Ocular and Serum Profiles of Inflammatory Molecules Associated With Retinitis Pigmentosa

Abstract

Purpose: To investigate the profiles and correlations between local and systemic inflammatory molecules in patients with retinitis pigmentosa (RP).

Methods: The paired samples of aqueous humor and serum were collected from 36 eyes of 36 typical patients with RP and 25 eyes of age-matched patients with cataracts. The concentration of cytokines/chemokines was evaluated by a multiplexed immunoarray (Q-Plex). The correlations between ocular and serum inflammatory molecules and their association with visual function were analyzed.

Results: The aqueous levels of IL-6, Eotaxin, GROα, I-309, IL-8, IP-10, MCP-1, MCP-2, RANTES, and TARC were significantly elevated in patients with RP compared to controls (all P < 0.05). The detection rate of aqueous IL-23 was higher in patients with RP (27.8%) compared with controls (0%). In patients with RP, Spearman correlation test demonstrated positive correlations for IL-23, I-309, IL-8, and RANTES between aqueous and serum expression levels (IL-23: ⍴ = 0.8604, P < 0.0001; I-309: ρ = 0.4172, P = 0.0113; IL-8: ρ = 0.3325, P = 0.0476; RANTES: ρ = 0.6685, P < 0.0001). In addition, higher aqueous IL-23 was associated with faster visual acuity loss in 10 patients with RP with detected aqueous IL-23 (ρ = 0.4119 and P = 0.0264). Multiple factor analysis confirmed that aqueous and serum IL-23 were associated with visual acuity loss in patients with RP.

Conclusions: These findings suggest that ocular and systemic inflammatory responses have a close interaction in patients with RP. Further longitudinal studies with larger cohorts are needed to explore the correlation between specific inflammatory pathways and the progression of RP.

Translational Relevance: This study demonstrates the local–systemic interaction of immune responses in patients with RP.

Introduction

Retinitis pigmentosa (RP) or rod-cone dystrophy is a group of inherited retinal degeneration with progressive photoreceptor cell loss caused by a variety of genetic mutations.^1,2 RP affects more than one million people worldwide and is a significant cause of visual impairment in developed countries.^2,3 Most genetic defects associated with RP are expressed exclusively in rod photoreceptor cells, causing rod dysfunction and cell death, followed by secondary cone cell death. Cell death and inflammation closely interact, and accumulating evidence indicates that neuroinflammation modulates the progression of rod and cone degeneration in RP.^4–7 

Inflammatory cytokines and chemokines produced in the retina play a role in activating retina-resident microglial cells, disrupting the blood–retinal barrier, and mediating the engulfment of peripheral immune cells.^4,8–11 Our previous study showed that numerous cytokines and chemokines, including interleukin (IL) 1α, IL-1β, IL-2, IL-4, IL-6, IL-8, IL-10, interferon (IFN) γ, growth-related oncogene (GRO) α, I-309, IFN-γ–inducible protein (IP) 10, monocyte chemotactic protein (MCP) 1, MCP-2, and activated regulated chemokine (TARC), which regulate both innate and adaptive immune responses, were markedly increased in the aqueous humor or vitreous of patients with RP.⁴ In addition, it was reported that aqueous flare values, a marker of inflammation and blood–ocular barrier breakdown, were increased in patients with RP and that higher levels of aqueous flare were associated with lower visual function and faster vision loss in patients with RP.^4,12,13 These findings suggest that neuroinflammation is deeply involved in the pathology of RP, although the causal relationship remains unclear. 

In addition to ocular neuroinflammation, RP is associated with the activation of systemic inflammatory responses. We recently demonstrated that inflammatory monocytes with high expression of CCR2 and CX3CR1 are increased in the peripheral blood of an RP mouse model and human patients and that these inflammatory monocytes contribute to cone cell death in RP model mice.⁶ In line with these findings, levels of IL-8 and regulated activation normal T-cell expressed and secreted (RANTES), which mediate monocyte activation and migration, were elevated in the serum of patients with RP.¹⁴ Associations between increased peripheral inflammation (e.g., proportions of inflammatory monocytes, C-reactive protein, and IL-8) and decreased visual function were also observed.^6,14,15 

These findings suggest that both systemic and ocular neuroinflammation are involved in RP, but the correlations between inflammatory molecules in the eye and serum have not been fully explored. Therefore, in this study, paired aqueous humor and serum samples from patients with RP were subjected to a multiplex enzyme-linked immunosorbent assay (ELISA)–based Q-Plex assay to analyze the correlations between neuroinflammatory molecules in the eye and peripheral blood and their relevance to visual function in patients with RP. 

Methods

Study Design

This retrospective observational study followed the principles outlined in the Declaration of Helsinki, and institutional review board/ethics committee approval was obtained from Kyushu University Hospital (Fukuoka, Japan). All enrolled participants were informed regarding the potential consequences of the study, and written informed consent was obtained from all participants. 

Participants

Thirty-six eyes of 36 patients with typical RP (15 male and 21 female) and 25 age-related patients with cataracts (10 male and 15 female) who underwent cataract surgery at Kyushu University Hospital between 2019 and 2024 were included. The aqueous humor samples were collected during cataract surgery, and the peripheral blood samples were collected on the same date. If both eyes were included in the study, the results of the right eye and its paired serum were subjected to analysis. 

The diagnosis of typical RP was based on a history of night blindness, progressive loss of peripheral vision or ring scotoma, markedly reduced electroretinogram responses, and attenuation of retinal vessels and bone spicule-like pigment clumping in the mid- and peripheral retina. The genetic inheritance patterns were determined based on the detected variants. Patients with other ocular diseases, such as glaucoma, age-related macular degeneration (AMD), or uveitis, and patients with a history of malignancy or taking anti-inflammatory medications were excluded from the study. 

Baseline characteristics such as smoking, alcohol habits, body mass index (BMI), medication, macular complications of RP, and systemic diseases, including hypertension, hyperlipidemia, diabetes mellitus, fatty liver disease, and autoimmune diseases, were retrospectively obtained from the electronic medical charts. 

Clinical Examination

Best-corrected visual acuity (BCVA) of patients was measured with the Landolt decimal VA chart (CV-6000: Tomey, Nagoya, Japan; or AVC-36: Kowa, Nagoya, Japan) at 5 m or with single Landolt test cards (HP-1258; Handaya, Tokyo, Japan). The values were converted into logarithm of the minimum angle of resolution (logMAR) units for statistical evaluation. The BCVA was defined as the minimum Landolt C letter that the patient could correctly recognize >60% (3/5) of the time. The vision loss slope was the rate of visual acuity (VA) change, which was calculated from consecutive logMAR values (at least 3 times) during the postsurgery period (from 1 month to 3 years after cataract surgery). All the participants were identified without posterior capsular opacification prior to collecting BCVA data. All patients underwent automated static perimetry tests with the Humphrey Field Analyzer (HFA) (Humphrey Instruments, San Leandro, CA, USA) using the central 10-2 Swedish Interactive Thresholding Algorithm Standard Program. The HFA results with insufficient reliability (fixation loss, 20%; false positive, 15%; or false negative, 33%) were excluded from the analysis. Detailed funduscopic examinations using fundus photography, optical coherence tomography, and fundus autofluorescence were also carried out on all patients. 

Patients with RP were classified into three groups based on the results of HFA10-2 tests within 6 months of cataract surgery, as previously described¹⁶: mild RP (mean deviation [MD], ≥−5), moderate RP (−25 < baseline MD < −5), and late RP (baseline MD ≤−25). 

Aqueous Humor and Serum Sample Collection

Aqueous humor samples were collected at the start of cataract surgery using a 30-gauge needle. Peripheral blood was collected on the day of cataract surgery. The aqueous humor samples were immediately stored at 4°C and transferred within 3 hours to a –80°C freezer for long-term storage. The serum was isolated by centrifugation for 5 minutes at 1200 × g, dispensed into 1.5-mL tubes, and frozen at –80°C within 1 hour from blood drawing. The samples were kept in a freezer until analyzed and thawed 30 minutes before examination. The freeze–thaw process was conducted only once for each sample. 

Multiplex ELISA-Based Chemiluminescent Assay

The concentrations of 15 cytokines (IL-1α, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-10, IL-12p70, IL-13, 1L-15, IL-17, IL-23, IFN-γ, and tumor necrosis factor [TNF] α, TNF-β) and 9 chemokines (Eotaxin, GRO-α, I-309, IL-8, IP-10, MCP-1, MCP-2, RANTES, and TARC) were measured by a multiplex ELISA-based Q-Plex Human Cytokine array (Quansys Biosciences, West Logan, UT, USA). The signals of the cytokine/chemokine arrays were determined using chemiluminescence and imaged with LAS 4010 CCD cameras (Fujifilm, Tokyo, Japan). 

Statistical Analyses

The data of visual acuity and BMI were presented as the arithmetic mean ± standard deviation. The ages of enrolled patients with RP and cataracts and the levels of cytokines/chemokines are shown as the median and interquartile range. 

If the detection rate was ≥80% in either aqueous humor or serum samples, statistical differences in the concentrations of cytokines/chemokines between the paired aqueous humor and serum samples were analyzed by Wilcoxon signed rank test, and those between RP and control groups were analyzed with the Mann–Whitney test. Differences in frequency were analyzed by the Pearson χ² test if the detection rate of a given cytokine was ≥50% in either aqueous humor or serum, as previously described.^14,17 

Correlations between cytokine or chemokine values in the aqueous humor or serum and visual function were analyzed by Spearman's rank test. Correlations between basic characteristics and cytokine/chemokine levels were analyzed by using Spearman's correlation test (for continuous variables) or Kendall's tau test (for categorical variables). 

All statistical analyses were performed using the statistical package IBM SPSS Statistics software (SPSS) version 26.0 (SPSS, Inc., Chicago, IL, USA). Values of P < 0.05 were considered to indicate statistical significance. The cytokine/chemokine data were input into GraphPad Prism (version 8.0; GraphPad Software, San Diego, CA, USA) to generate the heatmaps and correlation scores between aqueous humor and serum samples. Visualization of complex networks was performed using Cytoscape software (version 3.9.1; The Cytoscape Consortium, San Diego, CA, USA). 

Hierarchical Cluster Analysis

The data of Spearman's correlation coefficients among cytokines/chemokines in the aqueous humor and serum were input into SPSS (version 26.0; SPSS, Inc.) to run a hierarchical cluster analysis.¹⁸ The average linkage (between-groups) method and the squared Euclidean distance measurement were used to get the agglomeration schedule, icicle plot, and dendrogram. The agglomeration schedule listed all the stages. The clustering process was stopped and added a termination line on the dendrogram after determining large difference between the coefficients of two consecutive stages. 

Multiple Factor Analysis

Multiple factor analysis (MFA) is a multivariate statistical technique that simultaneously analyzes multiple groups of variables to explore relationships and patterns across data sets.¹⁹ To achieve a comprehensive understanding of the correlations of the grouped cytokines/chemokines in aqueous humor and serum, as well as visual progression and disease severity, MFA of 35 variables from the detectable cytokines/chemokines, visual acuity loss, and mean deviation was performed. MFA distributed seven groups to the global plot (six groups of inflammatory molecules were obtained through hierarchical cluster analysis): group A (5 continuous variables: aqueous RANTES, TARC, IP-10, and MCP-1 and serum RANTES), group B (3 continuous variables: aqueous IL-6 and serum MCP-1 and TARC), group C (7 continuous variables: aqueous IL-23, Eotaxin, GROα, IL-8, I-309, serum IL-23, and IFN-γ), group D (5 continuous variables: serum Eotaxin, GROα, I-309, IP-10, and MCP-2), group E (11 continuous variables: serum IL-10, IL-17, IL-2, IL-8, IL-4, IL-15, IL-6, IL-12, IL-1a, IL-13, and IL-5), group F (aqueous MCP-2 and serum TNFα), and group VAloss_MD (2 continuous variables: visual acuity loss and mean deviation). The FactoMineR (v2.10) package (Agrocampus Ouest and the University of Rennes 1, Rennes, Brittany, France) was used for MFA, and the graphics were obtained using the factoextra (v1.0.7) package (Kassambara, A., France). 

Results

Patient Demographics

Table 1 shows the baseline clinical characteristics of the participants. Thirty-six patients diagnosed with typical RP and 25 age-matched patients with cataracts underwent cataract surgery and were followed up for 3 years after surgery. The sex, age, visual acuity, smoking and alcohol habits, macular complications of RP, BMI, and systemic disease were not significantly different between the two groups. The median age was 65 years (range, 41–81) and 69 years (range, 53–80) in the patients with RP and controls, respectively. The causative gene was identified in 8 of 36 patients with RP. Only one eye (the right eye if both eyes were included) of each patient was analyzed. 

Table 1.

 

Characteristics of All Participating Patients With Retinitis Pigmentosa

 

Differential Inflammatory Profiles Between Patients With RP and Controls

Tables 2 and 3 show the concentrations and detection rates of cytokines/chemokines in the paired aqueous humor and serum samples from patients with RP and controls (Fig. 1). 

Table 2.

 

Values and Detection Rates of Cytokines in the Aqueous Humor and Serum by Multiplex Bead Immunoassay in the Controls and Patients With Retinitis Pigmentosa

 

Table 3.

 

Values and Detection Rates of Chemokines in the Aqueous Humor and Serum by Multiplex Bead Immunoassay in the Controls and Patients With Retinitis Pigmentosa

 

Figure 1.

 

Heatmap analysis of aqueous and serum cytokines and chemokines in RP and control groups. Heatmap analysis of cytokines and chemokines using GraphPad Prism software. P < 0.05 is considered statistically significant. *Representing significantly upregulated molecules in aqueous humor and serum of RP group compared to controls.

In cytokines (Table 2), the detection rate of IL-6 in aqueous humor was upregulated in patients with RP compared to controls (92% vs. 44%, P = 0.0001). Additionally, the detection rate of aqueous IL-23 was higher in patients with RP (27.8%) compared to controls (0%). The expression level of serum TNF-α was upregulated in patients with RP compared with controls (TNF-α: P = 0.0004). 

In chemokines (Table 3), all the detected chemokines in aqueous humor were upregulated in patients with RP compared with controls (P < 0.05). 

Figure 1 shows the heatmap of cytokine and chemokine expression in the aqueous humor and serum of patients with RP and controls. These patients with RP were classified into mild, moderate, and severe disease stages based on the HFA10-2 results. 

Associations Between Ocular and Systemic Inflammatory Profiles in Patients With RP

The correlations between ocular and serum levels for each cytokine or chemokine were analyzed by Spearman's rank correlation test. Expression levels of IL-23, I-309, IL-8, and RANTES showed significant associations between paired samples of aqueous humor and serum from patients with RP (IL-23: ⍴ = 0.8604, P < 0.0001; I-309: ⍴ = 0.4172, P = 0.0113; IL-8: ⍴ = 0.3325, P = 0.0476; RANTES: ⍴ = 0.6685, P < 0.0001; Fig. 2). In control samples, there was no significant correlation between aqueous humor and serum expression of all cytokines and chemokines. 

Figure 2.

 

Correlations of cytokines/chemokines between aqueous humor and serum in patients with RP. The scatterplot shows the relation between the levels of IL-23, I-309, IL-8, and RANTES in aqueous humor to the corresponding levels in serum.

Correlations among the concentrations of inflammatory cytokines/chemokines in the aqueous humor and serum of patients with RP are shown in Supplementary Figure S1. Figure 3 visualizes the complex and significant relationships among ocular and systemic inflammatory molecules in patients with RP. Hierarchical cluster analysis of these molecules revealed six closely related cluster groups (Supplementary Fig. S2). One cluster included aqueous humor MCP-1, RANTES, TARC, IP-10, and serum RANTES, and these molecules were negatively correlated with serum IL-1β, IL-4, IL-6, IL-8, IL-12, IL-13, IL-15, GROα, I-309, and MCP2 (Figs. 4A, 4B). On the other hand, aqueous IL-23, I-309, IL-8, Eotaxin, GROα, and serum IL-23 and IFN-γ formed another cluster, as shown in Figures 4C–E. 

Figure 3.

 

Visualizing network of all molecules. Molecules whose levels in aqueous humor were significantly correlated with those in serum were visualized using Cytoscape software (version 3.9.1). Molecules in aqueous humor are shown in yellow, and those in serum are purple. Spearman's correlation scores are presented with the edge width. Positive correlations are presented in red and negative correlations are presented in blue.

Figure 4.

 

Correlation networks of MCP-1, RANTES, I-309, IL-23, and IL-8. The concentrations of cytokines/chemokines in aqueous humor and serum were significantly associated with the aqueous levels of MCP-1 (A), RANTES (B), I-309 (C), IL-8 (D), and IL-23 (E).

Relationships Between Inflammatory Molecules and Visual Function in Patients With RP

The correlations between inflammatory molecules and visual function were analyzed in the patients with RP. Among inflammatory molecules in the aqueous humor and serum, Spearman's correlation analysis showed that higher aqueous humor IL-23 was associated with a faster rate of VA loss (IL-23: ⍴ = 0.4119, P = 0.0264; Fig. 5). In addition, we conducted MFA of six clusters identified with hierarchical clustering analysis. The correlations between these clusters and the vision loss were visualized in the dimensions in Figures 6A and 6B. The first two dimensions of the MFA explained approximately 37.73% of the variability, 22.97% and 14.76% for the first dimension (Dim 1) and second dimension (Dim 2), respectively. We identified VA loss accounting for almost 37.18% of the variance on Dim 2, which largely summarized the vision progression. MFA depicts positively correlated variables located together within the same quadrant, whereas negative ones are positioned on opposite sides of the plot origin (opposed quadrants). The data showed that aqueous and serum IL-23 positively correlated with VA loss and MD due to their close position and same direction on the factor map (Fig. 6B). 

Figure 5.

 

Relationships between cytokines in aqueous humor and best-corrected visual acuity loss. The scatterplot shows the correlation between aqueous levels of IL-23 and best-corrected visual acuity loss.

Figure 6.

 

MFA for cytokines/chemokines, VA loss, and MD. (A) A global groups plot visualized. (B) The correlation between quantitative variables and dimensions. MFA in R software using the “FactoMineR” package.

MFA also showed that the molecules of group C (aqueous IL-23, Eotaxin, GROα, IL-8, and I-309 and serum IL-23 and IFN-γ) and group E (serum IL-10, IL-17, IL-2, IL-8, IL-4, IL-15, IL-6, IL-12, IL-1a, IL-13, and IL-5) were closely located with VA loss. In contrast, the MCP-1 network (group A: aqueous RANTES, TARC, IP-10, and MCP-1 and serum RANTES) and groups B and D were in the far and opposite position to vision loss in the MFA analysis, indicating the limited and negative correlations with vision loss (Fig. 6B). 

To explore potential covariates associated with aqueous IL-23, differences in age, BMI, sex, baseline BCVA, and MD were analyzed between patients with RP with or without aqueous IL-23 detected (Supplementary Table S1), and there were no significant differences between the groups. 

Associations Between Inflammatory Molecules, Macular Complications, and Other Background Factors in RP

We next analyzed the relationships between cytokine/chemokine levels and macular complications in patients with RP. The levels of IL-8 and I-309 in aqueous humor and I-309 in serum were significantly correlated with macular complications (epiretinal membrane, cystoid macular edema, vitreomacular traction syndrome, and/or macular hole) in patients with RP (aqueous IL-8: ⍴ = 0.428, P = 0.003; aqueous I-309: ⍴ = 0.333, P = 0.021; serum I-309: ⍴ = 0.308, P = 0.033). 

Regarding other factors, aqueous IFN-γ and serum IL-2, IL-6, IL-8, and IL-10 were associated with autosomal dominant causative genes in patients with RP (IFN-γ: ⍴ = 0.482, P = 0.004; IL-2: ⍴ = 0.317, P = 0.036; IL-6: ⍴ = 0.334, P = 0.037; IL-8: ⍴ = 0.319, P = 0.023; IL-10: ⍴ = 0.343, P = 0.025). Serum I-309 was correlated with age (I-309: ⍴ = 0.443, P = 0.007), and serum MCP-1 and TARC were associated with alcohol habit (MCP-1: ⍴ = 0.355, P = 0.011; TARC: ⍴ = 0.314, P = 0.025) (Supplementary Table S2). 

Discussion

In the present study, we investigated the profiles and correlations of ocular and serum cytokines/chemokines in patients with RP. Using multiplex immunoassay, our study showed that (1) the aqueous levels of IL-6 and all the chemokines were significantly increased in patients with RP; (2) the levels of I-309, IL-8, RANTES, and IL-23 in aqueous humor were significantly correlated with those in serum; (3) in 10 patients with RP with detected aqueous IL-23, higher levels of IL-23 in aqueous humor were associated with a faster rate of visual acuity loss; and (4) aqueous and serum IL-23 in patients with RP showed a close association with vision loss in MFA. These findings suggest an interaction between ocular and systemic inflammation in patients with RP. 

IL-23 is a cytokine that can drive the differentiation of Th17 lymphocytes, thereby producing IL-17.²⁰ It has been shown that the IL-23/Th17 pathway is involved in chronic inflammation in uveitis, which regulates the innate immunity and causes dysfunction of RPE cells.²¹ Accumulations of Th17-related cytokines in peripheral blood, like TGF-β, IL-6, IL-23, and IL-17, are associated with Vogt–Koyanagi–Harada and Behçet diseases.^22,23 I-309, also known as CCL1, is a chemokine expressed by various immune cells and stimulates Th2 lymphocytes and monocytes through CCR8 receptors.^24,25 IL-8, also called CXCL8, is a mediator of innate immune response that potently activates neutrophils and monocytes, and it is induced by inflammatory cytokines such as TNF-a, IL-1b, and IL-17.^26–28 Using 36 paired samples, our present study demonstrated that there was positive ocular–systemic correlation for each of IL-23, I-309, IL-8, and RANTES. This implies the presence of inflammation not only in the eye but also in the peripheral blood, suggesting systematic chronic inflammation involving both innate and adaptive immunity that may underlie the neuroinflammation observed in RP. Among these molecules, we also found that higher IL-23 levels in aqueous humor and serum were correlated with faster progression of vision loss, although aqueous IL-23 was detected in only 10 of 36 patients with RP. It has been shown that anti–IL-23 and IL-17 biologic agents are clinically effective for psoriasis and other autoimmune/inflammatory diseases.^29,30 The importance of peripheral immune responses has been highlighted in other retinal diseases such as AMD, and the relevance and differences of neuroinflammation among disease conditions warrant further investigation.^31,32 

MCP-1, also called CCL2, which belongs to the CC chemokine family, regulates the migration and infiltration of leukocytes, including monocytes and macrophages.³³ Higher MCP-1 levels in the vitreous fluid compared with the serum of patients with diabetic retinopathy have been shown to lead to chronic inflammation in the diabetic retina due to the high cell permeability of the blood–retinal barrier.^34–36 In a mouse model of RP, knockout of Mcp-1 or its receptor Ccr2 reduced the engulfment of peripheral monocytes and attenuated retinal degeneration.^6,37 In this study, we found that aqueous MCP-1 levels were extremely high in patients with RP. This suggests that locally produced MCP-1 may activate retina-resident microglia and facilitate the engulfment of peripheral immune cells, thereby inducing neuroinflammation in RP. In addition, our MFA identified that the aqueous MCP-1, IP-10, RANTES, and TARC and serum RANTES were positively associated but negatively correlated with other detected inflammatory molecules. This indicates that the MCP-1 and MCP-1 network may locally modulate systemic neuroinflammation in RP. 

This pilot study has several limitations. First, the small sample size and short follow-up period after surgery made it difficult to determine the correlations between inflammatory molecules and disease progression. In the present study, we demonstrated that aqueous IL-23 was significantly correlated with vision loss, yet it was detected in only 10 of 36 patients. From the standard deviation measures, the required sample size of patients with RP was calculated to be 202 to achieve a power of 80% and a significance level of 5%. 

Second, statistical differences of inflammatory molecules across RP severity were not observed in this study, and thus it was difficult to determine the cause–effect relationships between inflammatory molecules and vision loss in patients with RP. There is a possibility that the elevated IL-23 levels in patients with RP in moderate and severe disease stages may be the result of the advanced disease at high risk of vision loss. 

Third, because our patients with visually significant cataracts were middle to older age, confounding factors other than RP, such as lifestyle-related disorders, might have affected their serum inflammatory profiles. 

In conclusion, our study demonstrated significant associations in the levels of IL-23, I-309, and IL-8 between the eye and peripheral blood, indicating a potential link between local and systemic immune responses in RP. Further investigation with a larger sample size and longer follow-up period will better characterize the relationships between neuroinflammation and disease progression in patients with RP. 

Acknowledgments

Supported by a grant from the Japanese Ministry of Education, Culture, Sports, Science, and Technology (#22H03242; YM); the Japan Agency for Medical Research and Development, Practical Research Project for Rare/Intractable Diseases, #JP22ek0109512h0002; and a Japan Intractable Disease (Nanbyo) Research Foundation grant (#2020C01; KF). SM is supported in part by gifts to the Mukai Fund, Massachusetts Eye and Ear. The sponsor or funding organization had no role in the design or conduct of this research. 

Disclosure: Y. Tao, None; M. Fukushima, None; S. Shimokawa, None; H. Zhao, None; A. Okita, None; K. Fujiwara, None; A. Takeda, None; S. Mukai, None; K.-H. Sonoda, None; Y. Murakami, None