Relationship Between Macular Curvature and Common Causative Genes of Retinitis Pigmentosa in Japanese Patients.

Purpose: To determine the relationship between the macular curvature and the causative genes of retinitis pigmentosa (RP).
Methods: We examined the medical records of the right eyes of 65 cases with RP (31 men and 34 women; average age, 47.6 years). There were 31 cases with the EYS variants, 11 cases with the USH2A variants, six cases with the RPGR variants, 13 cases with the RP1 variants, and four cases with the RP1L1 variants. The mean curvature of Bruch's membrane was calculated within 6 mm of the fovea as the mean macular curvature index (MMCI, 1/µm). We used multiple linear regression analysis to determine the independence of the causative genes contributing to the MMCIs after adjustments for age, sex, axial length, and width of the ellipsoid zone.
Results: The median MMCI was -31.2 × 10-5/µm for the RPGR eyes, -16.5 × 10-5/µm for the RP1L1 eyes, -13.0 × 10-5/µm for the RP1 eyes, -9.8 × 10-5/µm for the EYS eyes, and -9.0 × 10-5/µm for the USH2A eyes. Compared with the EYS gene as the reference gene, the RPGR gene was significantly related to the MMCI values after adjusting for the other parameters (P = 5.30 × 10-6). In contrast, the effects of the other genes, USH2A, RP1, and RP1L1, were not significantly different from that of the EYS gene (P = 0.26, P = 0.49, and P = 0.92, respectively). Conclusions: The RPGR gene had a stronger effect on the steep macular curvature than the other ciliopathy-related genes.

Retinitis pigmentosa (RP) is the most common hereditary retinal degenerative disease worldwide. It is characterized by a reduction in vision, visual field constriction, and night blindness, which develop due to the dysfunction and death of the rod photoreceptors. These changes lead to the death of the cone cells and a reduction of the visual acuity.,

To date, 89 genes causing nonsyndromic RP have been registered in the Retinal Information Network (https://sph.uth.edu/retnet/. Accessed on November 28, 2019). The recent adaptation of high-throughput DNA sequencing technologies has accelerated the identification of the causative genes of Japanese patients with RP, and these studies have revealed the major causative genes in this population.– Some of these major genes were categorized into the ciliopathy-related genes (e.g., the USH2A, RPGR, RP1, and RP1L1 genes),– and it was also suggested that the function of the EYS protein is associated with the cilia.,,

Recently, several studies have reported that steep macular curvatures were observed in the eyes of patients with various inherited retinal disorders, including retinal ciliopathy, Joubert syndrome, Leber's congenital amaurosis, and RP without high myopia.– In our earlier study, we established that the mean macular curvature index (MMCI) can be an additional parameter of the ocular anatomy that can be used to quantify the degree of macular curvature. The MMCI was determined from the spectral-domain optical coherence tomography (SD-OCT) images. We showed that the macular curvature of eyes with RP was steeper than that of normal eyes. We also found a significant relationship between the age, axial length (AL), and the width of the ellipsoid zone (EZ) and the steepness of the macular curvature in the RP cases.

Considering the genetic and clinical heterogeneity of RP, comparisons of the MMCIs associated with the causative genes is important for further understanding of the pathology of RP.

Therefore, the aim of this study was to assess the relationship between the macular curvature and the causative genes for RP.

Methods

Subjects

We reviewed the medical records of the right eye of 72 cases with typical RP from two facilities, Nagoya University Hospital (n = 37) and Kyushu University Hospital (n = 35), which were examined between 2002 and 2019. These cases had been genetically diagnosed with the five most frequent causative genes of RP (EYS, USH2A, RPGR, RP1, and RP1L1) in the Japanese population. The clinical diagnosis was based on the history of night blindness, ring scotoma and/or constriction of the visual fields, and severe rod-cone dysfunction or nonrecordable electroretinograms. In addition, an attenuation of the retinal vessels and bone spicule-like pigment clumping in the mid-peripheral and peripheral retina were detected by experienced ophthalmologists. We excluded seven cases (four EYS eyes, one USH2A eye, one RPGR eye, and one PR1 eye) due to incomplete clinical data, such as the visual acuity (n = 2), AL (n = 1), and SD-OCT images (n = 4). In the end, we studied 65 cases.

This retrospective study was approved by the ethics committee of each hospital (Nagoya University 16-0538-3, Kyushu University 2019-136) and was conducted in accordance with the tenets of the Declaration of Helsinki on biomedical research involving human subjects. The institutional review boards also waived the need for a written informed consent from each patient because the study design was a retrospective chart examination.

Measurements of Ocular Parameters

The best-corrected visual acuity (BCVA) was measured on the same day as the SD-OCT images were taken. For the statistical analyses, we converted the decimal value to logMAR units. We used an IOLMaster (Carl Zeiss Meditec, Dublin, CA, USA) to measure the AL of the eyes.

We selected the horizontal scanned OCT images consisting of 100 averaged images with the eye-tracking system functioning for the analyses. To adjust the size of each image, we corrected for the differences in the pixel resolution between the longitudinal and transverse directions. We measured the width of the EZ between the borders where the EZ band touched the upper surface of the retinal pigment epithelium with the built-in calipers using the Heidelberg Eye Explorer software (Heidelberg Engineering, Heidelberg, Germany). If the entire length of the EZ line exceeded the size of the OCT image, the borders of the EZ line were set to be that of the OCT image.

Measurement of Macular Curvature

We calculated the MMCI as an objective index of the macular curvature as described in detail. Briefly, we quantitatively analyzed the reflective line corresponding to Bruch's membrane across the fovea (yellow line in Fig. 1A in Komori et al.) using MATLAB software (The MathWorks, Inc., Natick, MA, USA). We marked 12 points on the Bruch's membrane line beginning from the fovea (yellow triangles in Fig. 1A in Komori et al.). The marks were separated by 760 µm in the OCT images. Using cubic spline interpolation, the approximate curvature of the marked points was calculated by the software (yellow and red lines in Fig. 1B in Komori et al.). The curvatures in the 6-mm range including the fovea were selected from the calculated curve. To reduce the effects of the optic nerve head, curvature values outside this range were not used. Using all measured values for the local curvature in 1-µm steps, the mean curvature between ±3 mm from the central fovea (red line in Fig. 1B in Komori et al.) was calculated. Plus values of MMCI indicated a convex shape and minus values indicated a concave shape.

Figure 1.

Fundus photographs and SD-OCT images of representative cases with RP for each causative gene. A: Case 1 is a 50-year-old female that was a case with EYS-related retinitis pigmentosa (RP) with an ellipsoid zone (EZ) of 635 µm. The mean macular curvature index (MMCI) was −0.3 × 10−5 µm−1. The axial length (AL) was 24.81 mm. B: Case 2 is a 48-year-old female with EYS-related RP with an EZ of 2762 µm. Bruch's membrane has a steeper curvature than that of Case 1. The MMCI was −36.0 × 10−5 µm−1. The AL was 24.52 mm. C: Case 3 is a 53-year-old male with USH2A-related RP and an EZ of 3266 µm. The MMCI was −8.0 × 10−5 µm−1. The AL was 23.81 mm. D: Case 4 is a 34-year-old male who is a case of RPGR-related RP with an EZ of 2333 µm. The OCT image shows a steep macular curvature. The MMCI was −47.8 × 10−5 µm−1. The AL was 25.83 mm. E: Case 5 is a 23-year-old female who is a case of RP1-related RP without an intact EZ (0 µm). This case had a relatively flat macular line. The MMCI was −5.62 × 10−5 µm−1. The AL was 21.81 mm. F: Case 6 is a 67-year-old male who is a case of RP1L1-related RP without an intact EZ (0 µm). The MMCI was −18.01 × 10−5 µm−1. The AL was 24.09 mm.

Genetic Diagnosis

Blood samples were collected for the genetic analyses. Genetic diagnosis was performed as described in detail.,, A summary of the genetic diagnosis of the RP cases is presented in Table 1.

Table 1.

Genetic Diagnosis of Patients With RP

ID	Age, y	Gender	Causative Genes	Genetic Inheritance Form	Variant 1	Zygosity 1	Variant 2	Zygosity 2
N-27	27	F	EYS	AR	p.(Ser1653fs)	Homo	—	—
N-38	54	M	EYS	AR	p.(Ser1653fs)	Hetero	p.(Tyr2935*)	Hetero
N-122	49	M	RPGR	XL	p.(G1u746fs)	Hemi	—	—
N-71	44	F	EYS	AR	p.(Ser1653fs)	Homo	—	—
N-109	59	M	USH2A	AR	p.(Ser4748Phe)	Hetero	p.(Cys934Trp)	Hetero
N-167	34	F	EYS	AR	p.(Ser1653fs)	Homo	—	—
N-168	29	M	USH2A	AR	p.(Thr3667Pro)	Hetero	p.(Val164Phe)	Hetero
N-169	51	F	EYS	AR	p.(AIa2736Pro)	Hetero	p.(GIn203*)	Hetero
N-180	61	F	EYS	AR	p.(Ser1653fs)	Hetero	p.(Gly2186G1u)	Hetero
N-183	70	F	EYS	AR	p.(Tyr2935*)	Hetero	p.(Asn404fs)	Hetero
N-208	48	F	USH2A	AR	p.(Ser5060Pro)	Hetero	p.(Ser4748Phe)	Hetero
N-224	39	M	EYS	AR	p.(Tyr2555fs)	Hetero	p.(Cys211Tyr)	Hetero
N-225	37	M	RPGR	XL	p.(G1u746fs)	Hemi	—	—
N-250	71	M	RP1	AD	p.(G1u661*)	Hetero	—	—
N-257	48	F	EYS	AR	p.(Ser1653fs)	Hetero	p.(AIa2498Pro)	Hetero
N-259	46	F	EYS	AR	p.(Gly2186G1u)	Homo	—	—
N-303	60	F	EYS	AR	p.(Asp498fs)	Hetero	p.(Cys211Tyr)	Hetero
N-37	57	F	EYS	AR	p.(Gly2799fs)	Hetero	p.(Arg1870Trp)	Hetero
N-39	31	F	RP1	AR	p.(Tyr1352Alafs*9)	Homo	—	—
NA0048	19	F	RP1	AR	p.(Tyr1352Alafs*9)	Homo	—	—
NA0070	38	F	RP1	AR	p.(Tyr1352Alafs*9)	Homo	—	—
N-34	47	M	EYS	AR	p.(Tyr2935*)	Hetero	p.(Gly2186G1u)	Hetero
N-75	64	F	USH2A	AR	p.(Cys934Trp)	Hetero	p.(G1n4371fs)	Hetero
NA0209	19	M	RP1	AR	p.(Tyr1352Alafs*9)	Hetero	p.(Cys1399Leufs*5)	Hetero
N-228	77	F	EYS	AR	p.(Gy2186G1u)	Hetero	p.(Tyr2935*)	Hetero
N-233	47	M	RPGR	XL	p.(G1u802fs)	Hemi	—	—
N-294	43	M	USH2A	AR	p.(Cys934Trp)	Homo	—	—
N-298	22	M	RPGR	XL	p.(Thr575fs)	Hemi	—	—
NA1048	23	M	RP1	AR	p.(Tyr1352Alafs*9)	Homo	—	—
N-1201	36	F	RP1	AR	p.(Tyr1352Alafs*9)	Hetero	p.(Cys1399Leufs*5)	Hetero
NAl209	29	F	RP1	AR	p.(Tyr1352Alafs*9)	Hetero	p.(Cys1399Leufs*5)	Hetero
OPH-783	35	M	RPGR	XL	p.?	Hemi	—	—
OPH-209	49	M	EYS	AR	p.(Ser1653fs)	Hetero	p.(Trp2640*)	Hetero
OPH-499	64	F	EYS	AR	p.(Ser1653fs)	Homo	—	—
OPH-39	69	M	RPGR	XL	p.(Gly718fs)	Hemi	—	—
OPH-753	48	F	USH2A	AR	p.(Cys934Trp)	Hetero	p.(G1u1985Lys)	Hetero
OPH-423	48	F	EYS	AR	p.(Ser1653fs)	Homo	—	—
OPH-985	42	M	EYS	AR	p.(Tyr2935*)	Hetero	p.(Gly2186G1u)	Hetero
OPH-616	43	M	EYS	AR	p.(Ser1653fs)	Homo	—	—
OPH-183	68	F	RP1L1	AR	p.(Arg658*)	Hetero	p.(G1u501*)	Hetero
OPH-302	54	M	USH2A	AR	p.(Arg1870Trp)	Hetero	p.(Gly2752Arg)	Hetero
OPH-570	54	M	USH2A	AR	p.(Pro5078Arg)	Hetero	p.(GIy2752Arg)	Hetero
OPH-908	40	M	EYS	AR	p.(Ser1653fs)	Homo	—	—
OPH-458	29	F	RP1	AD	p.(Arg872fs)	Hetero	—	—
OPH-698	53	F	EYS	AR	p.(Ser1653fs)	Homo	—	—
OPH-16	44	M	USH2A	AR	p.(Gly268Arg)	Homo	—	—
OPH-182	23	F	RP1	AR	p.(Cys1399fs)	Homo	p.(Ser2118Asn)	Homo
OPH-980	53	M	EYS	AR	p.(Ser1653fs)	Homo	—	—
OPH-613	47	M	EYS	AR	p.(Tyr2935*)	Homo	—	—
OPH-159	58	M	RP1	AD	p.(Arg872fs)	Hetero	—	—
OPH-767	51	M	EYS	AR	p.(Cys211Tyr)	Homo	p.(Leu2938Met)	Hetero
OPH-43	52	F	USH2A	AR	p.(Pro560Ala)	Homo	p.(Gly2752Arg)	Homo
OPH-552	46	F	EYS	AR	p.(Gln3101fs)	Homo	—	—
OPH-129	68	M	RP1L1	AR	p.(Ala1009fs)	Homo	—	—
OPH-824	53	F	EYS	AR	p.(Tyr2935*)	Homo	—	—
OPH-864	50	F	EYS	AR	p.(Tyr2935*)	Homo	—	—
OPH-459	36	F	USH2A	AR	p.(Tyr3701*)	Hetero	p.(lle3620Thr)	Hetero
OPH-465	66	M	RP1L1	AR	p.(Arg658*)	Homo	—	—
OPH-293	59	F	EYS	AR	p.(Ser1653fs)	Hetero	p.(Met12Thr)	Hetero
OPH-279	50	F	EYS	AR	p.(Ser1653fs)	Homo	—	—
OPH-327	52	M	RP1L1	AR	p.(Arg658*)	Homo	—	—
OPH-791	42	M	RP1	AD	p.(GIn689*)	Hetero	—	—
OPH-222	78	M	RP1	AD	p.(Arg872fs)	Hetero	—	—
OPH-617	50	F	EYS	AR	p.(Ser1653fs)	Homo	—	—
OPH-51	43	F	EYS	AR	p.(Tyr2935*)	Hetero	p.(Ser69fs)	Hetero

The noncanonical splice site variant (c.28+5G>A) in RPGR was detected in OPH-783. NA0209 & NA1209 and NA0048 & NA1048 were siblings. AD, autosomal dominant; AR, autosomal recessive; Hetero, heterozygous; Hemi, hemizygous; Homo, homozygous; XL, X-linked.

Statistical Analyses

We determined the significance of the differences in the age, BCVA, AL, EZ width, and MMCI for the five causative genes by Kruskal-Wallis tests. We also compared the MMCIs of each variant for each gene by Kruskal-Wallis tests. We used multiple linear regression analyses to determine the independence of the causative genes contributing to the MMCIs. To adjust for the clinical factors, we included the sex (male or female), age (years, continuous), EZ width (µm, continuous), and AL (mm, continuous) as covariates in the analysis. The MMCIs of the EYS eyes were defined as the reference value because the EYS is the most common causative gene of RP in the Japanese population,, and the distribution of MMCI values in EYS eyes was similar to that of all the RP cases in our previous study (Table 2). We quantified the effects of the other genes relative to that of the EYS value. The MMCI values were converted to absolute square numbers to treat them as normal distributions. We considered a P value of <0.05 to be statistically significant. The R software version 3.4.4 was used for all statistical analyses (available in the public domain at http://www.R-project.org/).

Table 2.

Characteristics of RP Cases Among Causative Genes

Parameter	EYS Eyes (n = 31)	USH2A Eyes (n = 11)	RPGR Eyes (n = 6)	RP1 Eyes (n = 13)	RP1L1 Eyes (n = 4)	P Value
Age, y	50.16 ± 10.00	48.28 ± 9.86	43.18 ± 15.95	38.18 ± 19.43	63.51 ± 7.44	0.01*
AL, mm	24.18 ± 1.00	24.34 ± 1.14	24.85 ± 1.82	23.46 ± 1.47	23.84 ± 1.00	0.19
BCVA, logMAR	0.38 ± 0.56	0.25 ± 0.31	0.71 ± 0.71	0.79 ± 0.81	0.56 ± 0.78	0.34
EZ width, µm	1465.7 ± 1419.7	1393.00 ± 965.61	1090.8 ± 1305.02	1244.00 ± 1797.80	2234.00 ± 2410.23	0.68
MMCI, × 10−5, µm−1	−15.54 ± 13.63	−10.35 ± 3.92	−35.81 ± 11.73	−14.71 ± 10.76	−13.84 ± 6.39	0.02*

Values are presented as mean ± SD (standard deviation).

*Statistical significance by Kruskal-Wallis test.

Results

Clinical and Genetic Characteristics of Patients With RP

The median BCVA of all the patients with RP was 0.30 logMAR units, and the median AL of all the patients with RP was 24.0 mm. In all of the eyes, the EZ was fully or partially disrupted. The median MMCI for all RP eyes was −12.6 × 10−5 µm−1.

The clinical and genetic characteristics of the patients with RP for each causative gene are presented in Table 1 and Table 2. There were 31 cases with the EYS variants (10 men and 21 women), 11 cases with USH2A variants (6 men and 5 women), 6 cases with RPGR variants (6 men), 13 cases with RP1 variants (6 men and 7 women), and 4 cases with RP1L1 variants (3 men and 1 woman). Of the 13 RP1-related patients with RP, there were eight ARRP cases, and seven of them had an Alu insertion in the RP1 gene, which has been reported as a frequent causative variant of patients with RP1-related RP in Japan. The common variant in EYS [p.(Ser1653fs)] was detected in 16 cases, and none of these patents were relatives., No significant differences were observed in the AL, BCVA, or the EZ width among the different causative genes (Table 2). However, we found significant differences in the age and MMCI among the different causative genes (P = 0.01 and P = 0.02, respectively).

Representative Fundus Photographs and SD-OCT Images of Patients With RP for Each Causative Gene

The fundus photographs and horizontally scanned SD-OCT images of six right eyes of representative patients with RP with different causative genes (cases 1–6) are shown in Figure 1. In the EYS eyes, one case had a flat macula (Fig. 1A) and another case had a steep macula (Fig. 1B). The USH2A (Fig. 1C), RP1 (Fig. 1E), and RP1L1 eyes (Fig. 1F) had relatively flat maculae. In contrast, the RPGR eye had a steep macula (Fig. 1D).

Distribution of MMCIs of RP Eyes Among Causative Genes

The distribution of the MMCIs of the RP eyes among the causative genes is shown in Figure 2. The MMCIs ranged from −52.62 × 10−5 µm−1 to −0.26 × 10−5 µm−1. Interestingly, the MMCI of all RPGR-related RP cases was less than −24 × 10−5 µm−1, indicating that the RPGR eyes had the steepest curvature. In contrast, all of the MMCIs of the USH2A eyes and RP1L1 eyes were greater than −20 × 10−5 µm−1, indicating that these eyes had a flatter macular curvature. In addition, we observed a wide range of MMCIs in the EYS and RP1 eyes (Table 2 and Fig. 2). The median MMCI was −31.2 × 10−5/µm for the RPGR eyes, −16.5 × 10−5/µm for the RP1L1 eyes, −13.0 × 10−5/µm for the RP1 eyes, −9.8 × 10−5/µm for the EYS eyes, and −9.0 × 10−5/µm for the USH2A eyes. The RPGR eyes had the steepest curvatures and were significantly steeper than that of the RP1L1 (P = 0.01), RP1 (P = 0.003), USH2A (P = 0.0002), and EYS (P = 0.003) eyes.

Figure 2.

Distribution of MMCIs of RP eyes among causative genes, EYS, USH2A, RPGR, RP1, and RP1L1 is shown. The box length represents the interquartile range, and the middle line depicts the median. MMCI, the mean macular curvature index; RP, retinitis pigmentosa.

Correlations Between MMCI and Clinical Parameters

The significance of the correlations between age, BCVA, AL, and EZ width and the MMCI was determined for all RP eyes. The MMCI was significantly correlated with the AL and the EZ width (P = 0.03 and P = 0.02, respectively), although the correlation coefficients were relatively low. On the other hand, the MMCI was not correlated with age or the BCVA (P = 0.87 and P = 0.63, respectively). The significant correlation observed between MMCI and AL was consistent with the results of our previous study.

Multiple Linear Regression Analysis for Independence of Causative Genes Contributing to MMCIs

Multiple linear regression analyses were performed to assess the independence of the causative genes contributing to the MMCIs after adjustment for age, sex, AL, and EZ width (Table 3). We used the effect of the EYS on MMCI as a reference and quantified the effects of the other genes relative to that of EYS. The MMCI values were converted to absolute square numbers to treat them as being normally distributed. Compared with the effect of EYS as the reference gene, RPGR significantly affected the MMCI values (P = 5.30 × 10−6). On the other hand, the effects of the other genes, USH2A, RP1, and RP1L1, were not significantly affected compared with the EYS gene (P = 0.26, P = 0.49, and P = 0.92, respectively).

Table 3.

Results of Multiple Linear Regression Analysis for Independence of Causative Genes Contributing to MMCIs

	RP Cases (n = 65)
Gene	Estimate	Standard Error	P Value
USH2A	3.92 × 10−5	3.47 × 10−5	0.26
RPGR	−2.42 × 10−4	4.80 × 10−5	5.30 × 10−6
RP1	−2.50 × 10−5	3.60 × 10−5	0.49
RP1L1	5.57 × 10−6	5.50 × 10−5	0.92

Results of multiple linear regression analysis for independence of causative genes contributing to the MMCIs. To adjust for the clinical factors, we included the sex (male or female), age (years, continuous), EZ width (µm, continuous), and AL (mm, continuous) as covariates in the analysis. The effect of the EYS gene on MMCI was defined as a reference. We quantified the effects of other genes. The MMCI values were converted to absolute square numbers to treat them as normal distributions.

Variant-Based Analyses

We compared the MMCIs of each homozygous variant of each AR gene, including p.(Ser1653fs), p.(Tyr2935*), p.(Gly2186Glu), p.(Cys211Tyr), and p.(Gln3101fs) in EYS; p.(Cys934Trp), p.(Gly268Arg), and p.(Pro560Ala) in USH2A; and p.(Arg658*) and p.(Ala1009fs) in RP1L1 (Table 1). For RPGR, we compared the MMCIs of each hemizygous variant [p.(Glu746fs), p.(Glu802fs), p.(Thr575fs), p.(Gly718fs)] and one splice-site variant (c.28+5G>A) (Table 1). We also compared the MMCIs of homozygous variants of RP1-related ARRP [p.(Tyr1352Alafs*9) and p.(Cys1399fs)] or heterozygous variants of RP1-related ADRP [p.(Glu661*), p.(Arg872fs), and p.(Gln689*)], because the inheritance pattern of this gene can be autosomal recessive or autosomal dominant (Table 1). Significant differences in the MMCIs of each variant were not observed for each gene (Kruskal-Wallis tests). We further examined the MMCI of EYS-related RP cases with the most frequent homozygous variants [p.(Ser1653fs) in EYS] and observed that MMCI varied even within the same variants (Supplementary Fig. S1).

Discussion

Our findings showed that the steepness of macular curvature in RP eyes differs among the different causative genes for RP. More specifically, the eyes with the RPGR gene had the steepest curvature compared to the other ciliopathy-related genes. Earlier studies have reported the findings in RPGR eyes complicated by high myopia– and on the relationship between RPGR-related RP and pathologic myopia., However, the macular curvature in these eyes was not mentioned. Our study showed that the macular curvature in eyes with RPGR-related RP was steeper than that in eyes with RP caused by other genes, and the effects of RPGR on MMCI were significant even after adjusting for AL. These findings indicate that the RPGR gene affected the macular curvature independently. Considering that RPGR-related X-linked RP is the target of current clinical trials of adeno-associated viral-mediated gene replacement therapy,,, the MMCI may be a useful parameter to evaluate the characteristics of this macula structure in the preoperative evaluations for the prevention of complications of subretinal injection, such as a macular hole. An observation of the MMCI before injection into the subretinal space could provide information on the appropriate angle, site, and injection dose in clinical trials.

We also found that the distribution of the MMCIs differed among the causative genes, although the MMCI of RP caused by each gene was lower than that of controls (–6.63 ± 5.63 × 10−5 µm−1). The EYS eyes had a wide range of MMCI, and in contrast, the USH2A and RP1L1 eyes had a flat macula and a narrow range of distribution of the MMCI. We could differentiate the effects on the specific phenotype of RP among causative genes using the MMCI.

Our results also indicated that there are morphologic differences among the ciliopathy-related genes. Although most of the genes examined, USH2A, RPGR, RP1, and RP1L1, encode proteins located in the photoreceptor cilia,– earlier studies have reported differences in the location of the encoded protein in the photoreceptors. Interestingly, the protein of RPGR is located only in the basal body of the cilium. Thus, we hypothesize that the abnormalities of the basal body of the cilia might lead to structural changes and steep macular curvature. However, the mechanisms that lead to this difference in MMCI among genes need additional experimental data.

In a previous study, Khan et al. reported their findings in cases of recessive early-onset retinal dystrophy with macular staphyloma caused by the C21orf2 gene, which encodes a protein that is localized to the photoreceptor primary cilium. However, reports on macular curvature in inherited retinal dystrophy, including RP, and detailed evaluations of the relationship between macular curvature and visual function are limited. Therefore, an investigation of the relationship between increased macular curvature and the vitreous body, vascular abnormalities, choroid, and the central visual function is necessary. The results of this study showed the steepest curvature in RPGR-related RP, which was previously described to have a severe course., This suggests that MMCI is somehow related to disease severity. If the steep curvature led to RP progression, the reduction of macular curvature might be a potential target for treatment. Increasing the number of cases, including cases with other causative genes, will also be important in future studies.

This study has several strengths. We collected a relatively large number of cases with genetically identified causative genes and quantified the macular curvature for each causative gene. In addition, we verified the effects of the genes on the MMCI after adjusting for other parameters.

There are also limitations in this study. This was a cross-sectional and retrospective analysis with potential selection biases. RP is a rare disease, and the number of genetically solved cases is limited. The RPGR-related RP is known to have a severe course, and it is assumed that the EZ line was shorter at a younger age. However, the number of subjects with RP caused by each gene in this study was small, and it is difficult to evaluate the clear trends by age and EZ line for each gene. In variant-based analysis, we could not detect significant differences in the phenotype-genotype correlation because most of the cases consisted of only one patient if we divided groups by each variant. There were several cases with the most frequent homozygous variant [p.(Ser1653fs) in EYS], and our results suggest that the MMCIs also varied among cases caused by the same causative variant. However, further comparisons with other variants were difficult; therefore, the small size of our sample was a limitation for variant-based statistical evaluations. In addition, most of the RPGR-related RP cases were caused by frameshift variants except for one case (OPH-783). Previous reports have suggested that frameshift variants of the RPGR cause more severe clinical alterations; therefore, this may have limited our results (Table 1). Further multicenter studies are needed to collect cases to reduce the selection bias.

In conclusion, our results suggest that the steepness of the macular curvature is significantly associated with the causative genes for RP. The RPGR gene had a stronger effect on the steep macular curvature than the other ciliopathy-related genes.