Long-term vitamin A supplementation in a preclinical mouse model for Rho^D190N-associated retinitis pigmentosa

Abstract

Retinitis pigmentosa (RP) is caused by one of many possible gene mutations. The National Institutes of Health recommends high daily doses of vitamin A palmitate for RP patients. There is a critical knowledge gap surrounding the therapeutic applicability of vitamin A to patients with the different subtypes of the disease. Here, we present a case report of a patient with RP caused by a p.D190N mutation in Rhodopsin (RHO) associated with abnormally high quantitative autofluorescence values after long-term vitamin A supplementation. We investigated the effects of vitamin A treatment strategy on RP caused by the p.D190N mutation in RHO by exposing Rhodopsin p.D190N (Rho^D190N/+) and wild-type (WT) mice to experimental vitamin A-supplemented and standard control diets. The patient’s case suggests that the vitamin A treatment strategy should be further studied to determine its effect on RP caused by p.D190N mutation in RHO and other mutations. Our mouse experiments revealed that Rho^D190N/+ mice on the vitamin A diet exhibited higher levels of autofluorescence and lipofuscin metabolites compared to WT mice on the same diet and isogenic controls on the standard control diet. Vitamin A supplementation diminished photoreceptor function in Rho^D190N/+ mice while preserving cone response in WT mice. Our findings highlight the importance of more investigations into the efficacy of clinical treatments like vitamin A for patients with certain genetic subtypes of disease and of genotyping in the precision care of inherited retinal degenerations.

Introduction

Affecting approximately one in 4000 individuals (1), retinitis pigmentosa (RP) is considered to be one of the leading causes of irreversible vision loss in the USA. Typically, patients with RP suffer from vision loss that progresses from the midperiphery of the retina into the central retina and in many cases into the macula and fovea (2,3). The most promising treatment for RP is gene therapy (4), which involves genetic repair of the pathogenic mutation. Given that gene therapy is gene- and mutation-specific, there is currently no cure that is broadly applicable to all RP patients. Thus, there has been great interest in the development of early therapeutic interventions that may prevent its progression, maintain healthy rod morphology or restore the loss of visual function.

Nutritional supplementation with vitamin A, lutein or docosahexaenoic acid, may slow disease progression (5–8). Vitamin A palmitate is the most widely used nutritional supplement (9). A study which documented that vitamin A supplementation delays disease progression, as measured by preservation of 30 Hz cone electroretinographic function, was later criticized for the robustness of the outcome measurement (9–14). Therefore, it is crucial to conduct further investigations to determine if vitamin A supplementation will benefit patients with different forms of RP.

To date, over 3000 mutations in more than 80 genes have been identified to cause RP (15,16). Almost one-third of all RP cases are caused by one of over 150 autosomal dominant mutations in the Rhodopsin (RHO) gene (2,17), making RHO-associated RP an attractive therapeutic target. The precise mechanism of photoreceptor degeneration caused by p.D190N mutations in RHO has been a topic of interest given the debilitating nature of RP (18–20). An investigation revealed that this mutation, which belongs to Class II, encodes a thermally unstable form of RHO that spontaneously isomerizes 11-cis-retinaldehyde (11-cis-RAL) in the absence of light (18–23). Studies examining the effect of vitamin A on Rho-associated RP found that a high vitamin A diet preserved functional activity in Rhodopsin p.T17M (Rho^T17M) mice but not in Rhodopsin p. P347S (Rho^P347S) mice, as demonstrated on electroretinography (ERG) tracings (24,25). Rho^T17M mice have a mutation which results in folding instability; therefore, it is hypothesized that by increasing the levels of 11-cis-RAL, vitamin A supplementation increases stability through saturating levels of rhodopsin-11-cis-RAL binding. In contrast, the Rho^P347S causes no currently recognized changes to protein stability, so increased levels of 11-cis-RAL are not expected to increase functional activity (24). The response of Rho^D190N to vitamin A supplementation has not yet been studied.

According to previous studies, both 11-cis-RAL and all-trans-retinaldehyde (atRAL) are precursors of bisretinoid lipofuscin, a byproduct that is steadily eliminated from the photoreceptor outer segment (OS) during membrane renewal (26). It has been suggested that bisretinoid fluorophores such as the pyridinium bisretinoid A2E cause cell toxicity that eventually leads to photoreceptor death (27,28). We hypothesize that for RP patients with the p.D190N mutation in RHO, the presence of the thermally unstable protein variant causes excessive isomerization of 11-cis-RAL, leading to bisretinoid lipofuscin toxicity and photoreceptor death. Vitamin A supplementation may accelerate lipofuscin accumulation and accelerate disease progression. The goal of this study is to investigate the effects of long-term vitamin A supplementation to Rho^D190N by studying the rate of ERG amplitude decline in patients as well as the monitoring of photoreceptor cell structure in mice. To carry out this project, we utilized a vitamin A-enriched ad libitum diet to address the effect of vitamin A in the Rhodopsin p.D190N (Rho^D190N/+)-associated RP mouse model. By addressing this knowledge gap, we hope to highlight the importance of personalized medicine and guide the future of RP clinical treatment.

Results

Quantitative fundus autofluorescence intensity in the proband of autosomal dominant RP with p.D190N mutation in RHO

A 52-year-old man with a p.D190N mutation in RHO presented with a history of worsening night blindness, decreased visual acuity and a reduced visual field beginning in his mid-40s (Fig. 1A and B). At presentation, short-wavelength fundus autofluorescence (SW-AF) images revealed a hyperautofluorescent ring centered at the fovea; the ring decreased in diameter over time (Fig. 1A and B). Before his first visit in 2010, the proband had been taking daily vitamin A palmitate (15 000 IU) without interruption for over 20 years (18). Quantitative AF intensity was determined from mean gray levels recorded within eight circularly arranged segments positioned at an eccentricity of approximately 7°–9° (marked with x’s in Fig. 1D, bottom row) (Fig. 1C and D). qAF levels were lower than the 95% confidence intervals for healthy eyes (Fig. 1C). Note the internal reference visible in the qAF images in Figure 1 (rectangles at top of qAF images in D, red arrow). The brightness of the reference is indicative of the use of a high sensitivity setting and thus low qAF. Nevertheless, qAF increased 1 and 2 years after the first visit. The proband discontinued vitamin A supplementation in 2012 (Fig. 1E, green arrow). For the next 2 years (4 years after the first visit), qAF continued to increase but then subsequently declined 6 and 9 years after the first visit (Fig. 1D and E). The latter reduction in qAF is consistent with progressive photoreceptor cell degeneration (29,30).

Figure 1

Increased SW-AF in patient with autosomal dominant retinitis pigmentosa (adRP) associated with p.D190N mutation in RHO taking vitamin A supplementation. (A, B) Color (white light) fundus photographs of the proband acquired at initial presentation (A) and at a follow-up visit after 9 years (B). The yellow circles delineate progressive degeneration within the macula. (C) The qAF intensities of the proband plotted at initial presentation (first visit) and at each follow-up visit (after 1, 2, 4, 6 and 9 years) are plotted together with mean qAF (black line) (±95% confidence intervals, dashed lines) acquired from 87 subjects with healthy eyes and self-identifying as white. (D) Color-coded qAF images acquired at initial presentation and at follow-up after 4 and 9 years. The lower images in (C) illustrate the measurement grid (outlined in red) utilized for image analysis. Mean gray levels were recorded from the internal reference (rectangle at the top) and from 8 circularly arranged segments situated at 7°–9° eccentricity (positions marked with X). (E) qAF values acquired at each visit are plotted. The green arrow indicates the year that the patient discontinued the vitamin A supplementation.

The patient’s family history includes an affected 77-year-old mother, and two affected sons (7 and 11 years old), who were confirmed by targeted genetic sequencing for the RHO^D190N/+ variant (Supplementary Material, Fig. S1A). Since the diagnosis, the affected sons were also taking daily vitamin A palmitate (10 000 IU) without interruption. The two affected sons exhibited early, large hyperautofluorescent rings on SW-AF imaging, demarcating the watershed zone. Spectral domain optical coherence tomography (SD-OCT) revealed bilateral foveal sparing, and disruption of the ellipsoid zone (EZ) line extending from the nasal and temporal ends of the watershed zone to the peripheries (Supplementary Material, Fig. S1B and C). Fundus images of the patient’s mother revealed a complete absence of the hyperautofluorescent ring on SW-AF and the EZ line on SD-OCT, indicating end stage of disease (Supplementary Material, Fig. S1D).

Photoreceptor cell loss is accentuated in Rho^D190N/+ mice fed a vitamin A-supplemented diet

To investigate the effects of vitamin A in Rho^D190N/+ mutants, the Rho^D190N/+ and wild-type (WT) mice were fed a vitamin A-supplemented diet, and their isogenic controls were fed a standard control diet. We examined photoreceptor viability by morphometric analysis. The nuclei in the outer nuclear layer (ONL) and inner and outer segment layer (IS/OS layer) were less in Rho^D190N/+ mice fed the vitamin A-supplemented diet than in those fed the standard control diet (Fig. 2A). Quantitative analysis revealed that vitamin A supplementation significantly decreased the thickness of the ONL only in Rho^D190N/+ mice (Fig. 2B). In contrast, diet did not affect ONL thickness in WT mice. Lastly, RHO expression levels were evaluated by immunoblotting retinal extracts. The findings here revealed reduced RHO expression in Rho^D190N/+ mice fed the vitamin A-supplemented diet as compared to those receiving the control diet (Fig. 2C).

Figure 2

Vitamin A supplementation exacerbated photoreceptor degeneration and loss in Rho^D190N/+ mice while preserves photoreceptor function in WT mice. (A) Histological analysis of the retina was used to evaluate photoreceptor survival at 13 months. ONL (yellow band) is visibly thinner in Rho^D190N/+ versus WT retina. Among Rho^D190N/+ mice, those given a vitamin A-supplemented diet exhibited diminished ONL and OSs than those given the standard control diet. H&E-stained sections were photographed ~500 microns from the optic nerve head (ONH) dorsal quadrant. (B) ONL thickness measurements, which were acquired at ~500 microns from ONH, were plotted for comparison of the two diets in each strain. Individual values (circles) and mean ± SEM are presented. ^*P < 0.05, n = 7. (C) Immunoblots of rhodopsin protein in retinal extracts from 16-month-old Rho^D190N/+ mice showed that vitamin A supplementation decreased RHO levels. β-actin, loading control. (D) Six-month-old Rho^D190N/+ mice were subject to a light treatment that bleached ~90% of the rhodopsin. Using ERG testing, rod-mediated b-wave amplitude as a fraction of the dark-adapted rod b-wave amplitude at the indicated times showed no significant differences between the two diet groups. Vitamin A-supplemented diet in Rho^D190N/+, n = 24; standard control diet in Rho^D190N/+, n = 16. Mean ± SD was presented. ^*P < 0.05. (E, F) ERG b-wave amplitudes recorded from Rho^D190N/+ and WT mice fed a vitamin A-supplemented diet or standard control diet at (E) 6 months and (F) 12 months. Individual values (mean value of both eyes in each sample) are presented as a single dot or triangle; mean ± SE; n = 18. ^*P < 0.05, ^**P < 0.01 and ^****P < 0.0001.

The vitamin A-supplemented diet diminished retinal function in Rho^D190N/+ mice

Having evaluated the morphological and biochemical changes, we next investigated whether vitamin A supplementation impacts photoreceptor activity using ERG testing. Rod-mediated b-wave amplitude as a fraction of the dark-adapted rod b-wave amplitude at the indicated times showed no significant differences between the two diet groups after bleaching ~90% of the rhodopsin in 6-month-old Rho^D190N/+ mice. The normalized b-wave amplitude was lower at the later time points in mice fed a vitamin A-supplemented diet (Fig. 2D). Neither b-wave amplitude nor latency achieved statistical significance.

To monitor retinal function and distinguish between rod and cone activity, we obtained scotopic, mixed maximum and photopic ERG recordings at 6 and 12 months of age. At 6 months, WT mice exhibited significantly higher scotopic (P < 0.0001), maximum (P < 0.0001) and photopic (Vit A: P < 0.0001; Ctrl: P < 0.01) b-wave amplitudes compared to Rho^D190N/+ mice on the same diet (Fig. 2E). Of note, vitamin A supplementation enhanced maximum and photopic b-wave amplitudes in WT mice (P < 0.05) but not in Rho^D190N/+ mice. At 12 months, vitamin A supplementation significantly decreased the maximum (P < 0.01) and photopic (P < 0.05) b-wave amplitudes in Rho^D190N/+ mice (Fig. 2F). Unexpectedly, the same vitamin A-supplemented diet had the opposite effect in the enhanced maximum (P < 0.05) and photopic (P < 0.05) b-wave amplitudes of WT mice. As expected, WT mice maintained significantly higher scotopic (P < 0.0001), maximum (P < 0.0001) and photopic (P < 0.0001) b-wave amplitudes than Rho^D190N/+ mice on the same diet. Overall, these results suggest that a vitamin A-supplemented diet may help preserve retinal function in aged WT mice, while accelerating retinal deterioration in Rho^D190N/+ mice.

Elevated retinoid levels in plasma and liver in mice on vitamin A-supplemented diet

To further investigate the change in retinoid and bisretinoid levels after administering vitamin A-supplemented diet to Rho^D190N/+ mutants and WT mice, samples (including plasma, liver and eyes) from mice at 13 months were collected and measured by ultra-performance liquid chromatography (UPLC) quantification.

Hepatic all-trans-retinol (atROL, P < 0.05) and all-trans-retinyl (atRE) palmitate (P < 0.01) levels were elevated in both Rho^D190N/+ and WT mice on the vitamin A-supplemented diet compared to levels in respective isogenic controls (Fig. 3A and B). However, there were no significant differences between the two mouse strains on the same diet. Consistent with these findings, mice fed the vitamin A-supplemented diet exhibited higher plasma atRE levels compared to those fed the standard control diets (Fig. 3C, P < 0.05). Interestingly, atRE levels were increased to a greater degree in WT mice than in Rho^D190N/+ mice (P < 0.05) (both on the vitamin A-supplemented diet), while both mouse strains exhibited similar levels of atRE on the standard control diet.

Figure 3

Vitamin A supplementation increased retinoid levels in the eyes of WT and Rho^D190N/+ mice. Bisretinoids are elevated in the eyes of Rho^D190N/+ mice receiving the vitamin A supplements but not in eyes of WT mice. (A, B) All-trans-retinol and atRE palmitate levels in the liver were measured by UPLC. n = 4, 4, 6, 5. (C) All-trans-retinyl palmitate levels in the plasma were measured by UPLC. n = 6, 4, 8, 3. (D, E) Representative UPLC chromatograms demonstrate the separation of retinoids extracted from eyes of WT and Rho^D190N/+ mice maintained on a standard control diet or vitamin A-supplemented diet. Dark-adapted eyes were used. The numbers indicate the identity of peaks based on comparison with authentic standards: 1: atROL; 2: atRE acetate (internal standard); 3: anti 11-cis-retinal-(O-ethyl) oxime; 4: anti all-trans-retinal-(O-ethyl) oxime; 5: syn 11-cis-retinal-(O-ethyl) oxime; 6: syn all-trans-retinal-(O-ethyl) oxime; 7: atRE palmitate. (F, G) Quantitation of ocular retinoid levels in WT and Rho^D190N/+ mice fed a vitamin A-supplemented or standard control diet. n = 6–10 samples/group; 1 eye/sample. (H, I) Quantification of bisretinoid levels in WT and Rho^D190N/+ mice receiving a vitamin A-supplemented diet versus standard control diet at 13 months. n = 3–4 samples/group; 3–5 eyes/sample. Individual values are presented as circles or triangles; mean values as horizontal lines. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001 and ^****P < 0.0001; t-test. Note the differences in the range on the y-axes.

Increased retinoid levels in the eyes of Rho^D190N/+ mice supplemented with vitamin A diet

Having established the systemic effects of vitamin A supplementation, we investigated whether it impacts retinoid levels in mouse eyes. Eyes from 13-month-old WT and Rho^D190N/+ mice were collected and subjected to UPLC. Retinoid quantitation was performed by adjusting chromatographic peak areas of atRAL (peaks 4 and 6), atROL (peak 1), atRE (peak 7) and 11-cis-RAL (peaks 3 and 5) to the internal standard (IS), atRE acetate (peak 2) (Fig. 3D and E). WT mice exhibited no difference in the four quantified retinoids between the vitamin A-supplemented diet and standard control diet (Fig. 3F). In contrast, a markedly, disproportionate increase in the levels of atROL (P < 0.001), atRE (P < 0.0001) and 11-cis-RAL (P < 0.001) was observed in Rho^D190N/+ mice fed a vitamin A-supplemented diet (Fig. 3G). Interestingly, WT mice showed significantly higher levels of atRAL, atROL, atRE and 11-cis-RAL than Rho^D190N/+ mice when both mouse strains were fed the standard control diet (Supplementary Material, Fig. S2A; P < 0.0001).

The vitamin A-supplemented diet increased A2E bisretinoids in RhoD190N/+ mice, but not in WT mice. Next, we examined whether higher retinoid levels were associated with increased bisretinoids in the eyes. The levels of four previously isolated and characterized bisretinoids were measured: A2-glycero-phosphoethanolamine (A2GPE), A2-dihydropyridine-phosphatidylethanolamine (A2-DHP-PE), atRAL dimer phosphatidylethanolamine (atRALdi-PE) and all-trans- and cis-isomers of A2E (31–34). No significant difference was observed in bisretinoid levels between WT mice fed the standard control diet and those fed the vitamin A-supplemented diet (Fig. 3H). Notably, vitamin A supplementation was associated with increased levels of total A2E (P < 0.05) and A2GPE (P < 0.001) in RhoD190N/+ mice (Fig. 3I). Consistent with the higher levels of ocular retinoids, WT mice fed the standard control diet also exhibited higher levels of A2GPE (P < 0.05), but not total A2E, A2-DHP-PE or atRALdi-PE, compared to RhoD190N/+ mice on the same diet (Supplementary Material, Fig. S2B).

The vitamin A-supplemented diet increased lipofuscin pigment accumulation in Rho^D190N/+ mice

In order to monitor changes in bisretinoid lipofuscin jointly with vitamin A supplementation, we also performed fundus AF imaging in Rho^D190N/+ mice at 7, 12 (not shown here) and 19 months (Fig. 4A). Irrespective of diet, the lipofuscin granules were brighter at 19 months than at 7 months. Increased brightness was particularly noticeable in mice fed the vitamin A-supplemented diet, suggesting that vitamin A supplementation exacerbates bisretinoid lipofuscin formation. Quantification of SW-AF was performed using 7–8 fundus images from different Rho^D190N/+ mice in each group. Although no difference was detected at 7 and 12 months, the AF intensities at 19 months were significantly higher in mice fed the vitamin A-supplemented diet than in those fed the standard control diet (Fig. 4B; P < 0.01).

Figure 4

Vitamin A supplementation increased the intensity of SW-AF, the number of hyperautofluorescent granules and pigment granules in RPE monolayer in Rho^D190N/+. (A) Representative SW-AF images (488 nm) acquired from Rho^D190N/+ mice at 7 and 19 months. (B) Fundus autofluorescence intensity was calculated in Rho^D190N/+ mice fed a vitamin A-supplemented or standard control diet at age 7 and 19 months. For each mouse, a single value (circles) was determined as the average of the two eyes. Mean ± SD per group is plotted. ^**P < 0.01. (C, D) Fluorescence microscopy analysis of granules in retinal pigment epithelial (RPE) cells in (C) Rho^D190N/+ and (D) WT mouse at 12 months. (E) Analysis of the high AF area was normalized by the length using Fiji software. Individual values of each sample (mean value of both eyes in each sample, 3 slices from each eye) were shown in the image as single dots or triangles; mean ± SE was presented; n = 3. ^*P < 0.05; ^**P < 0.01. (F–H) A membranous vacuole (V) appears in the cytoplasm, and the basal infoldings (BI) are dilated in the RPE of Rho^D190N/+ mice fed with a standard control diet. (Scale bar: 2 μm.) BrM, bruch’s membrane. (I–K) More severe membranous vacuoles and lipofuscin pigment granules (LG) appear in the cytoplasm of RPE cells in Rho^D190N/+ mice fed with the vitamin A-supplemented diet. The basal infoldings are fewer and dilated. (Scale bar: 2 μm.) (i, j and k) Partial enlargement of images I, J and K respectively. (Scale bar: 500 nm.) The irregularly shaped, homogeneous electron-dense bodies in the RPE cytoplasm indicate lipofuscin pigment granules. PR, photoreceptors; N, nuclear.

To further verify that the Rho^D190N/+ variant leads to disproportionately higher levels of lipofuscin following vitamin A supplementation, we performed fluorescent microscopy (Fig. 4C–E). Notably, qualitative analysis of retinal sections revealed that Rho^D190N/+ mice supplemented with vitamin A exhibited larger areas of lipofuscin pigment granules in the retinal pigment epithelium (RPE) than their isogenic controls (Fig. 4C). Moreover, quantitative fluorescence analysis demonstrated that vitamin A supplementation enhanced the brightness of lipofuscin pigment granules in Rho^D190N/+ mice (Fig. 4E, P = 0.0271). In WT mice, vitamin A supplementation appeared to increase the number of lipofuscin granules (Fig. 4D), but there was no significant difference in brightness between the two diet groups (Fig. 4E, P = 0.3859). Electron microscopy was used to further validate the changes within RPE cells, as lipofuscin granules can be easily observed by their homogenous electron density and sharp demarcations (Fig. 4F–K) (35,36). Greater numbers of lipofuscin granules were found in Rho^D190N/+ mice that were fed the vitamin A-supplemented diet (Fig. 4I–K) than those fed the standard control diet (Fig. 4F–H).

Discussion

RP is a group of various inherited retinal dystrophies. Forms of RP caused by loss of function mutations may be treated by gene supplementation with a single vector. However, gene supplementation therapy will not alleviate toxic gain of function RP phenotypes (37). Due to the large number of affected mutations in different genes, current gene therapy is a promising approach, but its therapeutic reach is limited (38). Nutritional supplementation with vitamin A, lutein or docosahexaenoic acid, is reported to slow disease progression (9,25,39). However, there are limitations to vitamin A treatment. High-dosage vitamin A treatment has yet to be widely tested in the various forms of RP (14). This report provides evidence supporting the need to carefully examine the effects of vitamin A supplementation for different genetic forms of RP caused by RHO-associated mutations.

Previous clinical studies suggest that vitamin A supplementation may help preserve retinal function in RP patients (9,11,40). Through a randomized, controlled, double-masked trial, Berson et al. assessed the therapeutic effects of vitamin A supplementation in a cohort of 601 RP patients using cone 30 Hz ERG amplitudes as the primary outcome measure (9). The landmark results of this study demonstrated that 15 000 international units (IU) per day of vitamin A delayed RP progression; these findings ultimately laid the foundation for the current clinical treatment of RP patients. It has since been suggested that daily vitamin A supplementation may have protective effects on photoreceptor function by preventing transient decreases in serum retinol concentration (41). However, our study found little to no improvement in visual function in the proband and the proband’s family members, suggesting vitamin A supplementation may provide only modest benefits for the patients with RP caused by a p.D190N mutation in RHO. Alternatively, vitamin A supplementation may cause faster visual chromophore regeneration rather than slowing of photoreceptor degeneration in this specific subtype of RP. This is one possible explanation of why the improvement shown in ERG results of patients is not also coupled with improved neurofunction (10).

Further complicating our ability to understand this process is the genetic and heterogeneity of RP, with over 3000 mutations and 80 genes implicated in this condition (1,15,42,43). Different genes, or different mutations in the same gene, may play different roles in RP progression. High levels of vitamin A in the diets of Rho^T17M mice were found to reduce the rate of decline of a-wave and b-wave amplitudes but failed to produce similar statistically significant effects in Rho^P347S mice (24). Therefore, vitamin A supplementation may influence heterogeneous forms of RP in different ways.

We highlight the clinical impact of this interaction in our case report of a 52-year-old RP patient who stated that he had been taking vitamin A supplements for over 20 years without halting of his disease progression. Genetic testing identified an autosomal dominant p.D190N mutation in RHO, and qAF was used to monitor disease progression. The qAF value of the patient increased and peaked ~1.5 years after the patient halted vitamin A supplementation. However, 4 years after the cessation of supplemental vitamin A, the patient’s qAF value began to decrease and continued to decline. At the last visit (7 years after stopping vitamin A treatment), the patient exhibited the lowest qAF intensity since his initial visit, 9 years prior. Of note, patients harboring the RHO^D190N mutation exhibit a hyperautofluorescent ring on SW-AF imaging (18,44). We have shown by quantitative fundus autofluorescence (qAF) that at some stages of RP, the SW-AF intensity is increased in the AF ring relative to intensities at the same location in healthy human eyes (30). This information in conjunction with our patient’s qAF results suggest that vitamin A supplementation may induce higher levels of bisretinoid lipofuscin accumulation in patients with the p.D190N mutation in RHO, but further studies are needed to corroborate this finding.

The proband is the only subject in our database who both had the p.D190N mutation in RHO and had taken vitamin A for a prolonged period. Due to the low prevalence of patients with the specific p.D190N mutation in RHO the sample size of this study is low. Furthermore, we were unable to compare the progression data collected from the proband to that of a similar aged patient with the p.D190N variant who was not taking vitamin A, because no such patient existed in our directory. The p.D190N variant is one of 150 RHO mutations that is related to autosomal dominant RP. Previous studies have shown that the p.D190N mutation decreases the Schiff base stability (45) and thermal stability (4,44) of dark-state RHO. This decrease in stability has been suggested to elevate dark spontaneous activation and decrease RHO content in OSs, thus leading to the characteristic night blindness observed in early stages of RP (22,23).

In healthy retinas, vitamin A serves as both an activator and substrate for RHO and it is essential to normal rod and dark adaptation functions (46). This mutual interaction contributes to vitamin A’s essential role in RHO function, linking RHO with benefits seen following vitamin A supplementation. However, it remains unclear whether and how an unstable variant of RHO, such as the p.D190N mutation, interacts with vitamin A during retinal degeneration.

Here, we addressed this knowledge gap by exposing Rho^D190N/+ and WT mice to experimental vitamin A-supplemented and standard control diets. Vitamin A supplementation elevated hepatic atROL and atRE levels as well as plasma atRE levels in both mouse models. These results support the accepted model, which suggests that ingested vitamin A is processed and stored in the liver, from whether it can be released into the circulation (46).

WT vitamin A-treated mice did not exhibit statistically significant elevated retinoids compared to WT mice on the standard control diet (Fig. 3F). However, Rho^D190N/+ mice fed a vitamin A-supplemented diet had statistically significantly increased levels of atROL, atRE and 11-cis RAL compared to Rho^D190N/+ mice fed a standard control diet (Fig. 3G). Interestingly, only Rho^D190N/+ mice demonstrated a corresponding increase in bisretinoids—specifically A2E and A2GPE when given a vitamin A-supplemented diet (Fig. 3I).

Bisretinoids form nonenzymatically in the photoreceptor OS. Reaction of one retinaldehyde with PE forms N-retinylidene-PE (NRPE) while reaction of a second retinaldehyde with NRPE leads to bisretinoid formation. Bisretinoid lipofuscin formation may be increased by excessive release of 11-cis-RAL by thermally unstable RHO^D190N having decreased affinity for the 11-cis-chromophore. The subsequent shedding and phagocytosis of OS membrane lead to the accumulation of the bisretinoids within the RPE (47). RPE lipofuscin is known to be composed of various members of a family of bisretinoids that includes A2E (48).

Bisretinoid accumulation has been associated with the pathogenesis of retinal degenerations. For instance, accelerated accumulation of bisretinoids in albino Abca4^−/− mice is associated with thinning of ONL, a finding indicative of photoreceptor cell death (49). In vitro cell studies have implicated A2E in RPE cell death (50) and retinal inflammation (51,52). Bisretinoids likely mediate cellular damage by initiating photooxidative processes, by activation of complement and by disruption of autophagy and phagosome trafficking (52–57).

Furthermore, previous studies demonstrate the importance of RPE in photoreceptor viability (58,59). In summary, our findings support the hypothesis that administering excess dietary vitamin A in the presence of the RHO^D190N variants exacerbates the accumulation of toxic bisretinoid lipofuscin fluorophores in the RPE cells, inducing photoreceptor death. Further study is warranted to elucidate the exact mechanism of this process.

Since vitamin A supplementation adversely affected cone-rod survival in Rho^D190N/+ mice, genotype-stratified patients should be added as an inclusion criterion for further vitamin A studies. Moreover, clinical studies have shown that excessive vitamin A supplementation may be toxic in young patients with RP and recessive Stargardt disease (10). Daily oral administration of deuterated vitamin A has been used to slow bisretinoid formation in recessive Stargardt disease (60) and should be explored further to determine if it is effective for other forms of photoreceptor cell degeneration. Altogether, our findings highlight the importance of genotyping and precision medicine in the context of retinal degenerative disorders.

Materials and Methods

Proband

A patient diagnosed with RP caused by a p.D190N mutation in RHO at the Applied Genetics Clinic at Edward S. Harkness Eye Institute, Columbia University Irving Medical Center was followed for 9 years. This study was approved by the Institutional Review Board (Protocol number AAAR8743). This analysis was of minimal risk to the patient and was performed retrospectively so no written consent was required. No personally identifiable images or of the proband are presented. We performed fundus pattern analysis using funduscopic examination, digital color fundus and ultrawide-field color fundus photography (Optos, Dunfermline, UK) following methods in Cho et al. (61). We used a confocal scanning laser ophthalmoscope to obtain SD-OCT images and qAF images (cSLO; Spectralis HRP + OCT, Heidelberg Engineering, Heidelberg, Germany) (61). The qAF was analyzed with a dedicated image analysis program written in IGOR (WaveMetrics, Inc., Lake Oswego, OR) software as previously described (30,62).

Animals

C57BL/6J mice were purchased from Jackson Laboratory (Bar Harbor, ME) to generate Rho^D190N/+ and WT control mice for this study. Human mutation knock-in mice, Rho^D190N/D190N, were established in our lab as previously described (19). All C57BL/6J and Rho^D190N/D190N mice were Rpe65^Met-450. All mice were maintained in the Columbia University Pathogen-free Animal Care Services Facility under a 12/12-h light/dark cycle, with food and water available ad libitum following methods previously described (63). All efforts were made to minimize animal numbers and suffering. The experiment on mice were performed in accordance with the Statement for the Use of Animals in Ophthalmic and Vision Research of the Association for Research in Vision and Ophthalmology. The International Animal Care and Use Committee (IACUC) at Columbia University approved all experiments prior to experimentation.

Vitamin A-supplemented diet

WT and Rho^D190N/+ mice were fed an adjusted vitamin A-supplemented diet (120 000 IU vitamin A/kg diet; NIH-31 modified; TD.06501, ENVIGO, Indianapolis, IN) or standard control diet (24 500 IU vitamin A/kg diet; NIH-31 modified; TD.96338, ENVIGO, Indianapolis, IN) from postnatal day 21 until sacrifice. According to the IACUC protocol, pups need to be separated from the mom at postnatal day 21.

Anesthesia of the mice

A mixture of ketamine hydrochloride [10 mg/100 g body weight (BW); Ketaset, Fort Dodge Animal Health, Fort Dodge, IA] and xylazine (1 mg/100 g BW; Anased, Lloyd Laboratory, Shenandoah, IA) (19) was injected intraperitoneally into mice with a 1 cc syringe. A heating pad was used to maintain the temperature of the mice during experiments.

Quantitation of retinoids

Mice were dark-adapted for over 12 h, and all tissue manipulations were performed under a dim red light. After anesthesia, mice were placed on ice, and their chest was opened following a previously described protocol (64). More than 700 μl of blood was drawn from the right ventricle using a 1 cc syringe pre-rinsed with 50 μl 100X EDTA. Plasma was collected after blood centrifugation for 10 min at 10 000g at 4°C. The upper liver was collected and measured. Eyes were enucleated and stored in 1.5 ml Eppendorf covered in foil paper. All samples were immediately frozen in liquid nitrogen and stored at −80°C for further processing. Frozen mouse eyecups (1 eye/sample) were homogenized and derivatized on ice using 1 ml of O-ethylhydroxylamine (100 mm) in Dulbecco’s phosphate-buffered saline (DPBS) (pH 6.5, without CaCl₂ and MgCl₂) using glass homogenizers on ice under dim red light. (65) After vortexing, samples were incubated at 4°C for 15 min. All-trans-retinyl acetate was used as an IS, and methanol (1 ml) was added to the samples. After vortexing, hexane extraction was performed (3 ml, 3 times). After centrifugation (1500g at 4°C for 5 min), the hexane phase was dried under argon gas and then resuspended in 20 μl of acetonitrile for UPLC analysis (65). For extraction of retinoids, plasma was recovered from −80°C. One milliliter of methanol and 120 μl of 1 M O-ethylhydroxylamine in DPBS (pH 6.5, without CaCl₂ and MgCl₂) were added to 200 μl of plasma (final concentration of O-ethylhydroxylamine, 100 mm). Extractions were performed following the procedure previously described for eyecups. The samples were then resuspended in 6 μl of acetonitrile for UPLC analysis. Frozen liver was placed in a homogenization tube, and 2 ml of DPBS (without CaCl₂ and MgCl₂) and glass beads (0.1 and 0.5 mm) were added. Samples were homogenized twice for 15 min in a Disruptor Genie (Scientific Industries Inc., Springfield, MA). The derivatization and extraction processes were performed as described above. For UPLC analysis, the samples were resuspended in 5 ml of acetonitrile. For quantitation of retinoids, a Waters Acquity UPLC-photodiode array system (Waters, Milford, MA) was used with a Charged Surface Hybrid C18 column (1.7 μm, 2.1 × 100 mm) as previously described (66). Samples (5 μl) were injected into the UPLC. Retinal (O-ethyl) oximes (11-cis-RAL and atRAL) were monitored at 360 nm, and atROL and atRE palmitate were monitored at 320 nm. Ultraviolet absorbance peaks were identified by comparison to external standards of synthesized 11-cis-RAL (O-ethyl) oxime, atRAL (O-ethyl) oxime, atROL and atRE palmitate.

Quantitation of bisretinoids

Mouse eyecups (1–5 eyes/sample) were homogenized and extracted in chloroform/methanol (1:1). The extract was concentrated by evaporation of solvent under argon gas and redissolved in chloroform/methanol (1:1) (for HPLC) and ethanol (for UPLC). For measurement of bisretinoids, 30 μl samples were injected into the HPLC (Alliance System, Waters Corp., Milford, MA) using an Atlantis dC18 (3 μm, 4.6 × 150 mm; Waters) column as previously described (34,67). For UPLC quantitation, 10 μl samples were injected into a Waters Acquity UPLC-MS system using an Acquity BEH Phenyl Column (1.7 μm, 2.1 × 100 mm; Waters) as previously described (47,66). Molar quantities per eye were calculated based on standard solutions with spectrophotometrically determined concentrations. Peak areas were calculated using Waters Empower Software, and the results were analyzed using Excel (Microsoft, Redmond, WA).

Fluorescence microscopy

Mice were sacrificed, and eyeballs were enucleated and fixed in 4% paraformaldehyde (PFA; Electron Microscopy Science, Hatfield, PA) for 5 min at room temperature. The anterior segments were removed, including the cornea, lens, iris and vitreous body. Eyes were fixed in 4% PFA for 40 min at 4°C, followed by 5% and 10% sucrose for 1 h each, at 4°C. Next, the eyes were incubated in 30% sucrose at 4°C overnight and then embedded in OCT compound (Sakura, Torrance, CA). Then, 10 μm cryostat sections were cut and mounted on slides (Superfrost Plus; Fisher Scientific, Pittsburgh, PA). Sections were warmed and washed twice with 1× PBS for 3 min. Next, coverslips were attached with mounting medium (S302380-2, Dako Mounting Medium, Agilent Technologies, Santa Clara, CA). Images of the retina sections were taken using a confocal microscope (excitation 488 nm, emission 525 nm, 40× oil objective lens; Nikon A1). Data were analyzed using Fiji (SciJava software ecosystem). The threshold was set using MaxEntropy, and the size of the area of interest was normalized by the length of RPE layer captured in the image.

Sections were also stained with hematoxylin and eosin (H&E) to investigate adverse effects of therapy. Mice were sacrificed, and eyeballs were enucleated, fixed with formalin and paraffin-embedded. H&E staining was performed as previously described.

Transmission electron microscopy

Mice were euthanized under anesthesia and fixed by vascular perfusion with 2% formaldehyde and 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.2). The anterior segments of the eye were removed (including the cornea, lens, iris and vitreous body), and the eyes were cut into small pieces and placed in the same fixative for 60 min following methods in Koch et al. (68). Next, the eyes were post-fixed with 1% OsO4 in cacodylate buffer for 60 min, dehydrated and then embedded in Lx-112 (Ladd Research Industries, Inc., Vermont, VA). Sections were cut at 90 nm, collected on grids, stained with uranyl acetate and lead citrate and examined using an electronic microscope (JEOL JEM-1200 EXII). Images were taken with Hamamatsu ORCA HR Camera (Naka-ku, Hamamatsu City, Shizuoka, Japan).

Electroretinography

Mice were dark-adapted overnight for more than 12 h and anesthetized as previously described (69). All procedures were performed under dim red light. A solution consisting of 2.5% phenylephrine hydrochloride (Paragon BioTek Inc., Portland, OR) and 1% tropicamide (Akorn Inc.) was administered topically to dilate the pupils of the mice. ERG was performed according to the International Society for Clinical Electrophysiology of Vision (ISCEV) standard, using the Espion E2 Electroretinography System (Diagnosys LLC, Lowell, MA). After 5 min of re-dark adaptation, the pulse intensity for the regular ERG test was 0.001 Cd/m² for the scotopic test, 3 Cd/m² for the maximal test and 30 Cd/m² for the photopic test. To test recovery of visual sensitivity after a bleach, mice were exposed to 800 cd/m² for 2 min, re-dark adapted and the rod photoreceptor response was recorded every 3 min using a pulse intensity of 1 Cd/m². A heating pad was used to maintain the temperature of the mice during the experiments. ERG recordings were performed at 6 and 12 months of age, which were at least 1 month apart from the fundus AF. This was done to avoid cataract formation, which can interfere with AF readings. Furthermore, this 1-month spacing strategy mitigates potential mouse death from frequent and not adequately spaced anesthetization procedures.

Multimodal fundus imaging in mice

AF images were obtained using a Spectralis optical coherence tomography scanning laser confocal ophthalmoscope (OCT-SLO Spectralis 2; Heidelberg Engineering, Heidelberg, Germany) as previously described (63,70). Mouse pupils were dilated using topical 2.5% phenylephrine hydrochloride and 1% tropicamide (Akorn Inc., Lake Forest, IL) before and after anesthesia. Mice were anesthetized by intraperitoneal injection of ketamine/xylazine as described above. A heating pad was used to keep body temperature at 40°C. Infrared reflectance (IR, 820 nm) and fundus AF (488 nm) images were acquired at a resolution of 768 × 768 pixels using a 55° wide-field lens mounted on a Heidelberg Spectralis cSLO. Fundus AF intensity was analyzed using ImageJ (https://imagej.nih.gov/ij/; provided in the public domain by the NIH, Bethesda, MD) and fundus AF intensity was determined within a ring of eight segments surrounding the optic disc.

Statistics

Data were analyzed using Prism 8 (GraphPad Software, LLC, San Diego, CA). Data from all the four groups (Rho^D190N/+ mice on vitamin supplemented diet vs. Rho^D190N/+ mice on standard control diet vs. control mice on the vitamin A-supplemented diet vs. control mice on the standard control diet) were analyzed using one-way analysis of variance (ANOVA). The threshold of the F-value was 0.05. A Student’s t-test was used to compare the two groups when F < 0.05 as indicated by a one-way ANOVA. P-value for t-test: ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001.

Sample-size estimation

Twenty-four Rho^D190N/+ mice were on a vitamin A-supplemented diet and 16 Rho^D190N/+ mice were on a standard control diet for the experiment measuring the recovery of rod sensitivity after light treatment (Fig. 2D). Eighteen mice per group were used to test the difference in ERG results for Rho^D190N/+ mice fed a vitamin A diet compared to WT control mice that were not fed a vitamin A diet (Fig. 2E). For the retinoid experiment, 6–10 mice were used per each group (Rho^D190N/+ mice fed vitamin A diet, Rho^D190N/+ mice fed control diet, WT mice fed vitamin A diet and WT mice fed control diet) (Fig. 3F and G). For the bisretinoid experiment, 3–4 samples were collected per group. 3–5 eyes from individual mice were measured per sample (Rho^D190N/+ mice fed vitamin A diet, Rho^D190N/+ mice fed control diet, WT mice fed vitamin A diet and WT mice fed control diet) (Fig. 3H and I). Three mice per group were used for AF analysis (Rho^D190N/+ mice fed vitamin A diet, Rho^D190N/+ mice fed control diet, WT mice fed vitamin A diet and WT mice fed control diet) (Fig. 4).

Ethics statement

The study was approved by the Institutional Review Board (Protocol number AAAR8743). Written consent was not required as the analysis was performed as a retrospective review and there was minimal risk to the patient. No information or identifiable images of the patient are presented. The IACUC at Columbia University approved all experiments prior to experimentation.

Acknowledgements

We thank Sarah R. Levi, Joseph Ryu, Lawrence Chan and Andrew Zheng and members of their laboratories for sharing ideas and for critically reading the manuscript.

Conflict of Interest statement. S.H.T. receives financial support from Abeona Therapeutics, Inc. and Emendo. He is also the founder of Rejuvitas and is on the scientific and clinical advisory board for Nanoscope Therapeutics and Medical Excellence Capital.

Funding

National Institutes of Health (grant numbers: 5P30CA013696, U01 EY030580, U54OD020351, R24EY028758, R24EY027285, 5P30EY019007, R01EY018213, R01EY024698, R01EY024091, R01EY12951, R01EY026682, and R21AG050437); the Schneeweiss Stem Cell Fund; New York State (grant number SDHDOH01-C32590GG-3450000); the Foundation Fighting Blindness New York Regional Research Center Grant (grant numbers PPA-1218-0751-COLU, and TA-NMT-0116-0692-COLU); Nancy & Kobi Karp; the Crowley Family Funds; The Rosenbaum Family Foundation; Alcon Research Institute; the Gebroe Family Foundation and unrestricted funds from Research to Prevent Blindness, New York, NY, USA.

Author notes

Hye Jin Kim and Chia-Hua Cheng authors contributed equally to this manuscript.

The authors wish it to be known that, in their opinion, the last Xiaorong Li, Janet R. Sparrow and Stephen H. Tsang authors should be regarded as Co-corresponding authors.