INTRODUCTION

Stargardt macular dystrophy (SMD) is the most common form of juvenile macular degeneration. SMD results from the production of defective rim proteins encoded by the ABCA4 gene, leading to the accumulation of di-retinoid-pyridinium-ethanolamine in the retinal pigment epithelium (RPE), RPE cell loss, and photoreceptor death.¹ Therapeutic options for SMD remain limited, although several treatment strategies are currently under investigation. Gene therapy, stem cell therapy and pharmacological investigations are currently at the preclinical or early clinical phase.^{2 3} RPE dysfunction and loss, followed by subsequent photoreceptor dysfunction and loss, have been implicated in the pathophysiology of SMD; therefore, RPE replacement has been suggested as a therapeutic treatment option for patients with SMD.^{1 4 5} Proper functioning of the RPE is important for maintaining the health and integrity of the outer retina, photoreceptors and choriocapillaris. Healthy RPE cells play many crucial roles in the retina, including transportation of nutrients (eg, glucose or vitamin A) from the

blood to the photoreceptors, secretion of growth factors, phagocytosis of the outer segments of photoreceptors, formation of the blood–retina barrier by tight junctions and establishment of the immune privilege of the eye.^{6 7}

Human embryonic stem cells (hESCs) have been suggested as potential therapeutic options for use in regenerative medicine.⁸ However, tumorigenicity, immunogenicity and ethical concerns have been barriers to the adoption of this approach.^{4 9–11} Schwartz et al ¹² reported the results of a phase I clinical trial using hESC-RPE cell suspensions for the treatment of dry age-related macular degeneration and SMD; their report included only the 1-year safety, tolerability, visual functional and anatomical changes. Mehat et al ¹³ recently reported 12-month results from a trial initiated by the same sponsor (Astellas Institute for Regenerative Medicine, Marlborough, MA, USA) in 12 patients with SMD. We also reported the preliminary (12-month follow-up) results of phase I/II clinical trials in two patients with age-related macular degeneration and two patients with SMD, using the same hESC-derived RPE cells (sponsored by CHA Biotech Co, Ltd., Republic of Korea) in 2015.⁴ In the present study, we examined the long-term (3-year) outcomes of a phase I clinical trial of patients with SMD who underwent subretinal transplantation of hESC-RPE cell suspensions.

RESULTS

Clinical trial results

The first patient was a 45-year-old man who began to lose his vision at the age of 26 years. He had three missense mutations in the ABCA4 gene: c.983A>T (Glu328Val), c.1933G>A (Asp645Asn), and c.3106G>A (Glu1036Lys); all were previously reported as Stargardt-associated mutations.¹⁵ The patient’s initial BCVA was counting fingers (one ETDRS letter) in the treated eye and 20/800 (four ETDRS letters) in the untreated eye. He began taking oral tacrolimus and MMF, 1 week prior to transplantation. In total, 50 000 hESC-RPE cells were transplanted into the subretinal space while the patient was under general anaesthesia. No unusual signs of inflammation were observed during the postoperative period. Mild vascular leakage on fluorescein angiography was observed in both eyes at the 7-week visit without any subjective symptoms; this finding did not change until 9 months after transplantation. At the 3-month visit, the BCVA of the treated eye was 20/800 (10 ETDRS letters), which represented an improvement of nine letters compared with baseline. In accordance with the protocol, MMF was continued until 9 months postoperatively. At 52 weeks postoperatively, the BCVA was 20/640 (13 letters) in the treated eye and 20/640 (13 ETDRS letters) in the untreated eye. The BCVA remained stable for 3 years postoperatively: 10 ETDRS letters in the treated eye and 11 ETDRS letters in the untreated eye (online supplementary table S4). No abnormal proliferations of transplanted cells were encountered and there were no signs of obvious retinal thickening/thinning or retinal oedema. No preretinal or subretinal pigmentation was noted after hESC-RPE cell transplantation during the 3-year follow-up period (figure 1). Subjective assessment of visual function according to the total NEI VFQ-25 score showed stable overall scores during the 3-year follow-up period (figure 3A). Intraclass correlation coefficients with 95% CIs were calculated to assess inter-grader agreement in semi-automated DDAF area measurements. The measurements demonstrated an intraclass correlation coefficient of 0.991 (95% CI 0.959 to 0.998). DDAF areas gradually increased in both eyes during the 3-year follow-up period (figure 2). The initial DDAF area was 9.96 mm² in the treated eye, which was larger than the corresponding area in the untreated eye (4.75 mm²). Annual progression of the DDAF was 1.49 mm² in the treated eye and 0.52 mm² in the untreated eye. The central scotoma depth on Goldmann visual field examination exhibited slightly enhanced depth at 3 years postoperatively. No clinically significant changes were observed in electroretinography or multifocal electroretinography findings during the 3-year follow-up period (figure 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Ophthalmic results of the first Stargardt macular dystrophy (SMD) patient.

Baseline (A, B) and 3-year postoperative (C, D) fundus photography and spectral domain-optical coherence tomography (SD-OCT) images. No obvious adverse event or pigmentation (arrow indicates the human embryonic stem cell-retinal pigment epithelium (hESC-RPE) cell suspension-injected retinotomy site) was noted after hESC-RPE cell suspension transplantation. Electroretinogram (ERG) examinations at baseline (E) and at 3 years postoperatively (F) showing no significant changes. Goldmann visual field examinations at baseline (G) and 3 years postoperatively (H) showed a central scotoma of slightly diminished size.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Changes in definitely decreased autofluorescence (DDAF) areas during the 3-year follow-up.

(A-D, M) Changes in the DDAF area in the first patient. The DDAF areas gradually increased in both eyes at 3 years after surgery (C, D) in contrast to the baseline values (A, B). The initial DDAF area was larger in the treated eye (B) than in the untreated eye (A). (E–H, N) The second patient had an extensive DDAF area at baseline in both eyes. The DDAF area expanded rapidly compared with that in the other patients. (I–L, O) The third patient had DDAF areas with homogeneous backgrounds in both eyes at baseline (I, J). The annual progression rates of the DDAF areas tended to be slow in the third patient compared with those in the first and second patients, who had DDAF areas with heterogenous backgrounds. TE, treated eye; UTE, untreated eye.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Functional changes during the 3-year follow-up.

(A) Changes in the National Eye Institute Visual Function Questionnaire 25 (NEI VFQ-25) scores during the 3-year follow-up. (B). Mean change in best-corrected visual acuity (ETDRS letters) compared with baseline.

Supplemental material

[bjophthalmol-2020-316225supp002.pdf]

The second patient was a 40-year-old man with advanced SMD. His vision had been impaired since the age of 17 years, and he had been diagnosed with SMD at the age of 20 years. He had two missense mutations: c.2894A>G (Asn965Ser) and c.4972A>C (Ser1658Arg). The first of these mutations, c.2894A>G (Asn965Ser), was previously reported as a Stargardt-associated mutation.¹⁵ The patient’s initial BCVA was 20/640 (13 ETDRS letters) in the treated eye and 20/250 (32 ETDRS letters) in the untreated eye. The patient began taking MMF and tacrolimus, 1 week prior to transplantation. In total, 50 000 RPE cells were transplanted into the subretinal space while the patient was under general anaesthesia. No unusual signs of inflammation were observed during the postoperative period. At the 3-month visit, the BCVA of the treated eye was 20/200 (34 ETDRS letters), which represented an improvement of 19 letters compared with baseline. In accordance with the protocol, MMF was continued until 9 months postoperatively. At 52 weeks postoperatively, the BCVA was 20/250 (32 letters) in the treated eye and 20/160 (41 ETDRS letters) in the untreated eye. The BCVA remained stable for 3 years postoperatively; 32 ETDRS letters in the treated eye and 38 ETDRS letters in the untreated eye (online supplementary table S4). Subjective assessment of visual function according to the total NEI VFQ-25 score showed improvement by 7.47 points, due to enhanced BCVA (figure 3). Subretinal pigmentation with hypofluorescence was observed at 4 weeks postoperatively, which increased until 6 weeks; it then remained stable during the 3-year follow-up period. Furthermore, multiple hyperautofluorescent spots and stippling patterns of autofluorescence were observed at the injection site (figure 4). DDAF areas also gradually increased in both eyes during the 3-year follow-up period (figure 2). Initial DDAF areas were 36.82 mm² in the treated eye and 32.03 mm² in the untreated eye. Annual progression of the DDAF was 2.63 mm² in the treated eye and 3.03 mm² in the untreated eye (figure 2). Visual field examination was unreliable because the patient had poor visual function and a large central scotoma at baseline; however, there were no obvious postoperative changes. Electroretinography and multifocal electroretinography findings for the injection site were stable during the 3-year follow-up period (figure 4).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Ophthalmic results of the second SMD patient.

Baseline (A, B) and 3-year postoperative (C, D) fundus photography and spectral domain-optical coherence tomography (SD-OCT) images. Subretinal pigmentation (B, inset, yellow arrows) was noted in the bleb area, which may represent engraftment of injected human embryonic stem cell-retinal pigment epithelium (hESC-RPE) cells (C and D). SD-OCT at baseline (B), and at 3 years postoperatively, (D) at sites showing an RPE monolayer. Fundus autofluorescence images (left panels of B and D) show stippling patterns of autofluorescence at 3 years after surgery (D, inset, yellow arrows). The retinotomy site for the hESC-RPE cell suspension injection is also denoted (A, C, D, inset, white arrow). There was no change in the electroretinograms (ERGs) from baseline (E) to 3 years postoperatively (F). The central scotoma depth was slightly increased in depth at 3 years after surgery (G, H).

The third patient was a 40-year-old man who had been clinically diagnosed with SMD at another tertiary hospital. He had experienced blurred vision in the central visual field of his right eye for 3 years, and in the left eye for 3 months. He had no family history of eye disease. Genetic analysis was not mandatory for enrolment in this trial; therefore, we performed hESC-RPE transplantation prior to genetic analysis. Several months postoperatively, genetic analysis revealed no mutations in ABCA4. Mutations in introns and regulatory elements were not detected because exome sequencing was used in this study. The patient’s initial BCVA was 20/160 (39 ETDRS letters) in the treated eye (right eye) and 20/160 (40 ETDRS letters) in the untreated eye (left eye). BCVA and visual field examination were repeated three times during the screening, and thorough discussions were held to determine which eye to treat. The right eye was selected for transplantation because the central scotoma was larger in the right eye than in the left eye. The patient began taking MMF and tacrolimus, 1 week prior to transplantation. In total, 50 000 RPE cells were transplanted into the subretinal space while the patient was under general anaesthesia. No unusual signs of inflammation were observed during the postoperative period and the patient’s visual acuity was stable in both eyes until 3 months postoperatively (45 ETDRS letters in the treated eye and 46 ETDRS letters in the untreated eye). At 19 weeks postoperatively, the patient presented to the outpatient clinic for an unscheduled visit due to upper respiratory symptoms (sore throat and cough); routine fundus examination revealed rhegmatogenous retinal detachment (RRD) involving the macula, due to a retinal hole in the superotemporal periphery of the treated eye. Retinal reattachment was achieved with immediate pars plana vitrectomy, endolaser photocoagulation, 10% C₃F₈ gas tamponade, and scleral encircling with a 3 mm silicone sponge (figure 5). The patient’s BCVA recovered to the preoperative level (38 ETDRS letters) at 5 weeks after retinal reattachment and remained stable (44 ETDRS letters) at 3-year post-transplantation. BCVA in the untreated eye was 20/125 (45 ETDRS letters) at 3-year post-transplantation (online supplementary table S4). The patient’s subjective visual status, as indicated by the NEI VFQ-25 score, did not decrease, despite the RRD (figure 3). Goldmann visual field examination revealed that the area of the central scotoma showed no obvious changes over 3 years in either eye. Reductions of all standard electroretinography amplitudes and prolongation of all implicit times were more evident in the treated eye than in the untreated eye after RRD. However, the treated eye recovered to preoperative values at 1 year after retinal reattachment; these values were maintained during the 3-year follow-up period after subretinal hESC-RPE cell implantation. Multifocal electroretinography amplitudes were reduced after RRD; however, they recovered to the preoperative level and remained stable during the 3-year follow-up period, despite cataract development. No subretinal pigmentation was noted in this patient; however, after cell injection, a small preretinal pigmentation (temporal to the fovea) developed over the epiretinal membrane, which had been observed at baseline. Otherwise, no clinically significant anatomical changes were observed on spectral domain-optical coherence tomography (figure 5). DDAF areas gradually increased in both eyes during the 3-year follow-up period. Initial DDAF areas were 6.74 mm² in the treated eye and 6.92 mm² in the untreated eye; these gradually increased in both eyes over 3 years (figure 2). The annual progression of DDAF was 0.72 mm² in the treated eye and 0.38 mm² in the untreated eye.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Ophthalmic results of the third SMD patient.

Baseline (A, B) and 3-year postoperative (C, D) fundus photography and spectral domain-optical coherence tomography (SD-OCT) images. No obvious pigmentation was noted after human embryonic stem cell-retinal pigment epithelium (hESC-RPE) cell transplantation. The retinotomy site for the hESC-RPE cell suspension injection is denoted (A, B, C, D, inset, white arrow). Large laser scars were noted around the retinal hole at the superior temporal periphery. Attached retinas with successful 360° scleral encircling and laser scars were noted (C). Electroretinogram (ERGs) at baseline (E) and 3 years postoperatively (F); there were no significant changes. Goldmann visual field examinations at baseline (G) and 3 years postoperatively (H) show a central scotoma of slightly diminished size.

DISCUSSION

The primary objectives of this study were to determine the long-term safety and tolerability of subretinal hESC-RPE cell suspension transplantation in human patients. We did not encounter any severe systemic adverse events, including events related to immunosuppression, during the 3-year follow-up period (online supplementary table S2). No severe ophthalmic adverse events were observed related to the hESC-RPE cells; however, RRD occurred in the third patient at 19 weeks after transplantation (online supplementary tables S1 and S3). In a study of 1862 patients who underwent sutureless 23-gauge or 25-gauge vitrectomy, Rizzo et al ¹⁸ observed a postoperative retinal detachment rate of 1.7% after a median of 49 days. In our patient, RRD occurred at 19 weeks after the first 23-gauge vitrectomy, which was later than the median duration reported by Rizzo et al. An atrophic retinal hole was observed at the peripheral retina, which may have been related to the initial surgical procedure (ie, pars plana vitrectomy); alternatively, it may have occurred spontaneously due to the thin, degenerated retina. The hole was deemed unrelated to subretinal transplantation of the hESC-RPE cell suspension, because it was located distant from the site of transplantation and there were no signs of proliferative vitreoretinopathy. Leung et al described a patient who developed retinal detachment and proliferative vitreoretinopathy after pars plana core vitrectomy and subretinal injection of autologous bone marrow-derived stem cells.¹⁹ They reported that the posterior hyaloid was adherent to the retina and substantial epiretinal membrane peeling occurred during retinal reattachment surgery. Our patient exhibited no evidence of tractional membranes; moreover, we found no RPE cell-related proliferative vitreoretinopathy manifestations (eg, vitreous pigment clumps or proliferative vitreoretinopathy membrane containing RPE cells), as observed in patients with proliferative vitreoretinopathy after trauma.^{19 20} Retinal detachment in this patient was presumably unrelated to the transplanted hESC-RPE cells.

The main concern in the application of hESC-derived cell therapy is the potential for tumorigenicity. Klimanskaya et al reported teratoma formation within 8 weeks after injection of hESCs into non-obese diabetic/severe-combined immunodeficient mice.⁹ All of our patients were followed up for at least 3 years; we found no abnormal proliferation suggestive of teratoma formation using colour fundus photography, autofluorescence imaging, and spectral domain-optical coherence tomography with 3 μm resolution. In addition, we observed no severe ocular inflammation or signs of obvious immune rejection (eg, fluid collection, oedema, fibrous membrane formation, persistent leakage on fluorescein angiography, greying of the graft, or loss of its pigmentation) in any of our patients (online supplementary table S3). A small preretinal pigmentation over the epiretinal membrane developed in the third patient at 3 months after transplantation. The pigmentation was presumed to originate from inadvertent preretinal hESC-RPE cell injection or reflux of transplanted hESC-RPE cells from the subretinal space. The epiretinal membrane was monitored closely, because a previous study indicated the occurrence of clinically significant epiretinal membrane requiring surgical treatment.⁴ However, pathological changes that could compromise visual function (eg, macular wrinkling or proliferative vitreoretinopathy) were not observed during the 3-year follow-up period in this study.

Subretinal pigmentation was observed at the injection site in one patient at 4 weeks after transplantation; it then remained stable for 3 years. This type of pigmentation was also reported in previous studies using hESC-RPE cell suspensions.^{4 13 21} In addition, multiple spots of enhanced autofluorescence were observed, which also remained stable for 3 years; similar observations were reported previously in similar studies of hESC-RPE cell suspensions.^{4 21} Small deposits were observed at the inner side of Bruch’s membrane on spectral domain-optical coherence tomography at areas of subretinal pigmentation.⁴ Currently, there is no scientific evidence in humans to indicate that these areas of subretinal pigmentation at subretinal hESC-RPE injected areas are the results of successful engraftment; we can only speculate on several possible mechanisms, based on the findings in previous studies. In a case report regarding histology findings in a patient with age-related macular degeneration patient who underwent hESC-RPE cell transplantation, the donor hESC-RPE cells survived in a preretinal pigmented membrane; this suggested that pigmentation in subretinal locations may be indicative of hESC-RPE cell engraftment.²² Another study showed that native adult RPE cells may proliferate and migrate within 72 hours after retinal detachment in cats.²³ Proliferation of host RPE cells or ingestion of released pigment from hESC-RPE by recipient cells may also explain the appearance of subretinal pigmentation after injection of hESC-RPE cell suspension.⁴ Although the same hESC-RPE cells were used, patients in a clinical trial in the UK showed the development of subretinal pigmentation in all study eyes¹³; this incidence was much higher than in other trial results from the USA,²¹ where only 13 of 18 eyes showed pigmentation. The differences in rates of subretinal pigmentation among these studies may be related to the different conditions involved in preparation and injection of the cells. However, the underlying reasons cannot be definitively determined, based on current knowledge.

In this study, the annual rates of progression of the DDAF area were 1.48 mm² in the treated eye and 0.52 mm² in the untreated eye of the first patient. The DDAF areas expanded rapidly in the second patient (2.63 and 3.03 mm²/year in treated and untreated eyes, respectively); this patient had an extensive DDAF area at baseline (36.82 and 32.03 mm² in treated and untreated eyes, respectively). These results were consistent with the previously reported natural course of SMD; the annual progression rate of DDAF was more rapid in patients who had larger initial DDAF areas.^{2 16} The slower annual progression rates in treated and untreated eyes of the third patient were also consistent with the natural course of SMD; the annual progression rate of DDAF was more rapid in patients who had DDAF with a heterogeneous background, compared with patients with a homogeneous background.^{2 16}

The sample size in this study was insufficient for statistical power calculation or definitive determination of efficacy. The visual acuity could be interpreted as a safety parameter to monitor adverse effects. Notably, there was no decline in visual acuity in the three treated eyes (figure 3B), although SMD is a progressive disease that causes a loss of 0.3 lines of BCVA per year.^{2 16} The second patient showed moderate BCVA improvement (+19 letters) in the study eye, compared with the untreated eye (+6 letters); the first patient showed stable BCVA in both eyes (+9 letters in the treated eye and +7 letters in the untreated eye) during the 3-year follow-up period. The patient who underwent RRD also had stable BCVA during the 3-year follow-up period. The BCVA improvement in the second patient was accompanied by an increase in total NEI VFQ-25 score of 7.47 points. These results were consistent with previous reports indicating that a mean change of at least 15 letters in visual acuity was correlated with a change of 5–10 points in NEI VFQ-25 score.^24–27 However, the improvement in VFQ-25 did not obviously reflect visual acuity in the first and third patients. The results suggested that the second patient could experience a difference in subjective visual function, based on visual acuity improvement of the worse eye. A previous study showed that spontaneous improvement in visual acuity in the worse-seeing eyes of patients with bilateral geographic atrophy and central scotoma could be related to deterioration in visual acuity of the better-seeing eye.²⁸ However, the second patient showed improvement of visual function in both eyes; the visual acuity of that patient’s better-seeing eye did not show deterioration. Humphrey visual field examinations were unreliable because of poor fixation and high false-negative rates in these patients with advanced SMD. Other functional examinations showed no clinically significant changes.

In summary, the results of this study suggest the long-term safety, tolerability, and feasibility of subretinal hESC-RPE transplantation. Further randomised, multicentre trials with larger numbers of patients are needed to confirm the efficacy and safety of hESC-RPE cell suspensions as viable treatment for patients with SMD.