permit an initial, albeit limited, assessment of whether the predicted functional recovery may be observed in patients.

ER-100: From Primate Rescue to the First Human Visual-Field Signal

Click to enlarge

What the Preclinical Program Led Us to Expect

Regulatory Impact previously examined the preclinical rationale and regulatory pathway in ER-100 Enters the Clinic: What FDA Will Need to See From Partial Epigenetic Reprogramming. Published on August 13, 2026, that analysis emphasized the need to connect improvements in retinal electrophysiology to durable and clinically meaningful human visual function. The October 2026 preliminary visual-field observations begin to address that anticipated translational question, but they do not yet establish a treatment effect. 42

The underlying therapeutic hypothesis is that optic nerve injury leaves a heterogeneous neuronal population rather than an immediate binary distinction between healthy and dead cells. Some retinal ganglion cells may remain anatomically present but function below their normal capacity. A therapy acting on their cellular state could potentially improve visual function without replacing neurons that have been eliminated. 567

The nonhuman primate evidence was particularly noteworthy because it did not rely exclusively on epigenetic biomarkers. In a laser-induced NAION-like model, investigators measured pattern electroretinography (pERG), a retinal electrical response with substantial contributions from RGC function. A disclosed post-injury treatment experiment reported a P50-N95 amplitude of 4.60 ± 0.24 µV after OSK treatment versus 2.89 ± 0.79 µV with vehicle at five weeks after injury (P = 0.039). That corresponds to approximately 59% higher amplitude in the treated group. 8

Separate characterization of the primate injury model described a reduction from approximately 6.9 µV before injury to approximately 2.6 µV afterward. Using the historical pre-injury amplitude as an illustrative reference, the increase from 2.89 µV with vehicle to 4.60 µV after OSK could represent recovery of approximately 43% of the injury-associated electrophysiologic deficit. This combines measurements from related but separate experimental datasets; it is a heuristic, not a formal treatment-effect estimate. 98

Investigators also reported, at the 2026 Association for Research in Vision and Ophthalmology meeting, perifoveal expression of OSK in retinal cells, partial reversal of injury-associated DNA methylation changes, pERG improvement of approximately 2 to 3 µV, and mitigation of retinal structural damage. These findings strengthened the proposed link between altered cellular state, retinal physiology and preservation of injured tissue. 10

Taken together, the preclinical program suggested three testable propositions: OSK can induce biological changes in injured retinal cells; those changes can coincide with improved neuronal physiology; and neuronal recovery may ultimately improve functional vision. The last proposition was the least directly established before human testing, because improvements in electrophysiology cannot be mapped directly onto what a patient is able to see. The October 2026 findings are relevant to that third proposition but do not establish the sequence of events linking treatment to visual recovery.

How Meaningful Is a Visual-Field Improvement Greater Than 4 dB?

Humphrey visual-field testing measures the light-detection threshold at multiple locations while a participant maintains fixation. Sensitivity is expressed in decibels (dB), on a logarithmic scale. A 4 dB increase corresponds to a threshold stimulus intensity approximately 2.5 times lower than before, assuming otherwise comparable testing conditions. Such a change at a reproducibly impaired location could represent a physiologically important gain in light-detection sensitivity, but does not represent a 2.5-fold improvement in overall vision. 12

Similarly, the company's statement that at least 28% of visual-field points improved by more than 4 dB does not mean that 28% of lost vision was restored. The registered study uses the Swedish Interactive Thresholding Algorithm (SITA) Standard 24-2 protocol in OAG, conventionally reporting 52 visual-field locations after excluding the blind-spot locations. Under that conventional point count, 28% corresponds to roughly 15 locations. The precise distribution, baseline thresholds, clinical importance and performance of the remaining test locations were not disclosed. 12

A coherent cluster of improving locations that corresponds to a pre-existing glaucomatous defect would be more persuasive than scattered changes across highly variable locations. The location of improvement also matters clinically: central or paracentral changes may have different functional implications from peripheral changes. No participant-level visual-field plots were provided in the October announcement. 2

The distinction between pointwise change and global mean deviation is important. The registered Phase 1 outcome description identifies change in visual-field mean deviation, whereas the announcement emphasized the percentage of individual points exceeding a specified improvement threshold. Improvement in a subset of points need not produce a clinically meaningful change in the visual field as a whole, particularly if other locations deteriorate or remain unchanged. The mean-deviation values and complete pointwise distributions have not been disclosed. 12

The Central Limitation: Automated Perimetry Is Variable

Automated perimetry is susceptible to test-retest variability, particularly in eyes with established glaucomatous visual-field loss. In a retrospective study involving test-retest fields from 4,044 glaucomatous eyes, Choi and colleagues reported a mean root mean square difference of 4.39 ± 2.55 dB across the 52 visual-field locations. This is a global measure of aggregate pointwise test-retest variation, not a significance threshold for any single location. Nevertheless, it demonstrates why a pointwise improvement of several decibels cannot, by itself, establish a treatment effect. 11

Perimetric learning is another relevant consideration. Heijl and Bengtsson prospectively tested previously inexperienced glaucoma patients and reported a mean 2.81 dB improvement in mean deviation between the first and second sessions. The effect was smaller on subsequent tests. ER-100 participants with longstanding glaucoma may already have substantial perimetric experience, so the magnitude observed in this historical study cannot simply be transferred to the Phase 1 cohort. The finding nevertheless illustrates the importance of multiple reliable baselines and repeated follow-up assessments. 12

The current release does not provide false-positive rates, fixation-loss indices, false-negative rates, retest confirmation, the number of baseline examinations or prior perimetry experience. Without those details, the finding is most accurately described as a preliminary pointwise visual-field signal with unresolved reproducibility and treatment attribution. 2

There is also uncertainty about analysis selection. The company did not specify whether the greater-than-4-dB threshold and the 28% criterion were prespecified, or how the result relates to the protocol's intended visual-field analyses. The selection of an exploratory cutpoint after inspecting a small dataset can exaggerate apparent evidence of efficacy; it should not be treated as a prospectively validated responder definition. 21

Does the Human Signal Match the Expected Preclinical Effect?

The preclinical and human findings are best compared as successive, nonidentical links in a translational hypothesis rather than as numerically interchangeable effect sizes. A 2 to 3 µV improvement in pERG cannot be converted into a particular decibel improvement in Humphrey visual-field sensitivity, nor vice versa. The tests interrogate different aspects of retinal and visual-system function, with different sources of variability. 138102

The strongest feature of the early human signal is that it was observed in a test of visual function, rather than solely as a molecular biomarker. If confirmed, such a measure could be closer to the eventual clinical objective than changes in DNA methylation alone. Yet the absence of pERG results leaves the proposed chain incomplete: induced OSK expression, modified neuronal state, improved retinal physiology and improved visual performance have not all been documented in the same treated participants. 21

The indication distinction also matters. Much of the strongest primate evidence arose from acute experimental optic nerve ischemia, whereas the initial human cohort involved chronic glaucoma. Glaucoma can involve long-standing and heterogeneous neuronal loss, and the amount of viable, dysfunctional neuronal substrate may differ substantially between eyes. The human observations are therefore suggestive but do not establish that the magnitude observed after acute experimental injury can be reproduced in chronic human disease. 812

The Therapeutic Substrate: Surviving but Dysfunctional Neurons

The concept of recovering visual function does not require replacement of retinal ganglion cells that have died. A more plausible substrate is a population of surviving neurons with impaired electrophysiologic activity and retained, at least partial, connectivity to downstream visual pathways. RGC dysfunction can precede detectable structural loss, and studies have reported changes in retinal electrophysiology following intraocular-pressure reduction in human glaucoma, supporting the broader principle that some dysfunction can be reversible. 56

ER-100 pursues that possibility through a mechanism different from conventional pressure reduction. Its biological effect requires living cells capable of expressing the delivered transcription factors. Consequently, the concept more naturally supports restoration of function in surviving cells, protection of threatened cells or preservation of viable axons than replacement of an eliminated neuronal population. 78

At Day 56, an improvement in visual-field sensitivity would not independently demonstrate axonal regeneration, retinal neuron replacement or reversal of cellular age. Even a confirmed functional response could reflect several mechanisms, including relief of neuronal dysfunction, effects on cellular stress pathways or changes unrelated to ER-100. Molecular target-engagement, electrophysiologic and structural data would be needed to discriminate among these explanations.

Patient selection may eventually depend on the balance between functional impairment and retained neuronal substrate. An eye with substantial measured visual loss but surviving cells may differ in treatment potential from an eye with profound anatomical loss. This remains a mechanistic hypothesis; no clinical biomarker has yet been validated to predict response to ER-100. 15

What the Initial Disclosure Does Not Establish

The registered study includes multiple ocular and visual assessments capable of linking the observed pointwise signal to broader measures of function and anatomy. The October 8 announcement did not provide the participant-level datasets required for that integrated analysis. 12

These omissions do not invalidate the reported observations, but they substantially limit their interpretation. A larger proportion of improved points is not necessarily stronger evidence of therapeutic benefit than a smaller but reproducible, anatomically consistent response accompanied by electrophysiologic improvement. The key is whether several independent measurements converge on a coherent treatment-associated change.

Why Day 112 and Longer-Term Follow-Up Matter

The Day 56 interim readout coincides with the planned period of systemic doxycycline administration. The registered trial includes assessments at Day 112 and longer-term follow-up. Evaluating whether functional changes persist after induction ends is particularly important for a therapy designed to cause a durable change in cellular state. 12

Sustained visual-field improvement after cessation of doxycycline would be compatible with a persisting biological effect, although causality would still need to be demonstrated. Improvement that disappears after induction might indicate that continued transgene expression, a treatment-associated condition or another transient factor is necessary for the observed change. Failure to reproduce the initial finding would increase concern that test variability contributed to the signal.

Durability must also be evaluated alongside safety. The AAV vector may persist beyond the intended window of OSK activation, and delayed ocular inflammation, immunologic events, inappropriate cell-state changes or other adverse outcomes cannot be excluded by a short initial observation period. The existing long-term follow-up framework is therefore relevant to both benefit and risk. 31

Regulatory Implications: From a Safety Milestone to an Efficacy Strategy

ER-100 falls within the U.S. Food and Drug Administration's gene therapy framework administered by the Center for Biologics Evaluation and Research (CBER). The first cohort supports dose escalation, an important Phase 1 outcome, but the small uncontrolled dataset is not sufficient to establish efficacy or to define a registrational endpoint. 2114

In glaucoma, the prospective therapeutic claim differs from that of established intraocular-pressure-lowering agents. A claim centered on neuronal restoration would require evidence of clinically meaningful visual improvement beyond the effects of background glaucoma care. In a future controlled study, that distinction would need to be addressed through population selection, stable concomitant therapy, appropriate controls and a defensible visual-function endpoint strategy. 141

Human pERG could serve as an important supportive pharmacodynamic measure, given its use in the preclinical program, but improvement in pERG alone is not interchangeable with clinically meaningful vision restoration. Likewise, improvement in a selected fraction of visual-field points has not been established as a sufficient standalone basis for an approval claim. Prospective analyses of field-wide indices, localized sensitivity changes, visual-function endpoints, and structural outcomes could clarify the most interpretable efficacy signal. 14113

The recommendation to increase the dose from 2 × 10¹¹ to 6 × 10¹¹ vector genomes creates an opportunity to characterize biological activity and safety across exposure levels. A coherent dose-response pattern could strengthen causal interpretation, although small, sequential, nonrandomized cohorts will remain vulnerable to baseline heterogeneity and cannot substitute for controlled evidence. The higher dose may also change ocular inflammatory or immunogenic risks. 21

From a product-development perspective, FDA would also examine reproducible vector manufacture, biologically meaningful potency assays, regulation of OSK expression, immunogenicity, biodistribution, shedding and long-term safety. These considerations are particularly material for a gene therapy whose desired pharmacology is controlled reprogramming rather than constitutive replacement of a missing gene product. The public materials do not specify a commercial potency strategy or pivotal-trial design. 1531

NAION May Offer a Different Test of the Hypothesis

The later NAION portion of the Phase 1 program will assess ER-100 in an acute optic neuropathy more closely aligned with the ischemic injury model used in the nonhuman primate experiments. NAION has no FDA-approved disease-directed pharmacologic treatment, making preservation or recovery of visual function an important unmet medical objective. 1168

The acute injury setting may provide a different population of threatened but viable neurons than chronic glaucoma, potentially changing the window in which treatment can be beneficial. This is a biological inference rather than an established predictor of ER-100 responsiveness. Natural-history variability, including spontaneous improvement in some patients, makes carefully designed controls essential to any later efficacy claim. 161

Accordingly, the initial OAG findings should not be interpreted as proof that the preclinical NAION treatment effect translates into humans. The indications will require separate analysis of timing, baseline injury severity, visual outcomes, safety and benefit-risk.

The Most Informative Next Disclosure

A future announcement showing a greater number of improved pointwise visual-field locations would be useful, but the more decisive advance would be participant-level evidence of reproducibility and agreement among independent measures. Particularly informative results would include repeated visual-field maps demonstrating coherent improvements, meaningful changes in prespecified global indices, concordant human pERG responses, supporting contrast-sensitivity or acuity findings, and a durable effect beyond doxycycline induction.

Establishing these relationships would also help define whether the treatment is principally neuroprotective, neurorestorative or both. Slower subsequent loss of retinal nerve fiber layer thickness would support a preservation hypothesis, whereas reproducible improvement in function after documented baseline injury would offer stronger support for restoration. Neither mechanism is established by the Day 56 disclosure. 12

Demonstration of clinical benefit would not, by itself, prove that retinal cells had become epigenetically younger. For a viable therapeutic development strategy, however, reproducible and meaningful patient benefit with acceptable safety is ultimately more important than assigning a numerical biological-age change to the treated tissue.

Conclusion

The first ER-100 human results are notable because the preliminary functional observation concerns a clinically relevant aspect of vision, rather than exclusively a molecular or retinal electrophysiologic biomarker. Two of three participants at the initial dose level exhibited localized visual-field improvements at Day 56, and early tolerability supported progression to a higher dose.

The direction of these observations is consistent with the preclinical hypothesis that surviving but injured retinal ganglion cells can recover function after controlled OSK expression. Nevertheless, their magnitude cannot be translated into a percentage of vision restored, and the disclosed pointwise changes require confirmation in a measurement system with substantial known variability.

The absence of human pERG measurements, participant-level mean-deviation data, structural findings and post-induction durability data leaves the central efficacy and mechanistic questions unresolved. The evidence supports continued clinical evaluation, not a conclusion that epigenetic rejuvenation or functional restoration has been demonstrated in humans.

If subsequent controlled evidence demonstrates sustained improvements in visual function after established optic nerve injury, ER-100 could validate a therapeutic approach distinct from conventional pressure reduction and provide important human evidence for partial epigenetic reprogramming. At present, the program has produced an acceptable early safety observation and a preliminary visual-function signal, while the evidence necessary to establish neurorestoration remains to be generated.