All trans-Retinal – 11 Amino Undecanoate

All-trans-retinal dimer formation alleviates the cytotoxicity of all- trans-retinal in human retinal pigment epithelial cells
Jie Li , Yanli Zhang , Xianhui Cai , Qingqing Xia , Jingmeng Chen , Yi Liao , Zuguo Liu , Yalin Wu *

Keywords:
All-trans-retinal
Metabolism
All-trans-retinal dimer
Human retinal pigment epithelial cells Cytotoxicity

1. Introduction

A B S T R A C T

Effective clearance of all-trans-retinal (atRAL) from retinal pigment epithelial (RPE) cells is important for avoiding its cytotoxicity. However, the metabolism of atRAL in RPE cells is poorly clariﬁed. The present study was designed to analyze metabolic products of atRAL and to compare the cytotoxicity of atRAL versus its derivative all-trans-retinal dimer (atRAL-dimer) in human RPE cells. We found that all-trans- retinol (atROL) and a mixture of atRAL condensation metabolites including atRAL-dimer and A2E were generated after incubating RPE cells with atRAL for 6 h, and the amount of atRAL-dimer was signiﬁcantly higher than that of A2E. In the eyes of Rdh8 Abca4 mice, a mouse model with defects in retinoid cycle that displays some symbolic characteristics of age-related macular degeneration (AMD), the level of atRAL-dimer was increased compared to wild-type mice, and was even much greater than that of A2E & isomers. The cytotoxicity of atRAL-dimer was reduced compared with its precursor atRAL. The latter could provoke intracellular reactive oxygen species (ROS) overproduction, increase the mRNA expression of several oxidative stress related genes (Nrf2, HO-1, and g -GCSh), and induce DCm loss in RPE cells. By contrast, the abilities of atRAL-dimer to induce intracellular ROS and oxidative stress were much weaker versus that of concentration-matched atRAL, and atRAL-dimer exhibited no toxic effect on mitochondrial function at higher concentrations. In conclusion, the formation of atRAL-dimer during atRAL metabolic process ameliorates the cytotoxicity of atRAL by reducing oxidative stress.

visual cycle, all-trans-retinal (atRAL) (Fig. 1 A), is closely associated with the pathogenesis of retinal macular degeneration (Chen et al.,

Age-related macular degeneration (AMD) is a leading cause of visual impairment in the developed world (Cheung and Eaton, 2013; Coleman et al., 2008; Lim et al., 2012). As an ocular disease with complicated etiology, a multitude of genetic and environ- mental factors are involved in AMD progression (Maeda et al., 2009c ). The recent years have witnessed great progress in AMD studies (Chen et al., 2012a, 2012b; Kohno et al., 2013; Maeda et al., 2009b, 2008, 2012; Masutomi et al., 2012; Sawada et al., 2014; Shiose et al., 2011; Sparrow et al., 2010). An increasing body of research suggests that the toxic effect of an intermediate of the 2012b; Li et al., 2015; Maeda et al., 2008; Sawada et al., 2014). Previous studies indicate that free atRAL can induce severe cytotoxicity if it is not properly cleared in the retina (Maeda et al., 2009b ). Accordingly, an elaborate program was evolved to eliminate atRAL: free atRAL molecules liberated from rhodopsin following photoexcitation reacts with phosphatidylethanolamine (PE) in disc membranes and generates the Schiff base product N- retinylidene-phosphatidylethanolamine (NRPE) (Molday et al., 2009; Rando, 2001; Sparrow et al., 2010), thus decreasing atRAL concentration and inhibiting its cytotoxicity. NRPE is then rapidly hydrolyzed and reduced into all-trans-retinol (atROL) (vitamin A) (Fig. 1 B) in the photoreceptor cell, and returns into the reaction of 11-cis-retinal regeneration in the retinal pigment epithelial (RPE) cell (Rando, 2001 ). Under some circumstances, NRPE can react with a second molecule of atRAL and eventually generate a series of

42 J. Li et al. / Toxicology 371 (2016) 41–48

We also found that the generation of atRAL-dimer signiﬁcantly relieved the cytotoxicity of atRAL in RPE cells through the reduction of the production of intracellular reactive oxygen species (ROS). This work provides further insights into atRAL metabolism in RPE cells. Even more importantly, a better understanding of the physiological role for the generation of lipofuscin bisretinoid atRAL-dimer was attained.

2. Materials and methods

2.1. Reagents and cell line

atRAL was purchased from Sigma-Aldrich (St. Louis, MO, USA) and prepared as a stock DMSO solution (20 mM) and stored in the dark at 20 C. atRAL-dimer was synthesized as previously described (Fishkin et al., 2005 ). atROL was synthesized by treating

Fig. 1. Structures of all-trans-retinal, all-trans-retinol, all-trans-retinal dimer and A2E.

bisretinoids, including A2E (Fig. 1 D), isoA2E, iisoA2E, and all-trans- retinal dimer (atRAL-dimer) (Fig. 1 C) (Li et al., 2013; Sparrow et al., 2012, 2010). Collectively, to our knowledge, at least three pathways are involved in the clearance of atRAL in the retina: (1) the retinol dehydrogenase (RDH) dependent clearance, in which atRAL is reduced into atROL and participates in the 11-cis-retinal regenera- tion (Rando, 2001; Saari, 2012); (2) the generation of bisretinoids such as A2E and atRAL-dimer (Sparrow et al., 2012, 2010); (3) since atRAL is light sensitive, this chromophore is readily isomerized, photo-oxidized, and photo-cleaved in the retina.
The integrity of RPE monolayer is essential for maintaining the normal visual transduction in the retina (Ford et al., 2011 ). But RPE cell is especially vulnerable to damage because of its unique physiological function, anatomic location, and relative high metabolic activity (He et al., 2008 ). What’sworse is that free atRAL which escapes reduction in the photoreceptor cell will eventually enter into RPE cells by phagocytosis or diffusion (Maeda et al., 2009b ). Therefore, delayed clearance of atRAL in the photoreceptors will result in its excess accumulation in RPE cells, and may cause oxidative stress mediated cytotoxicity (Chen et al., 2012b; Li et al., 2015; Maeda et al., 2008; Saari, 2012). Plenty of evidence has demonstrated that RPE cell is responsible for the esteriﬁcation, isomerization and oxidation of atROL, which is transported from the photoreceptor cell or taken up from blood, and is indispensable for the regeneration of 11-cis-retinal (Batten et al., 2004; Moiseyev et al., 2005; Parker and Crouch, 2010; Saari, 2012). Nevertheless, so far there is little study on the metabolic processing of atRAL in RPE cells.
atRAL-dimer, a unique atRAL condensation product with an aldehyde group, is scarce in the wild-type mouse retina, whereas it is abundant in the retina of Stargardt’sdisease mouse model Abca4 mice and Rdh8 Abca4 mice that assemble the typical characteristics of retinopathies in AMD patients (Kim et al., 2007; Maeda et al., 2008). A previous study has found that the quantity of atRAL-dimer-PE, which is a conjugate derivative of atRAL-dimer with PE, in the retina of Abca4 mice are even more abundant than A2E (Fishkin et al., 2005; Kim et al., 2007). The photo-activity of atRAL-dimer is signiﬁcantly higher than that of A2E, which is considered to play an even more critical role in the pathogenesis of retinal macular degeneration (Kim et al., 2007 ). However, the conﬁrmative evidence to clarify the role of atRAL-dimer is still missing.
In this study, the accumulation of atRAL in RPE cells were simulated in vitro and in vivo, and its metabolism was analyzed by reverse-phase high performance liquid chromatography (HPLC). We observed that the buildup of atRAL-dimer was selectively augmented in the metabolites of abnormally accumulated atRAL.

atRAL with 1 eq of sodium borohydride (NaBH4) in dry ethanol. HPLC grade acetonitrile was purchased from Fisher scientiﬁc (Fair Lawn, NJ, USA). HPLC grade triﬂuoroacetic acid was obtained from Aladdin (Shanghai, China). All other chemical reagents were AR grade and were purchased from Sinopharm (Shanghai, China).
A human RPE cell line ARPE-19, which displays the differenti- ated phenotype of RPE cells, was purchased from FuDan IBS Cell Center (Shanghai, China). ARPE-19 cells were maintained in Dulbecco’s Modiﬁed Eagle Medium (DMEM) with 10% fetal bovine serum (PAN biotech, Germany) and 1% penicillin/streptomycin in a humidiﬁed incubator at 37 C and 5% CO2. ARPE-19 cells seeded in 6-well or 96-well cell culture plates (Nunc, Shanghai) attained 70– 80% conﬂuence and were used for the subsequent experiments.
2.2. Analysis of cell viability

Cytotoxicity was detected by MTT (Solabio, Beijing) assay. Brieﬂy, after incubation of ARPE-19 cells with atRAL or atRAL- dimer for indicated time intervals, cells were incubated with 0.5 mg/ml MTT for 4 h at 37 C. Then the medium was removed and 150 ml DMSO were added to each well. After shaking for 10 min, the optical density (OD) value in each well was spectrophotometrically measured at 490 nm with a microplate reader (Multiskan FC, Thermo scientiﬁc). Cell viability was presented as a proportion of control optical density.

2.3. Animals

Two pairs of Rdh8/Abca4 knockout (Rdh8 Abca4 ) mice were generously provided by Prof. Krzysztof Palczewski (Case Western Reserve University, Cleveland, Ohio) as a gift, and were bred and raised in the laboratory animal center of Zhejiang University. C57BL/6 mice were obtained from Shanghai SLAC Laboratory Animal Co. Ltd., and raised under 12 h on-off cyclic lighting with an in-cage illuminance of 60–90 lx. Experiments with animals were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee of Zhejiang University and adherent to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.

2.4. Tissue extraction and HPLC analysis

Murine eyes were extracted in accordance with previously reported method (Li et al., 2013 ). Brieﬂy, tissues (6 or 8 eyes/ sample) were homogenized in 1 ml phosphate buffered saline (PBS) in the presence of an equal volume of 50% methanolic chloroform using a glass tissue grinder. After centrifugation at 3000 g for 5 min, the organic layer was moved into a 25 ml round bottom ﬂask. Extraction was repeated with addition of 3 ml chloroform for three times. After the organic phases were pooled,and the combined solvents was removed in a rotary evaporator, residues in the bottom ﬂask were re-dissolved with 50% methanolic chloroform and transferred into a 1.5 ml centrifuge tube, and dried under argon gas. The extract was then redissolved in 50% methanolic chloroform and centrifuged at 13000 g for 5 min. The supernatant was examined by reverse phase HPLC (Waters Corp. Milford, MA) equipped with 2695 separation module and a 2998 photodiode array detector.
For chromatographic detection, an analytical scale Atlantis dC18 (3 mm, 4.6 mm 150 mm, Waters) column was used for analysis of RPE lipofuscin bisretinoids. The gradient mobile phase was composed of acetonitrile and water in the presence of 0.1% triﬂuoroacetic acid: 75–90% acetonitrile (0–30 min); 90–100% acetonitrile (30–40 min); 100% acetonitrile (40–100 min) with a ﬂow rate of 0.5 ml/min. Photodiode array detection was monitored at 430 nm. Tissue extraction and HPLC analysis were carried out under dim red light. Integrated peak areas (mV s) were determined by Empower 3 software (Waters Corp., Milford, MA).

2.5. Metabolisms of atRAL in ARPE-19 cells
Cells were incubated with 15 mM atRAL for 6 h, and then washed with PBS for three times and harvested for HPLC analysis. The cell lysate was extracted with 50% methanolic chloroform. For compound elution, an Atlantis dC18 (3 mm, 4.6 150 mm, Waters) reverse-phase column was used for the stationary phase, and a gradient of acetonitrile in water with 0.1% triﬂuoroacetic acid was set for the mobile phase: 85–100% acetonitrile, 0.8 ml/min 15 min; 100% acetonitrile, 0.8–1.2 ml/min 15–20 min; 100% aceto- nitrile, 1.2 ml/min, 20–40 min. Detections by photodiode array were set at 325 nm and 430 nm.

2.6. Determination of mitochondrial transmembrane potential

Disruption of the mitochondrial inner transmembrane poten- tial (DCm) is an early event of apoptosis (Zhang et al., 2012 ). To measure DCm, a JC-1 staining kit (Beyotime, Haimen, China) was used for the analysis following the manufacturer’s instructions with minor modiﬁcations. Brieﬂy, ARPE-19 cells were seeded on 96-well culture plates and maintained in cell-culture incubator for 24 h. Cells were exposed to atRAL or atRAL-dimer for 6 h, washed with PBS and then incubated with JC-1 for 30 min at 37 C. After washed with JC-1 staining buffer for three times, cells were observed and photographed under an inverted ﬂuorescence microscope (ECLIPSE Ti, Nikon).

2.7. Measurement of ROS in RPE cells

The production of total ROS was estimated using dichloro- dihydro-ﬂuorescein diacetate (DCFH-DA) (Beyotime, Haimen, China) staining assay. ARPE-19 cells were seeded on 96-well culture plates and maintained in cell-culture incubator for 24 h.

After cells were exposed to atRAL or atRAL-dimer for 6 h, the medium was removed, and 100 ml DMEM that contains 1 mM of ROS ﬂuorescent probe DCFH-DA was added in each well. After incubation for 30 min, cultures were washed twice with PBS and subsequently photographed by the ﬂuorescence microscope (ECLIPSE Ti, Nikon).

2.8. RNA extraction, cDNA synthesis and quantitative real-time PCR

ARPE-19 cells were seeded on 6-well culture plates and maintained in cell-culture incubator for 24 h. After cells were exposed to atRAL or atRAL-dimer for 6 h, the medium was removed, and washed with PBS twice. Total cellular RNA was puriﬁed using TRIzol reagent (Invitrogen Inc., Carlsbad, CA), and complementary DNA (cDNA) was synthesized using a ReverTra Ace qPCR RT kit (TOYOBO). Real-time PCR was performed using Brilliant SYBR Green qPCR Master Mix (TakaRa) on a CFX96 real- time PCR detection system (Bio-Rad). Reactions without template served as negative controls. Each real-time PCR was performed on 3 different experimental samples, and each sample was performed in triplicate. The relative expression level of each target gene was calculated by the 2 method using GAPDH as a loading control (Livak and Schmittgen, 2001 ). The primer sequences (Shanghai Sunny Biotechnology, Shanghai) used for real-time PCR are listed in Table 1.

2.9. Statistical analysis

All experiments were performed at least three times. Data were expressed as the means SEM and analyzed by using two-tailed and unpaired Student’st test or One-way ANOVA, followed by Newman–Keuls test for multiple comparisons. P value of <0.05 was considered signiﬁcantly different. The statistic and graphic software GraphPad Prism (Version 5, GraphPad Software, Inc., La Jolla, CA) was used for all statistical and graphic analysis. 3. Results 3.1. Metabolism of atRAL in RPE cells 15 mM atRAL were incubated with ARPE-19 cells in the dark for 6 h. The cells were harvested, followed by analysis of its hydrophobic extract by reverse-phase HPLC while monitoring eluents at 325 nm. We found two new peaks exhibiting absorbance maxima of 350 nm in front of the atRAL peak, their retention time (Rt) being 6.9 min and 7.6 min, respectively. By comparison with the standard HPLC proﬁle of atROL (vitamin A) (Fig. 2 A), we conﬁrmed that the 7.6-min-peak was atROL. Monitoring at 430 nm, we readily detected the atRAL-dimer (Rt = 27.6 min) in the extract. In addition, trace amounts of A2E and its isomer (Fig. 2 B, Top inset left) were also found to be produced in RPE cells. HPLC quantiﬁcation of A2E and atRAL-dimer in atRAL-treated RPE cells Fig. 2. Metabolism of atRAL in RPE cells. RPE cells were harvested after incubation with atRAL for 6 h. (A) Representative reverse-phase HPLC chromatograms (monitored at 325 nm) was generated with extracts of atRAL-treated RPE cells, atROL, and atRAL; Top insets, UV–vis absorbance spectra of atROL and atRAL. (B) Representative reverse-phase HPLC chromatograms (monitored at 430 nm) were generated with extracts of atRAL-treated RPE cells, RPE cells and atRAL-dimer standard; Top inset left, HPLC chromatograms of A2E and its isomers; Top inset right, UV–vis absorbance spectrum of atRAL-dimer. abs, absorbance; mAU, milliabsorbance unit. (C) HPLC quantiﬁcation of A2E and atRAL- dimer in atRAL-treated RPE cells. Each value represents mean SEM (n = 3). indicated that the latter was easier to accumulate in RPE cells (2.1 0.5 pmole/10 cells vs 5.7 0.6 pmole/10 cells) (Fig. 2 C). 3.2. atRAL-dimer is accumulated in greater abundance than A2E & isomers in Rdh8 Abca4 mouse retina In the retina of wild-type (C57BL/6) mice, there were little lipofuscin bisretinoid components, such as A2E and its isomers (A2E & isomers) and atRAL-dimer (Figs. 3 A, B), but contents of A2E & isomers and atRAL-dimer were signiﬁcantly increased in the eyes of Rdh8 Abca4 mice (Figs. 3 A, B), consistent with previous reports (Kim et al., 2007 ). We then quantiﬁed the amounts of A2E & isomers and atRAL-dimer in C57BL/6 and Rdh8 Abca4 mice, respectively. The levels of A2E & isomers in the eyes of wild-type mice was 5.5 1.0 pmole/eye, but raised up to 20.4 2.3 pmole/eye in Rdh8 Abca4 mice (Fig. 3 C). The amount of atRAL-dimer was only 2.9 0.7 pmole/eye in wild-type mice, but increased to 61.8 5.5 pmole/eye in the Rdh8 Abca4 mice, exhibiting about 19-fold increase in the double-gene knockout mouse eyes (Fig. 3 D). 3.3. atRAL-dimer is less toxic than atRAL in RPE cells Cytotoxicities of atRAL or atRAL-dimer in RPE cells were measured by MTT assay. We found that atRAL reduced cell viability in a time- and concentration-dependent manner (Fig. 4 A), whereas atRAL-dimer did not cause a signiﬁcant decline of RPE cell viability at comparable concentrations (Fig. 4 B). However, after 72 h, the cytotoxicity of atRAL-dimer began to appear, and was in a concentration-dependent manner (Fig. 4 B). Mitochondrial trans- membrane potential (DCm) of RPE cells was detected by JC-1 staining. All the cells under normal condition exhibited a bright red ﬂuorescence, while atRAL concentration-dependently increased the number of cells with green ﬂuorescence within 6 h, suggesting that it caused DCm loss dependent on the concentration of atRAL (Fig. 4 C). Interestingly, 20 or 30 mM of atRAL-dimer did not signiﬁcantly affect mitochondrial transmembrane potential within Fig. 3. atRAL-dimer is accumulated in greater amount than A2E & isomers in eyes of mice. (A) a representative reverse-phase HPLC chromatogram (monitored at 430 nm) was generated with atRAL-dimer, Top inset, UV–vis absorbance spectrum of atRAL-dimer; (B) Representative reverse-phase HPLC chromatograms (monitored at 430 nm) were generated with hydrophobic extracts of eyes from 6-month-old C57BL/6 (wild-type) or Rdh8 Abca4 mice; eight eyes/sample in the wild-type and six eyes/sample in the Rdh8 Abca4 , respectively. Top insets, UV–vis absorbance spectra of atRAL-dimer; Quantiﬁcation of A2E & isomers (C) or atRAL-dimer (D) in the eyes of wild-type or Rdh8 Abca4 mice; Each value represents mean SEM (n = 3). abs, absorbance; mAU, milli- absorbance unit. **, P <0.01; ***, P <0.001. Fig. 4. atRAL-dimer causes less cytotoxicity than atRAL in RPE cells. (A) Cell viability, 6, 12 and 24 h after atRAL exposure (0, 10, 15, 20 and 30 mM), was evaluated by MTT assay. Each value represents mean SEM (n = 4). (B) Cell viability, 6, 12, 24, 48 and 72 h after atRAL-dimer exposure (0, 10, 15, 20 and 30 mM), was evaluated by MTT assay. Each value represents mean SEM (n = 4). (C) Mitochondrial function in response to atRAL or atRAL-dimer exposure was determined by staining with the DCm sensing dye JC-1. After ARPE-19 cells were treated with atRAL (0, 10, 20 mM) or atRAL-dimer (0, 20, 30 mM) for 6 h, changes of DCm were observed under an inverted ﬂuorescent microscope. Scale bar, 100 mm. Fig. 5. The oxidative stress of RPE cells stimulated by atRAL-dimer was signiﬁcantly weaker than that of atRAL. (A) Intracellular ROS, 6 h after exposure of cells to atRAL and atRAL-dimer (0, 10 and 20 mM), was visualized by ﬂuorescence microscopy. Scale bar, 100 mm. (B) The mRNA expression levels of oxidative stress-related genes Nrf2 (a), HO-1 (b) and g -GCSh (c) in cells untreated or treated with 15 mM atRAL or atRAL-dimer for 6 h were quantiﬁed by qRT-PCR. Expression levels of mRNA were normalized to GAPDH mRNA levels, and expressed as fold increases changes in treated cells versus control cells. Each value represents mean SEM (n = 3), each performed in triplicate. *, P <0.05; **, P <0.01; ***, P <0.001. ns, statistically nonsigniﬁcant different (P > 0.05).

6 h (Fig. 4 C). These results suggest that atRAL-dimer is less toxic than atRAL.

3.4. atRAL-dimer causes less cellular oxidative stress than atRAL

The intracellular ROS was assayed after atRAL or atRAL-dimer treatment by DCFH-DA staining. Compared to the control group, atRAL caused a concentration-dependent increase in ROS produc- tion after 6 h incubation (Fig. 5 A). However, at the same concentration, the ROS ﬂuorescence signal generated by atRAL- dimer treatment was very weak (Fig. 5 A). Furthermore, we determined the level of oxidative stress related genes, such as Nrf2, HO-1, and g -GCSh, by qRT-PCR. We found that 6-h treatment of 15 mM of atRAL signiﬁcantly upregulated the expression levels of Nrf2, HO-1, and g -GCSh (about 1.5-fold, 25.7-fold and 3.0-fold of the control, respectively). While at the same condition, atRAL- dimer could not remarkably affect the expression levels of Nrf2 and g -GCSh (only approximately 1.1-fold and 1.4-fold of that in the control group) (Fig. 5 B,a and c), except that it up-regulated the mRNA level of HO-1 for about 7.2-fold compared with the control (Fig. 5 B, b).

formation dependents on an alkaline environment, which might be achieved by some cell membrane binding basic proteins (Fishkin et al., 2005 ). From this perspective, A2E production seems to be much easier than that of atRAL-dimer in the cell. On the contrary, in the current study, we found that RPE cells were prone to generate more atRAL-dimer than A2E (Figs. 2 B, C). Though both atRAL-dimer and A2E are products of NRPE and atRAL, their biosynthetic pathways are different (Fishkin et al., 2005 ). For instance, the elimination of the amino group of PE in atRAL-dimer precursor might not be an enzymatic process, whereas A2PE hydrolysis requires phospholipase activity (Fishkin et al., 2005; Liu et al., 2000; Sparrow et al., 2008). This might contribute to the fact that atRAL-dimer production is more efﬁcient than that of A2E under the physiological condition.
Small amount of atRAL-dimer was found in wild-type mouse retina, but much lower than the content of A2E (Figs. 3 B–D). When the intraocular atRAL was accumulated, for example, in Abca4 mice, an animal model of Stargardt’sdisease, atRAL-dimer-PE was more abundant than A2E (Kim et al., 2007). Accordingly, we found that the accumulation of atRAL-dimer in the retina was also much more than A2E & isomers in Rdh8 Abca4 mice (Fig. 3 B).

Furthermore, the increase folds of atRAL-dimer content in Rdh8

4. Discussion

Effective clearance of atRAL in the retina is of great importance for the vision maintenance (Chen et al., 2012b; Maeda et al., 2009b, 2008). Free atRAL is cytotoxic and could cause retinal degeneration when accumulated in the retina, while it also functions as an indispensable intermediate of the visual cycle (Chen et al., 2012b; Maeda et al., 2009b, 2008; Sawada et al., 2014). Most of atRAL in the retina will be converted to atROL by RDHs using NADPH as a cofactor (Maeda et al., 2008 ). However, the most important two RDHs for reducing the bulk of atRAL in the retina, RDH8 and RDH12, are mainly distributed in photoreceptors rather than RPE cells (Chen et al., 2012a ), therefore the elimination of atRAL in RPE cells remains elusive.
In this study, a human RPE cell line ARPE-19, which sustains the morphology and functions of RPE cells (Dunn et al., 1996 ), was employed to simulate the excess accumulation of atRAL in vitro, and the clearance process of atRAL was further explored afterwards. Our results demonstrated that the majority of intracellular atRAL was reduced into atROL (Fig. 2 A), which should be due to the catalytic functions of RDHs, indicating that the elimination of atRAL in RPE cells was also mainly an enzymatic process. Although RDHs resident in RPE cells are mainly responsible for 11-cis-retinol oxidation, some of them, for example, RDH11, also have the activity towards atRAL (Parker and Crouch, 2010 ). However, our observation is in contradiction with a previous report which suggests that ARPE-19 cells lack RDH activity toward retinal (Haeseleer et al., 2002 ).
We also found that some atRAL was transformed into atRAL- dimer and other bisretinoids (Fig. 2 B). Presently, the generation of bisretinoids is mainly considered as a non-enzymatic synthetic process, as so far no enzyme has been identiﬁed to be involved in this process (Sparrow et al., 2010 ). Therefore, RPE cells themselves might not have the ability to actively modulate its formation. The reaction rate of bisretinoids formation is positively related to the concentration of the reactants, thus the increase of free atRAL in the retina resulted in excess accumulation of bisretinoids. As a result, in the retina of Rdh8 Abca4 mice in which atRAL clearance is disrupted, we observed that the deposition of atRAL- dimer, A2E and its isomers (A2E & isomers) were sharply increased (Figs. 3 B–D).
Previous reports suggested that A2E and its precursor A2PE could be generated in solution under ostensibly neutral conditions (Liu et al., 2000; Parish et al., 1998), while the atRAL-dimer

Abca4 mice were signiﬁcantly higher than that of A2E & isomers in wild-type mice (Figs. 3 C, D), which suggests that the buildup of atRAL-dimer might be selectively increased after the abnormal accumulation of atRAL in the retina.
The balance between generation and degradation determines the intracellular content of atRAL-induced metabolites. Our results indicate that the production of atRAL-dimer is more than A2E when atRAL is accumulated in cells (Figs. 2 B, C). However, the total amount of atRAL-dimer was much less than that of A2E under physiological condition (Fig. 3 B), which suggested that the atRAL- dimer production rate might be much lower than its removal rate in the normal eye. Plenty of evidence has demonstrated that the bisretinoids deposited in the RPE defy further degradation as lysosomal enzymes might not be able to recognize them (Sparrow et al., 2010 ). Nevertheless, it was also reported that atRAL-dimer and A2E are highly photosensitive, and liable to be photo cleaved in vitro (Kim et al., 2007; Sparrow et al., 2002; Wu et al., 2010). Owing to the relative higher light intensity and oxygen content in the retina, atRAL-dimer and A2E seem easy to be oxidized, cleaved into certain hydrophilic molecules and excreted from RPE cells. Moreover, given the fact that atRAL-dimer is more light-sensitive than A2E (Kim et al., 2007; Maeda et al., 2009a), the retina may possess stronger ability to efﬁciently eliminate atRAL-dimer after illumination. In Rdh8 Abca4 mice, atRAL clearance was disrupted and free atRAL concentration was sharply raised up in the retina, resulting in the production of huge amount of atRAL- dimer and A2E (Maeda et al., 2008 ). Indeed, light-induced degradation might be the only way for their clearance, but the amount of atRAL-dimer eliminated by light-driven cleavage is limited as only ﬁnite amount of light can reach the retina daily. Thus, not only that the photo-cleavage effect could not cut down the accumulation of the bisretinoids, but also because the formation of atRAL-dimer was far more efﬁcient than that of A2E, a higher concentration of atRAL-dimer was detected in the retina of Rdh8 Abca4 mice (Figs. 3 B–D).
atRAL could provoke intracellular ROS overproduction, cause oxidative stress, and induce mitochondrial dysfunction in RPE cells (Figs. 4 and 5), which were in accordance with previously reported studies (Chen et al., 2012b; Li et al., 2015). In the present study, we found that the abilities of atRAL-dimer to induce ROS and oxidative stress were much weaker even at the same concentration of atRAL (Fig. 5 ). Mitochondria might be a target cellular organelle for atRAL, while atRAL-dimer exhibited no toxic effect on mitochondrial function even at higher concentrations (Fig. 4 C). The toxic effect of atRAL is mainly due to its active aldehyde group ( CHO) (Maeda et al., 2009b ). Similarly, atRAL-dimer also contains a chemically active aldehyde group, which can also react with biological macromolecules. For example, it appears to readily react with PE and subsequently forms atRAL-dimer-PE (Kim et al., 2007). Thus, the active aldehyde group of atRAL-dimer may contribute to our observations that long term atRAL-dimer incubation also caused viability inhibition in RPE cells (Fig. 4 B). Nevertheless, due to the steric hindrance caused by the alkyl chain attached to the a-C and the electronic effect induced by the alkyl group, the aldehyde group of atRAL-dimer is more stable than that of atRAL, which signiﬁcantly weakens its cytotoxicity (Figs. 4 and 5). Furthermore, that two molecules of atRAL condense to form one molecule atRAL- dimer could efﬁciently decrease the concentration of free atRAL in the retina. Taken together, the rapid conversion of atRAL into atRAL-dimer under a certain concentration can effectively alleviate its cytotoxicity by reducing free atRAL concentration and relieving atRAL induced-oxidative stress in RPE cells. This could be an effective way to protect the retina from the toxicities induced by the excess atRAL accumulation.
Collectively, our study demonstrated that the metabolic path- ways of atRAL in RPE cells include both RDHs-dependent reduction and non-enzymatic bisretinoids generation. We found that atRAL- dimer rather than A2E was more easily to be deposited in vitro and in vivo when atRAL was accumulated. The generation of atRAL- dimer could be an effective way to alleviate the cytotoxicity of free atRAL by reducing oxidative stress in RPE cells. This work provides us a further understanding for the existence of atRAL-dimer in the retina.

Conﬂict of interest
The authors declare that there are no conﬂicts of interest. Acknowledgements

The present study was supported by China National Natural Science Foundation Grants 81570857 (Yalin Wu) and 81271018 (Yalin Wu); the Fundamental Research Funds for the Central Universities grant (Yalin Wu); and Medical Science and Technology Program of Zhejiang Province 2016KYA195 (Jie Li).

References
Batten, M.L., Imanishi, Y., Maeda, T., Tu, D.C., Moise, A.R., Bronson, D., Possin, D., Van Gelder, R.N., Baehr, W., Palczewski, K., 2004. Lecithin-retinol acyltransferase is essential for accumulation of all-trans-retinyl esters in the eye and in the liver. J. Biol. Chem. 279 (11), 10422–10432.
Chen, C., Thompson, D.A., Koutalos, Y., 2012a. Reduction of all-trans-retinal in vertebrate rod photoreceptors requires the combined action of RDH8 and RDH12. J. Biol. Chem. 287 (29), 24662–24670.
Chen, Y., Okano, K., Maeda, T., Chauhan, V., Golczak, M., Maeda, A., Palczewski, K., 2012b. Mechanism of all-trans-retinal toxicity with implications for stargardt disease and age-related macular degeneration. J. Biol. Chem. 287 (7), 5059– 5069.
Cheung, L.K., Eaton, A., 2013. Age-related macular degeneration. Pharmacotherapy 33 (8), 838–855.
Coleman, H.R., Chan, C.-C., Ferris, F.L., Chew, E.Y., 2008. Age-related macular degeneration. Lancet 372 (9652), 1835–1845.
Colombrita, C., Lombardo, G., Scapagnini, G., Abraham, N.G., 2003. Heme oxygenase- 1 expression levels are cell cycle dependent. Biochem. Biophys. Res. Commun. 308 (4), 1001–1008.
Dunn, K.C., Aotaki-Keen, A.E., Putkey, F.R., Hjelmeland, L.M., 1996. ARPE-19, a human retinal pigment epithelial cell line with differentiated properties. Exp. Eye Res. 62 (2), 155–169 .
Fishkin, N.E., Sparrow, J.R., Allikmets, R., Nakanishi, K., 2005. Isolation and characterization of a retinal pigment epithelial cell ﬂuorophore: an all-trans- retinal dimer conjugate. Proc. Natl. Acad. Sci. U. S. A. 102 (20),
7091–7096.

Ford, K.M., Saint-Geniez, M., Walshe, T., Zahr, A., D’Amore, P.A., 2011. Expression and role of VEGF in the adult retinal pigment epithelium. Invest. Ophthalmol. Vis. Sci. 52 (13), 9478–9487.
Haeseleer, F., Jang, G.F., Imanishi, Y., Driessen, C.A., Matsumura, M., Nelson, P.S., Palczewski, K., 2002. Dual-substrate speciﬁcity short chain retinol dehydrogenases from the vertebrate retina. J. Biol. Chem. 277 (47), 45537– 45546.
He, S., Yaung, J., Kim, Y.H., Barron, E., Ryan, S.J., Hinton, D.R., 2008. Endoplasmic reticulum stress induced by oxidative stress in retinal pigment epithelial cells. Graefes Arch. Clin. Exp. Ophthalmol. 246 (5), 677–683 .
Hu, Y., Lu, W., Chen, G., Zhang, H., Jia, Y., Wei, Y., Yang, H., Zhang, W., Fiskus, W., Bhalla, K., Keating, M., Huang, P., Garcia-Manero, G., 2010. Overcoming resistance to histone deacetylase inhibitors in human leukemia with the redox modulating compound beta-phenylethyl isothiocyanate. Blood 116 (15), 2732– 2741.
Kim, S.R., Jang, Y.P., Jockusch, S., Fishkin, N.E., Turro, N.J., Sparrow, J.R., 2007. The all- trans-retinal dimer series of lipofuscin pigments in retinal pigment epithelial cells in a recessive Stargardt disease model. Proc. Natl. Acad. Sci. U. S. A. 104 (49), 19273–19278 .
Kohno, H., Chen, Y., Kevany, B.M., Pearlman, E., Miyagi, M., Maeda, T., Palczewski, K., Maeda, A., 2013. Photoreceptor proteins initiate microglial activation via Toll- like receptor 4 in retinal degeneration mediated by all-trans-retinal. J. Biol. Chem. 288 (21), 15326–15341.
Li, J., Yao, K., Yu, X., Dong, X., Gan, L., Luo, C., Wu, Y., 2013. Identiﬁcation of a novel lipofuscin pigment (iisoA2E) in retina and its effects in the retinal pigment epithelial cells. J. Biol. Chem. 288 (50), 35671–35682.
Li, J., Cai, X., Xia, Q., Yao, K., Chen, J., Zhang, Y., Naranmandura, H., Liu, X., Wu, Y., 2015. Involvement of endoplasmic reticulum stress in all-trans-retinal-induced retinal pigment epithelium degeneration. Toxicol. Sci. 143 (1), 196–208.
Lim, L.S., Mitchell, P., Seddon, J.M., Holz, F.G., Wong, T.Y., 2012. Age-related macular degeneration. Lancet 379 (9827), 1728–1738.
Liu, J., Itagaki, Y., Ben-Shabat, S., Nakanishi, K., Sparrow, J.R., 2000. The biosynthesis of A2E, a ﬂuorophore of aging retina, involves the formation of the precursor, A2-PE, in the photoreceptor outer segment membrane. J. Biol. Chem. 275 (38), 29354–29360.
Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2 method. Methods 25, 402–408.
Maeda, A., Maeda, T., Golczak, M., Palczewski, K., 2008. Retinopathy in mice induced by disrupted all-trans-retinal clearance. J. Biol. Chem. 283 (39), 26684–26693.
Maeda, A., Golczak, M., Maeda, T., Palczewski, K., 2009a. Limited roles of Rdh8, Rdh12, and Abca4 in all-trans-retinal clearance in mouse retina. Invest. Ophthalmol. Vis. Sci. 50 (11), 5435–5443.
Maeda, A., Maeda, T., Golczak, M., Chou, S., Desai, A., Hoppel, C.L., Matsuyama, S., Palczewski, K., 2009b. Involvement of all-trans-retinal in acute light-induced retinopathy of mice. J. Biol. Chem. 284 (22), 15173–15183.
Maeda, T., Maeda, A., Matosky, M., Okano, K., Roos, S., Tang, J., Palczewski, K., 2009c. Evaluation of potential therapies for a mouse model of human age-related macular degeneration caused by delayed all-trans-retinal clearance. Invest. Ophthalmol. Vis. Sci. 50 (10), 4917–4925.
Maeda, T., Golczak, M., Maeda, A., 2012. Retinal photodamage mediated by all-trans- retinal. Photochem. Photobiol. 88 (6), 1309–1319.
Masutomi, K., Chen, C., Nakatani, K., Koutalos, Y., 2012. All-trans retinal mediates light-induced oxidation in single living rod photoreceptors. Photochem. Photobiol. 88 (6), 1356–1361.
Moiseyev, G., Chen, Y., Takahashi, Y., Wu, B.X., Ma, J.X., 2005. RPE65 is the isomerohydrolase in the retinoid visual cycle. Proc. Natl. Acad. Sci. U. S. A. 102 (35), 12413–12418. .
Molday, R.S., Zhong, M., Quazi, F., 2009. The role of the photoreceptor ABC transporter ABCA4 in lipid transport and Stargardt macular degeneration. Biochim. Biophys. Acta 1791 (7), 573–583 .
Nguyen, T., Sherratt, P.J., Huang, H.C., Yang, C.S., Pickett, C.B., 2003. Increased protein stability as a mechanism that enhances Nrf2-mediated transcriptional activation of the antioxidant response element. Degradation of Nrf2 by the 26S proteasome. J. Biol. Chem. 278 (7), 4536–4541 .
Parish, C.A., Hashimoto, M., Nakanishi, K., Dillon, J., Sparrow, J., 1998. Isolation and one-step preparation of A2E and iso-A2E, ﬂuorophores from human retinal pigment epithelium. Proc. Natl. Acad. Sci. U. S. A. 95 (25), 14609–14613.
Parker, R.O., Crouch, R.K., 2010. Retinol dehydrogenases (RDHs) in the visual cycle. Exp. Eye Res. 91 (6), 788–792.
Rando, R.R., 2001. The biochemistry of the visual cycle. Chem. Rev. 101 (7), 1881– 1896.
Saari, J.C., 2012. Vitamin A metabolism in rod and cone visual cycles. Annu. Rev. Nutr. 32, 125–145 .
Sawada, O., Perusek, L., Kohno, H., Howell, S.J., Maeda, A., Matsuyama, S., Maeda, T., 2014. All-trans-retinal induces Bax activation via DNA damage to mediate retinal cell apoptosis. Exp. Eye Res. 123, 27–36.
Shiose, S., Chen, Y., Okano, K., Roy, S., Kohno, H., Tang, J., Pearlman, E., Maeda, T., Palczewski, K., Maeda, A., 2011. Toll-like receptor 3 is required for development of retinopathy caused by impaired all-trans-retinal clearance in mice. J. Biol. Chem. 286 (17), 15543–15555.
Sparrow, J.R., Zhou, J., Ben-Shabat, S., Vollmer, H., Itagaki, Y., Nakanishi, K., 2002. Involvement of oxidative mechanisms in blue-light-induced damage to A2E- laden RPE. Invest. Ophthalmol. Vis. Sci. 43 (4), 1222–1227 .
Sparrow, J.R., Kim, S.R., Cuervo, A.M., Bandhyopadhyayand, U., 2008. A2E, a pigment of RPE lipofuscin, is generated from the precursor, A2PE, by a lysosomal enzyme activity. Recent Adv. Retin. Degeneration 613, 393–398.
Sparrow, J.R., Wu, Y., Kim, C.Y., Zhou, J., 2010. Phospholipid meets all-trans-retinal: the making of RPE bisretinoids. J. Lipid Res. 51 (2), 247–261.
Sparrow, J.R., Gregory-Roberts, E., Yamamoto, K., Blonska, A., Ghosh, S.K., Ueda, K., Zhou, J., 2012. The bisretinoids of retinal pigment epithelium. Prog. Retin. Eye Res. 31 (2), 121–135.
Wu, Y., Yanase, E., Feng, X., Siegel, M.M., Sparrow, J.R., 2010. Structural characterization of bisretinoid A2E photocleavage products and implications for age-related macular degeneration. Proc. Natl. Acad. Sci. U. S. A. 107 (16), 7275– 7280.
Zhang, X., Wu, J., Dou, Y., Xia, B., Rong, W., Rimbach, G., Lou, Y., 2012. Asiatic acid protects primary neurons against C2-ceramide-induced apoptosis. Eur. J. Pharmacol. 679 (1–3), 51–59.All trans-Retinal