To assess safety of gene therapy in G11778A Leber hereditary optic neuropathy (LHON).
Phase 1 clinical trial.
Setting : single institution. Participants : Patients with G11778A LHON and chronic bilateral visual loss >12 months (group 1, n = 11), acute bilateral visual loss <12 months (group 2, n = 9), or unilateral visual loss (group 3, n = 8). Intervention : unilateral intravitreal AAV2(Y444,500,730F)-P1ND4v2 injection with low, medium, high, and higher doses to worse eye for groups 1 and 2 and better eye for group 3. Outcome Measures : Best-corrected visual acuity (BCVA), adverse events, and vector antibody responses. Mean follow-up was 24 months (range, 12-36 months); BCVAs were compared with a published prospective natural history cohort with designated surrogate study and fellow eyes.
Incident uveitis (8 of 28, 29%), the only vector-related adverse event, resulted in no attributable vision sequelae and was related to vector dose: 5 of 7 (71%) higher-dose eyes vs 3 of 21 (14%) low-, medium-, or high-dose eyes ( P < .001). Incident uveitis requiring treatment was associated with increased serum AAV2 neutralizing antibody titers (p=0.007) but not serum AAV2 polymerase chain reaction. Improvements of ≥15-letter BCVA occurred in some treated and fellow eyes of groups 1 and 2 and some surrogate study and fellow eyes of natural history subjects. All study eyes (BCVA ≥20/40) in group 3 lost ≥15 letters within the first year despite treatment.
G11778A LHON gene therapy has a favorable safety profile. Our results suggest that if there is an efficacy effect, it is likely small and not dose related. Demonstration of efficacy requires randomization of patients to a group not receiving vector in either eye.
L eber hereditary optic neuropathy (LHON) is a maternally inherited mitochondrial DNA (mDNA) disease associated with severe visual loss. , Approximately 95% of LHON is caused by 1 of 3 pathogenic point mDNA mutations coding for the respiratory chain subunits of the nicotinamide adenine dinucleotide ubiquinone-oxidoreductase (complex I) genes: 3460G>A ND1, 11778G>A ND4, and 14484T>C ND6. Carriers of LHON genotypes have incomplete penetrance, and only a portion of LHON genotype carriers, more likely males than females, will manifest the disease with a low annual conversion rate. Affected individuals usually are young adults and have severe bilateral visual loss typically starting in 1 eye followed by the fellow eye within days to months. A presymptomatic stage before visual loss, consisting of optic disc edema and retinal ganglion cell (RGC) layer thinning, often occurs. A prospective natural history study found spontaneous partial recovery of visual acuity of 3 lines or more in approximately 18% of patients occurring months to years after the onset of visual loss. No established effective treatment of LHON is yet available.
We report the safety and efficacy results of our phase 1 G11778A LHON gene therapy that included 3 study groups in which each participant was treated with a unilateral intravitreal injection of low, medium, high, or higher dose of gene therapy vector AAV2(Y444,500,730F)-P1ND4v2. The treatment uses allotopic expression by delivering a normal nuclear-encoded ND4 gene (OMIM 516003 .0001) into the nuclei of RGCs via an adeno-associated virus vector. The nuclear version of the mitochondrial gene ND4 subunit is recoded for expression and is imported into mitochondria to normalize the defective adenosine triphosphatase synthesis of RGCs and improve their survival.
study design and intervention
This phase 1, open-label gene therapy clinical trial for G11778A LHON (NCT02161380) was approved by the University of Miami Institutional Review Board and the United States Food and Drug Administration and was conducted at the Bascom Palmer Eye Institute (Miami, FL). The study was designed as a 3 + 3 dose escalation of unilateral intravitreal gene therapy injection of AAV2(Y444,500,730F)-P1ND4v2 for each of the 3 groups of G11778A LHON patients with visual loss.
Three groups of patients with G11778A mDNA and visual loss were enrolled. The chronic bilateral group (group 1) had onset of visual loss in 1 eye of ≥12 months and ≥6 months in the more recently affected eye with best-corrected visual acuity (BCVA) reduced to ≤35 Early Treatment Diabetic Retinopathy Study (ETDRS) letters (Snellen equivalent, 20/200) in each eye. The acute bilateral group (group 2) had onset of visual loss in both eyes of ≤12 months to a BCVA of ≤35 ETDRS letters. The unilateral acute group (group 3) had visual loss in 1 eye to ≤35 ETDRS letters, with mildly impaired but good acuity of 70 letters or more (Snellen equivalent, 20/40) in the contralateral eye. Mean follow-up was 24 months (median: 36 months, range: 12-36 months). To avoid possible confounding effect of idebenone, a washout period of at least 1 week was used before gene therapy injection if the patient was taking idebenone at baseline 1 examination (n = 6).
The treatment consisted of a single unilateral intravitreal 100 µL injection of AAV2(Y444,500,730F)-P1ND4v2 to the worse eye for groups 1 and 2, and to the better eye for group 3. Each group independently had dose escalations from low, to medium, to high, and to higher dose. Table 1 summarizes the number of patients enrolled and the AAV2(Y444,500,730F)-P1ND4v2 doses received. The study was initially designed with treatment of low (5.0 × 10 8 viral genomes [vg]/eye), medium (2.5 × 10 9 vg/eye), and high (2.4 × 10 10 vg/eye) doses of AAV2(Y444,500,730F)-P1ND4v2 manufactured by the University of Florida (UF). Manufacturing of the high dose was changed to Children’s Hospital of Pennsylvania (CHOP) because UF was unable to provide the vector at the high dose. The CHOP high-dose vector was 2.4 × 10 9 vg/eye; CHOP determination of the UF low dose was 1.2 × 10 8 vg/eye and that of the UF medium dose was 5.8 × 10 8 vg/eye. Given that the CHOP high dose was similar to the UF medium dose as initially determined by UF, the Food and Drug Administration (FDA) recommended labeling the CHOP 2.4 × 10 9 vg/eye dose as the high dose and a CHOP 1.0 × 10 10 vg/eye vector as a higher dose. Because of discrepancies between UF and CHOP dose assessment, the FDA Center for Biologics Evaluation and Research required injection of 3 chronic bilateral patients at the high dose before proceeding to the higher dose in 3 acute bilateral patients. No safety concerns emerged in these high-dose chronic bilateral patients, and administration of higher-dose investigational product (IP) began in the acute bilateral group. Subsequently, the data safety monitoring committee (DSMC) and the FDA approved higher-dose injection of 3 chronic bilateral patients in addition to the planned 3 unilateral acute patients. However, the higher-dose injections were suspended after 2 chronic bilateral, 3 acute bilateral, and 2 acute unilateral patients were injected due to DSMC concerns related to incident uveitis requiring treatment in eyes injected with the higher-dose vector.
|Study Group||Expected Doses (vg/eye) by Manufacturing Vector Core Certificate of Activity and Protocol Dilution Procedures|
|UF Low 5.0 × 10 8 a vg/eye||UF Medium 2.5 × 10 9 b vg/eye||CHOP High 2.4 × 10 9 vg/eye||CHOP Higher 1.0 × 10 10 vg/eye|
|1: Chronic bilateral||N = 3||N = 3||N = 3||N = 2|
|2: Acute bilateral||N = 3||N = 3||None||N = 3|
|3: Acute unilateral||N = 3||N = 3||None||N = 2|
a 1.2 × 10 8 vg/eye CHOP titers.
The primary safety outcome measure for groups 1 and 2 was a 15-ETDRS-letter loss of visual acuity in the injected eye from the value recorded on the day before vector injection (the second baseline visit). For group 3, the primary outcome measure was a 3-line drop in acuity in the injected eye, relative to the worst vision of the fellow eye recorded during the study. The first baseline visit occurred at study entry after determination of a qualifying visual acuity and group assignment. The change in BCVA during follow-up was determined from the second baseline measurement to guard against regression to the mean. In addition to the first and second baseline visits, serial ETDRS BCVA was performed by a masked certified ophthalmic technician on postinjection days 1, 7, 30, 60, 90, 180, 270, 365, 548, 730, and 1096. Additional testing included periodic Humphrey 30-2 white-on-white standard visual field, steady-state pattern electroretinography, and spectral-domain optical coherence tomography assessment of peripapillary retinal nerve fiber layer (RNFL) and retinal ganglion cell layer (GCL) thicknesses. The RNFL and GCL thicknesses were measured as described previously and were collected only at baseline, day 7, day 60, month 12, and annually thereafter.
The neutralizing antibody (Nab) titer was reported as the highest serum or anterior chamber sample dilution that inhibited self-complementary AAV2(Y444,500,730F)-smCBA-mCherry transduction by 50% or more compared with no serum control. Quantitative SYBR green polymerase chain reaction (PCR) analysis was performed using the bGH poly A primers (bovine growth hormone polyadenylation signal): F1-5′-AGCCTCGACTGTGCCTTCTAGTT-3′ and R1-5′0-GGGTTCCTGCTATTGTCTTCCCAA-3′ with SsoAdvanced universal SYBR Green Supermix (cat. no. 172-5271; SsoAdvanced; BioRad) according to a standard protocol as described previously.
Longitudinal analysis of changes was performed with generalized estimating equation (GEE) analysis, with logistic link function for dichotomous variables and linear link function for continuous variables. To avoid reporting partial results in a dose-group cohort for which follow-up is still being accrued, we used the maximum common follow-up for each dose group. This phase 1 study was not randomized and did not include an uninjected control group. Therefore, we compared changes from baseline with those observed in our previously published natural history cohort. For these analyses, we used the 26 of 44 natural history patients who would have been eligible for this gene therapy trial at baseline. We assigned the worse eye of natural history patients as the study eye to act as a surrogate for the treated eye in the current study, and we assigned the better eye as the fellow eye. When both eyes of eligible natural history patients had identical baseline acuities, we randomly selected the study eye. Of note, visits of our natural history patients occurred every 6 months, limiting the follow-up times that could be compared directly with the gene therapy patients.
Changes from baseline in the study eyes vs the fellow eyes were compared between the gene therapy subjects and the natural history cohort stratified by onset group with 2-factor analysis of variance. Missing and out-of-window visits became a substantial issue during 2020, due to the Covid-19 pandemic. Linear interpolation was approved by the DSMC as a method for estimation of 14 missed visits and to adjust 14 out-of-window visits to the window midpoint. This procedure was applied only to follow-up visits between month 6 and month 36. No measurements were estimated for missed visits unless bracketed by 2 existing measurements. The dataset used in these analyses was closed on April 20, 2021.
patient demographics and follow-up
A total of 28 patients (5 females, 23 White, 4 White Hispanic, and 1 Native American) aged 16 to 56 years were treated (details in Supplemental Table 1). Twenty-seven (96%) patients completed the final month 36 visit or are being actively followed beyond the month 18 visit (details in Supplemental Table 2). One patient was lost to follow-up after the 180-day visit.
Participants experienced transient ocular irritation due to the injection procedure, and 1 patient required treatment for elevated postinjection IOP that resolved within 24 hours. The only IP-related adverse event was incident uveitis, which was not related to the patient group but was significantly related to vector dose: 1 of 9 (11%), low dose; 1 of 9 (11%), medium dose; 1 of 3 (33%), high dose; and 5 of 7 (71%), higher dose ( P = .006, exact test of trend in proportions) ( Table 2 ). Four of the total eight uveitis cases resolved spontaneously without treatment. All 4 cases of incident uveitis that required treatment occurred in eyes injected with the higher-dose vector.
|Participant||Group||Dose||Inj. MM/YYYY||Onset MM/YYYY||Resolution MM/YYYY||Vision Sequelae||Treatment|
|2-4-27||2: Acute||Higher||6/2019||6/2019||12/2019||None||Topical prednisolone, brimonidine/timolol, cyclopentolate, oral prednisone|
|3-4-28||3: Unilateral||Higher||10/2019||12/2019||05/2020||None||Topical prednisolone|
|1-4-29||1: Chronic||Higher||12/2019||1/2020||Ongoing a||None||Topical prednisolone, timolol|
|1-4-31||1: Chronic||Higher||03/2020||03/2020||12/2020||None||Topical prednisolone, brimonidine/timolol, bimatoprost, oral acetazolamide|
a Medications discontinued in 6/2021; 1+ anterior chamber cells continuing at 24-month visit in 12-2021 with stable or decreased laser flare meter indicating no clinically significant inflammation causing breakdown of blood aqueous barrier and low risk of future complications.
For purposes of assessing IP-related loss of vision in unilateral participants injected in their better-seeing eyes, a visual acuity severe adverse event was defined as loss of acuity in the injected eye equivalent to 15 ETDRS letters worse than the worst fellow-eye acuity measured at any study visit. This occurred in 2 patients: one at 9 months after injection, which resolved at month 18, and the other at 6 months after injection, which was still ongoing at month 36. Neither of these 2 patients’ eyes lost peripheral visual field (assessed by confrontation), and the DSMC and the FDA concurred with the study team that these changes were likely due to the natural history of visual loss in acute onset LHON. No unexpected concerning findings appeared on dilated examination for any patients at any follow-up visit. Blood and urine samples collected according to protocol were scrutinized and showed no indication of IP-related toxicity. A single patient reported an in-patient hospitalization during follow-up that was required to treat a pre-existing parathyroid gland condition.
The complete Nabs data appear in Supplemental Table 3 and are summarized in Table 3 as the means and standard deviations of logarithmic viral vector titers. The change in Nabs was derived by subtracting the follow-up visit measurements from the preinjection baseline. There were small increases in Nabs among patients without uveitis on the order of 0.24 to 0.6 log units compared with an average increase of 0.3 to 3 log units in Nabs among cases with uveitis requiring treatment. Fitting a GEE model to these data (AR1 correlation structure) suggested a larger increase ( P = .026, 2-degree-of-freedom comparison of 3-category uveitis factor) in Nabs of patients with uveitis requiring treatment than in patients with no uveitis ( P = .007); however, patient numbers were very small, and this was a post hoc analysis.
|Uveitis||No Uveitis||Uveitis With No Symptom||No Uveitis||Uveitis With No Symptom||No Uveitis||Uveitis With No Symptom||No Uveitis||Uveitis With No Symptom||Uveitis Treated Symptomatic|
|Mean (SD)||2.28 (1.79)||4.31 (N/A)||1.23 (1.04)||1.90 (N/A)||0.70 (0.00)||4.31 (N/A)||0.70 (0.00)||0.70 (N/A)||1.45 (0.90)|
|Mean (SD)||2.35 (1.73)||4.31 (N/A)||1.23 (1.04)||4.31 (N/A)||1.30 (0.85)||4.31 (N/A)||0.70 (0.00)||0.70 (N/A)||1.75 (1.24)|
|Mean (SD)||2.51 (1.81)||4.31 (N/A)||1.23 (0.88)||4.31 (N/A)||1.30 (0.85)||4.31 (N/A)||0.70 (0.00)||3.11 (N/A)||3.71 (0.00)|
|Mean (SD)||2.59 (1.72)||1.47 (0.97)||4.31 (N/A)||1.00 (0.43)||4.31 (N/A)||1.30 (0.85)||0.70 (N/A)||3.11 (1.20)|