First-in-Human Gene Therapy Trial of AAV8-hCARp.h CNGB3 in Adults and Children With CNGB3-associated Achromatopsia

AAV8-hCARp

][4] It is primarily caused by sequence variants in genes that are critical to phototransduction in cone photoreceptors; most cases are associated with changes in either CNGA3 or CNGB3 , which encode the αand β-subunits of the cyclic nucleotide-gated channels, respectively. 2 , 5Variants in CNGB3 account for approximately half of all cases of ACHM. 4 Other genes associated with ACHM include GNAT2 , PDE6C , PDE6H , and ATF6 . 3 , 5-7he majority of those affected have complete ACHM, with no function in all 3 classes of cone photorecep-tor (long-, middle-, and short-wavelength sensitive), 1 , 4 , 7 , 8 which is associated with marked PA and nystagmus, typically from birth, and VA of ≤20/200.Incomplete ACHM, in which there is residual function in ≥1 cone subtype, occurs far less frequently and is associated with residual color discrimination, less severe impairment of VA ( ≥20/40 to ≤20/120), and milder or absent PA and nystagmus. 3 , 7 , 8here is currently no effective treatment available for ACHM; management strategies aim to mitigate the adverse impact on quality of life (eg, from PA and reduced VA). 3 , 9 The marked PA, being both uncomfortable and further degrading VA/contrast sensitivity (CS), can result in significant limitations in the ability to perform activities of daily living. 10In a survey of people with ACHM, 38% named PA as the symptom they would most want to improve above all other aspects of visual impairment, and 73% stated that it affected their employment prospects. 11These limitations are incompletely mitigated by countermeasures such as sunglasses and tinted contact lenses. 11NGB3-associated ACHM has several characteristics that make it an attractive candidate for gene therapy.Effective improvement in cone photoreceptor function would provide a clear and reliable outcome measure.3][14] Moreover, in preclinical models of CNGB3 deficiency, gene supplementation improved cone function and morphology. 5 , 15here are no peer-reviewed publications of clinical trials of gene therapy for the treatment of adults or children with ACHM associated with CNGB3 ; however, a phase 1/2 clinical trial investigating gene therapy for ACHM associated with CNGA3 has demonstrated early promise. 16he phase 1/2 clinical trial described herein was conducted to assess the safety and efficacy of AAV8-hCARp.hCNGB3 gene replacement in adults and children with CNGB3 -associated ACHM during a 6-month follow-up.

METHODS
This open-label, nonrandomized, phase 1/2 trial (Clinical-Trials.govIdentifier: NCT03001310) was conducted at one site in the United Kingdom (Moorfields Eye Hospital NHS Foundation Trust, London, United Kingdom) and one site in the United States (University of Michigan Kellogg Eye Center, Ann Arbor, Michigan, USA).The study protocol was reviewed by an independent ethics committee or institutional review board, as appropriate, and the study was conducted in accordance with the principles of the Declaration of Helsinki.All participants or their legal representatives provided consent, or parental permission and assent were obtained.
Eligible participants were ≥3 years of age with a diagnosis of ACHM, confirmed by a study-associated retina specialist, and had accredited laboratory-confirmed homozygous or compound heterozygous missense or null disease-causing variants in CNGB3 .The study timeline is presented in Supplementary Figure S1.Children were enrolled only after determination of the maximum tolerated dose in adults during the dose-escalation phase.Children were defined as ≥3 years of age and < 16 years in the United Kingdom or < 18 years in the United States, as per medical legislation.Evidence of relative photoreceptor preservation at the macula, as assessed by spectral-domain optical coherence tomography (SD-OCT) and/or adaptive optics scanning light ophthalmoscopy, was required.Exclusion criteria are presented in the Supplementary Methods.All participants were followed for 6 months after surgery before entering a separate long-term follow-up study (ClinicalTrials.govIdentifier: NCT03278873).
• STUDY TREATMENT: AAV8-hCARp.hCNGB3 is a recombinant serotype 2/8 adeno-associated viral vector containing a human CNGB3 complementary DNA driven by a 0.4-kb fragment of the human promoter.To determine the maximum tolerated dose of AAV8-hCARp.hCNGB3 , adult participants were enrolled in a dose-escalation protocol (Supplementary Figure S2) and administered 1 of 3 doses of AAV8-hCARp.hCNGB3 in a total volume of up to 0.5 mL: low dose, 0.1 × 10 12 vg/mL; intermediate dose, 0.3 × 10 12 vg/mL; or high dose, 1.0 × 10 12 vg/mL.Once the maximum tolerated dose was established, 12 additional participants (children) were enrolled in a confirmatory safety phase.If a dose-limiting event (DLE; Table 1 ) occurred in the first child within the first 6 weeks, the independent data monitoring committee convened to confirm the dose levels administered to future children.
Efficient transduction of the cone photoreceptor cells required AAV8-hCARp.hCNGB3 to be administered to the subretinal space.The recombinant vector was delivered under general anesthesia to the worse-seeing eye (as determined by the participant and investigator using ocular dominance or VA) as a suspension of viral vector particles injected subretinally under direct observation using an operating microscope.This procedure involved 3-port pars plana vitrectomy followed by injection of the viral vector suspension using a 41-gauge subretinal cannula through ≥1 retinotomies into the subretinal space, with the first retinotomy superior to the macula, close to the superotemporal vascular arcade, with the aim of including the central macula in the bleb created.Tapered oral and topical steroids were prescribed pre-and postoperatively to protect against intraocular inflammatory responses.See the Supplementary Methods for more information.
• ASSESSMENT SCHEDULE: Detailed baseline functional and structural assessments of both eyes were performed, including ocular examination, retinal imaging, psychophys-

244
AMERICAN JOURNAL OF OPHTHALMOLOGY MONTH 2023 anterior chamber cells 3 + (26-50 cells in a field size of 1 × 1-mm slit beam), anterior chamber flare 3 + (mar ked, ir is and lens details hazy), or vitreous haze 3 + that failed to improve by 2 steps (or to grade 0) during a 6-week period. 47al testing, and participant-related outcome assessments.Baseline serology, blood pressure, hematology, and biochemistry were also undertaken.Body mass index, height, weight, pulse, respiratory rate, oxygen saturation, and temperature were measured at baseline; day 2; and weeks 1, 2, and 4. A full clinical ocular examination was conducted on day 1, including VA and intraocular pressure.Color fundus photography, fundus autofluorescence imaging, SD-OCT, and ocular examination were performed at baseline and every postoperative visit (including days 1 and 3).Measurements of CS, reading, static perimetry, microperimetry, color vision, PA, full-field stimulus testing, and vision-related quality of life (VRQoL) were taken at baseline and weeks 12 and 24.Adaptive optics scanning light ophthalmoscopy was completed for a defined subset of participants (n = 8) in both eyes at baseline and week 24.Blood samples were collected to measure immune response at baseline and then 4 weeks, 3 months, and 6 months after surgery.
• OUTCOME MEASURES: The primary end point was safety and tolerability of subretinal administration of AAV8-hCARp.hCNGB3 , defined as the absence of a DLE within 6 weeks of administration.Clinical assessment of intraocular inflammation and anatomic integrity was conducted using slitlamp biomicroscopy and retinal imaging, including SD-OCT.
Efficacy was the secondary objective and was explored (1) on an individual basis, (2) by dose, and (3) by adult or child cohort, examining any change from baseline in assessments of visual and retinal function and functional vision.
8][19] Reading ability, including reading acuity, maximum reading rate, and critical print size, was assessed with MNREAD cards 20 and International Reading Speed Texts. 21olor vision was assessed using the American Optical Hardy-Rand-Rittler (HRR) plates test. 22This evaluation was performed at baseline and postsurgery at weeks 12 and 24.The direction of improvement would be an increase in the number of plates correctly identified by the participant.4][25] These tests measure the amount of saturation required to discriminate a colored target from a series of gray distractors, with a reduction in needed saturation as a marker of colorsensitivity improvement.
Retinal sensitivity was determined by static perimetry using a custom 185-point grid (Octopus 900; Haag-Streit AG) and full-field stimulus testing, a psychophysical threshold method to evaluate light sensitivity. 26The retinal locus of fixation and bivariate contour ellipse area were also determined using microperimetry, as captured by the Macular Integrity Assessment device (CenterVue).Fullfield electroretinogram (ERG) and pattern ERG were performed according to the International Society for Clinical Electrophysiology of Vision (ISCEV) standards [27][28][29] to assess both generalized retinal (rod and cone systems) and isolated macular function.Modified ISCEV protocols were used as necessary in younger children using internationally recognized protocols.
The degree of light sensitivity (ie, PA) was investigated by video analysis of participant response to a uniform light stimulus (3 study-independent reviewers evaluated the videos to assess monocular palpebral aperture narrowing in response to a light stimulus) and by a participantreported symptomatology questionnaire (assessing severity of PA with and without aids, and its impact on daily activities on an 11-point [0-10] rating scale). 11Participant response (ie, no change, improved, or worsened) was determined by agreement of 2 of 3 reviewers.Further details of this assessment, as well as for palpebral aperture measurements, are presented in the Supplementary Methods.
• STATISTICAL ANALYSES: Given that the primary outcome of this study was safety and tolerability, and given the exploratory nature of the efficacy assessments, this study was not powered for formal statistical analysis.Because of the rare nature of ACHM and the small study sample, no formal sample size calculations were performed, and no formal hypothesis testing was conducted.Furthermore, the evaluation of response was based on clinical interpretation as standards of clinical meaningfulness have not yet been established for most outcome assessments in ACHM.

RESULTS
AAV8-hCARp.h CNGB3 was administered to 23 participants (20 in the United Kingdom, 3 in the United States).Following completion of dose escalation in 11 adults, 12 children were treated in the confirmatory safety phase.The first child, who received the high dose of 1.0 × 10 12 vg/mL, developed a DLE of ocular inflammation.As a result, all subsequent doses were either the intermediate dose (0.3 × 10 12 vg/mL) or an additional "other" dose (0.6 × 10 12 vg/mL).The additional dose of 0.6 × 10 12 vg/mL was selected in consultation with the independent data monitoring committee and administered to maximize therapeutic effect while maintaining the safety profile of AAV8-hCARp.hCNGB3 .
Twenty-two of the 23 participants had a single retinotomy performed during administration, and 1 participant had 2 retinotomies.Adult and child participants received subretinal injection volumes ranging from 0.2 to 0.5 mL (mean, 0.35 mL); the total number of viral genomes delivered ranged from 0.3 to 4.5 × 10 11 .No surgical complications were observed during dose administration.
• SAFETY: AAV8-hCARp.hCNGB3 demonstrated an expected and manageable adverse event (AE) profile.Three (13.0%) participants experienced 1 primary safety outcome event, as defined in Table 1 (all severe inflammation: 1 in the intermediate-dose group, 2 in the high-dose group), that met the Standardization of Uveitis Nomenclature Working Group criteria, also defined in Table 1 .One participant experienced severe inflammation on day 19, and the 2 other participants experienced severe inflammation on day 43 after surgery.All events were resolved before the end of the study (6 months posttreatment).
Intraocular inflammation was the principal AE related to the study therapy.Nine (39.1%) participants (5 adults, 4 children) experienced an intraocular inflammation AE considered related to the gene therapy itself; 2 of these (n = 2/23; 8.7%) were considered serious AEs (SAEs; 1 in the intermediate-dose group, 1 in the high-dose group) and were DLEs.More severe events of inflammation tended to occur at higher dose levels, whereas those at the lowest dose were less severe and of shorter duration.Inflammation was reported both in the anterior chamber and in the intermediate and posterior ocular segments in the form of intermediate uveitis and panuveitis.
All cases resolved with an extended course of topical and/or systemic corticosteroid treatment.No baseline factors appeared to be predictive of ocular inflammation.Both participants with an SAE of inflammation completed the study per protocol and participated in the long-term followup study.One of these 2 participants also experienced an SAE of ectopic pregnancy that was managed by emergency surgery; this resolved and was considered unrelated to the gene therapy.
The most common treatment-emergent AEs ( Table 2 ) were eye disorders.All participants reported ≥1 ocular AE, as well as ≥1 AE considered "definitely related" to the surgery, including conjunctival hemorrhage (20 [87.0%]), lenticular opacities (15 [65.2%]; 9 children, 6 adults), and reduced VA (15 [65.2%]).VA reduction was transient and related to the surgical procedure; VA returned to baseline for all affected participants during the first 4 weeks of followup.
SD-OCT imaging revealed a minimal mean increase in central retinal thickness (CRT) from baseline to week 24 postintervention at the group level (1.2 μm).This finding was not secondary to edema or fibrosis and is likely not clinically significant.The manufacturer's technical specifications include a 1.8-μm estimated SD of repeatability of CRT volume scans. 38See Supplementary Figure S3 for line plots of CRT change from baseline over time for individual participants.Additionally, no safety signals were noted from the color fundus photography or fundus autofluorescence imaging.
There were no clinically significant changes in laboratory tests, vital signs, or physical examination.No participants had an AE leading to study discontinuation.

246
AMERICAN JOURNAL OF OPHTHALMOLOGY MONTH 2023 • EFFICACY: There was no systematic pattern of change from baseline to week 24 for any specific efficacy assessment between the treated and untreated eye across the entire study population, within either cohort (adult or child), or with respect to dose.However, favorable changes were observed for individual participants across several assessments, including ≥1 domain of color vision (lvvCCT; n = 6/23), PA (investigator assessed; n = 11/20), and the VRQoL questionnaires at week 24 ( ≥1 domain of the IVI; n = 21/23).See Supplementary Table S2 for a summary of responses for the treated vs untreated eyes at baseline vs week 24 for bivariate contour ellipse area, CRT, microperimetry, HRR, UCDT, and lvvCCT.
Pattern ERGs were unrecordable or undetectable at baseline and follow-up.Additionally, no participant had any significant treatment-related change in full-field ERG measures beyond expected intersession variability.
For color vision, 6 of 23 (26%) participants had improvements in ≥1 axis using the lvvCCT, as assessed by comparison of 24-week measurement with baseline.There were no systematic differences between the treated and untreated eyes across all participants ( Table 3 ; Supplementary Figures S4-S6).
A qualitative assessment of PA by graders blinded to which eye had been treated (video review of palpebral aperture narrowing in response to a light stimulus) detected improvement in comparison to baseline in the treated eye in  3 and 4).
Ratings of the more general HRQoL were equivocal.On the EQ-5D-5L and the visual analog scale (EQ-VAS), no changes from baseline were observed at week 24 across the domains of mobility, self-care, and anxiety/depression.Among child participants, a reduction was observed in the percentage of participants reporting problems on the EQ-5D-Y domains of mobility (2/12; 16.7%), usual activities (2/12; 16.7%), and anxiety/depression (2/12; 16.7%).No change was observed in the pain (1/12; 8.3%) dimension, and 1 (8.3%) participant reported problems with selfcare at week 24.A small improvement in change from baseline in mean ± SD EQ-VAS score was observed at week 24 (4.2 ± 8.7).
PRO item-level post hoc analyses revealed ceiling effects at baseline in > 25% and up to 100% of study participants for 20 of 28 items of the IVI, 17 of 24 items of the IVI-C, and 5 of 5 items of the EQ-5D, an indicator that many items in each instrument may not be relevant to the target population and would inherently limit the ability of the PRO instrument to detect treatment effects.PRO item and domain analyses are shown in Supplementary Tables S3 to S6.
One 21-year-old female participant (01-009) demonstrated improvement at week 24 in color vision and positive changes in functional vision based on observed improvements on the IVI reading and accessing information domain.She had evidence at baseline of residual color vision and better VA than average, suggesting incomplete ACHM.Measures of color vision for this participant are shown in Supplementary Figure S8.

DISCUSSION
In this first-in-human study, subretinal delivery of AAV8-hCARp.hCNGB3 gene therapy in adults and children with ACHM associated with CNGB3 was found to have an AE profile that is anticipated and manageable.There were no consistent changes from baseline in any efficacy assessments across the dose or age cohorts at week 24.
As expected, given surgical delivery, ocular AEs were common. 39Ocular inflammation is an anticipated response to the injection of viral protein and is known to occur in ocular gene therapeutic applications with other similar agents. 40Most ocular AEs were temporally related to the surgery, were mild in severity, and resolved with minimal intervention.There was a trend toward more severe cases of intraocular inflammation occurring at higher doses.Inflammation appeared to be dose related and was reduced through appropriate dose selection and prophylactic use of corticosteroids.Overall, the safety results support further exploration of efficacy in future studies.
Improvements of varying magnitude were observed in several assessments at the individual participant level in both children and adults (eg, in color vision and PA).Based on a clinical interpretation of change, the assessments that most frequently suggested benefit were PA and color vision testing and the participant-reported scores on the IVI questionnaire.One participant demonstrated a notable improvement in color discrimination.
An analysis of 4 children with CNGA3 -or CNGB3associated ACHM receiving gene therapy in clinical studies NCT03758404 and NCT03001310 evaluated treatmentrelated changes in cone function against untreated participants. 40Two of the 4 demonstrated cone-mediated signals in the visual cortex and significant improvements in psychophysical measures of cone-mediated visual function.These results represent the opportunity for gene therapy to activate dormant cone-signaling pathways in ACHM. 41verall, our findings demonstrate short-term safety and tolerability, with the positive trends observed in select participants, meriting further investigation.
There are several limitations and areas for development, some of which are inherent to the small sample sizes of trials in ultra-rare diseases.Additionally, the lack of a concurrent control group limited the generalizability of the findings, notably regarding the interpretations of end points such as PROs.Most of the efficacy assessments performed in this study are neither conducted in routine clinical practice nor fully validated.In addition, international-standard suprathreshold assessments may not be sufficiently sensitive to identify small or close-to threshold responses and do not exclude the possibility of a therapeutic response at week 24.This study attempted to identify sensitive end points, but 24 weeks may be of insufficient duration to determine lateronset benefit.The results of the study should be interpreted in the context of these limitations and considered in the design of forthcoming trials.
Although there are no approved treatments for ACHM, several small-scale gene therapy studies for AAV8h CNGA3 and AAV2tYF-PR1.7-hCNGA3 have demonstrated preliminary safety and efficacy. 16Assessing novel end points will be important to provide treatment options that address the limitations of VRQoL and visual function assessments for people with ACHM.There are no precedents for clinical development of a therapeutic agent for ACHM or for its regulatory approval.As such, limited information was available to guide optimal trial design, inclusion criteria, or end points.
In this study, concerns that a lack of cortical plasticity might preclude a benefit to vision in older participants led to a decision to enroll younger participants in the safety-confirmation expansion phase. 42The youngest participant included in this study was 5 years old, which may be beyond the critical period of visual development.Therefore, there may not be significant differences in cortical plasticity between the ages of participants included in this study, and, thus, studying gene therapy in even younger children may identify more meaningful improvements in vision. 41Nevertheless, age did not appear to be a determinative factor in measurable response in our trial, with arguably the most promising response being in a young adult, likely with incomplete ACHM.Future studies should also work toward a better understanding of factors such as age, baseline severity of illness, phenotype, and genotype.
In addition, it will be necessary to further refine and validate relevant efficacy measures that address clinically meaningful treatment effects.Reduction in PA, a frequent and debilitating symptom, is one of the priorities for this participant population, 11 , 43 with further investigation needed to better understand how PA can be optimally and objectively quantified and potentially correlated with improvements in visual function and PROs.The data from our trial, along with findings from qualitative and quantitative evaluations of PROs in an ACHM population, indicate that developing or modifying PRO measures specific to ACHM are required to address the ceiling effects and lack of content validity of existing PRO measures.
Initial reports of participant feedback to investigators after the intervention indicated that they were, for example, better able to "see colors," "go outside without having to wear protective eyewear," "distinguish between black and white," and "see more clearly when exposed to high light levels."This was reflected in post hoc analyses of item 3 of the PA questionnaire (ie, 3 fewer participants reported "increased difficulty seeing due to PA" at week 24 compared to baseline).
Similarly, tailoring VA and CS measurements to the specific abilities and deficits of adults or children with ACHM may help capture possible treatment effects.Adults or children with ACHM have greater functional difficulty in photopic conditions. 8Best-corrected VA (BCVA) in this study was evaluated under standard conditions with room lights turned off and chart lights turned on, with BCVA unchanged over the course of the study.Assessing BCVA at varying levels of ambient luminance merits further exploration. 44Regarding CS, the Pelli-Robson test may be insensitive in this population because this test presents only 1 spatial frequency and is limited in terms of contrast resolution.To address these limitations, contrast-detection thresholds across multiple spatial frequencies may indicate improvements with greater precision.
In conclusion, the results of this study demonstrate an acceptable safety profile for AAV8-hCARp.hCNGB3 in adults and children with ACHM.Indications of benefit in individual participants merit further investigation of efficacy and provide valuable data to optimize the design of future trials.
Funding/Support: This study was funded by MeiraGTx UK II Ltd in collaboration with Janssen Research & Development, LLC.The sponsors participated in the design of the study; management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.Financial Disclosures: Michel Michaelides receives consulting fees from MeiraGTx/Janssen; receives support for attending meetings and/or travel from MeiraGTx/Janssen; has stock or stock options in MeiraGTx; reports receipt of equipment, materials, drugs, medical writing, gifts, or other services from MeiraGTx/Janssen; and is supported by grants from the National Institute for Health Research (NIHR) Biomedical Research Centre at Moorfields Eye Hospital NHS Foundation Trust, UCL Institute of Ophthalmology, and The Wellcome Trust (099173/Z/12/Z).Cagri G. Besirli receives consulting fees from Janssen; receives clinical trial support from Janssen; has stock or stock options in Johnson & Johnson; and receives payment or honoraria from Vindico Medical.Neruban Kumaran receives consulting fees from MeiraGTx.Anastasios Georgiadis receives support for attending meetings and/or travel from MeiraGTx; and has stock or stock options in, and is an employee of MeiraGTx.Alexander J. Smith receives consulting fees from MeiraGTx; and receives funding from the Medical Research Council, United Kingdom.Caterina Ripamonti receives consulting fees from MeiraGTx.Robert H. Henderson receives consulting fees from Novartis and Stoke Therapeutics; receives payment or honoraria from Novartis, Medscape, and STELLAR; receives support for attending meetings and/or travel from Novartis; and is on a data safety monitoring board or advisory board for the University of Tübingen, Novartis, and the University of Oxford.Penny Fleck has stock or stock options in and is an employee of Janssen.Eddy Anglade has stock or stock options in and is an employee of Janssen.Xiangwen Dong has stock or stock options in and is an employee of Janssen.George Capuano has stock or stock options in and is an employee of Janssen.Wentao Lu has stock or stock options in and is an employee of Janssen.Pamela Berry has stock or stock options in Janssen and GSK, and is an employee of Janssen.Thomas Kane receives consulting fees from MeiraGTx.Stuart Naylor has stock or stock options in and is an employee of MeiraGTx.Michalis Georgiou receives consulting fees from MeiraGTx/Janssen.Robin R. Ali receives funding from the Medical Research Council, United Kingdom.Alexandria Forbes has stock or stock options in and is an employee of MeiraGTx.James Bainbridge receives funding from the National Institute for Health Research, United Kingdom.Nashila Hirji, Sui Chien Wong, Serena Zaman, Irene Gottlob, Anthony G. Robson, Alberta Thiadens, and Angelos Kalitzeos have no financial disclosures to report.All authors attest that they meet the current ICMJE criteria for authorship.Author Contributions: Conceptualization: M.M., N.H., N.K., A.G., A.J.S., I.G., R.

TABLE 2 .
Treatment-emergent AEs (Specific AEs > 10%)a-d AE = adverse event, SAE = serious adverse event.a All data are displayed as n (%).b Any AE occurring on or after the initial administration of gene therapy was considered treatment emergent.c Incidence (%) is based on the number of participants in each group, not the number of events.d AEs are coded using the Medical Dictionary for Regulatory Activities , version 22.1.

TABLE 3 .
Summary of Trivector lvvCCT and Change From Baseline Over Time