Neovascular age-related macular degeneration (nAMD) and diabetic macular edema (DME) are the leading causes of visual impairment and blindness globally.1,2 While nAMD and DME are distinct disease entities, angiogenesis and vascular leakage play central roles in the development and progression of both conditions. Current standard-of-care treatment for nAMD and DME consists of intravitreal (IVT) anti-VEGF therapy with proven efficacy in multiple large clinical trials.3, 4, 5, 6, 7, 8 Despite these advances, anti-VEGF therapy is limited by short-term durability and inconsistent dosing frequency. Moreover, a substantial subset of patients exhibits a suboptimal or declining response over time, and disease progression may continue, highlighting the need for alternative or adjunctive interventions targeting non-VEGF pathways.9
Roundabout receptors (Robo) and their slit ligands (slit ligand for roundabout protein) regulate diverse biological process across multiple cell types, including leukocytes, neurons, vascular smooth muscle cells, and endothelial cells.10 Roundabout receptor 4 is expressed specifically on endothelial cells, and Robo4 stimulation has been implicated in the reduction of both vascular leakage and angiogenesis in various disease models, including neovascular eye disease.10, 11, 12, 13, 14 Mechanistically, Robo4 stimulation causes the inhibition of Rac1,12 a Rho family guanosine triphosphatase, which can be activated in endothelial cells by VEGF and other angiogenic factors. Additionally, endothelial Robo4 expression has been detected in the eyes of patients with neovascular retinopathies, including within the endothelium of choroidal fibrovascular tissue in patients with age-related macular degeneration,15 indicating that Robo4 may regulate the development of choroidal neovascularization (CNV) in patients with this condition. Therefore, Rac1 inhibition may be an effective target to inhibit angiogenesis in patients with retinopathy.
Collectively, these findings support Rac1 inhibition via Robo4 activation as a biologically plausible therapeutic strategy for retinal vascular disease.
Agonistic monoclonal antibody against Robo4 (DS-7080a) is an agonistic anti-Robo4 antibody that inhibits Rac1 signaling via Robo4 stimulation. Preclinical studies of DS-7080a in primates did not identify any treatment-related adverse effects.16 Additionally, in vitro nonclinical studies of DS-7080a have demonstrated that it inhibited the migration of human umbilical vein endothelial cells induced by angiogenic factors that had been detected in the eyes of patients with nAMD.15In vivo studies also illustrated the ability of DS-7080a to block the development of vascular leakage in laser-induced CNV in cynomolgus monkeys.16,17 Coadministration of DS-7080a plus ranibizumab via IVT injection in monkeys did not result in additional safety concerns.16 Therefore, we hypothesized that DS-7080a might safely reduce pathological vascular leakage, alone or in combination with ranibizumab. The trial was a phase I, first-in-human study designed to evaluate the safety and tolerability of DS-7080a as monotherapy or adjunctive with ranibizumab in patients with nAMD and DME.
In addition, we exploratively evaluated the effects of treatments on best-corrected visual acuity (BCVA) and the central subfield thickness (CST) on spectral-domain OCT (SD-OCT).
Methods
Study Design
The study was a phase I, first-in-human, multicenter, open-label, multiple dose-escalation and dose-expansion clinical trial in subjects with nAMD or DME. The trial was conducted at 12 centers in the United States between July 2015 and January 2018. The trial was divided into 3 parts: part 1 (first phase of the study [P1]) and part 2 (second phase of the study [P2]) involved subjects with nAMD receiving dose escalation and dose expansion, respectively, while part 3 (third phase of the study [P3]) involved subjects with DME receiving dose expansion only.
Part 1 consisted of 3 sequential ascending DS-7080a dose levels (1.0, 2.0, and 4.0 mg) each consisting of 3 to 6 subjects following a traditional “3 + 3” design for a total of 9 to 18 subjects. The purpose of P1 was to assess the safety and tolerability of different doses of DS-7080a and to establish the maximum tolerated dose (MTD) to be used in P2 and P3.
In P2, 27 subjects were randomized and assigned to one of the following 3 treatment arms: monotherapy with DS-7080a at MTD (4.0 mg), monotherapy with ranibizumab (0.5 mg), or treatment with both DS-7080a (4.0 mg) and ranibizumab (0.5 mg). In P3, 20 subjects with DME were randomized to one of the 2 treatment arms: monotherapy with DS-7080a at MTD (4.0 mg) or monotherapy with ranibizumab (0.3 mg).
The rationale for using 3 treatment arms in P2 was to allow the comparison of DS-7080a with ranibizumab and to explore any potential additive or synergistic effect of combining DS-7080a with ranibizumab in nAMD, where anti-VEGF therapy is the standard of care. In contrast, P3, which enrolled patients with DME, was designed with 2 arms (DS-7080a vs. ranibizumab) to streamline evaluation and focus primarily on the safety and tolerability of DS-7080a in a smaller target population.
The study consisted of 8 visits and included a 14-day screening phase, an 84-day treatment phase, and a 28-day follow-up phase. All patients received 3 treatments at baseline/day 1, day 29, and day 57. As this was a phase I, first-in-human trial, no formal power calculation was performed.
The primary objective was to establish the MTD and to assess the safety and tolerability of DS-7080a, consistent with early-phase exploratory trial design.
This trial was designed as an open-label study, meaning that patients and treating investigators were aware of treatment assignment. However, all imaging endpoints were evaluated by a centralized independent reading center, and the sponsor's data analysts were masked to treatment allocation during image grading. Although the clinical study was completed in January 2018, analysis and reporting were delayed due to extended internal review processes and the coronavirus disease 2019 pandemic, which disrupted timelines for data analysis and manuscript preparation.
The overall study design is illustrated in Figure 1.Figure 1The overall study design and structure for each part of the trial. AE = adverse event; DME = diabetic macular edema; DS-7080a = agonistic monoclonal antibody against Robo4; MTD = maximum tolerated dose; nAMD = neovascular age-related macular degeneration.
Eligibility Criteria
Participants were eligible for the nAMD cohort if they were ≥50 years of age and met the following characteristics: active primary subfoveal CNV lesions secondary to age-related macular degeneration; CNV ≥50% of total lesion size in the study eye; CST >315 μm on SD-OCT as assessed by the independent reading center; and ETDRS letter score for BCVA (ETDRS-BCVA) ≤49 in the study eye and ≥49 in the fellow eye for P1 or ETDRS-BCVA letter score between 28 and 78 in the study eye for P2.
Participants were eligible for the DME cohort if they were ≥18 years of age and had the following characteristics: diagnosis of type 1 or type 2 diabetes mellitus and a glycosylated hemoglobin A1c of ≥6.2% and ≤12.0%; presence of intraretinal or subretinal fluid due to DME with CST of ≥335 μm; leakage observed on fluorescein angiography in the study eye confirmed by the reading center; and ETDRS-BCVA letter score between 25 and 78 in the study eye.
Ethical Considerations
The study was registered and is publicly available under the ClinicalTrials.gov website with study ID NCT02530918.18 Institutional Review Board approval was obtained from the Alpha Institutional Review Board. The study was conducted in accordance with the CONsolidated Standards of Reporting Trials recommendations, the Declaration of Helsinki, Good Clinical Practice guidelines, and the Health Insurance Portability and Accountability Act. All participants provided written informed consent prior to enrollment.
All imaging assessment was performed at the centralized reading center by masked graders. All patient data were deidentified and stored in a password-secure browser only accessible to the study team.
Study Drug
Agonistic monoclonal antibody against Robo4 is a full-length immunoglobulin G2 obtained from mice immunized with the extracellular domain of human recombinant Robo.19 Agonistic monoclonal antibody against Robo4 was provided to the site as a frozen sterile solution in a glass vial containing 0.3 mL of an 80 mg/mL DS-7080a solution formulated at pH 5.3 in 10 mM acetate buffer, 100 mg/mL trehalose, and 0.01 w/v% polysorbate 20. It was aseptically diluted with 5% dextrose injection and administered by IVT injection of 50 μL for each dose (1.0, 2.0, or 4.0 mg) in the study eye of each subject.
Ranibizumab was administered as IVT injections of 0.5 mg (nAMD) or 0.3 mg (DME) aqueous solution containing 10 mM histidine HCl; 10% α,α-trehalose dihydrate; 0.01% polysorbate 20 with pH 5.5. Topical anesthesia and 5% povidone iodine antisepsis were administered in the study eye before IVT injection. In case of combination treatment, ranibizumab was administered ≥30 to 60 minutes prior to DS-7080a injection.
Study drugs were prepared and injected intravitreally according to the instructions from the manufacturer.
Rescue Criteria
Subjects could be rescued at the discretion of the principal investigator based on the following criteria: (1) a decrease in BCVA in the study eye of ≥10 letters; (2) an increase of 100 μm in CST on SD-OCT due to disease worsening; or (3) any other clinical concern related to subject safety.
Rescue treatment was mandatory when predefined criteria were met (≥10-letter decrease in BCVA, ≥100 μm increase in CST, or other health-related concerns), with final decision at the investigator's discretion to account for individual patient circumstances. All rescue interventions were documented, and patients who required rescue were discontinued from study drug but continued to be followed for safety assessments.
Assessments
Spectral-domain OCT, ETDRS-BCVA, and adverse events (AEs) assessments were performed at screening, all visits during the treatment period (day 01, 08, 15, 29, and 57), and during follow-up (day 85 and 113).
Spectral-domain OCT scans were captured and assessed using Spectralis OCT (Heidelberg Engineering). Best-corrected visual acuity was assessed according to the ETDRS protocol by certified examiners. Adverse events queries were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events Version 4.0.
Fluorescein angiography was performed at screening and day 85, and fundus photography was captured at screening and days 1, 29, 57, 85, and 113 visits. Blood and urine samples were collected at all visits for evaluation of DS-7080a systemic concentration. Physical examination and electrocardiogram were performed at screening and days 29, 85, and 113.
Outcome Parameters
Safety and tolerability of DS-7080a were assessed by monitoring the incidence, nature, frequency, and severity of systemic and ocular AE, as well as fundus photography findings, vital sign measurements, physical examinations and clinical laboratory test results, electrocardiogram recordings, and antidrug antibody titers.
Treatment-emergent AEs (TEAEs) were defined as any AE that occurred or worsened after the first administration of the study drug and within 28 days of the last dose. Serious AEs (SAEs) were defined according to National Cancer Institute Common Terminology Criteria for Adverse Events criteria as events that were life-threatening, resulted in hospitalization, disability, or death, or were considered medically significant.
Tolerability was assessed using the National Cancer Institute Common Terminology Criteria for Adverse Events definitions and subspecifications for TEAEs and SEAs.
The exploratory efficacy endpoints included: CST measured on SD-OCT and visual acuity measured as ETDRS-BCVA letter score.
Statistical Methods
While statistical comparisons between dose cohorts were not made in P1, the trial aimed to establish MTD and demonstrate the safety and tolerability of the study drug.
Treatment groups within the cohort were compared by exploratory inferential statistics. The statistical summaries were based on the full analysis set (FAS), which included all randomized patients with >1 dose of study drug and >1 measurement of efficacy endpoints. Patients who discontinued prior to receiving a dose or prior to first efficacy measurement were not included in the FAS.
Given the early-phase nature of this trial and small sample size, a full intention-to-treat analysis including all randomized subjects was not performed. This approach may introduce attrition bias, and we acknowledge this as a limitation.
Changes from baseline in CST and ETDRS-BCVA letter score were summarized by treatment and visit using descriptive statistics. Reported parameters included number of subjects (n), mean, standard deviation (SD), median, and range. Comparisons for the mean change from baseline between treatment arms were performed using a repeated measure analysis of covariance model with baseline value of the continuous variable and treatment as factors.
Ranibizumab treatment was used as the standard reference. Normality of continuous outcomes (BCVA, CST) was assessed using the Shapiro–Wilk test. The mean treatment difference, standard error, P value, and the 90% confidence interval on the mean difference were presented for each comparison by visit.
Confidence intervals were reported at the 90% level in alignment with the exploratory nature of this phase I trial, which focused on safety and tolerability rather than hypothesis testing.
Given the small sample size, results were interpreted cautiously, and variability was considered in descriptive summaries. The primary analyses were exploratory and employed repeated-measures analysis of covariance models; however, nonparametric summaries (medians, ranges) were also reviewed to ensure robustness. Due to limited enrollment, mixed-effects modeling was not feasible, but the potential influence of outliers was acknowledged when interpreting results.
This level of precision is commonly employed in early-phase studies to provide a descriptive estimate of treatment effects without overstating statistical significance.
Results
Baseline Demographics
All enrolled subjects were included in the safety, dose-limiting toxicity, and FAS data analysis sets. In P1, all subjects were White adults aged 52-89 years, with a mean ± SD) age of 67.0 ± 7.21, 72.0 ± 14.93, and 64.7 ± 11.02 years for the 1.0 mg, 2.0 mg, and 4.0 mg dose cohorts, respectively. Most subjects in P2 were also White adults aged 61-90 years, and mean ± SD age was comparable across treatment groups.
The safety analysis set for P3 included 20 subjects aged 46-74 years with a mean ± SD age of 59.6 ± 7.68 years.
Of the 36 subjects enrolled in P1 and P2, 35 subjects completed the study, and of the 20 enrolled subjects in P3, 17 completed the study. Twenty-five subjects (69%) across treatment groups in P1 and P2 had prior treatment for nAMD. Baseline BCVA and CST values showed some imbalance between treatment groups, particularly in P2, where mean BCVA was lower and CST slightly higher in certain arms.
These differences should be considered when interpreting comparative efficacy outcomes. Table 1 summarizes the patient demographics and baseline characteristics.Table 1Baseline Demographics of Enrolled Subjects in Parts 1, 2, and 3 of the Study Reported as n (%)Baseline DemographicsPart 1 (Dose Escalation)Part 2 (Dose Expansion)Part 3 (Dose Expansion)DS-7080a OnlyDS-7080a4 mg Only (N = 9)RBZ0.5 mg Only (N = 9)DS-7080a 4 mg + RBZ 0.5 mg (N = 9)DS-7080a4 mg Only (N = 10)RBZ0.3 mg Only (N = 10)1 mg (N = 3)2 mg (N = 3)4 mg (N = 3)Sex Male2 (66.7)0 (0.0)2 (66.7)4 (44.4)4 (44.4)5 (55.6)4 (40.0)6 (60.0) Female1 (33.3)3 (100.0)1 (33.3)5 (55.6)5 (55.6)4 (44.4)6 (60.0)4 (40.0)Ethnicity Not Hispanic/Latino3 (100.0)3 (100.0)1 (33.3)8 (88.9)9 (100.0)9 (100.0)4 (44.4)7 (70.0) Hispanic or Latino0 (0.0)0 (0.0)2 (66.7)1 (11.1)0 (0.0)0 (0.0)5 (55.6)3 (30.0)Race White3 (100.0)3 (100.0)3 (100.0)9 (100.0)7 (77.8)9 (100.0)8 (88.9)10 (100.0) African American0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0) Asian0 (0.0)0 (0.0)0 (0.0)0 (0.0)1 (11.1)0 (0.0)0 (0.0)0 (0.0) American Indian/Alaskan Native0 (0.0)0 (0.0)0 (0.0)0 (0.0)1 (11.1)0 (0.0)1 (11.1)0 (0.0) Other0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)Age (yrs) Mean67.07264.780.479.675.959.560.1 SD7.214.911.04.95.80.17.48.0 Median65.066.070.081.079.076.061.060.0 Range61-7561-8952-7273-8670-9061-9046-6748-74DS-7080a = agonistic monoclonal antibody against Robo4; RBZ = ranibizumab; SD = standard deviation.
Safety Analysis
Table 2 summarizes the safety of DS-7080a in patients with nAMD and DME, the details of which are described later in the article. Some AEs were classified both as severe TEAEs and as SAEs, reflecting regulatory reporting definitions rather than distinct clinical events. For example, acute mountain sickness, pancreatitis, and sepsis were each reported once per subject but coded both as severe TEAEs and as SAEs to avoid the impression of double counting.
No subject in any cohort died during the study. All cases of withdrawal due to drug-related TEAEs have been detailed later in the article. Specifically, 3 ocular TEAEs (blurred vision, decreased visual acuity, and visual distortion) that occurred in a single subject receiving DS-7080a monotherapy led to treatment discontinuation, and one additional case of transient visual distortion in a subject in the combination arm led to discontinuation.
The vitreous hemorrhage reported in 1 subject was judged by the investigator as unrelated to DS-7080a or ranibizumab, as the event occurred in the context of proliferative diabetic retinopathy progression.Table 2Safety and Efficacy Analyses Sets for nAMD and DME Subjects in Parts 1, 2, and 3 of the StudyParametersPart 1 (Dose Escalation)Part 2 (Dose Expansion)Part 3 (Dose Expansion)DS-7080a OnlyDS-7080a4 mg Only (N = 9)RBZ0.5 mg Only (N = 9)DS-7080a 4 mg + RBZ 0.5 mg (N = 9)DS-7080a4 mg Only (N = 10)RBZ0.3 mg Only (N = 10)1 mg (N = 3)2 mg (N = 3)4 mg (N = 3)Safety analysis TEAEs Subjects reporting any TEAE0 (0.0)3 (100.0)2 (66.7)5 (55.6)4 (44.4)6 (66.7)6 (60.0)5 (50.0) Total number of TEAEs01221311252025 TEAEs with outcome of death Subjects with any TEAE with outcome of death0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0) Total number of TEAEs with outcome of death00000000 TEAEs leading to dose D/C Subjects reporting any TEAE leading to dose D/C0 (0.0)0 (0.0)0 (0.0)1 (11.1)0 (0.0)0 (0.0)0 (0.0)0 (0.0) Total number of TEAEs leading to dose D/C00010000 Severe TEAEs Subjects reporting a severe TEAE0 (0.0)0 (0.0)0 (0.0)0 (0.0)1 (11.1)1 (11.1)2 (20.0)1 (10.0) Subjects reporting a severe ocular TEAE0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)1 (10.0)0 (0.0) Total number of severe TEAEs00001221 Treatment-related TEAEs Subjects reporting a TEAE related to Rx0 (0.0)0 (0.0)0 (0.0)1 (11.1)0 (0.0)1 (11.1)0 (0.0)1 (10.0) Subjects reporting an Rx-related ocular TEAE0 (0.0)0 (0.0)0 (0.0)3 (33.3)0 (0.0)1 (11.1)0 (0.0)1 (10.0) Subjects reporting an Rx-related severe TEAE0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0) Total Rx-related TEAEs00030102 SAEs Subjects reporting an SAE0 (0.0)0 (0.0)0 (0.0)1 (11.1)1 (11.1)1 (11.1)2 (20.0)1 (10.0) Subjects reporting an Rx-related SAE0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0)0 (0.0) Total number of SAEs00011221Efficacy analysis CST, study eye (μm) Mean445.0385.0529.3519.6518.1473.1506.2503.9 SD86.6124.0137.2154. 5210.7140.372.9192.8 Median396.0385.0592.0499.0413.0398.0509.0430.0 Range394–545261–509372–624383–856368–1012326–777397–633337–-908 ETDRS-BCVA (letters) Mean37.319.042.059.251.652.658.860.2 SD5.713.74.415.817.612.2411.59.5 Median39.016.040.064.059.052.058.062.5 Range31–427–3439–4732–7324–7338–7134–7544–69BCVA = best-corrected visual acuity; CST = central subfield thickness D/C = discontinuation; DME = diabetic macular edema; DS-7080a = agonistic monoclonal antibody against Robo4; nAMD = neovascular age-related macular degeneration; RBZ = ranibizumab; Rx = treatment; SAE = serious adverse event; SD = standard deviation; TEAE = treatment-emergent adverse event.All variables (i.e., TEAE, SAE, etc.) are reported as n (%).
Rescue Criteria
In total, 3 subjects met predefined rescue criteria and were treated accordingly. One subject in the DS-7080a monotherapy arm was rescued after experiencing a ≥10-letter BCVA decrease, and 2 subjects in the DME cohort (one in each treatment arm) were rescued due to CST increase and vision decline. In all cases, the study drug was discontinued, and patients continued in safety follow-up.
TEAEs in nAMD Subjects
A total of 63 TEAEs were reported by 20 subjects (55.6%) across P1 and P2. In P1, 12 TEAEs were reported by 3 (100.0%) subjects who received 2.0 mg doses of DS-7080a and 2 (66.7%) subjects who received 4.0 mg doses of DS-7080a. However, there were no SAEs, no TEAEs rated as severe, and no TEAEs were determined to be related to study drug.
No dose-limiting toxicities were observed, and the 4.0-mg dose was therefore established as the MTD for P2 and P3.
In P2, 49 TEAEs were reported by 5 (55.6%) subjects in the DS-7080a monotherapy group, 4 (44.4%) subjects in the ranibizumab monotherapy group, and 6 (66.7%) subjects in the combination group. Overall, of the 27 subjects in P2, 2 (7.4%) experienced a total of 3 TEAEs that were rated as severe, but they were nonocular and not related to the study drug. These include acute mountain sickness in 1 subject and acute pancreatitis and severe sepsis in the second subject.
Two subjects (7.4%) experienced 4 TEAEs that were determined to be related to the study drug. Three of these related TEAEs occurred in 1 subject in the DS-7080a monotherapy group and consisted of blurred vision, decreased visual acuity (13 letters), and visual distortion, which led to treatment discontinuation. Rescue therapy was provided and the subject recovered from this AE.
The fourth of these related TEAEs occurred in 1 subject from the combination group who reported visual distortion (“seeing black”) for 3 minutes after the second injection (DS-7080a). Additionally, 4 SAEs unrelated to the study drug were experienced by 3 subjects (11.1%). These consisted of urinary tract infection, acute mountain sickness, acute pancreatitis, and severe sepsis.
TEAEs in DME Subjects
In P3, 11 (55.0%) subjects reported a total of 45 TEAEs; 6 (60.0%) subjects in the DS-7080a group reported 20 TEAEs, and 5 (50.0%) subjects in the ranibizumab group reported 25 TEAEs. Three subjects (15.0%) had a TEAE rated severe (2 DS-7080a-treated subjects, 1 ranibizumab-treated subject). These were not related to the study drug.
In the DS-7080a treatment group, 1 subject had increased intraocular pressure and a second had worsening of coronary artery disease. In the ranibizumab treatment group, 1 subject experienced renal failure. One ranibizumab-treated subject (10%) reported 2 TEAEs which were related to the study drug.
These were an incidence of subconjunctival hemorrhage and ocular soreness. A total of 2 DS-7080a-treated subjects (20%) and 1 ranibizumab-treated subject (10%) experienced 3 SAEs cumulatively. No TEAEs caused a discontinuation of study drug in either P3 treatment group.
These TEAEs included urinary tract infection, worsening hypertension, and vitreous hemorrhage.
Efficacy Analysis
Table 2 summarizes the changes in BCVA and CST for patients with nAMD and DME subjects, as detailed later in the article. In nAMD and DME subjects, DS-7080a monotherapy or cotreatment with ranibizumab resulted in minimal or no improvements in BCVA and CST, while ranibizumab alone resulted in substantial improvements in both metrics.
BCVA Change in Subjects with nAMD
A total of 27 subjects were enrolled in P2 of the trial, and 1 subject from the DS-7080a monotherapy arm received rescue therapy and was discontinued from the study. The baseline BCVA mean ± SD of the DS-7080a, ranibizumab, or DS-7080a and ranibizumab treatment groups was 59.2 ± 15.79, 51.6 ± 17.59, and 52.6 ± 12.24 letters, respectively, as shown in Figure 2. Agonistic monoclonal antibody against Robo4 treatment resulted in minimal change in BCVA score from baseline to day 57 and 113.
Treatment with ranibizumab alone resulted in a mean ± SD increase in BCVA score of 51.6 ± 17.59 to 56.6 ± 17.41 letters from baseline to the end-of-treatment visit. Treatment with both DS-7080a and ranibizumab, similar to DS-7080a monotherapy, maintained a near-baseline BCVA score. Only at day 113 was BCVA improvement statistically greater in the ranibizumab monotherapy group relative to the DS-7080a monotherapy group (P = 0.0074) and the combination group (P = 0.0249).
There was a high degree of intersubject variability. At day 113, BCVA increase of >5 letters was observed in 2 subjects in the DS-7080a monotherapy group.Figure 2Best-corrected visual acuity and CST changes from baseline in nAMD and DME subjects for parts 2 and 3 of the study. BCVA = best-corrected visual acuity; CST = central subfield thickness; DME = diabetic macular edema; DS-7080a = agonistic monoclonal antibody against Robo4; nAMD = neovascular age-related macular degeneration.
CST Change in Patients with nAMD
The mean ± SD CST at baseline in the study eye of the DS-7080a, ranibizumab, and DS-7080a with ranibizumab groups was 519.6 ± 154.46 μm, 518.1 ± 210.75 μm, and 473.1 ± 137.9 μm, respectively. The mean ± SD CST change at day 57 from baseline in the study eye of the DS-7080a, ranibizumab, and DS-7080a with ranibizumab groups was 6.4 ± 65.02 μm, –71.4 ± 128.84 μm, –14.7 ± 55.28 μm, respectively. At day 113 from baseline, the mean ± SD change in CST was –58.4 ± 128.22 μm, –23.1 ± 53.70 μm, and 14.0 ± 58.80 μm, respectively, as shown in Figure 2.
At day 8 and day 57, the DS-7080a treatment group differed significantly from ranibizumab treatment (P = 0.0306 and P = 0.0488, respectively). A trend was observed in the ranibizumab treatment group relative to the DS-7080a treatment group and the combination after baseline until day 57. The retinal thickness returned toward baseline level in the ranibizumab group approximately 56 days after the last treatment.
The retinal thickness in the DS-7080a group did not show any significant change throughout the study; however, the greatest mean reduction was observed at day 113. Figure 3 provides a balanced view of CST decrease (first column) and increase (second column) observed in 2 different cases of nAMD receiving DS-7080a monotherapy, reflecting intersubject variability.Figure 3A balanced view of CST changes under DS-7080a monotherapy in 2 cases of neovascular age-related macular degeneration: the first column shows CST decrease while the second column shows CST increase on day 113 from baseline, illustrating the variability of individual responses observed in this early-phase trial. CST = central subfield thickness; DS-7080a = agonistic monoclonal antibody against Robo4.
BCVA Change in Subjects with DME
A total of 20 subjects were enrolled in P3 of the study, and 2 subjects in the DS-7080a group discontinued participation for personal reasons. The baseline BCVA mean ± SD of the DS-7080a and ranibizumab treatment groups was 58.8 ± 11.46 and 60.2 ± 9.46 letters, respectively. The mean BCVA change at day 57 from baseline in the study eye of the DS-7080a and ranibizumab was 0.8 ± 9.22 letters and 10.5 ± 7.38 letters, respectively, and at day 113, it was 3.96 ± 2.13 and 7.04 ± 2.00 letters, respectively, as shown in Figure 2.
The greatest mean BCVA improvement in DS-7080a-treated subjects was at day 113 with 4.08 letters gain; however, changes were variable and did not demonstrate a consistent trend throughout the study visits. Treatment with ranibizumab resulted in a mean increase in BCVA score of up to 10.5 letters at day 57. Only at day 57 did ranibizumab treatment result in a significantly greater change in BCVA score relative to DS-7080a treatment (P = 0.0275).
Visual improvement formed a consistent trend in patients treated with ranibizumab. High intersubject variability was observed in both treatment groups at all visits.
CST Change in Subjects with DME
The mean ± SD baseline retinal thickness in the study eye of the DS-7080a and ranibizumab treatment groups was 506.2 ± 72.95 μm and 503.9 ± 192.85 μm, respectively. The mean CST change at day 57 from baseline in the study eye of the DS-7080a and ranibizumab groups was 26.8 ± 168.75 μm and –130.7 ± 159.33 μm, respectively, and at day 113, it was –71.5 ± 68.98 μm and –117.0 ± 151.11 μm, respectively (Fig 3). Ranibizumab treatment resulted in a significantly greater reduction in retinal thickness compared with subjects treated with DS-7080a from day 8 to day 85 of this study at the 0.05 level in pairwise comparisons.
Ranibizumab treatment maintained a mean of 100 to 200 μm reduction in retinal thickness from baseline. In contrast, the mean CST change from baseline in the DS-7080a treatment group demonstrated minimal and variable increase up until day 113.
Discussion
The purpose of this phase I, first-in-human trial was to evaluate the safety and tolerability of DS-7080a in patients with nAMD via IVT injection. The drug demonstrated to be safe and was generally well tolerated in subjects with nAMD up to the 4-mg dose. Only 2 subjects reported TEAEs related to the study drug consisting of blurred vision, decreased visual acuity, and visual distortion.
As explained earlier, the vitreous hemorrhage observed during the trial was attributed to underlying diabetic retinopathy rather than study drug exposure, consistent with the known natural history of advanced disease. While unlikely to be treatment-related, this event underscores the importance of careful monitoring in clinical populations at risk for spontaneous hemorrhage. For clarity, certain events were counted both as severe TEAEs and as SAEs in alignment with regulatory definitions.
These were not distinct occurrences, but single events reported under both categories. Only 2 subjects (approximately 8% of the total study population) reported drug-related TEAEs consisting of blurred vision, decreased visual acuity, and visual distortion. Although these events were transient and resolved after discontinuation, their occurrence emphasizes the importance of continued safety monitoring in future trials.
The study also assessed the clinical efficacy of DS-7080a alone or in combination with ranibizumab in reducing CST and improving visual acuity. At day 57 in the nAMD cohort, a statistically significant decrease in CST was observed in the group receiving ranibizumab monotherapy, while no statistically significant change was observed in the DS-7080a monotherapy and combination groups. Agonistic monoclonal antibody against Robo4 did not show any efficacy as monotherapy or additional benefits in combination with ranibizumab in terms of CST reduction.
As established, ranibizumab monotherapy resulted in an overall trend toward improvement in BCVA; however, the observed improvement was not statistically significant until day 113. Meanwhile, DS-7080a did not demonstrate consistent visual acuity improvement. Additionally, the outcomes in the combination group were less favorable than in the ranibizumab only group, which further confirms that DS-7080a is not effective in reducing CST and improving visual acuity in patients with nAMD.
The absence of additional benefit in the combination arm compared with ranibizumab monotherapy suggests that DS-7080a did not provide additive efficacy either. This finding reinforces the conclusion that DS-7080a, despite preclinical promise, did not demonstrate functional or anatomic improvement in this phase I study.
On day 57 in the DME cohort, a statistically significant improvement in BCVA and CST reduction was observed in the ranibizumab group, while similar changes were not observed in the DS-7080a group. The efficacy data from both disease cohorts demonstrated that DS-7080a monotherapy did not provide evidence of consistent BCVA improvement or CST reduction patients with nAMD and DME. However, the study is a phase I trial, and the sample size is small; therefore, any efficacy results warrant cautious interpretation.
The data in the study are not consistent with preclinical studies which had concluded that DS-7080a blocked the development of vascular leakage in laser-induced CNV, a primate model of nAMD. Additionally, the pharmacokinetic activity of DS-7080a was not measured in aqueous humor or vitreous humor after injection; therefore, target engagement and Robo4 pathway activation could not be confirmed. As a result, it is not known whether the suboptimal efficacy reflects insufficient ocular exposure, inadequate target engagement, or intrinsic limitations of Robo4 modulation in human disease.
Future DS-7080a trials should incorporate ocular pharmacokinetic and pharmacodynamic assessments before conclusions about clinical efficacy can be made.
Although the study did not meet the secondary endpoints in terms of BCVA improvement and CST reduction, some patients in the nAMD cohort showed fluid reduction at day 113 after being treated with DS-7080a alone. Specifically, out of 8 subjects enrolled in the DS-7080a monotherapy group, 4 subjects showed fluid reduction at day 113, including complete resolution of subretinal fluid and intraretinal fluid in 2 cases, as shown in Figure 3 (first column). It is not known whether the effects were due to the study drug; however, the fluid reduction indicated that the drug might provide some benefits in patients with nAMD.
While a few subjects in the DS-7080a monotherapy group demonstrated resolution of subretinal or intraretinal fluid, these observations were not prespecified efficacy endpoints and should be considered exploratory. Such findings may generate hypotheses for future trials but cannot be interpreted as evidence of treatment effect in the present study.
It is important to acknowledge that CST and BCVA, while commonly used in retinal clinical trials, do not always demonstrate strong correlation. Prior studies have shown only a modest association between anatomic changes on OCT and functional visual outcomes, underscoring that CST is an imperfect surrogate endpoint for BCVA.3,6,7 In the present trial, CST was included as a structural endpoint to complement BCVA, particularly given the early-phase safety study design. However, the limited concordance between CST reduction and visual improvement highlights the need for additional functional and patient-reported outcome measures to better capture treatment benefit.4,9 In addition, reliance on CST as a surrogate endpoint has inherent limitations.
While changes in retinal thickness are sensitive to treatment effects, they do not consistently predict functional improvement, and the clinical relevance of small CST changes is uncertain. Furthermore, the inconsistency between our trial findings and supportive preclinical studies underscores the limitations of animal models in predicting human outcomes as nonclinical models may overestimate therapeutic efficacy due to species-specific differences in ocular pharmacokinetics, disease mechanisms, and drug–target interactions.
This study had several strengths. Patients were randomly assigned to treatments, and all staff at investigative sites, and the sponsor, were masked to treatment groups. Safety and unexpected AEs were evaluated and reported by certified staff using regulatory protocols.
Standardized measures of efficacy were assessed by an independent and experienced reading center. All images were reviewed and assessed by the reading center using formalized reading protocols. Although the trial was open-label, masking of image assessment at the reading center minimized potential bias in evaluation of anatomic endpoints.
The study also had several limitations. The trial is a phase I study with a sample size. Because no formal power calculation was conducted, the study could not detect efficacy differences between treatment arms.
In addition, use of an FAS rather than a strict intention-to-treat approach may have introduced attrition bias, although this is a common feature of early-phase safety-focused trials with small sample sizes. Although we assessed normality, the small sample size also limited the statistical power to apply more advanced models, such as mixed-effects analyses. Therefore, efficacy results should be considered exploratory and interpreted with caution.
Another limitation of the study is baseline imbalance between arms, especially in P2. Differences in BCVA and CST at enrollment may have also influenced observed outcomes, further complicating interpretation in this small sample. Agonistic monoclonal antibody against Robo4 was hypothesized to decrease macular edema; however, central macular edema was not included as an inclusion criterion.
Using CST as an efficacy endpoint could also not account for edema reduction outside of the central macula. Moreover, DS-7080a protein levels and activity were not measured in the aqueous humor or vitreous humor; therefore, whether DS-7080a actually stimulated Robo4 signaling is not known. Additionally, ocular pharmacokinetic and pharmacodynamic activity was not assessed, preventing the confirmation of target engagement and Robo4 pathway activation in vivo.
This limitation leaves uncertainty as to whether the lack of efficacy reflects insufficient pharmacologic effect despite adequate exposure. Lastly, the use of 90% confidence intervals, while appropriate for exploratory phase I studies, may be perceived as less conservative than 95% confidence intervals and should be interpreted with caution.
In conclusion, although DS-7080a was generally well tolerated, several drug-related TEAEs, including ocular events leading to discontinuation, were observed. Agonistic monoclonal antibody against Robo4 does not show efficacy in terms of BCVA improvement and CST decrease when administered alone or in conjunction with ranibizumab.