Cysteinyl leukotriene receptor 1 modulates autophagic activity in retinal pigment epithelial cells

The retinal pigment epithelium (RPE), which is among the tissues in the body that are exposed to the highest levels of phagocytosis and oxidative stress, is dependent on autophagy function. Impaired autophagy and continuous cellular stress are associated with various disorders, such as dry age-related macular degeneration (AMD), a disease for which effective therapies are lacking. Cysteinyl leukotriene receptor (CysLTR) 1 is a potential modulator of autophagy; thus, the aim of this study was to investigate the role of CysLTR1 in autophagy regulation in the RPE cell line ARPE-19. The polarized ARPE-19 monolayer exhibited expression of CysLTR1, which was colocalized with β-tubulin III. In ARPE-19 cells, autophagic activity was rhythmically regulated and was increased upon CysLTR1 inhibition by Zafirlukast (ZK) treatment. H2O2 affected the proautophagic regulatory effect of ZK treatment depending on whether it was applied simultaneously with or prior to ZK treatment. Furthermore, mRNA levels of genes related to the leukotriene system, autophagy and the unfolded protein response were positively correlated. As CysLTR1 is involved in autophagy regulation under basal and oxidative stress conditions, a dysfunctional leukotriene system could negatively affect RPE functions. Therefore, CysLTR1 is a potential target for new treatment approaches for neurodegenerative disorders, such as AMD.

In addition to playing immune-related roles, leukotrienes, which regulate noninflammatory activities, are synthesized by diverse nonhematopoietic cells, such as retinal pigment epithelial (RPE) cells 4,5 . It has been postulated that leukotrienes (LTB 4 and LTC 4 ) are involved in the phagocytosis of disks shed by RPE cells in Xenopus laevis 4 . Recently, it was reported that inhibition of ALOX5 by PEDF-R peptides increases the survival of RPE cells undergoing oxidative stress 5 . This augmented survival of RPE cells could be explained by the recent finding that LTC 4 induces an intracellular death-triggering mechanism in the late phase of the unfolded protein response (UPR) caused by excessive endoplasmic reticulum (ER) stress 6 . Upon UPR activation, the membrane-associated CysLTR1 and CysLTR2 are internalized, and CsyLTR1-and CysLTR2-mediated cell death is inhibited by leukotriene receptor antagonists 6 . The UPR induces three distinct pathways via eukaryotic translation initiation factor 2-alpha kinase 3 (EIF2AK3) [PERK], activating transcription factor 6 (ATF6) and inositol-requiring enzyme 1 (IRE1; gene: ERN1), leading to activation of the transcription factors ATF4, ATF6 and XBP1, respectively. These three transcription factors have been reported to be important regulators of autophagy under moderate UPR activity [7][8][9] . Therefore, autophagy is recognized as a cellular prosurvival mechanism 10,11 . As the inhibition of intracellular LTC 4 receptor signaling keeps the cell in a prosurvival UPR state and moderate UPR activity induces autophagy to cope with ER stress, it seems obvious that leukotriene receptor antagonists could have the potential to regulate autophagy. Recently, Hu et al. demonstrated that blockage of CysLTR1 signaling leads to amelioration of liver injury through activation of autophagy upon aluminum overload 12 .
Autophagy is an intracellular process through which misfolded or long-lived proteins, lipid droplets, invading microorganisms and damaged organelles are degraded and recycled 13 and an adaptive process that provides nutrients and energy in response to stress, such as starvation or hypoxia. At least three forms of autophagy have

CysLTR1 expression in ARPE-19 cells. As Reynolds et al. already evaluated the protein expression of
CysLTR1 in dispersed ARPE-19 cells 34 , the first aim of the present study was to confirm CysLTR1 expression and determine the localization of the receptor in polarized ARPE-19 cells. Immunofluorescence (IF) analysis revealed basolateral CysLTR1 expression in polarized ARPE-19 monolayers (Supplementary Fig. 1) and localization of CysLTR1 to class III β-tubulin-positive cytoskeleton structures (microtubules) (Fig. 1, arrowheads). However, in individual cells of the monolayer CysLTR1 was internalized and located near the nucleus (Fig. 1d, arrow) and was not localized to class III β-tubulin-positive cytoskeleton structures. A specific fluorescence signal was absent in the secondary antibody only control (Fig. 1e).
As ER stress induced by brefeldin A increases CysLTR1 expression 6 , we investigated the effect of H 2 O 2 as an ER stress inducer on CysLTR1 mRNA and protein expression in polarized ARPE-19 cells. Polarized ARPE-19 cells were left untreated or treated with a nonlethal dose of H 2 O 2 (300 µM) for 3 h to induce oxidative stress 28,35 . The mRNA expression of CysLTR1 in polarized ARPE-19 cells was very low under basal conditions (p < 0.005, normalized to glucuronidase beta (GUSB)) and was below the detection limit in 9 of 15 samples (Fig. 2a). However, H 2 O 2 treatment significantly increased CysLTR1 expression levels (mean = 0.023, normalized to GUSB) (Fig. 2a). Western blot analysis confirmed that CysLTR1 protein expression was stable under basal conditions and was not increased upon H 2 O 2 treatment for 3 h (Fig. 2b,c, Supplementary Fig. 2a,b). The mRNA expression of the two other CysLT receptors, CysLTR 2 and GPR17, in ARPE-19 cells was not detectable by qPCR (n = 4, data not shown).

Rhythmic expression of LC3B and UPR transcription factors. Basal autophagic flux in RPE cells
exhibits a circadian rhythm in vivo 36 . Furthermore, ex vivo cultured ARPE-19 cells express clock and phagocytosis genes in a rhythmic manner 37 . An excellent short summary of the circadian rhythm and clock genes can be found in a review by Cox and Tagahashi 38 . Based on these findings, the potential rhythmic regulation of the autophagic process in polarized ARPE-19 cells was investigated. The LC3-II (lipidated LC3) level is used as a marker of autophagic activity 39,40 . Therefore, LC3-I (unlipidated LC3) and LC3-II levels were analyzed by western blotting (Fig. 3, Supplementary Fig. 2c,d). Cells were not synchronized by serum shock to avoid unknown effects on the intrinsic regulation of basal autophagic activity. Nevertheless, synchronization of three different cell batches was achieved by simultaneous thawing, medium exchange, splitting and polarization. As not all cell batches (n = 9) were synchronized, the data obtained from all time-course experiments of LC3-I and LC3-II expression were combined retrospectively (Fig. 3b). For each time-course experiment, cells were seeded in 12-well plates and expanded until cell confluence was reached. Afterwards, the cells were polarized for 7 days under low-serum conditions. Thus, all generated monolayers used for a single experiment were synchronized ( Supplementary Fig. 4). Polarized cells were harvested every 4 or 8 h within a time period of 20-52 h. There was a clear time-dependent alteration of LC3-I and LC3-II protein expression in ARPE-19 cells (Fig. 3a-d). The combined data indicated an autophagic flux period of approximately ≥ 48-56 h (Fig. 3a-d). LC3-I and LC3-II www.nature.com/scientificreports/ protein levels were regulated concordantly over time ( Fig. 3a-d). The expression of the autophagy-related genes BECN1, MAP1LC3B and SQSTM1 at the time points exhibiting the highest and lowest expression levels within 40 h were significantly different ( Fig. 3e-g). The expression levels of the UPR transcription factors ATF4, ATF6 and spliced XBP1 (XBP1s), similar to those of autophagic genes, changed within 40 h, but the difference in expression at the time points exhibiting the highest and lowest levels was only significant for ATF6 ( Fig. 3h-j). The reference gene GUSB exhibited stable expression over 40 h and was unaffected by rhythmic autophagic activity ( Supplementary Fig. 3).
Increased autophagic activity upon leukotriene receptor inhibition. As polarized ARPE-19 cells express the CysLTR1 protein and because H 2 O 2 , a known autophagy modulator 28 , increases CysLTR1 mRNA levels, cells were left untreated (3 h) or treated with 100 nM CysLTR1 receptor antagonist zafirlukast (ZK), a nonlethal dose of H 2 O 2 (300 µM) or a combination of both for 3 h. Rhythmic regulation of LC3-I and LC3-II (see Fig. 3) was not considered for cell treatments; nevertheless, the cells exposed to each treatment were compared to a time-matched control sample (all generated monolayers used for a single experiment were synchronized, Supplementary Fig. 4). To prevent LC3-II degradation, the cells were cotreated with 10 µg/ml E64d and 10 µg/ml pepstatin A. Increased levels of lipidated LC3-II in the presence of lysosomal degradation inhibitors indicate amplified autophagic flux 39,40 . LC3-I and LC3-II protein expression levels were analyzed by western blotting (Fig. 4, Supplementary Fig. 6). H 2 O 2 treatment of ARPE-19 cells did not affect LC3-I levels, but LC3-II levels were significantly increased in ARPE-19 cells (Fig. 4a-c) compared to the untreated control sample. Compared to the untreated controls, ZK treatment significantly increased LC3-II levels in unchallenged (no oxidative stress) and 300 µM H 2 O 2 -treated cells (Fig. 4b,c) but not on LC3-I levels (Fig. 4a, c). The optimal ZK concentration for cell treatment was determined by a small preliminary experiment, it was found that 100 nM ZK effectively activated autophagic activity through CysLTR1 inhibition ( Supplementary Fig. 5). Additionally, the effect of CysLTR1 inhibition by ZK (100 nM) treatment on caspase 3/7 activity in untreated and 300 µM H 2 O 2 -treated ARPE-19 cells was investigated after a period of 3, 6, 24 or 48 h. Interestingly, caspase 3/7 activity in polarized ARPE-19 cells was significantly increased after 24 and 48 h under basal and oxidative stress (300 µM H 2 O 2 ) conditions. CysLTR1 inhibition led to an overall significant reduction in caspase 3/7 activity under basal cell conditions but not in cells exposed to oxidative stress (300 µM H 2 O 2 ) (Supplementary Fig. 7a,b).
As H 2 O 2 treatment induced CysLTR1 gene expression and CysLTR1 inhibition during challenge with H 2 O 2 for 3 h further increased autophagic flux, the impact of CysLTR1 inhibition on autophagic activity following short and long periods of H 2 O 2 exposure was determined. Polarized ARPE-19 cells were challenged with 300 µM H 2 O 2 for 3 or 24 h. Afterwards, the cells were incubated in the absence of H 2 O 2 for an additional 3 h without or with 10 or 100 nM ZK. ZK treatment without prior H 2 O 2 challenge served as a control "no oxidative stress" condition. ZK (100 nM) led to a significant increase in LC3-II but not LC3-I in unchallenged ARPE-19 cells compared to untreated control samples (Fig. 7a-c, Supplementary Fig. 9). The lower concentration of ZK (10 nM) had no significant effect on LC3-I and LC3-II expression (Fig. 7a-c). Interestingly, compared to control treatment or treatment with 300 µM H 2 O 2 for 3 h followed by culture in fresh medium without treatment for  (Fig. 7a-c). The lower ZK concentration (10 nM) had no effect on LC3-I and LC3-II expression following a short period of H 2 O 2 exposure (Fig. 7a-c). Compared to control treatment with 300 µM H 2 O 2 for 24 h followed by culture in fresh medium without treatment for 3 h, CysLTR1 inhibition by 10 nM ZK but not 100 nM ZK led to a significant decrease in LC3-I levels and a trend toward reduced LC3-II levels ( Fig. 7a-c).

Discussion
Autophagy dysfunction and cellular stress increase with age and are associated with various diseases, including neurodegenerative disorders 43,44 . RPE cells are exposed to high levels of phagocytosis and oxidative stress; hence, functional autophagy is essential for ensuring the integrity of the RPE 22 . Accordingly, impaired autophagy in RPE cells is a main characteristic of dry AMD, the most common cause of blindness in individuals over 60 years of age in Western countries 26,27 . Thus, the interplay between ER stress and autophagy is a highly relevant topic of current research 45 . CysLTR1 has the potential to be an essential modulator of the autophagic process 12 and to play a crucial role during ER stress 6 ; therefore, we hypothesize that there is a strong association between CysLTR1 activity, ER stress and autophagy in RPE cells. ARPE-19 cells are able to generate polarized monolayers to mimic a functional RPE in vitro 46 . It should be noted that ARPE-19 cells are of male origin and that the CYSLTR1 gene is located on the X chromosome; thus, a sex difference cannot be excluded and should be considered in future experiments using primary RPE cells. Although rather low CysLTR1 mRNA levels were found under basal conditions, stable protein expression was observed in ARPE-19 cells by IF and western blot analysis. Interestingly, CysLTR1 was located basolaterally in the polarized monolayer in vitro and was localized to the microtubules of the cells. In vivo, the microtubules of RPE cells are involved in POS phagocytosis and are important for the intracellular transport of (auto)phagosomes [47][48][49] . A correlation between leukotriene activity in RPE cells of Xenopus laevis and POS phagocytosis was already postulated in 1989 by Birkle et al. 4 . Taken together, these data indicate the involvement of CysLTR1 in phagocytic processes or phagosome transport. In individual cells CysLTR1 was localized intracellularly near the nucleus, as reported by Dvash et al. 6 . Whether the membrane-bound or intracellular localization of CysLTR1 indicates that the receptor plays a role in distinct mechanisms or has a functional dependency should be clarified in  www.nature.com/scientificreports/ explain the observed discrepancy between mRNA and protein levels following H 2 O 2 treatment. Nevertheless, the induction of CysLTR1 mRNA upon H 2 O 2 treatment highlights the potential importance of the leukotriene system in ER/oxidative stress and autophagy. Numerous cellular functions are regulated by a daily rhythm, known as the circadian rhythm. Hence, the mammalian retina contains a circadian clock system, and following POS shedding, the phagocytosis and recycling of POSs by RPE cells is time-dependent 50 . Thus, basal autophagy in murine RPE cells is rhythmically regulated, as visualized by peaks of LC3-II levels and the number of autophagosomes over a period of time 36 . Interestingly, ARPE-19 cells exhibit rhythmic expression of clock and phagocytosis genes in vitro 37 . In line with this previous study, rhythmic regulation of autophagic activity in the absence of external stimuli was observed in the present study. Furthermore, autophagic genes and UPR transcription factors are expressed in a comparable rhythm. It was previously reported that basal UPR activity in the mouse liver is cyclic 42 . The interplay between autophagy and UPR activity is the focus of research, especially research on neurodegenerative diseases 51 . We observed positive correlations between almost all investigated autophagy-and UPR-related genes, which further highlights a fundamental dependency of these two systems in RPE cells. Milicevic et al. reported rhythmic expression of clock and phagocytosis genes lasting < 24 h, which is different from the rhythm of ≥ 48-56 h observed in the present study. This discrepancy may be explained by the serum shock used to synchronize the cell rhythm applied by Milicevic et al. 37 . Nevertheless, ARPE-19 cells follow an intrinsic rhythmic cell metabolism that is independent of external stimuli. Thus, dynamic autophagic activity should be considered when working with ARPE-19 cells in vitro. Whether the membrane-bound or intracellular localization of CysLTR1 is associated with the rhythmic regulation of phagocytosis, phagosome transport, autophagy or basal UPR activity is of interest for future studies. In particular, as basal mRNA levels of ALOX5 and CyLTR1 are positively correlated with rhythmically regulated expression of autophagy-and UPR-related genes, this correlation indicates a time-dependent activity of the leukotriene system in human RPE cells. www.nature.com/scientificreports/ Although different batches of ARPE-19 at different passage numbers (P9-15) were treated at different time points within the autophagic rhythm in the present study, LC3-II protein levels were consistently higher in ZK-treated cells than in time-matched controls. An increase in LC3-II levels in the presence of lysosomal degradation inhibitors is indicative of increased autophagic flux 39,40 . Therefore, the increase in LC3-II protein levels following ZK treatment in the current experiments indicates that CysLTR1 is involved in autophagic flux regulation. However, due to variations in the gene expression of autophagy-and UPR-related genes during the intrinsic autophagic rhythm, two groups roughly representing two phases within the autophagic rhythm were defined based on the expression of these genes and analyzed. This biphasic analysis revealed differences in gene expression following inhibition of CysLTR1 signaling. In cells expressing low levels of autophagy-/UPR-related genes, ZK treatment did not significantly affect the expression of autophagic genes. In contrast, CysLTR1 inhibition in cells expressing high levels of autophagy-/UPR-related genes significantly reduced the expression of various autophagy-/UPR-related genes. This reduction could be a consequence of a negative feedback loop, as CysLTR1 was inhibited in the presence of high levels of basal autophagy and UPR-related gene expression 52,53 . Overall, CysLTR1 inhibition modulated the expression of autophagy-and UPR-related genes and completely abolished the correlation between ALOX5 expression and the aforementioned autophagy-/UPR-related gene expression, suggesting self-regulation of CysLTR1 signaling. The rhythm-independent regulation of LC3-II and the rhythm-dependent modulation of gene expression by CysLTR1 inhibition could be explained by a direct effect on autophagosome formation and additional modulation of transcriptional processes. Recently, PERK and ATF4 were reported to modulate autophagy in two distinct ways, namely, at the transcriptional level and through direct regulation of autophagosome formation at the protein level 7 . The combination of these data with our findings suggests a potential interplay between CysLTR1 signaling and the PERK-ATF4 axis that represents a promising target for future studies.
In the present study, H 2 O 2 was used to mimic cellular oxidative stress 28,35,54,55 . Cotreatment with H 2 O 2 and ZK resulted in increased autophagic activity; however, treatment of ARPE-19 cells with ZK following prior exposure to H 2 O 2 did not increase autophagy but induced a trend toward reduced autophagic activity. CysLTR1 inhibition significantly reduced caspase 3/7 activity in unstressed cells, which highlights the relationship between CysLTR1 activity and apoptosis under basal cell conditions. Interestingly, under oxidative stress, CysLTR1 inhibition had no effect on caspase 3/7 activity. These data likely indicate that CysLTR1 signaling is endogenously regulated upon/during an oxidative stress response and that CysLTR1 plays a dual role under basal and oxidative stress conditions, which should be investigated in future studies.
In summary, ARPE-19 cells exhibited rhythmic basal autophagic activity in the absence of external stimuli. Furthermore, CysLTR1 was mostly expressed basolaterally in RPE cells, and inhibition of CysLTR1 resulted in an increase in autophagic activity. The expression of leukotriene-related genes was correlated with that of autophagy-and UPR-related genes under basal and cellular stress conditions, and oxidative stress drastically increased CysLTR1 mRNA levels and altered the potential of CysLTR1 inhibition to regulate autophagic activity. CysLTR1 inhibition by ZK or Montelukast has been used to treat asthma since the nineties 3 , but the potential of CysLTR1 as a pharmacological target for the treatment of neurodegenerative disorders, such as Alzheimer's disease (AD), has become a topic of interest because the leukotriene system targets AD pathologies on multiple levels 56 . Similarly, dry AMD is a multifactorial neurodegenerative disease with limited treatment possibilities 57 . Most interestingly, the leukotriene system was recently described as a potential target for therapeutic approaches for the exudative form of late AMD (wet AMD), as diverse leukotriene inhibitors reduce choroidal neovascularization in murine wet AMD models 58 . Because inhibition of the leukotriene system also induces autophagy, inhibits cell death and exerts immunomodulatory effects, it represents a potential pharmaceutical target for the treatment of not only wet but also dry AMD 1,5,6,12,58 . Therefore, gaining a fundamental understanding of physiological and pathophysiological leukotriene-dependent effects in AMD-associated cellular systems seems to be a promising strategy for the development of new therapeutic strategies. www.nature.com/scientificreports/ were separated into two groups of low and high autophagy-/UPR-related gene expression roughly representing roughly two phases of the autophagic rhythm based on the expression of a combination of seven genes (see below). The median of the mean expression of BECN1, MAP1LC3B, SQSTM1, mTOR, ATF4, ATF6 and XBP1 normalized to the expression of GUSB was used for group classification (< median = low basal autophagy-/UPRrelated gene expression, ≥ median = high basal autophagy-/UPR-related gene expression).

Western blot analysis.
Polarized ARPE-19 monolayers were generated and treated in 12-well plates. Proteins were isolated with RIPA lysis buffer (Santa Cruz Biotechnology, USA), separated by SDS page using 10% (receptor analysis) or AnykD (LC3B analysis) TGX stain-free gels (Bio-Rad) and transferred to a PVDF membrane (Amersham Hybond, GE Healthcare, IL, USA) by wet electroblotting (Bio-Rad). Protein expression was normalized to the amount of total loaded protein (Image Lab 6.0.1, Bio-Rad). For detection of CysLTR1, the membrane was blocked with EveryBlot (Bio-Rad) for 10 min at room temperature and incubated overnight with an anti-CysLT1 antibody (ab151484, Abcam) diluted 1:500 in EveryBlot blocking solution at room temperature. CysLTR1 was visualized using an anti-rabbit antibody conjugated to HRP (Agilent, CA, USA) and Clarity Western ECL Substrate (Bio-Rad). For detection of LC3-I and LC3-II, the membrane was blocked with 5% milk powder for 1 h at room temperature and incubated overnight with a recombinant anti-LC3B antibody [EPR18709] (ab192890, Abcam, UK) diluted 1:1000 in blocking solution at 4 °C. LC3B was visualized using an anti-rabbit antibody conjugated to HRP and Clarity Western ECL Substrate. Chemiluminescence was detected using the ChemiDoc XRS + system (Bio-Rad). Full-length blots are presented in the supplementary material.
Documentation. IF images were taken by using a confocal laser-scanning unit (Axio Observer Z1 attached to LSM710, Zeiss, Germany; 40 × oil immersion objective lens, numerical aperture 1.30, Zeiss). The single optical section mode of the confocal microscope was used for image acquisition to document pixel colocalization of different channels with appropriate filter settings for Alexa Fluor 488 (495 nm excitation), Alexa Fluor 555 (555 nm excitation) and DAPI (345 nm excitation).  GUSB  AGC GAG TAT GGA GCA GAA AC  TGA TCC AGA CCC AGA TGG TA   MAP1LC3B  CGT CGG AGA AGA CCT TCA AG  CTG CTT CTC ACC CTT GTA TCG   SQSTM1  TGA AAC ACG GAC ACT TCG G  TCA GGA AAT TCA CAC TCG GATC   BECN1  ACG AGT GTC AGA ACT ACA AACG TTT CCA CAT CTT CCA GCT CC   mTOR  TTC GTG CCT GTC TGA TTC TC  ATC CCG ATT CAT GCC CTT C   ATF4  ATG GGT TCT CCA GCG ACA AG  GGC ATC CAA GTC GAA CTC CT   ATF6  GCC GCC GTC CCA GAT ATT A  CCG AGT TCA GCA AAG AGA GC   DDIT3  GTT AAA GAT GAG CGG GTG GC  GCT TTC AGG TGT GGT GAT GTATG   EIF2AK3  ACG ATG AGA CAG AGT TGC GA  TGC TAA GGC TGG ATG ACA CC   ERN1  CGG CCT CGG GAT TTT TGG AA  TGC CAT CAT TAG GAT CTG GGAG   HSPA5  TTG GAG GTG GGC AAA CAA AG  GTC TTT GGT TGC TTG GCG TT   PPP1R15A  GAC TGC AAA GGC GGC TCA