biomolecules

Article The Interplay between Retinal Pathways of Cholesterol Output and Its Effects on Mouse Retina

Alexey M. Petrov , Artem A. Astafev, Natalia Mast, Aicha Saadane, Nicole El-Darzi and Irina A. Pikuleva * Department of Ophthalmology and Visual Sciences, Case Western Reserve University, Cleveland, OH 44106, USA; [email protected] (A.M.P.); [email protected] (A.A.A.); [email protected] (N.M.); [email protected] (A.S.); [email protected] (N.E.-D.) * Correspondence: [email protected]; Tel.: +1-216-368-3823



Received: 13 November 2019; Accepted: 10 December 2019; Published: 12 December 2019


Abstract: In mammalian retina, cholesterol excess is mainly metabolized to oxysterols by cytochromes P450 27A1 (CYP27A1) and 46A1 (CYP46A1) or removed on lipoprotein particles containing apolipoprotein E (APOE). In contrast, esterification by sterol-O-acyltransferase 1 (SOAT) plays only a minor role in this process. Accordingly, retinal cholesterol levels are unchanged / / / / in Soat1− − mice but are increased in Cyp27a1− −Cyp46a1− − and Apoe− − mice. Herein, we / / / / / / characterized Cyp27a1− −Cyp46a1− −Soat1− − and Cyp27a1− −Cyp46a1− −Apoe− − mice. In the former, retinal cholesterol levels, anatomical gross structure, and vasculature were normal, yet the / / / electroretinographic responses were impaired. Conversely, in Cyp27a1− −Cyp46a1− −Apoe− − mice, retinal cholesterol levels were increased while anatomical structure and vasculature were unaffected with only male mice showing a decrease in electroretinographic responses. Sterol profiling, qRT-PCR, proteomics, and transmission electron microscopy mapped potential compensatory mechanisms / / / / / / in the Cyp27a1− −Cyp46a1− −Soat1− − and Cyp27a1− −Cyp46a1− −Apoe− − retina. These included decreased cholesterol biosynthesis along with enhanced formation of intra- and extracellular vesicles, possibly a reserve mechanism for lowering retinal cholesterol. In addition, there was altered / / / abundance of proteins in Cyp27a1− −Cyp46a1− −Soat1− − mice that can affect photoreceptor function, survival, and retinal energy homeostasis (glucose and fatty acid metabolism). Therefore, the levels of retinal cholesterol do not seem to predict retinal abnormalities, and it is rather the network of compensatory mechanisms that appears to determine retinal phenotype.

Keywords: CYP27A1; CYP46A1; SOAT1; APOE; cholesterol; oxysterols; vesicular traffic; glucose metabolism; fatty acid oxidation; synaptic function; inflammation; photoreceptors; cilia; age-related macular degeneration

1. Introduction Cholesterol is abundant in mammalian retina, a multilayered light-sensitive tissue lining the back of the eye. The neurosensory retina is underlined by the retinal pigment epithelium (RPE) and then choroid, which nourish the retina. Evidence suggests that retinal cholesterol dyshomeostasis is linked to age-related macular degeneration (AMD), a cause of vision loss in ~200 million older adults worldwide [1]. First, in aging humans, cholesterol accumulates on either side of the RPE and represents a significant component of drusen and subretinal drusenoid deposits [2–6], the AMD hallmarks. Second, genetic studies identified several cholesterol-related genes (CETP, ABCA1, LIPC, LPL, VLDLR, LRP6, and APOE, which encodes apolipoprotein E) as risk factors for AMD and suggested that these gene effects are independent of plasma lipid profiles [7–11]. Lastly, the two prospective clinical trials of statins, drugs prescribed for reduction of plasma cholesterol, showed a decreased risk

Biomolecules 2019, 9, 867; doi:10.3390/biom9120867 www.mdpi.com/journal/biomolecules Biomolecules 2019, 9, 867 2 of 22

of progressionBiomolecules 2019 of, non-advanced 9, x AMD in specific patient populations [12,13]. Moreover, evidence2 of 22 was obtained that intensive statin therapy may even remove drusen in early stages of AMD [13]. These resultsdecreased underscore risk theof progression importance of of non-advanced understanding AM theD in mechanisms specific patient by which populations retinal [12,13]. cholesterol is managed.Moreover, evidence was obtained that intensive statin therapy may even remove drusen in early Thestages major of AMD pathways [13]. These of results cholesterol underscore input the andimportance output ofin understanding the retina seemthe mechanisms to be identified by (Figurewhich1)[14 retinal–20]. cholesterol Input is dependent is managed. on cholesterol biosynthesis in di fferent retinal cells and uptake The major pathways of cholesterol input and output in the retina seem to be identified (Figure from the systemic circulation; the latter is likely from the choriocapillaris, as cholesterol delivery 1) [14−20]. Input is dependent on cholesterol biosynthesis in different retinal cells and uptake from fromthe the systemic inner retinal circulation; blood the vessels latter is islikely currently from the speculative choriocapillaris, [20]. as Output cholesterol includes delivery metabolism from the to oxysterolsinner retinal by the blood cytochrome vessels is P450 currently enzymes speculative CYP27A1 [20]. and Output CYP46A1, includes the metabolism photoreceptor to oxysterols phagocytosis, by and etheffl uxcytochrome to lipoprotein P450 enzymes particles CYP27A1 (LP), which and CYP46A1 circulate, the in photoreceptor the intraretinal phag space.ocytosis, LP and then efflux deliver cholesterolto lipoprotein to different particles retinal (LP), cells which as well circulate as the in systemic the intraretinal circulation, space. probably LP then deliver through cholesterol the RPE [to14 –20]. Intracellulardifferent esterificationretinal cells as by well sterol-O-acyltransferase as the systemic circulation, 1 (SOAT1, probably named through previously the RPE as[14 acyl-CoA−20]. cholesterolIntracellular acyltransferase esterification 1 orby ACAT1,sterol-O-acyltransferase Text S1) represents 1 (SOAT1, themeans namedof previously cholesterol as storageacyl-CoA and a formcholesterol of cholesterol acyltransferase output. Cholesterol 1 or ACAT1, esters Text form S1) represents lipid droplets, the means cellular of cholesterol organelles, storage which and provide a a uniqueform separation of cholesterol of the output. aqueous Cholesterol and organic esters form phases lipid of droplets, the cell [cellular21]. Cholesterol organelles, esterswhich areprovide found in a unique separation of the aqueous and organic phases of the cell [21]. Cholesterol esters are found the photoreceptors outer segments, drusen, and Bruch’s membrane [2,3,22]. in the photoreceptors outer segments, drusen, and Bruch’s membrane [2,3,22].

FigureFigure 1. Schematic1. Schematic representation representation ofof thethe known chol cholesterol-relatedesterol-related pathways pathways contributing contributing to to cholesterolcholesterol homeostasis homeostasis in in the the retina. retina. Local Local biosynthesisbiosynthesis accounts accounts for for the the majority majority (~72%) (~72%) of retinal of retinal cholesterolcholesterol input, input, at least at least in in mice, mice, with with the the remaining remaining cholesterol cholesterol (~28%) (~28%) being being taken taken up by up the by retina the retina fromfrom the systemicthe systemic circulation circulation [ 23[23].]. A small small fraction fraction of ofretinal retinal cholesterol cholesterol (up to (up 15%) to is 15%) esterified is esterified by by SOAT1SOAT1 [24[24,25],25] andand isis storedstored in the form of of lipid lipid droplets. droplets. Retinal Retinal choles cholesterolterol output output includes includes metabolismmetabolism by CYP27A1by CYP27A1 and and CYP46A1, CYP46A1, the the photoreceptorphotoreceptor phagocytosis, phagocytosis, and and removal removal on lipoprotein on lipoprotein particlesparticles containing containing APOE. APOE. The The relative relative quantitative quantitative contribution of of these these three three pathways pathways to the to total the total cholesterol output in the retina is currently unknown. 27HC and 24HC are 27-hydroxycholesterol and cholesterol output in the retina is currently unknown. 27HC and 24HC are 27-hydroxycholesterol and 24-hydroxycholesterol, respectively, the products of CYP27A1 and CYP46A1; EC, esterified 24-hydroxycholesterol, respectively, the products of CYP27A1 and CYP46A1; EC, esterified cholesterol, cholesterol, the product of SOAT1; LP, lipoprotein particles. the product of SOAT1; LP, lipoprotein particles. To ascertain retinal significance of cholesterol removal by metabolism and LP, we previously Tocomprehensively ascertain retinal characterized significance the retina of cholesterol of Cyp27a1 removal−/−Cyp46a1 by−/− metabolism mice and mice and lacking LP, we APOE, previously a / / comprehensivelymajor apolipoprotein characterized in the retina the retina[22,26,27]. of Cyp27a1In the former,− −Cyp46a1 total cholesterol− − mice was and increased mice lacking up to 2- APOE, a majorfold, apolipoprotein and all of the cholesterol in the retina surplus [22,26 was,27]. converted In the former, to cholesterol total cholesterol esters, which was were increased mainly up to 2-fold,accumulated and all of in the the cholesterolphotoreceptor surplus outer segments was converted [22]. In addition, to cholesterol there we esters,re focal which accumulations were mainly accumulatedof free (unesterified) in the photoreceptor cholesterol outer in the segments Cyp27a1 [−/22−Cyp46a1]. In addition,−/− retina, there impairments were focal in accumulationsthe retinal function (as assessed by electroretinographic recordings,/ ERG),/ and pathologic changes in retinal of free (unesterified) cholesterol in the Cyp27a1− −Cyp46a1− − retina, impairments in the retinal −/− functionblood (as vessels assessed [26]. byIn Apoe electroretinographicmice, retinal cholesterol recordings, was also ERG), increased and pathologic (up to 3-fold). changes However, in retinal cholesterol excess remained/ mainly unesterified and did not form focal deposits as it was mostly blood vessels [26]. In Apoe− − mice, retinal cholesterol was also increased (up to 3-fold). However, present outside retinal cells on the LP circulating in the intraretinal space [27]. Remarkably, the cholesterol excess remained mainly unesterified and did not form focal deposits as it was mostly Apoe−/− retina did not have any abnormalities of clinical relevance in structure or in blood vessels, / presentlikely outside because retinal of the cells compensatory on the LP circulating responses intriggered the intraretinal by APOE space absence. [27]. These Remarkably, responses the were Apoe − − retina did not have any abnormalities of clinical relevance in structure or in blood vessels, likely Biomolecules 2019, 9, 867 3 of 22 because of the compensatory responses triggered by APOE absence. These responses were identified and included changes in the retinal cholesterol biosynthesis and abundance of proteins involved in cellular cytoskeleton maintenance, vesicular traffic, pro- and anti-inflammatory signaling, as well as / / / iron homeostasis [27]. We also generated Cyp27a1− −Cyp46a1− −Soat1− − mice and established that retinal cholesterol esterification is mediated by SOAT1 [22]. While this genotype had normal levels of retinal cholesterol, signs of retinal degeneration were seen in 6-month old animals—a shortened length of the photoreceptor outer segments as well as increased apoptosis in the photoreceptors [22]. The purpose of the present work was to gain insight into the interplay between retinal pathways / / / of cholesterol output by finishing ophthalmic characterizations of Cyp27a1− −Cyp46a1− −Soat1− − / / / mice and generating as well as characterizing Cyp27a1− −Cyp46a1− −Apoe− − mice. We obtained evidence supporting the activation of a previously unrecognized vesicle-mediated mechanism for retinal cholesterol removal in both triple knockouts and suggest that this mechanism serves as a reserve means of retinal cholesterol maintenance. In addition, we mapped retinal proteins and processes in / / /– Cyp27a1− − Cyp46a1− − Soat1− mice that could be linked to retinal cholesterol-related pathways and be important for retinal structure, function, and metabolism.

2. Materials and Methods

2.1. Animals Both female and male mice were used, which were 6–12 month old. All animals were free rd8 / of the Crbl mutation. C57BL/6J and Apoe− − mice on the C57BL/6J background were from the / Jackson Laboratory (catalog nos. 000664 and 002052, respectively). Soat1− − mice were from the laboratory of Dr. T.Y. Chang (Dartmouth College, Hanover, NH, USA) [28] and were rederived / / on the C57BL/6J background. Cyp27a1− − Cyp46a1− − mice were generated previously in this / / laboratory [26] and were on a mixed strain background (C57BL/6J; 129S6/SvEv). Cyp27a1− −Cyp46a1− − / / / / / mice were crossed with Soat1− − or Apoe− − mice to generate Cyp27a1− −Cyp46a1− −Soat1− − and / / / Cyp27a1− −Cyp46a1− −Apoe− − mice, respectively. All animals were maintained on a standard 12-h light (~10 lux)-dark cycle and were provided regular rodent chow and water ad libitum. All animal experiments were approved by Case Western Reserve University’s IACUC and conformed to recommendations of the American Veterinary Association Panel on Euthanasia (ethical approval number 2014-0154).

2.2. In Vivo Characterizations Ultra-high resolution spectral-domain optical coherence tomography (SD-OCT), fluorescein angiography (FA), and ERG were carried out as described [26,29]. We used an 840HHP SD-OCT system (Bioptigen, Durham, NC, USA), a scanning laser ophthalmoscope (Spectralis HRA+OCT, Heidelberg Engineering, Heidelberg, Germany), and an Electrophysiological System UTAS E-3000 (LKC Technologies Inc., Gaithersburg, MD, USA), respectively. Images for FA were acquired after a bolus (0.1 mL) intraperitoneal injection of 1.0% sodium fluorescein (Akorn, Lake Forest, IL, USA) in phosphate-buffered saline.

2.3. Retinal Sterol Content and Gene Expression Retinal isolation and processing for sterol quantifications were as described [25,26]. Sterol content (total and unesterified, the former representing a sum of esterified and unesterified cholesterol) was measured in individual retinas by isotope dilution gas chromatography-mass spectroscopy. Deuterated sterol analogs were used as internal standards. For the quantifications of gene expression, retinas from 4–8 mice per genotype were combined, and total RNA was isolated by the TRIzol Reagent (Life Technologies, Grand Island, NY, USA) according to the manufacturer’s instructions. Quantitative RT-PCRs (qRT-PCR) were performed as described [26,30] in triplicate on an LightCycler 480 instrument (Roche Life Science, Penzberg, Germany) using cDNA (obtained from total RNA by SuperScript Biomolecules 2019, 9, 867 4 of 22

Reverse Transcriptase, Invitrogen Carlsbad, CA, USA), a pair of gene-specific primers (Table S1), and a FastStart Universal SYBR Green Master (Rox) (Roche Life Science); β-actin served as a reference gene. ∆∆Ct Changes in relative mRNA levels were calculated by the 2− method [31].

2.4. Retinal Proteomics The label-free approach was used as described [27]. Briefly, three biological replicates per genotype / / / (Cyp27a1− −Cyp46a1− −Soat1− − or C57BL/6J, 6–8 month old male mice), each containing two retinas from two different mice, were frozen, cryo-pulverized, and lysed with 3% SDS. Samples were then cleaned of detergent, placed into 50 mM Tris-HCl, pH 8.0, containing 8 M urea, and alkylated with iodoacetaminde. The 8 M urea was next diluted to 4 M with 50 mM Tris-HCl, pH 8.0, and samples were concentrated. Proteins (20 µg of total protein) were sequentially digested with lysyl endopeptidase (Wako Chemicals, Richmond, VA, USA) and trypsin, and an equal amount of internal standard (Pierce Retention Time Calibration Mixture 88321, Thermo Fisher Scientific, Waltham, MA, USA) was added to each sample. Digested proteins (400 ng per sample) were analyzed by LC/MS/MS carried out on a LTQ-Orbitrap Velos mass spectrometer (Thermo Fisher Scientific) equipped with a nanoAcquityTM ultra-high-pressure liquid chromatography system (Waters). Blank injections were run after each sample to minimize carryover between samples. Full-scan MS spectra (m/z 380–1800) were acquired at a resolution of 60,000 at m/z 400 followed by 20 data-dependent MS/MS scans generated in the ion-trap detector by collision-induced dissociation of the peptide ions. The retention time, peak intensity, and mass accuracy of five peptides from the internal standard were monitored in all samples. The LC/MS/MS raw data were acquired using the Xcalibur software (Thermo Fisher Scientific) and imported into Rosetta Elucidator™ (RosettaCommons, Burlington, VT, USA). The peak list (.dta) files were searched by Mascot (Matrix Science) against the mouse Uniprot database (538,585 sequences). The search results were imported back into Elucidator, and differences in relative protein abundance were calculated by the PEAKS software (Bioinformatics Solutions Inc, Waterloo, ON, Canada) based on unique peptides [32]. Proteins with non-significant changes (p 0.05) in abundance between the ≥ genotypes were excluded from the subsequent proteomics analysis, as are the proteins with less than 2 unique peptides/protein and a 1.5-fold change in the relative abundance, even if this change was significant. Protein grouping was based on the protein function described in the literature (Text S1).

2.5. Transmission Electron Microscopy This was carried out using a 1200EX transmission electron microscope (JEOL Ltd. Tokyo, Japan). Tissue fixation was as described [27,33] and included sequential incubations in 0.1 M Na cacodylate buffer, pH 7.4, containing first 3% glutaraldehyde and then 1% OsO4, followed by incubations with 1% tannic acid in 0.05 M Na cacodylate, pH 7.4, and 1% para-phenylenediamine in 70% ethanol.

2.6. Statistical Analyses All images are representative of studies in three to five animals per genotype unless otherwise indicated. All quantitative data represent the mean SD or the mean SEM; the sample size is ± ± indicated in each figure or figure legend. Data were analyzed either by a two-tailed, unpaired Student’s t-test, two-way repeated measures ANOVA, two-way ANOVA with Bonferroni correction, or two-way ANOVA with a Dunnett’s multiple comparisons test. The GraphPad Prism software (GraphPad, San Diego, CA, USA) was used. Statistical significance was defined as * p 0.05; ** p 0.01; *** p 0.001. ≤ ≤ ≤ 3. Results

3.1. In Vivo Characterizations

/ / / / / We examined 6- and 12- month old Cyp27a1− −Cyp46a1− −Soat1− − and Cyp27a1− −Cyp46a1− − / / Apoe− − by using SD-OCT, FA, and ERG. We also included studies of Soat1− − mice as their retinal phenotype has not been reported. None of the three genotypes appeared to have any gross abnormalities Biomolecules 2019, 9, 867 5 of 22

in theBiomolecules retinal structure 2019, 9, x as assessed by SD-OCT (Figure2). Also, the three genotypes did5 of not 22 have major abnormalities in the retinal blood vessels as evaluated by FA (Figure2). However, studies by abnormalities in the retinal structure as assessed by SD-OCT (Figure 2). Also, the three/ genotypes/ did / ERG revealed age-dependent changes in overall retinal function. In Cyp27a1− − Cyp46a1− − Soat1− − mice ofnot both have sexes,major abnormalities dark-adapted in (scotopic)the retinal blood ERGs vessels were as decreased evaluated atby 6 FA months, (Figure and2). However, this decrease studies by ERG revealed age-dependent changes in overall retinal function. In Cyp27a1−/− Cyp46a1−/− was more pronounced at 12 months (Figure3A). In addition, male Cyp27a1 / Cyp46a1 / Soat1 / Soat1−/− mice of both sexes, dark-adapted (scotopic) ERGs were decreased at −6 −months, and− − this − − mice haddecrease a decrease was more in pronounced the b- wave at of12 light-adaptedmonths (Figure (photopic)3A). In addition, ERGs, male also Cyp27a1 with− a/−Cyp46a1 progression−/− at / 1 year.Soat1 In Soat1−/− mice− − hadmice, a decrease ERGs in were the b- decreased wave of light-adapted as well, especially (photopic) in ERGs, 12-month also with old males,a progression suggesting / that lackat 1 ofyear. the In SOAT1-mediated Soat1−/− mice, ERGs cholesterolwere decreased esterification as well, especi canally a ffinect 12-month retinal old function. males, suggesting In Cyp27a1 − − / / −/− Cyp46a1that− lack−Apoe of the− − SOAT1-mediatedmice, the ERGs ofcholesterol female animals esterification were can essentially affect retinal una fffunction.ected, and In Cyp27a1 only male mice had statisticallyCyp46a1−/−Apoe significant−/− mice, the reductions ERGs of female in both animals scotopic were and essentially photopic unaffected, ERGs at and 12 monthsonly male (Figure mice 3C). Thus, hadin vivo statisticallyevaluations significant revealed reductions that the in both major scot eopicffect and of the photopic two triple ERGs geneat 12 months ablations (Figure was 3C). on retinal Thus, in vivo evaluations revealed that the major effect of the two triple gene ablations was on retinal function, but not on retinal structure and vasculature. function, but not on retinal structure and vasculature.

/ / / Figure 2. In vivo imaging of 6- and 12-month old C57BL/6J (A), Cyp27a1− −Cyp46a1− −Soat1− − Figure 2. In vivo imaging of 6- and 12-month old C57BL/6J (A), Cyp27a1−/−Cyp46a1−/−Soat1−/− (B), (B), Soat1 / (C), and Cyp27a1 / Cyp46a1 / Apoe / (D) mice. The three rows in each panel are Soat1−−/−− (C), and Cyp27a1−−/−Cyp46a1− −/−Apoe− − −/− (D−) −mice. The three rows in each panel are representativerepresentative images images (n =(n3–5 = 3–5 mice mice/group)/group) involvinginvolving SD-OCT SD-OCT (upper (upper panels) panels) and andFA (middle FA (middle and and lowerlower panels). panels). The The SD-OCT SD-OCT panels panels show show a a fundusfundus image image and and two two cross-sections cross-sections of the of retina the retina (from (from left toleft right), to right), the latterthe latter is a is Doppler a Doppler flow. flow. The The FAFA panels show show an an early, early, intermediate, intermediate, and late and stage late stage fundus fluorescence (from left to right) with the laser beam being focused either on the retinal pigment epithelium (RPE) or ganglion cell layer (GCL); the post-injection time is indicated in the top right corner of each image. No sex-based differences were detected, hence only images of male mice are shown. Biomolecules 2019, 9, x 6 of 22

fundus fluorescence (from left to right) with the laser beam being focused either on the retinal pigment epithelium (RPE) or ganglion cell layer (GCL); the post-injection time is indicated in the top right corner of each image. No sex-based differences were detected, hence only images of male mice are Biomoleculesshown.2019 , 9, 867 6 of 22

/ / / Figure 3. Electroretinography responses in 6- and 12-month old Cyp27a1− −Cyp46a1− −Soat1− − / / / / −/− −/− −/− (FigureA), Soat1 3. −Electroretinography− (B), and Cyp27a1 responses− −Cyp46a1 in− 6-−Apoe and −12-month− (C) mice, old which Cyp27a1 wereCyp46a1 comparedSoat1 with those (A), Soat1in C57BL−/− (B/6J), and mice. Cyp27a1 The results−/−Cyp46a1 are presented−/−Apoe−/− as(C) means mice, whichSEM were of the compared measurements with those in 5–19 in C57BL/6J animals. ± *mice.p 0.05; The results ** p 0.01; are presented *** p 0.001 as means as assessed ± SEM by of repeated the measurements measures two-way in 5–19 animals. ANOVA, * p where ≤ 0.05; flash ** p ≤ ≤ ≤ ≤intensity 0.01; *** was p ≤ used0.001 asas aassessed repeated by variable repeated and measures interaction two-way between ANOVA, sex and where genotype flash was intensity evaluated was usedfor significance. as a repeated variable and interaction between sex and genotype was evaluated for significance.

3.2. Retinal Retinal Sterol Content Three retinal retinal sterols sterols (cholesterol, (cholesterol, lathosterol, lathosterol, and and desmosterol) desmosterol) as well as well as serum as serum cholesterol cholesterol were / / / / / / / werequantified quantified in Cyp27a1 in Cyp27a1−/−Cyp46a1− −Cyp46a1−/−Soat1−−−/−Soat1, Cyp27a1− −, Cyp27a1−/−Cyp46a1− −−Cyp46a1/−Apoe−/−,− −andApoe Soat1− −,− and/− mice Soat1 and− − comparedmice and comparedwith those within other those knockout in other lines knockout determined lines by determined us previously by us[22,26,27]. previously In both [22 ,serum26,27]. / / / andIn both retina, serum andthe retina,levels the of levels total of totalcholesterol cholesterol in in Cyp27a1Cyp27a1−−−/−Cyp46a1−/−−Soat1Soat1−/− − and / / / / / Cyp27a1−/−Cyp46a1−Cyp46a1−/−−Apoe−Apoe−/− −mice− mice were were more more similar similar to to those those in Soat1Soat1−−/− and ApoeApoe−−/− mice, / / respectively, than in Cyp27a1− −Cyp46a1− − mice (Figure4A,B, Tables S2 and S3). Yet serum cholesterol did not seem to be an important contributor to retinal cholesterol as the sex differences in the levels / / / / of this sterol in Apoe− − and Cyp27a1− −Cyp46a1− −Apoe− − mice were not translated into similar sex differences in the levels of total retinal cholesterol. Further, retinal lathosterol and desmosterol Biomolecules 2019, 9, x 7 of 22 respectively, than in Cyp27a1−/−Cyp46a1−/− mice (Figure 4A,B, Tables S2 and S3). Yet serum cholesterol Biomoleculesdid not seem2019, 9to, 867be an important contributor to retinal cholesterol as the sex differences in the 7levels of 22 of this sterol in Apoe−/− and Cyp27a1−/−Cyp46a1−/−Apoe−/− mice were not translated into similar sex differences in the levels of total retinal cholesterol. Further, retinal lathosterol and desmosterol levels levels serve as markers for cholesterol biosynthesis in neurons and astrocytes, respectively [34,35]. serve as markers/ for cholesterol/ / biosynthesis in neurons and astrocytes, respectively [34,35]. In In Cyp27a1− −Cyp46a1− −Soat1− − mice, retinal lathosterol but not desmosterol content was decreased −/− −/− −/− Cyp27a1 Cyp46a1 Soat1 mice,/ retinal lathosterol/ but not/ desmosterol content was decreased as as compared to those in Cyp27a1− −Cyp46a1− − and Soat1− − mice (Figure4C,D, Tables S4 and S5). compared to those in Cyp27a1−/−Cyp46a1−/− and Soat1−/− mice (Figure 4C,D, Tables S4 and S5). This This compensatory decrease is consistent with the known sensitivity of cholesterol biosynthesis to compensatory decrease is consistent with the known sensitivity of cholesterol biosynthesis to changes changes in intracellular levels of unesterified cholesterol [36] and can be a reason for why total in intracellular levels of unesterified cholestero/ l [36] and/ can /be a reason for why total retinal retinal cholesterol was normal in Cyp27a1− −Cyp46a1− −Soat1− − mice. Similarly, a decrease in −/− −/− −/− cholesterol was normal in Cyp27a1 Cyp46a1 Soat1/ mice. Similarly,/ / a decrease in retinal retinal cholesterol biosynthesis was observed in Cyp27a1− −Cyp46a1− −Apoe− − mice, whose levels of cholesterol biosynthesis was observed in Cyp27a1−/−Cyp46a1−/−Apoe−/− mice, whose levels of both both retinal lathosterol and desmosterol were approximately the mean between the sterol levels in −/− retinal/ lathosterol and /desmosterol/ were approximately the mean between the sterol levels in Apoe Apoe− − and Cyp27a1− −Cyp46a1− − mice and mirrored the total cholesterol levels in the retina of the and Cyp27a1−/−Cyp46a1−/− mice and mirrored the total cholesterol levels in the retina of the three three genotypes. Thus, serum and retinal sterol profiling identified a decrease in retinal cholesterol genotypes. Thus, serum and retinal sterol profiling identified a decrease in retinal cholesterol biosynthesis as a compensatory response in both triple knockout genotypes. biosynthesis as a compensatory response in both triple knockout genotypes.

/ Figure 4. Serum (A) and retinal (B–D) sterols. WT1, wild type mice (C57BL/6J) for the Soat1− −, / / / / / / / −/− −/− ApoeFigure− − 4., Cyp27a1Serum (A− )− andCyp46a1 retinal− − (Soat1B–D) −sterols.−, and WT1, Cyp27a1 wild− −typeCyp46a1 mice −(C57BL/6J)−Apoe− − forgenotypes; the Soat1 WT2,, Apoe wild, −/− −/− −/− −/− / −/− /−/− typeCyp27a1 mice (C57BLCyp46a1/6J; 129S6Soat1/SvEv), and for Cyp27a1 the Cyp27a1Cyp46a1− −Cyp46a1Apoe− − genotype.genotypes; The WT2, results wild are presentedtype mice as(C57BL/6J; means SD129S6/SvEv) of the individual for the Cyp27a1 measurements−/−Cyp46a1 in 3–7−/− genotype. animals/genotype The results/sex; are number presented of mice as means equals ± ± theSD numberof the individual of retinas; measurements animals were 6–8-monthsin 3–7 animals/genotype/sex; old. Statistical significance number of was mice assessed equals by the two-way number ANOVAof retinas; with animals Bonferroni were correction. 6–8-months Black old. lines Statis andtical asterisks significance indicate was statistically assessed significantby two-way di ffANOVAerences betweenwith Bonferroni sexes of thecorrection. same genotype; Black lines the resultsand asterisks of the other indicate comparisons statistically are summarizedsignificant differences in Tables S2–S5.between ** psexes0.01; of ***thep same0.001. genotype; the results of the other comparisons are summarized in Tables ≤ ≤ S2–S5. ** p ≤ 0.01; *** p ≤ 0.001. 3.3. Retinal Gene Expression 3.3. Retinal Gene Expression / / / Two groups of retinal genes were quantified in Cyp27a1− −Cyp46a1− −Soat1− − and / / / Cyp27a1Two− −Cyp46a1groups −of− Apoeretinal− − micegenes and were compared quantified with that in in wildCyp27a1 type−/− (WT)Cyp46a1 C57BL−/−Soat1/6J animals−/− and (FigureCyp27a15).− The/−Cyp46a1 first included−/−Apoe− the/− mice genes and related compared to cholesterol with that biosynthesis in wild (typeHmgcr (WT)) and C57BL/6J its transcriptional animals regulation(Figure 5). ( Srebp2The )first as wellincluded as apolipoprotein-mediated the genes related to cholesterolcholesterol transportbiosynthesis (Abca1 (Hmgcr, Apoa) 1andand its2, Apoa4transcriptional, Apob, Apoc3 regulation, Apod, Apoe (Srebp2, and) asApoj well). Theas apolipoprotein-mediated second were the genes related cholesterol to inflammation transport (Abca1 (Ccl2,, Cox2Apoa, Il-61 and, and 2,Tnf Apoa4α), whose, Apob expression, Apoc3, Apod is controlled, Apoe, and by Apoj NF-).κ B,The which second in turn were is negativelythe genes regulatedrelated to byinflammation liver X receptors (Ccl2, (LXRs,Cox2, Il-6 transcription, and Tnfα), factors)whose expression when they is are controlled activated by by NF- oxysterolsκB, which produced in turn is bynegatively CYP27A1 regulated and CYP46A1 by liver [37 X]. receptors Statistically (LXRs, significant transcription changes factors) were detected when they in the are expression activated ofby almostoxysterols every produced studied by gene. CYP27A1 However, and theseCYP46A1 changes [37]. wereStatistically mostly significant sex-specific changes and therefore were detected likely didin the not expression contribute of to almost retinal every cholesterol studied levels gene. thatHowever, were similarthese changes in female were and mostly male sex-specific animals of theand / / / / / / Cyp27a1therefore− −likelyCyp46a1 did− not− Soat1 contribute− − and to Cyp27a1 retinal− choles−Cyp46a1terol− levels−Apoe that− − genotypes.were similar Only in femaleApoa1 and andApoa2 male / / / / / / in Cyp27a1− −Cyp46a1− −Soat1− − mice as well as Apob and Apoj in Cyp27a1− −Cyp46a1− −Apoe− − Biomolecules 2019, 9, x 8 of 22

Biomoleculesanimals of2019 the, 9 Cyp27a1, 867 −/−Cyp46a1−/− Soat1−/− and Cyp27a1−/−Cyp46a1−/−Apoe−/− genotypes. Only Apoa18 of 22 and Apoa2 in Cyp27a1−/−Cyp46a1−/−Soat1−/− mice as well as Apob and Apoj in −/− −/− −/− miceCyp27a1 showedCyp46a1 expressionApoe changes mice that showed are common expression for both changes sexes. that The are ambiguous commonresults for both of thesexes. retinal The geneambiguous expression results justified of the subsequent retinal gene use expression of retinal proteomicsjustified subsequent to compare use the of relative retinal proteinproteomics levels. to compare the relative protein levels.

Figure 5. Retinal expression of the genes related to cholesterol and inflammation in female (A) and male Figure 5. Retinal expression of the genes related to cholesterol and inflammation in female (A) and (B) mice. The results are presented as means SD of triplicate measurements in pooled samples from male (B) mice. The results are presented as means± ± SD of triplicate measurements in pooled samples 6–8 month old C57BL/6J female mice (n = 7 mice, 13 retinas), C57BL/6J male mice (n = 8 mice, 14 retinas), from 6–8 /month old C57BL/6J/ / female mice (n = 7 mice, 13 retinas), C57BL/6J male/ mice (n =/ 8 mice, /14 Cyp27a1− −Cyp46a1− −Soat1− − female mice (n = 5 mice, 10 retinas), Cyp27a1− −Cyp46a1− −Soat1− − retinas), Cyp27a1−/−Cyp46a1−/−Soat1−/− female mice (n = 5 mice, 10 retinas), male mice (n = 6 mice, 12 retinas), Cyp27a1 / Cyp46a1 / ApoE / female mice (n = 5 mice, 10 retinas), −/− −/− −/− − − − − − − −/− −/− −/− Cyp27a1 Cyp46a1/ Soat1/ male mice/ (n = 6 mice, 12 retinas), Cyp27a1 Cyp46a1 ApoE female and Cyp27a1− −Cyp46a1− −ApoE− − male mice (n = 4 mice, 8 retinas). * p 0.05; ** p 0.01; mice (n = 5 mice, 10 retinas), and Cyp27a1−/−Cyp46a1−/−ApoE−/− male mice (n = 4 mice,≤ 8 retinas).≤ * p ≤ *** p 0.001 vs. the C57BL/6J retina as assessed by two-way ANOVA with the post-hoc Dunnet’s test. 0.05;≤ ** p ≤ 0.01; *** p ≤ 0.001 vs. the C57BL/6J retina as assessed by two-way ANOVA with the post- ND, not detectable. hoc Dunnet’s test. ND, not detectable. 3.4. Retinal Proteomics

3.4. Retinal Proteomics / / / This was conducted on Cyp27a1− −Cyp46a1− −Soat1− − vs. WT male mice to gain insight into whetherThis there was areconducted additional on mechanisms Cyp27a1−/−Cyp46a1 that contribute−/−Soat1−/ to− vs. the WT normalization male mice ofto retinal gain insight cholesterol into levelswhether besides there aare decrease additional in cholesterol mechanisms biosynthesis. that contribute We alsoto the sought normalizatio to mapn the of retinal mechanisms cholesterol that impairlevels besides retinal function,a decrease when in cholesterol the retinal biosynthes cholesterolis. levels We also are normal.sought to The map label-free the mechanisms analysis wasthat used,impair which retinal confirmed function, unaltered when the retinal retinal expression cholesterol of levels APOA4 are and normal. APOE The (Table label-free1) suggested analysis by thewas mRNAused, which measurements confirmed (Figure unaltered5) but retinal did not expression show altered of APOA4 abundance and APOE of APOA1, (Table APOA2, 1) suggested and APOJ, by the / / / whosemRNA gene measurements expression (Figure was aff 5)ected but did in thenot Cyp27a1show altered− −Cyp46a1 abundance− −Soat1 of APOA1,− − vs. APOA2, WT retina and of APOJ, male micewhose (Figure gene 5expression). Proteins was encoded affected by Srebp2in the ,Cyp27a1Hmgcr, Abca1−/−Cyp46a1, Apob−,/−Apoc3Soat1−,/− and vs. ApodWT retinawere notof male detected. mice Nevertheless,(Figure 5). Proteins collectively, encoded the by data Srebp2 obtained, Hmgcr suggested, Abca1, thatApob cholesterol, Apoc3, and transport Apod were mediated not detected. by the majorNevertheless, retinal apolipoproteins collectively, the did data not obtained seem to sugge be significantlysted that cholesterol affected by transport simultaneous mediated absence by the of CYP27A1,major retinal CYP46A1, apolipoproteins and SOAT1. did Innot addition, seem to we be identifiedsignificantly 43 proteinsaffected withby simultaneous a significantly absence different of abundance.CYP27A1, CYP46A1, Out of these and proteins, SOAT1. theIn addition, abundance we ofidentified 41 proteins 43 proteins was increased with a andsignificantly that of 3 proteinsdifferent / / / wasabundance. decreased Out in of the these Cyp27a1 proteins,− − Cyp46a1the abundance− −Soat1 of− 41− vs.proteins WT was retina increased (Table2 ).and Protein that of grouping 3 proteins basedwas decreased on function in the (Text Cyp27a1 S1) suggested−/−Cyp46a1 that−/−Soat1 at least−/− vs. 6 WT biological retina (Table processes 2). Protein could grouping be affected based in the on / / / Cyp27a1function− (Text−Cyp46a1 S1) −suggested−Soat1− − retina:that at (1) least cytoskeletal 6 biological dynamics processes and vesicle could formation be affected (11 proteins, in the FigureCyp27a16);− (2)/−Cyp46a1 energy− homeostasis/−Soat1−/− retina: (9 proteins, (1) cytoskeletal Figure7 );dynamics (3) genetic and information vesicle formation transfer (DNA(11 proteins,/RNA processing,Figure 6); (2) transcription, energy homeostasis regulation (9 proteins, of transcription Figure 7); and (3) translation, genetic information 9 proteins); transfer (4) inflammation (DNA/RNA (7processing, proteins); (5)transcription, synaptic function regulation (4 proteins); of transcription and (6) ubiquitinationand translation, (3 9 proteins). proteins); Since (4) inflammation changes in the (7 vesicleproteins); formation (5) synaptic can be function detected (4 by proteins); TEM, we and evaluated (6) ubiquitination retinal ultrastructure (3 proteins). next. Since changes in the vesicle formation can be detected by TEM, we evaluated retinal ultrastructure next. Biomolecules 2019, 9, 867 9 of 22 Biomolecules 2019, 9, x 9 of 22

/ −/− /−/− −/− FigureFigure 6.6. DiDifferentiallyfferentially expressed expressed proteins proteins in in the the Cyp27a1 Cyp27a1− −Cyp46a1Cyp46a1− −Soat1Soat1− − vs.vs. wild wild type type retina ofof pertinencepertinence toto thethe formationformation of of vesicles vesicles and and protrusions. protrusions. Vesicular Vesicular formation formation via via endosomal endosomal pathway pathway is shownis shown inblue in blue font font and arrowsand arrows and that and via that the via secretory the secretory pathway pathway is shown inis orangeshown fontin orange and arrows; font and the formationarrows; the of formation membrane of protrusions membrane (Pt) protrusions is indicated (Pt) in is purple indicated font andin purple arrows. font Protein and arrows. full names Protein and allfull the names supporting and allreferences the supporting are provided references in the are Text prov S1.ided In in the the endosomal Text S1. In pathway, the endosomal SEPT6 (an pathway, F-actin bindingSEPT6 (an protein) F-actin is binding required protein) for the is maturation required for of the early maturation endosomes of early (EE) intoendosomes multi-vesicular (EE) intobodies multi- (MVB),vesicular whereas bodies RAB4B(MVB), (awhereas small GTPase)RAB4B (a and small PAFAH1B2 GTPase) (anand endosome-associatedPAFAH1B2 (an endosome-associated phospholipase) participatephospholipase) in material participate delivery in frommaterial EE todeliver recyclingy from endosomes EE to (RE)recycling and subsequently endosomes to(RE) plasma and membranesubsequently via to vesicles plasma (Ve). membrane Increased via expressionvesicles (Ve). of SEPT6Increased and expression PAFAH1B2 of alongSEPT6 with and aPAFAH1B2 decreased expressionalong with ofa decreased RAB4B point expression to an increased of RAB4B MVB poin butt to notan increased RE formation MVB and but thereforenot RE formation an increased and exosometherefore (Ex) an release.increased In theexosome secretory (Ex) pathway, release. CHGAIn the (asecretory neurosecretory pathway, protein) CHGA is present(a neurosecretory in all of the pathway-relatedprotein) is present organelles: in all of endoplasmic the pathway-rela reticulumted (ER),organelles: Golgi complexendoplasmic (GC), andreticulum secretory (ER), granules Golgi (SG).complex CHGA (GC), expression and secretory initiates granules SG formation (SG). CHGA and increases expression a number initiates of SG SG. formation SRP19 is and responsible increases for a translationnumber of andSG. targetingSRP19 is ofresponsible the membrane for translat and secretoryion and proteins targeting to of ER the and membrane their further and routing secretory via theproteins secretory to ER pathway. and their When further PAFAH1B2 routing via is the bound secretory to GC, pathway. it maintains When GC PAFAH1B2 integrity and is bound therefore to GC, the functionalityit maintains ofGC the integrity secretory pathway.and therefore Increased the abundancefunctionality of SRP19,of the CHGA, secretory and PAFAH1B2pathway. Increased points to anabundance increase inof membraneSRP19, CHGA, traffi ckingand PAFAH1B2 via the secretory points pathway. to an increase Lastly, in Pt membrane formation involvestrafficking TAGLN2, via the TAGLN3,secretory SEPT,pathway. and Lastly, GPM6A6. Pt formation TAGLN2 andinvolves TAGLN3 TAGLN2, increase TAGLN3, G-actin SEPT, polymerization. and GPM6A6. SEPT6 TAGLN2 acts on actinand TAGLN3 assembly increase indirectly G-actin via cortactin, polymerization. a regulator SEPT6 of actin acts polymerization on actin assembly and indirectly a substrate via for cortactin, HDAC6, whicha regulator deacetylates of actin cortactinpolymerization and increases and a itssubstrat bindinge for to HDAC6, actin. HDAC6 which wasdeacetylates found in cortactin extracellular and vesiclesincreases (EV) its derivedbinding fromto actin. Pt. GPM6AHDAC6 (awas transmembrane found in extracellular protein) inducesvesicles clustering(EV) derived of lipid from rafts Pt. andGPM6A facilitates (a transmembrane the Pt formation protein) independent induces ofclustering actin. Pt of formation lipid rafts can and also facilitates be indirectly the Pt a ffformationected by TINAGL1independent (an of extracellular actin. Pt formation matrix can protein) also be and indirectly ATP1A2 affected (a transmembrane by TINAGL1 protein),(an extracellular which canmatrix be associatedprotein) and with ATP1A2 the actin (a transmembr cytoskeleton.ane Increased protein), expressionwhich can be of associated TAGLN2, TAGLN3,with the actin SEPT, cytoskeleton. TINAGL1, andIncreased ATP1A2 expression raises a possibilityof TAGLN2, that TAGLN3, Pt formation SEPT, could TINAGL1, be increased and ATP1A2 as well raises and could a possibility increase that the formationPt formation of extracellularcould be increased vesicles as (EV).well and could increase the formation of extracellular vesicles (EV). Biomolecules 2019, 9, 867 10 of 22 Biomolecules 2019, 9, x 10 of 22

/ / / FigureFigure 7.7. DiDifferentiallyfferentially expressedexpressed proteinsproteins inin thethe Cyp27a1 Cyp27a1−−/−−Cyp46a1−/−Soat1−Soat1−/− vs.− vs. wild wild type type retina retina of ofpertinence pertinence to toenergy energy and and protein protein homeostasis. homeostasis. Gl Glucoseucose production production and and utilization utilization are indicated inin blueblue fontfont andand arrows; fattyfatty acidacid productionproduction andand ββ-oxidation-oxidation areare inin orangeorange fontfont andand arrows;arrows; andand aminoamino acidacid metabolismmetabolism isis inin purplepurple fontfont andand arrows. Protein full names and all the supporting references areare providedprovided inin thethe TextText S1.S1. InIn glucoseglucose productionproduction andand utilization,utilization, GAAGAA (a(a lysosomallysosomal enzyme)enzyme) catalyzescatalyzes glucoseglucose releaserelease from from glycogen. glycogen. Simultaneously, Simultaneously, in the in cytoplasm, the cytoplasm, glucose-1-phosphate glucose-1-phosphate (1P) is generated (1P) is bygenerated GALE action,by GALE which action, catalyzes which the finalcatalyzes step inthe the final major step pathway in the ofmajor galactose pathway metabolism. of galactose Both glucosemetabolism. and glucose-1P Both glucose are and fuel glucos for glycolysis,e-1P are which fuel for produces glycolysis, energy which in the produces form of energy ATP. Pyruvate, in the form the endof ATP. product Pyruvate, of glycolysis, the end enters product the mitochondriaof glycolysis, andenters is convertedthe mitochondria to acetyl and CoA is by converted the PDH complexto acetyl containingCoA by the PDHX. PDH complex This PDH-dependent containing PDHX. reaction This is the PDH- ratedependent limiting step reaction for the is utilization the rate limiting of pyruvate step under aerobic conditions. Further, in the fatty acid-related pathways, free fatty acids can be released in for the utilization of pyruvate under aerobic conditions. Further, in the fatty acid-related pathways, lysosomes from the breakdown of ceramides by ASAH1 and then reach the mitochondria, where they free fatty acids can be released in lysosomes from the breakdown of ceramides by ASAH1 and then enter β-oxidation. The last step in β-oxidation is catalyzed by ACAA2 to yield acetyl CoA. Acetyl CoA reach the mitochondria, where they enter β-oxidation. The last step in β-oxidation is catalyzed by from both pyruvate and fatty acid β-oxidation enters the tricarboxylic acid cycle (TCA) to produce ATP. ACAA2 to yield acetyl CoA. Acetyl CoA from both pyruvate and fatty acid β-oxidation enters the Lastly, in amino acid metabolism, the amino acid levels in the lysosomes are sensed by the protein tricarboxylic acid cycle (TCA) to produce ATP. Lastly, in amino acid metabolism, the amino acid complex including LAMTOR3 and V-ATPase (a proton pump). In the cytoplasm, 3P-hydroxypyruvate levels in the lysosomes are sensed by the protein complex including LAMTOR3 and V-ATPase (a (produced from the glycolysis intermediate) is converted to 3P-serine by PSAT1. In the mitochondria, proton pump). In the cytoplasm, 3P-hydroxypyruvate (produced from the glycolysis intermediate) is homocysteine thiolactone (a by-product of protein biosynthesis) could be converted to homocysteine converted to 3P-serine by PSAT1. In the mitochondria, homocysteine thiolactone (a by-product of by BPHL. Both 3P-serine and homocysteine are the precursors of cysteine used for protein synthesis. protein biosynthesis) could be converted to homocysteine by BPHL. Both 3P-serine and homocysteine Catabolism of proteins produces urea, which can be transported from cells by SLC14A1. The expression are the precursors of cysteine used for protein synthesis. Catabolism of proteins produces urea, which of all the enzymes indicated in this figure was increased in the Cyp27a1 / Cyp46a1 / Soat1 / vs. WT can be transported from cells by SLC14A1. The expression of all the enzymes− − indicated− − in− this− figure retina, thus raising a possibility of increased energy production as well as altered protein homeostasis. was increased in the Cyp27a1−/−Cyp46a1−/−Soat1−/− vs. WT retina, thus raising a possibility of increased 3.5. TEMenergy production as well as altered protein homeostasis. Postfixation with osmium, tannic acid, and p-phenylenediamine was used, which preserves 3.5. TEM and enhances the visualization of membranous structures (tannic acid) and osmium-treated neutralPostfixation lipids (p-phenylenediamine) with osmium, tannic acid, [33]. and This p-phenylenediamine approach revealed was several used, diwhichfferences preserves between and / / / / / / WT,enhances Cyp27a1 the visualization− −Cyp46a1− of−Soat1 membranous− −, and Cyp27a1structures− −(tannicCyp46a1 acid)− − andApoe osmium-treated− − mice. First, neutral undigested lipids / / / material(p-phenylenediamine) was observed in[33]. some This of theapproach RPE cells revealed of Cyp27a1 several− −Cyp46a1 differences− −Soat1 − −betweenmice but WT, not −/− / −/− −/− / / −/− −/− −/− WTCyp27a1 or Cyp27a1Cyp46a1− −Cyp46a1Soat1 ,− and−Apoe Cyp27a1− − miceCyp46a1 (Figure8ApoeA–C), probablymice. First, a undigested reflectionof material impaired was phagocytosisobserved in atsome the phagolysosomal of the RPE stagecells [38of]. Second,Cyp27a1 chromatin−/−Cyp46a1 condensation−/−Soat1−/− mice was morebut not abundant WT inor −/− −/− −/− / / / / / / theCyp27a1 photoreceptorCyp46a1 nucleiApoe of mice Cyp27a1 (Figure− −Cyp46a1 8A–C), −probably−Soat1− a− reflectionthan Cyp27a1 of impaired− −Cyp46a1 phagocytosis− −Apoe− at− micethe phagolysosomal and was essentially stage absent[38]. Second, in WT chroma mice (Figuretin condensation8D–F). This was observation more abundant is consistent in the −/− −/− −/− −/− −//− −/− / / withphotoreceptor increased nuclei photoreceptor of Cyp27a1 apoptosisCyp46a1 documentedSoat1 than previously Cyp27a1 in Cyp27a1Cyp46a1− −ApoeCyp46a1 mice− − Soat1and was− − / / / miceessentially [22]. absent Third, Cyp27a1in WT mice− −Cyp46a1 (Figures− − Soat18D–F).− − Thismice observation had increased, is consistent as compared with to increased WT and / / / −/− −/− −/− Cyp27a1photoreceptor− −Cyp46a1 apoptosis− −Apoe documented− − mice, withpreviously the presence in Cyp27a1 of multivesicularCyp46a1 bodiesSoat1 at mice both [22]. apical Third, and Cyp27a1−/−Cyp46a1−/−Soat1−/− mice had increased, as compared to WT and Cyp27a1−/−Cyp46a1−/−Apoe−/− mice, with the presence of multivesicular bodies at both apical and basal Biomolecules 2019, 9, 867 11 of 22 Biomolecules 2019, 9, x 11 of 22 Biomolecules 2019, 9, x 11 of 22 basalsides sides of the of theRPE RPE (Figure (Figure 9A–H,9A–H, O–Q) O–Q) in agreement in agreement with with the change the change in vesicular in vesicular traffic tra suggestedffic suggested by −/− / −/− −/−/ / bythesides the proteomics proteomics of the RPE analysis (Figure analysis (Figure9A–H, (Figure O–Q) 6).6 Similarly,). in Similarly, agreement Cyp27a1 Cyp27a1 with theCyp46a1 change− −Cyp46a1Soat1 in vesicular− −miceSoat1 traffichad− − moremice suggested frequent hadmore by frequentextracellularthe proteomics extracellular vesicles analysis vesicles in the(Figure outer in the6). nuclear outerSimilarly, nuclearlayer Cyp27a1 (Figure layer−/ − (Figure9I–M,Cyp46a1 R–T),9I–M,−/−Soat1 and R–T),− /−extensive mice and had extensive budding more frequent budding of the ofcaveoli-like theextracellular caveoli-like structures vesicles structures inin the bloodouter in the vesselnuclear blood pericytes layer vessel (Figure and pericytes endothelium 9I–M, and R–T), endothelium of and the ganglionextensive of cell thebudding layer ganglion (Figure of the cell layer10).caveoli-like (Figure Thus, 10structuresintracellular). Thus, in intracellular the and blood extracellular vessel and pericytes extracellular ve andsicular endothelium vesicular traffic tra seemedofffi thec seemedganglion to be tocell beincreased layer increased (Figure in in Cyp27a110). Thus,/ −/−Cyp46a1 intracellular−//−Soat1 −/−and /relative extracellular to WT or Cyp27a1 vesicular−/−Cyp46a1 /traffic−/ −Apoeseemed/−/− mice to and / be directed increased toward in Cyp27a1− −Cyp46a1− −Soat1− − relative to WT or Cyp27a1− −Cyp46a1− −Apoe− − mice and directed −/− −/− −/− −/− −/− −/− towardbothCyp27a1 choriocapillaries both choriocapillariesCyp46a1 Soat1 and inner andrelative retina inner to blood retinaWT or vessels. bloodCyp27a1vessels. Cyp46a1 Apoe mice and directed toward both choriocapillaries and inner retina blood vessels.

Figure 8. Retinal ultrastructure as assessed by TEM. Cross sections through the retinal pigment FigureepitheliumFigure 8. 8.Retinal Retinal (RPE) ultrastructure ultrastructureand outer nuclear asas assessedassessed layer byby(ONL) TEM.TEM. are Cross shown. sections sections Orange through through dashed the the retinalellipse retinal pigmentdenotes pigment epitheliumundigestedepithelium (RPE) material;(RPE) and and outer cyan outer nuclear arrowheads nuclear layer (ONL)layerdenote (ONL) are chromatin shown. are shown. Orange condensation. dashedOrange ellipseAdashed, B, D denotes , ellipseE: Images undigesteddenotes are material;representativeundigested cyan material; arrowheads of three cyan male denote arrowheads mice chromatinper genotype. denote condensation. chromatinC, F: Only onecondensation.A, B male, D, Eanimal: Images A , wasB, are Dimaged., representativeE: Images All mice are of threewererepresentative male 7-months mice per ofold. three genotype. Scale male bars: mice C, 2 μ F:m per Only (A genotype.,B one,C); 5 male μm C, ( animalD F:,E Only,F). was one imaged. male animal All mice was wereimaged. 7-months All mice old. Scalewere bars: 7-months 2 µm ( Aold.,B ,ScaleC); 5 µbars:m ( D2 ,μEm,F ).(A,B,C); 5 μm (D,E,F).

Figure 9. Retinal ultrastructure as assessed by TEM. Colored squares and rectangles in A, C, E, G, I, L, O,Figure and R 9.denote Retinal regions ultrastructure that are as shown assessed at aby higher TEM. magnificationColored squares in andB, Drectangles, F, H, J, inK, AM, ,CN, E, ,P G, Q, I, S, andLFigure,T O., Cyan and 9. RetinalR arrowheads denote ultrastructure regions denote that areas some assessed shown of the at by a vesicles higherTEM. Colored magnification in the RPE squares; purple in and B, D arrowsrectangles, F, H, J denote, K in, M A, ,N someC,, PE, ,Q G of, ,S I the,, L, O, and R denote regions that are shown at a higher magnification in B, D, F, H, J, K, M, N, P, Q, S, vesiclesand T in. Cyan the extracellular arrowheads denote space of some the ONL.of the Avesicles–M: Images in the areRPE representative; purple arrows of denote three malesome mice of the per vesiclesand T. Cyan in the arrowheads extracellular denote space ofsome the ofONL. the A-vesicles M: Images in the are RPE representative; purple arrows of threedenote male some mice of perthe genotype. O–T: Only one male animal was imaged. All mice were 7-months old. Scale bars: 2 µm vesicles in the extracellular space of the ONL. A- M: Images are representative of three male mice per (A,C,E,G,J,K,M,N,O,S,T); 5µm (I,L,R); 0.5 µm (B,D,F,H,P,Q). Biomolecules 2019, 9, x 12 of 22

Biomoleculesgenotype.2019, 9, 867 O-T: Only one male animal was imaged. All mice were 7-months old. Scale bars: 2 μm 12 of 22 (A,C,E,G,J,K,M,N,O,S,T); 5μm (I,L,R); 0.5 μm (B,D,F,H,P,Q).

Figure 10. Retinal ultrastructure as assessed by TEM. Colored squares in A, E, I denote regions in the Figure 10. Retinal ultrastructure as assessed by TEM. Colored squares in A, E, I denote regions in the neurovascular units of the ganglion cell layers that are shown at a higher magnification in B–D, F–H, neurovascular units of the ganglion cell layers that are shown at a higher magnification in B–D, F–H, J–L. Orange arrowheads denote some of the vesicles. En, vascular endothelium; P, pericyte. A–H: J–L. Orange arrowheads denote some of the vesicles. En, vascular endothelium; P, pericyte. A–H: I L ImagesImages are are representative representative of of three three mice mice per per genotype.genotype. I––L: :Only Only one one animal animal wa wass imaged. imaged. All Allmice mice werewere 7-month 7-month old old males. males. Scale Scale bars: bars: 2 2µ μmm ((AA,,EE,,I); 1 μµmm ( (BB––DD,F,F––HH,J,–JL–).L ). 4. Discussion 4. Discussion In theIn the present present work, work, we we continued continued to to investigate investigate retinal retinal mechanisms mechanisms that that maintain maintain cholesterol cholesterol homeostasishomeostasis and and could could be importantbe important for retinalfor retinal structure structure and and function function [22 ,[22,26,27,29,30,39,40].26,27,29,30,39,40]. Only Only three enzymesthree enzymes (CYP27A1, (CYP27A1, CYP46A1, CYP46A1, and SOAT1) and SOAT1) metabolize metabolize intracellular intracellular (unesterified) (unesterified) cholesterol cholesterol in the / / retinain (Figurethe retina1) (Figure with retinal 1) with cholesterol retinal choles contentterol content being almost being almost doubled doubled in Cyp27a1 in Cyp27a1− −Cyp46a1−/−Cyp46a1− −−/−mice / −/− −//− −/− / −/− / andmice normal and innormal Soat1 in− −Soat1mice (Figuremice (Figure4)[ 22 4),26 [22,26].]. Hence Hence by by generating generating Cyp27a1 Cyp27a1− −Cyp46a1Cyp46a1−Soat1−Soat1 − − −/− / −/− / mice,mice, we we expected expected that that their their retinalretinal cholesterol cholesterol will will still still be beincreased increased as in as Cyp27a1 in Cyp27a1Cyp46a1− −Cyp46a1 mice − − but be unesterified as in Soat1−/− mice (Figure/ 4B). Yet, the retinal cholesterol levels were normal in mice but be unesterified as in Soat1− − mice (Figure4B). Yet, the retinal cholesterol levels Cyp27a1−/−Cyp46a1−/−Soat1/ −/− mice (Figure/ 4B), thus/ suggesting that there are compensatory responses were normal in Cyp27a1− −Cyp46a1− −Soat1− − mice (Figure4B), thus suggesting that there are −/− −/− −/− compensatoryin the retina, responses which lowered in the the retina, sterol whichlevels. Similarly, lowered by the generating sterol levels. Cyp27a1 Similarly,Cyp46a1 by generatingApoe mice, we expected an additive genotype effect due to simultaneous blockage of the two major Cyp27a1 / Cyp46a1 / Apoe / mice, we expected an additive genotype effect due to simultaneous pathways− − of retinal− − cholesterol− − output - cholesterol removal via metabolism (CYP27A1 and blockage of the two major pathways of retinal cholesterol output - cholesterol removal via metabolism CYP46A1) and LP containing APOE which accepts cholesterol effluxed from cells. However, the (CYP27A1Cyp27a1 and−/−Cyp46a1 CYP46A1)−/−Apoe and−/− retina LP containing had lower, APOE not higher, which total accepts cholesterol cholesterol levels than effluxed the Apoe from−/− cells. / / / However,retina (Figure the Cyp27a1 4B), also− −Cyp46a1 an indication− −Apoe of −the− retinacompensatory had lower, responses. not higher, Therefore, total cholesterolto map these levels / thancompensatory the Apoe− − responses,retina (Figure we 4firstB), alsoconducted an indication retinal sterol of the profiling compensatory and quantified responses. cholesterol Therefore, to mapprecursors these compensatorylathosterol and responses, desmosterol. we firstThe conductedlevels of lathosterol retinal sterol were profiling decreased and quantifiedin the cholesterolCyp27a1 precursors−/−Cyp46a1−/− lathosterolSoat1−/− retina and as compared desmosterol. to Soat1 The−/− levelsand Cyp27a1 of lathosterol−/−Cyp46a1 were−/− retina, decreased and both in the / / / −/− −/− −/− / / −/− / Cyp27a1sterols− −wereCyp46a1 decreased− −Soat1 in the− − Cyp27a1retina asCyp46a1 comparedApoe to Soat1 retina− − asand compared Cyp27a1 to− the−Cyp46a1 Apoe −retina− retina, / / / and(Figure both sterols 4C,D). wereThese decreased decreases suggested in the Cyp27a1 that the− −retinalCyp46a1 cholesterol− −Apoe biosynthesis− − retina as is compared decreased toin the / Apoe− − retina (Figure4C,D). These decreases suggested that the retinal cholesterol biosynthesis is / / / / / / decreased in Cyp27a1− −Cyp46a1− −Soat1− − and Cyp27a1− −Cyp46a1− −Apoe− − mice, and represents a compensatory response to the triple gene ablations. Next, we assessed the apolipoprotein-mediated retinal cholesterol traffic by measuring the mRNA for different apolipoproteins (Figure5) and conducting retinal proteomics on Biomolecules 2019, 9, 867 13 of 22

/ / / Cyp27a1− −Cyp46a1− −Soat1− − vs. WT mice (Table1; Table2). Changes in the gene expression were sex-specific (Figure5) and for apolipoproteins were not translated into changes in the protein expression as exemplified by APOAs, APOE, and APOJ, the major retinal apolipoproteins (Table1). Rather, retinal / / / proteomics pointed to enhanced vesicle formation in the Cyp27a1− −Cyp46a1− −Soat1− − vs. WT retina as an expression of 7 proteins from the endosomal, secretory, and protrusion formation pathways (CHGA, HDAC6, PAFAH1B, SEPT6, SRP19, TAGLN2, and TAGLN3, Figure6)[ 41–50] was increased in the former (Table2). Hence, retinal vesicle formation was investigated by TEM and found to be most / / / / / / prominent in the Cyp27a1− −Cyp46a1− −Soat1− − retina followed by Cyp27a1− −Cyp46a1− −Apoe− − / / / and then WT retina (Figure9; Figure 10). Notably, in the Cyp27a1 − −Cyp46a1− −Soat1− − retina, vesicles were detected at both apical and basal RPE sides, possibly a reflection of the bi-directional traffic to the intraretinal and choroidal vascular networks. Traffic to the intraretinal vasculature was further supported by abundant vesicles in the intraretinal space of the outer nuclear layer and neurovascular units of the ganglion cell layer (Figures9 and 10). Also, the expression of HDAC6, which was found in the apically secreted extracellular vesicles derived from the oxidatively stressed RPE [51], was / / / increased 3-fold (Table2) in the Cyp27a1 − −Cyp46a1− −Soat1− − retina.

Table 1. Proteins detected by the label-free approach that were analyzed for gene expression in the / / / Cyp27a1− −Cyp46a1− −Soat1− − (TKO) and C57BL/6J (WT) mouse retinas.

№ of Identified Sequence Peptide Intensity (a.u.) 106 TKO/WT, Protein × p Peptides Coverage (%) TKO WT Protein Ratio APOA 1 7 36 3.55 1.63 3.63 0.69 0.98 0.94 ± ± APOA 2 1 10 0.95 0.55 0.32 0.16 2.99 0.13 ± ± APOA4 1 3 0.09 0.15 0.18 0.17 0.47 0.49 ± ± APOE 7 24 1.52 0.21 1.56 0.26 0.98 0.86 ± ± APOJ 7 20 3.17 0.11 2.66 0.19 1.12 0.56 ± ±

/ / Table 2. Proteins with significant changes in abundance ( 1.5-fold) in the Cyp27a1− −Cyp46a1− − / ≥ Soat1− − (TKO) vs. C57BL/6J (WT) retina. Protein grouping is by process and shows each protein only in one group despite the involvement in multiple processes. Proteins with a decreased expression are in bold.

Peptide Intensity (a.u.) № of Peptides With № of Sequence TKO/WT, Protein 106 Significant Changes in Unique Coverage × Protein Abundance Peptides (%) TKO WT Ratio Cytoskeletal organization, vesicular and secretory pathway ASAP1 2 2 3 1.0 0.1 0.6 0.1 1.6 ± ± ATP1A2 28 12 35 4.9 1.7 1.6 0.5 3.1 ± ± CHGA 2 2 9 2.6 0.1 1.6 0.5 1.6 ± ± HDAC6 3 3 5 0.9 0.2 0.3 0.2 3.1 ± ± PAFAH1B2 5 5 42 13.3 1.7 8.8 1.2 1.5 ± ± PI4KA 10 10 6 1.5 0.2 0.9 0.1 1.7 ± ± SEPT6 11 6 37 6.5 0.2 4.1 0.8 1.6 ± ± SRP19 3 3 35 1.2 0.2 0.6 0.2 2.1 ± ± TAGLN2 9 6 54 2.5 0.4 0.9 0.2 2.8 ± ± TAGLN3 12 9 71 8.2 0.7 5.5 1.0 1.5 ± ± TINAGL1 3 3 12 0.7 0.1 0.3 0.1 2.2 ± ± Energy Homeostasis ACAA2 15 15 59 8.8 1.4 5.9 0.7 1.5 ± ± ASAH1 7 7 22 2.1 0.3 1.4 0.2 1.5 ± ± BPHL 3 3 19 1.1 0.2 0.5 0.1 2.1 ± ± GAA 3 3 6 1.0 0.2 0.7 0.1 1.5 ± ± GALE 3 3 15 1.1 0.1 0.6 0.1 1.9 ± ± LAMTOR3 2 2 31 1.1 0.1 0.4 0.1 2.8 ± ± PDHX 6 6 17 3.2 0.5 1.8 0.1 1.8 ± ± PSAT1 18 18 59 9.1 1.8 4.8 1.1 1.9 ± ± SLC14A1 4 4 9 2.0 0.5 1.1 0.2 1.8 ± ± Biomolecules 2019, 9, 867 14 of 22

Table 2. Cont.

Peptide Intensity (a.u.) № of Peptides With № of Sequence TKO/WT, Protein 106 Significant Changes in Unique Coverage × Protein Abundance Peptides (%) TKO WT Ratio DNA/RNA Processing, Transcription, Regulation of Transcription and Translation EEF1D 9 8 47 7.6 1.6 4.8 0.5 1.6 ± ± GTF2E1 3 3 9 0.5 0.0 0.3 0.1 1.7 ± ± NIPBL 2 2 1 0.4 0.0 0.1 0.2 3.2 ± ± NR2E3 6 6 26 1.9 0.5 0.9 0.3 2.1 ± ± NXF1 3 3 9 1.1 0.1 0.6 0.2 1.6 ± ± PDS5A 3 3 5 0.4 0.1 0.2 0.0 2.0 ± ± RORB 4 4 11 0.9 0.2 0.5 0.1 1.8 ± ± RPA3 2 2 33 0.5 0.1 0.1 0.2 3.7 ± ± U2AF2 10 10 36 16.2 0.3 10.2 2.2 1.6 ± ± Inflammation ANP32B 6 5 24 10.7 0.8 6.9 1.2 1.6 ± ± CD59A 2 2 21 3.0 0.6 1.6 0.4 1.9 ± ± OGFR 3 3 7 0.5 0.0 0.3 0.1 2.0 ± ± PRKRA 3 3 15 1.5 0.1 0.8 0.3 1.9 ± ± SERPINA1E 10 4 35 0.7 0.7 3.6 1.4 0.2 ± ± TRAFD1 3 3 8 1.5 0.2 0.9 0.2 1.8 ± ± WDFY1 4 4 12 1.4 0.4 0.4 0.1 3.5 ± ± Synaptic function GAP43 10 10 48 3.3 0.1 2.2 0.2 1.5 ± ± GPM6A 7 7 23 28.8 6.9 50.9 6.2 0.6 ± ± KCTD8 2 2 8 0.7 0.1 0.4 0.1 1.7 ± ± RAB4B 4 2 25 0.5 0.2 1.4 0.3 0.4 ± ± Ubiquitination CUL1 9 9 18 1.2 0.1 0.8 0.2 1.5 ± ± STUB1 6 6 25 1.3 0.3 0.6 0.2 2.0 ± ± UCHL5 5 5 28 1.9 0.3 1.2 0.3 1.6 ± ±

Extracellular vesicles are generally enriched with cholesterol as compared to other cellular membranes, and their release is usually enhanced from cells enriched with cholesterol [52]. Therefore, recently, extracellular vesicles were suggested to be a part of cellular machinery ensuring cholesterol homeostasis [52]. Yet, in vivo support for cholesterol removal via extracellular vesicles is still scarce and requires more evidence. We blocked the major pathways of retinal cholesterol output (Figure1) and created unique conditions, in which a compensatory downregulation of retinal cholesterol biosynthesis was probably not sufficient to normalize retinal cholesterol levels. Hence the formation of vesicles, possibly a reserve mechanism, was increased, and likely contributed to lowering of retinal cholesterol. If so, the present work appears to be the first in vivo study in which increased vesicle formation could be correlated with cholesterol lowering. We also identified an organ, the retina, where this mechanism could be operative. Vesicle formation requires energy [53], and remarkably, retinal proteomics identified 5 proteins (ACAA2, ASAH1, GAA, GALE, and PDHX, Figure7)[ 54–58], which were upregulated in the / / / Cyp27a1− −Cyp46a1− −Soat1− − vs. WT retina (Table2) and participate in glucose production and utilization as well as fatty acid β-oxidation (Table S1). If energy homeostasis is indeed affected / / / in the Cyp27a1− −Cyp46a1− −Soat1− − retina, there should be an increased production of ATP due to glycolysis and increased acetyl CoA availability as the latter is necessary for the tricarboxylic acid cycle, which provides NADPH for ATP production during mitochondrial (Figure7) or extra mitochondrial oxidative phosphorylation [59,60]. Also, extracellular vesicles contain proteins [61], therefore enhanced vesicle formation will lead to retinal protein loss, if this loss is not compensated for, a possible reason for an increase in abundance of the overwhelming majority of proteins in / / / the Cyp27a1− −Cyp46a1− −Soat1− − vs. WT retina (Table1). The expression of LAMTOR3, PSAT, BPHL, and SLC14A1 [62–65], which are involved in protein homeostasis (Figure7, Suppl. T1), was / / / increased in the Cyp27a1− −Cyp46a1− −Soat1− − retina (Table2) and could be the consequence of Biomolecules 2019, 9, 867 15 of 22 the enhanced extracellular vesicle formation. Thus not only the formation of extracellular vesicles / / / seemed to be enhanced in the Cyp27a1− −Cyp46a1− −Soat1− − retina (Figure6) but also the processes which are required (i.e., ATP production) and affected (i.e., protein homeostasis and ubiquitination) by vesicle formation. The present work provides insights into potential mechanisms that could underlie the / / / deteriorations in retinal function in Cyp27a1− −Cyp46a1− −Soat1− − mice (Figure3). Indeed, increased expression of HDAC6, a α-tubulin deacetylase, can perhaps lead to ciliary destabilization and potentially impair the traffic between the photoreceptor inner and outer segments [66,67]. In turn, these potential effects on the cilia can be compensated, fully or in part, by an increased expression of the proteins involved in rhodopsin traffic (ASAP1 and PI4KA), protein ubiquitination (CUL1, STUB1, and UCHL5) as well as the photoreceptor gene transcription (NR2E3 and RORβ)[68–76]. In addition, impaired ERG function could reflect a balance between the negative (decreased expression GPM6A and RAB4B) and positive (increased expression of GAP43 and KCTD8) effects on synaptic processes [77–80]. Inflammation may lead to the photoreceptor loss and thereby affect the ERG as well. The pro-inflammatory signaling is indicated by the upregulation of PRKRA and WDFY, which participate in the activation of NF-κB[81,82], a transcription factor and key regulator of immune and other functions, as well as a decreased abundance of SERPINA1E (an acute phase protein), which can attenuate microglia-mediated retinal degeneration [83]. The anti-inflammatory response could be due to the upregulation of ANP32B, ASAH1, CD59A, OGFR, and TRAFD1, proteins that play different roles (Figure 11, Table S1) but all affecting inflammation [84–89]. Thus, retinal proteomics also identified the non cholesterol-related proteins and processes that can be affected by blockage of the pathways of retinal cholesterol metabolism and storage. Further studies are required to / / / target more precisely the major contributors to changes in the Cyp27a1− −Cyp46a1− −Soat1− − retina. / / / In Cyp27a1− −Cyp46a1− −Apoe− − mice, retinal function was dependent on sex and was normal in female animals (Figure3), in agreement with our previous findings that APOE absence elicits strong compensatory mechanisms that minimize retinal impact of this protein absence [27]. Future work will / / / map these responses by retinal proteomics in Cyp27a1− −Cyp46a1− −Apoe− − mice for comparison to / Apoe− − results [27]. We propose the following model (Figure 12) to unify the available data and to identify the questions / / / that remain unanswered to completely explain changes in the Cyp27a1− −Cyp46a1− −Soat1− − retina. The absence of SOAT1 can increase the availability for the retina of fatty acids, which are used as the enzyme substrates, whereas the absence of CYP27A1 and CYP46A1 abolishes the production of 27-hydroxycholesterol and 24-hydroxycholesterol, which are activating ligands for LXRs. LXRs are important transcription factors that play essential roles in the regulation of cholesterol metabolism and also integration of glucose and fatty acid metabolism as well as immune and inflammatory responses [90]. Accordingly, glucose utilization and fatty acid β-oxidation could be increased in the / / / Cyp27a1− −Cyp46a1− −Soat1− − retina as a compensatory response to decreased LXR activation and/or increased fatty acid availability. This will lead to an increased energy production, which could be used to form extracellular vesicles, an energy-consuming process and alternate means to eliminate excess cellular cholesterol when the compensatory responses from cholesterol biosynthesis and efflux to LP are either not sufficient or not operative. In addition, decreased LXR activation could promote proinflammatory processes and affect retinal function. Apparently, other processes could also become / / / affected in the Cyp27a1− −Cyp46a1− −Soat1− − retina as suggested by retinal proteomics (Figure 11). Future studies are needed to address these outstanding questions. Biomolecules 2019, 9, 867 16 of 22 Biomolecules 2019, 9, x 16 of 22

/ / / Figure 11. Differentially expressed proteins in the Cyp27a1− −Cyp46a1− −Soat1− − vs. WT retina of −/− −/− −/− pertinenceFigure 11. toDifferentially retinal structure expressed and function. proteins Pathological in the Cyp27a1 processesCyp46a1 that couldSoat1 be triggered vs. WT by retina changes of inpertinence protein abundance to retinal structure are on the and left-hand function. side; Pathological all other processesprocesses arethat on could the right-handbe triggered side. by changes Protein fullin protein names abundance and all the supportingare on the left-hand references side; are providedall other processes in the Text are S1. on See the Discussion right-hand for side. explanation. Protein HDAC6full names deacetylates and all theα-tubulin, supporting a main references microtubule are component,provided in andthe therebyText S1. promote See Discussion microtubule for disassembly.explanation. HDAC6 PRKRA deacetylates and WDFY participateα-tubulin, a in main the activationmicrotubule of NF-component,κB, a transcription and thereby factor promote and keymicrotubule regulator disassembly. of many important PRKRA cellular and WDFY processes participate including in the immune activation response, of NF-κ synapticB, a transcription plasticity, memory,factor and cytokine key regulator production, of many and important other. SERPINA1 cellular processes is a serine including peptidase immune inhibitor, response, which synaptic can act asplasticity, an anti-inflammatory memory, cytokine agent. production, GPM6A and and RAB4Bother. SERPINA1 are involved is a in serine synaptogenesis. peptidase inhibitor, ASAP1 plays which a principalcan act asrole an anti-inflammatory in vesicular transport agent. of rhodopsinGPM6A an andd RAB4B other rhodopsin-likeare involved in receptors synaptogenesis. from the ASAP1 Golgi complexplays a principal to cilia. PI4KArole in vesicular is an important transport enzyme of rhodopsin for phototransduction and other rhodopsin-like and its downstream receptors productsfrom the canGolgi aff ectcomplex ASAP1. to CUL1cilia. PI4KA is an E3 is ubiquitin an important ligase, en whichzyme can for interact phototransduction with STUB1 andand protect its downstream cells from degenerationproducts can affect and apoptosis. ASAP1. CUL1 STUB1 is an is E3 another ubiquiti E3n ubiquitin-proteinligase, which can interact ligase which with STUB1 targets and misfolded protect chaperonecells from substratesdegeneration towards and apoptosis. proteasomal STUB1 degradation is another promoting E3 ubiquitin-protein neuronal survival ligase under which conditions targets ofmisfolded oxygen andchaperone glucose substrates deprivation. towards UCHL5 proteaso is a deubiquitinatingmal degradation enzyme promoting that neuronal inhibits apoptosis survival andunder plays conditions a role in of maintenance oxygen and of gl synapticucose deprivation. function. ANP32B UCHL5 is is a multifunctionala deubiquitinating protein enzyme involved that ininhibits immunomodulation, apoptosis and plays regulation a role in of maintenance transcription, of and synaptic regulation function. of apoptosis; ANP32B is it a has multifunctional pro-survival activityprotein whichinvolved correlates in immunomodulation, with its ability to inhibitregulation caspase of transcription, 3. ASAH1 is aand lysosomal regulation ceramidase, of apoptosis; whose it upregulationhas pro-survival could activity rescue which RPE fromcorrelates oxidative with stress;its ability its deficiency to inhibit leadscaspase to retinal3. ASAH1 inflammation is a lysosomal and severeceramidase, visual whose impairment. upregulation CD59A, could a membrane rescue RPE protein, from oxidative is known stress; to protect its deficiency cells from leads membrane to retinal attackinflammation complex, and hence severe suppressing visual impairment. inflammation CD59A, and a NF-membraneκB activation; protein, in is theknown retina, to protect CD59A cells can counteractfrom membrane retinal attack cell degeneration complex, hence induced suppressing by ischemia inflammation reperfusion and injury. NF-κ OGFRB activation; is a receptor, in the retina, which canCD59A regulate can counteract immune response retinal andcell reducedegeneration astroglyosis, induced a sign by inischemia many cases reperfusion of inflammation. injury. OGFR TRAFD1 is a isreceptor, an interferon- which and can LPS-inducible regulate immune protein response and a negative and reduce feedback astroglyosis, regulator ofa NF-signκ Bin signaling, many cases which of limitsinflammation. excessive TRAFD1 immune is response. an interferon- NR2E3 and is LPS-in a photoreceptor-specificducible protein and nuclear a negative receptor, feedback which regulator plays a criticalof NF-κ roleB signaling, in retinogenesis which and limits regulation excessive of theimmune genes involvedresponse. in NR2E3 phototransduction is a photoreceptor-specific in mature retina. RORnuclearβ is receptor, an orphan which receptor plays expressed a critical in role the in retina retinogenesis during embryogenesis and regulation and of necessary the genes for involved both cone in andphototransduction rod differentiation. in mature GAP43 retina. is a synaptic RORβ is protein an orphan associated receptor with expressed increased in synaptogenesis the retina during and axonalembryogenesis growth. KCTD8and necessary is the auxiliary for both GABA(B) cone and receptor rod differentiation. subunit that can GAP43 modulate is a receptorsynaptic response protein andassociated agonist-dependent with increased desensitization. synaptogenesis Altered and abundanceaxonal growth. of all theKCTD8 proteins is the indicated auxiliary in thisGABA(B) figure canreceptor affect subunit survival that and can function modulate of retinal receptor cells. response and agonist-dependent desensitization. Altered abundance of all the proteins indicated in this figure can affect survival and function of retinal cells.

We propose the following model (Figure 12) to unify the available data and to identify the questions that remain unanswered to completely explain changes in the Cyp27a1−/−Cyp46a1−/−Soat1−/− Biomolecules 2019, 9, x 17 of 22 retina. The absence of SOAT1 can increase the availability for the retina of fatty acids, which are used as the enzyme substrates, whereas the absence of CYP27A1 and CYP46A1 abolishes the production of 27-hydroxycholesterol and 24-hydroxycholesterol, which are activating ligands for LXRs. LXRs are important transcription factors that play essential roles in the regulation of cholesterol metabolism and also integration of glucose and fatty acid metabolism as well as immune and inflammatory responses [90]. Accordingly, glucose utilization and fatty acid β-oxidation could be increased in the Cyp27a1−/−Cyp46a1−/−Soat1−/− retina as a compensatory response to decreased LXR activation and/or increased fatty acid availability. This will lead to an increased energy production, which could be used to form extracellular vesicles, an energy-consuming process and alternate means to eliminate excess cellular cholesterol when the compensatory responses from cholesterol biosynthesis and efflux to LP are either not sufficient or not operative. In addition, decreased LXR activation could promote proinflammatory processes and affect retinal function. Apparently, other processes could also Biomolecules 2019 9 become affected, , 867in the Cyp27a1−/−Cyp46a1−/−Soat1−/− retina as suggested by retinal proteomics (Figure17 of 22 11). Future studies are needed to address these outstanding questions.

Figure 12. Proposed model unifying different processes in the Cyp27a1 / Cyp46a1 / Soat1 / retina. Figure 12. Proposed model unifying different processes in the Cyp27a1− −−/−Cyp46a1− −/−Soat1−−/− retina. Processes with an experimental support (in parenthesis) are in black font and those that are putative Processes with an experimental support (in parenthesis) are in black font and those that are putative are in gray font. See Discussion for explanation. 27HC and 24HC, 27- and 24-hydroxycholesterol, are in gray font. See Discussion for explanation. 27HC and 24HC, 27- and 24-hydroxycholesterol, respectively. GC-MS, gas chromatography-mass spectrometry; ERG, electroretinography; TEM, respectively. GC-MS, gas chromatography-mass spectrometry; ERG, electroretinography; TEM, transmission electron microscopy. transmission electron microscopy. 5. Conclusions 5. Conclusions / / / We comprehensively characterized the retina of Cyp27a1− −Cyp46a1− −Soat1− − and We / comprehensively/ / characterized the retina of Cyp27a1−/−Cyp46a1−/−Soat1−/− and Cyp27a1− −Cyp46a1− −Apoe− − mice and identified some of the homeostatic responses to the blockage −/− −/− −/− ofCyp27a1 the majorCyp46a1 pathwaysApoe of retinal mice cholesterol and identified output some and storage. of the homeostatic The major cholesterol-related responses to the blockage response includedof the major the upregulationpathways of of retinal the pathways cholesterol producing output intracellularand storage. and The extracellular major cholesterol-related vesicles, which seemsresponse to beincluded a reserve the mechanism upregulation for retinal of the cholesterol pathways removal. producing The non-cholesterol-related intracellular and extracellular responses pertainedvesicles, which to retinal seems function, to be a reserve photoreceptor mechanism viability, for retinal and possibly cholesterol retinal remo glucoseval. The as wellnon-cholesterol- as fatty acid metabolism.related responses The data pertained obtained to retinal provide function, insight intophotoreceptor the links between viability, retinal and possibly cholesterol retinal homeostasis glucose andas well other as importantfatty acid retinalmetabolism. processes. The Notably,data obtained the content provide of total insight retinal into cholesterol the links doesbetween not seem retinal to predictcholesterol retinal homeostasis abnormalities, and andother it isimportant rather the retinal network processes. of compensatory Notably, mechanismsthe content thatof total appears retinal to determinecholesterol thedoes retinal not phenotype. seem to predict retinal abnormalities, and it is rather the network of compensatory mechanisms that appears to determine the retinal phenotype. Supplementary Materials: The following are available online at http://www.mdpi.com/2218-273X/9/12/867/s1. SupportingSupplementary Text S1:Materials: Gene abbreviations, The following protein are available abbreviations, online andat www.mdpi.com/xxx/s1. functions of proteins indicated Supporting in the Text present T1: work;Gene Supplementalabbreviations, Tableprotein S1: Listabbrev of primersiations, forand quantitative functions real-timeof proteins PCR; indicated Supplemental in the Table present S2: Statistical work; analysesSupplemental by two-way Table S1: ANOVA List of with prim Bonferroniers for quantitative correction real-time of the serum PCR; total Supplemental cholesterol Table content S2: in Statistical different genotypes;analyses by Supplemental two-way ANOVA Table with S3: Statistical Bonferroni analyses correction by two-way of the serum ANOVA total with cholesterol Bonferroni content correction in different of the retinalgenotypes; total Supplemental cholesterol content Table inS3: di Statisticalfferent genotypes; analyses Supplementalby two-way ANOVA Table S4: with Statistical Bonferroni analyses correction by two-way of the ANOVA with Bonferroni correction of the retinal free lathosterol content in different genotypes; Supplemental Tableretinal S5. total Statistical cholesterol analyses content by in two-way differentANOVA genotypes; with Supplemental Bonferroni correctionTable S4: Statistical of the retinal analyses free desmosterolby two-way contentANOVA in with different Bonferroni genotypes. correction of the retinal free lathosterol content in different genotypes; Supplemental Table S5. Statistical analyses by two-way ANOVA with Bonferroni correction of the retinal free desmosterol Author Contributions: Conceptualization, A.M.P., N.M., I.A.P.; investigation, A.A.A., N.M., A.S., and N.E.-D.; writing,content in A.M.P., different A.A.A., genotypes. and I.A.P.; funding acquisition, I.A.P. Funding: This work was supported in part by NIH grants EY018383 and EY011373 (I.A.P.) and unrestricted grant from Research to Prevent Blindness. Irina A. Pikuleva is a Carl F. Asseff Professor of Ophthalmology. Acknowledgments: The authors thank the Visual Sciences Research Center Core Facilities (supported by National Institutes of Health Grant P30 EY11373) for assistance with mouse breeding (Heather Butler and Kathryn Franke), animal genotyping (John Denker), tissue sectioning (Catherine Doller), and microscopy (Anthony Gardella). We are also grateful to Hisashi Fujioka (Electron Microscopy Core facility) for help with studies of retinal ultrastructure, to Danie Schlatzer (Proteomics and Small Molecule Mass Spectrometry Core) for conducting retinal label free analysis, and to Neal Peachey for comments on the manuscript and assistance with the ERG studies. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Wong, W.L.; Su, X.; Li, X.; Cheung, C.M.; Klein, R.; Cheng, C.Y.; Wong, T.Y. Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: A systematic review and meta-analysis. Lancet Glob. Health 2014, 2, e106–e116. [CrossRef] Biomolecules 2019, 9, 867 18 of 22

2. Curcio, C.A.; Millican, C.L.; Bailey, T.; Kruth, H.S. Accumulation of cholesterol with age in human bruch’s membrane. Investig. Ophthalmol. Vis. Sci. 2001, 42, 265–274. [PubMed] 3. Curcio, C.A.; Presley, J.B.; Malek, G.; Medeiros, N.E.; Avery, D.V.; Kruth, H.S. Esterified and unesterified cholesterol in drusen and basal deposits of eyes with age-related maculopathy. Exp. Eye Res. 2005, 81, 731–741. [CrossRef][PubMed] 4. Oak, A.S.W.; Messinger, J.D.; Curcio, C.A. Subretinal drusenoid deposits: Further characterization by lipid histochemistry. Retina 2014, 34, 825–826. [CrossRef][PubMed] 5. Curcio, C.A. Antecedents of soft drusen, the specific deposits of age-related macular degeneration, in the biology of human macula. Investig. Ophthalmol. Vis. Sci. 2018, 59, AMD182–AMD194. [CrossRef] 6. Curcio, C.A. Soft drusen in age-related macular degeneration: Biology and targeting via the oil spill strategies. Investig. Ophthalmol. Vis. Sci. 2018, 59, AMD160–AMD181. [CrossRef] 7. Chen, W.; Stambolian, D.; Edwards, A.O.; Branham, K.E.; Othman, M.; Jakobsdottir, J.; Tosakulwong, N.; Pericak-Vance, M.A.; Campochiaro, P.A.; Klein, M.L.; et al. Genetic variants near timp3 and high-density lipoprotein-associated loci influence susceptibility to age-related macular degeneration. Proc. Natl. Acad. Sci. USA 2010, 107, 7401–7406. [CrossRef] 8. Fritsche, L.G.; Chen, W.; Schu, M.; Yaspan, B.L.; Yu, Y.; Thorleifsson, G.; Zack, D.J.; Arakawa, S.; Cipriani, V.; Ripke, S.; et al. Seven new loci associated with age-related macular degeneration. Nat. Genet. 2013, 45, 433–439. 9. Neale, B.M.; Fagerness, J.; Reynolds, R.; Sobrin, L.; Parker, M.; Raychaudhuri, S.; Tan, P.L.; Oh, E.C.; Merriam, J.E.; Souied, E.; et al. Genome-wide association study of advanced age-related macular degeneration identifies a role of the hepatic lipase gene (lipc). Proc. Natl. Acad. Sci. USA 2010, 107, 7395–7400. [CrossRef] 10. Yu, Y.; Reynolds, R.; Fagerness, J.; Rosner, B.; Daly, M.J.; Seddon, J.M. Association of variants in the lipc and abca1 genes with intermediate and large drusen and advanced age-related macular degeneration. Investig. Ophthalmol. Vis. Sci. 2011, 52, 4663–4670. [CrossRef] 11. Seddon, J.M. Macular degeneration epidemiology: Nature-nurture, lifestyle factors, genetic risk, and gene-environment interactions - the weisenfeld award lecture. Investig. Ophthalmol. Vis. Sci. 2017, 58, 6513–6528. [CrossRef] 12. Guymer, R.H.; Baird, P.N.; Varsamidis, M.; Busija, L.; Dimitrov, P.N.; Aung, K.Z.; Makeyeva, G.A.; Richardson, A.J.; Lim, L.; Robman, L.D. Proof of concept, randomized, placebo-controlled study of the effect of simvastatin on the course of age-related macular degeneration. PLoS ONE 2013, 8, e83759. [CrossRef] 13. Vavvas, D.G.; Daniels, A.B.; Kapsala, Z.G.; Goldfarb, J.W.; Ganotakis, E.; Loewenstein, J.I.; Young, L.H.; Gragoudas, E.S.; Eliott, D.; Kim, I.K.; et al. Regression of some high-risk features of age-related macular degeneration (amd) in patients receiving intensive statin treatment. EBioMedicine 2016, 5, 198–203. [CrossRef] [PubMed] 14. Albert, A.D.; Boesze-Battaglia, K. The role of cholesterol in rod outer segment membranes. Prog. Lipid. Res. 2005, 44, 99–124. [CrossRef][PubMed] 15. Tserentsoodol, N.; Gordiyenko, N.V.; Pascual, I.; Lee, J.W.; Fliesler, S.J.; Rodriguez, I.R. Intraretinal lipid transport is dependent on high density lipoprotein-like particles and class b scavenger receptors. Mol. Vis. 2006, 12, 1319–1333. [PubMed] 16. Tserentsoodol, N.; Sztein, J.; Campos, M.; Gordiyenko, N.V.; Fariss, R.N.; Lee, J.W.; Fliesler, S.J.; Rodriguez, I.R. Uptake of cholesterol by the retina occurs primarily via a low density lipoprotein receptor-mediated process. Mol. Vis. 2006, 12, 1306–1318. 17. Fliesler, S.J.; Bretillon, L. The ins and outs of cholesterol in the vertebrate retina. J. Lipid Res. 2010, 51, 3399–3413. [CrossRef] 18. Curcio, C.A.; Johnson, M.; Rudolf, M.; Huang, J.D. The oil spill in ageing bruch membrane. Br. J. Ophthalmol. 2011, 95, 1638–1645. [CrossRef] 19. Fliesler, S.J. Lipids and lipid metabolism in the eye. J. Lipid Res. 2010, 51, 1–3. [CrossRef] 20. Pikuleva, I.A.; Curcio, C.A. Cholesterol in the retina: The best is yet to come. Prog. Retin. Eye Res. 2014, 41, 64–89. [CrossRef] 21. Guo, Y.; Cordes, K.R.; Farese, R.V.; Walther, T.C. Lipid droplets at a glance. J. Cell Sci. 2009, 122, 749. [CrossRef][PubMed] Biomolecules 2019, 9, 867 19 of 22

22. Saadane, A.; Mast, N.; Dao, T.; Ahmad, B.; Pikuleva, I.A. Retinal hypercholesterolemia triggers cholesterol accumulation and esterification in photoreceptor cells. J. Biol. Chem. 2016, 291, 20427–20439. [CrossRef] [PubMed] 23. Lin, J.B.; Mast, N.; Bederman, I.R.; Li, Y.; Brunengraber, H.; Bjorkhem, I.; Pikuleva, I.A. Cholesterol in mouse retina originates primarily from in situ de novo biosynthesis. J. Lipid Res. 2016, 57, 258–264. [CrossRef] [PubMed] 24. Bretillon, L.; Thuret, G.; Gregoire, S.; Acar, N.; Joffre, C.; Bron, A.M.; Gain, P.; Creuzot-Garcher, C.P. Lipid and fatty acid profile of the retina, retinal pigment epithelium/choroid, and the lacrimal gland, and associations with adipose tissue fatty acids in human subjects. Exp. Eye Res. 2008, 87, 521–528. [CrossRef][PubMed] 25. Mast, N.; Reem, R.; Bederman, I.; Huang, S.; DiPatre, P.L.; Björkhem, I.; Pikuleva, I.A. Cholestenoic acid is an important elimination product of cholesterol in the retina: Comparison of retinal cholesterol metabolism with that in the brain. Investig. Ophthalmol. Vis. Sci. 2011, 52, 594–603. [CrossRef] 26. Saadane, A.; Mast, N.; Charvet, C.D.; Omarova, S.; Zheng, W.; Huang, S.S.; Kern, T.S.; Peachey, N.S.; Pikuleva, I.A. Retinal and nonocular abnormalities in cyp27a1(-/-)cyp46a1(-/-) mice with dysfunctional metabolism of cholesterol. Am. J. Pathol. 2014, 184, 2403–2419. [CrossRef] 27. Saadane, A.; Petrov, A.; Mast, N.; El-Darzi, N.; Dao, T.; Alnemri, A.; Song, Y.; Dunaief, J.L.; Pikuleva, I.A. Mechanisms that minimize retinal impact of apolipoprotein e absence. J. Lipid Res. 2018, 59, 2368–2382. [CrossRef] 28. Meiner, V.L.; Cases, S.; Myers, H.M.; Sande, E.R.; Bellosta, S.; Schambelan, M.; Pitas, R.E.; McGuire, J.; Herz, J.; Farese, R.V., Jr. Disruption of the acyl-coa:Cholesterol acyltransferase gene in mice: Evidence suggesting multiple cholesterol esterification enzymes in mammals. Proc. Natl. Acad. Sci. USA 1996, 93, 14041–14046. [CrossRef] 29. Omarova, S.; Charvet, C.D.; Reem, R.E.; Mast, N.; Zheng, W.; Huang, S.; Peachey, N.S.; Pikuleva, I.A. Abnormal vascularization in mouse retina with dysregulated retinal cholesterol homeostasis. J. Clin. Investig. 2012, 122, 3012–3023. [CrossRef] 30. Saadane, A.; Mast, N.; Trichonas, G.; Chakraborty, D.; Hammer, S.; Busik, J.V.; Grant, M.B.; Pikuleva, I.A. Retinal vascular abnormalities and microglia activation in mice with deficiency in cytochrome p450 46a1-mediated cholesterol removal. Am. J. Pathol. 2019, 189, 405–425. [CrossRef] 31. Pfaffl, M.W. A new mathematical model for relative quantification in real-time rt-pcr. Nucleic Acids Res. 2001, 29, e45. [CrossRef][PubMed] 32. Zhang, J.; Xin, L.; Shan, B.; Chen, W.; Xie, M.; Yuen, D.; Zhang, W.; Zhang, Z.; Lajoie, G.A.; Ma, B. Peaks db: De novo sequencing assisted database search for sensitive and accurate peptide identification. Mol. Cell. Proteomics 2012, 11, M111.010587. [CrossRef][PubMed] 33. Guyton, J.R.; Klemp, K.F. Ultrastructural discrimination of lipid droplets and vesicles in atherosclerosis: Value of osmium-thiocarbohydrazide-osmium and tannic acid-paraphenylenediamine techniques. J. Histochem. Cytochem. 1988, 36, 1319–1328. [CrossRef][PubMed] 34. Pfrieger, F.W.; Ungerer, N. Cholesterol metabolism in neurons and astrocytes. Prog. Lipid Res. 2011, 50, 357–371. [CrossRef][PubMed] 35. Mast, N.; Bederman, I.R.; Pikuleva, I.A. Retinal cholesterol content is reduced in simvastatin-treated mice due to inhibited local biosynthesis albeit increased uptake of serum cholesterol. Drug Metab. Dispos. 2018, 46, 1528–1537. [CrossRef][PubMed] 36. Brown, M.S.; Goldstein, J.L. Cholesterol feedback: From schoenheimer’s bottle to scap’s meladl. J. Lipid Res. 2009, 50, S15–S27. [CrossRef] 37. Zelcer, N.; Tontonoz, P. Liver x receptors as integrators of metabolic and inflammatory signaling. J. Clin. Investig. 2006, 116, 607–614. [CrossRef] 38. Ferrington, D.A.; Sinha, D.; Kaarniranta, K. Defects in retinal pigment epithelial cell proteolysis and the pathology associated with age-related macular degeneration. Progr. Ret. Eye Res. 2016, 51, 69–89. [CrossRef] 39. Zheng, W.; Mast, N.; Saadane, A.; Pikuleva, I.A. Pathways of cholesterol homeostasis in mouse retina responsive to dietary and pharmacologic treatments. J. Lipid Res. 2015, 56, 81–97. [CrossRef] 40. Zheng, W.; Reem, R.E.; Omarova, S.; Huang, S.; DiPatre, P.L.; Charvet, C.D.; Curcio, C.A.; Pikuleva, I.A. Spatial distribution of the pathways of cholesterol homeostasis in human retina. PLoS ONE 2012, 7, e37926. [CrossRef] Biomolecules 2019, 9, 867 20 of 22

41. Kim, T.; Loh, Y.P. Chromogranin a: A surprising link between granule biogenesis and hypertension. J. Clin. Investig. 2005, 115, 1711–1713. [CrossRef][PubMed] 42. Zhang, X.; Yuan, Z.; Zhang, Y.; Yong, S.; Salas-Burgos, A.; Koomen, J.; Olashaw, N.; Parsons, J.T.; Yang, X.J.; Dent, S.R.; et al. Hdac6 modulates cell motility by altering the acetylation level of cortactin. Mol. Cell 2007, 27, 197–213. [CrossRef][PubMed] 43. Bechler, M.E.; Doody, A.M.; Ha, K.D.; Judson, B.L.; Chen, I.; Brown, W.J. The phospholipase a(2) enzyme complex pafah ib mediates endosomal membrane tubule formation and trafficking. Mol. Biol. Cell 2011, 22, 2348–2359. [CrossRef][PubMed] 44. Bechler, M.E.; de Figueiredo, P.; Brown, W.J. A pla1-2 punch regulates the golgi complex. Trends Cell. Biol. 2012, 22, 116–124. [CrossRef][PubMed] 45. Traikov, S.; Stange, C.; Wassmer, T.; Paul-Gilloteaux, P.; Salamero, J.; Raposo, G.; Hoflack, B. Septin6 and septin7 gtp binding proteins regulate ap-3- and escrt-dependent multivesicular body biogenesis. PLoS ONE 2014, 9, e109372. [CrossRef] 46. Hu, J.; Bai, X.; Bowen, J.R.; Dolat, L.; Korobova, F.; Yu, W.; Baas, P.W.; Svitkina, T.; Gallo, G.; Spiliotis, E.T. Septin-driven coordination of actin and microtubule remodeling regulates the collateral branching of axons. Curr. Biol. 2012, 22, 1109–1115. [CrossRef] 47. Hainzl, T.; Huang, S.; Sauer-Eriksson, A.E. Structure of the srp19 rna complex and implications for signal recognition particle assembly. Nature 2002, 417, 767–771. [CrossRef] 48. Kim, H.R.; Kwon, M.S.; Lee, S.; Mun, Y.; Lee, K.S.; Kim, C.H.; Na, B.R.; Kim, B.N.R.; Piragyte, I.; Lee, H.S.; et al. Tagln2 polymerizes g-actin in a low ionic state but blocks arp2/3-nucleated actin branching in physiological conditions. Sci. Rep. 2018, 8, 5503. [CrossRef] 49. Na, B.R.; Kwon, M.S.; Chae, M.W.; Kim, H.R.; Kim, C.H.; Jun, C.D.; Park, Z.Y. Transgelin-2 in b-cells controls t-cell activation by stabilizing t cell - b cell conjugates. PLoS ONE 2016, 11, e0156429. [CrossRef] 50. de las Heras, R.; Depaz, I.; Jaquet, V.; Kroon, P.; Wilce, P.A. Neuronal protein 22 colocalises with both the microtubule and microfilament cytoskeleton in neurite-like processes. Brain Res. 2007, 1128, 12–20. [CrossRef] 51. Shah, N.; Ishii, M.; Brandon, C.; Ablonczy, Z.; Cai, J.; Liu, Y.; Chou, C.J.; Rohrer, B. Extracellular vesicle-mediated long-range communication in stressed retinal pigment epithelial cell monolayers. Biochim. Biophys. Acta Mol. Basis. Dis. 2018, 1864, 2610–2622. [CrossRef][PubMed] 52. Pfrieger, F.W.; Vitale, N. Cholesterol and the journey of extracellular vesicles. J. Lipid Res. 2018, 59, 2255–2261. [CrossRef][PubMed] 53. Stachowiak, J.C.; Brodsky, F.M.; Miller, E.A. A cost-benefit analysis of the physical mechanisms of membrane curvature. Nat. Cell Biol. 2013, 15, 1019–1027. [CrossRef][PubMed] 54. Yang, S.Y.; He, X.Y.; Schulz, H. Fatty acid oxidation in rat brain is limited by the low activity of 3-ketoacyl-coenzyme a thiolase. J. Biol. Chem. 1987, 262, 13027–13032. [PubMed] 55. Li, C.M.; Park, J.H.; He, X.; Levy, B.; Chen, F.; Arai, K.; Adler, D.A.; Disteche, C.M.; Koch, J.; Sandhoff, K.; et al. The human acid ceramidase gene (asah): Structure, chromosomal location, mutation analysis, and expression. Genomics 1999, 62, 223–231. [CrossRef][PubMed] 56. Adeva-Andany, M.M.; Gonzalez-Lucan, M.; Donapetry-Garcia, C.; Fernandez-Fernandez, C.; Ameneiros-Rodriguez, E. Glycogen metabolism in humans. BBA Clin. 2016, 5, 85–100. [CrossRef] 57. Beerens, K.; Soetaert, W.; Desmet, T. Udp-hexose 4-epimerases: A view on structure, mechanism and substrate specificity. Carbohydr. Res. 2015, 414, 8–14. [CrossRef] 58. Stacpoole, P.W. The pyruvate dehydrogenase complex as a therapeutic target for age-related diseases. Aging Cell 2012, 11, 371–377. [CrossRef] 59. Calzia, D.; Panfoli, I.; Heinig, N.; Schumann, U.; Ader, M.; Traverso, C.E.; Funk, R.H.; Roehlecke, C. Impairment of extramitochondrial oxidative phosphorylation in mouse rod outer segments by blue light irradiation. Biochimie 2016, 125, 171–178. [CrossRef] 60. Bruschi, M.; Petretto, A.; Caicci, F.; Bartolucci, M.; Calzia, D.; Santucci, L.; Manni, L.; Ramenghi, L.A.; Ghiggeri, G.; Traverso, C.E.; et al. Proteome of bovine mitochondria and rod outer segment disks: Commonalities and differences. J. Proteome Res. 2018, 17, 918–925. [CrossRef] 61. Zaborowski, M.P.; Balaj, L.; Breakefield, X.O.; Lai, C.P. Extracellular vesicles: Composition, biological relevance, and methods of study. Bioscience 2015, 65, 783–797. [CrossRef][PubMed] Biomolecules 2019, 9, 867 21 of 22

62. Bar-Peled, L.; Sabatini, D.M. Regulation of mtorc1 by amino acids. Trends Cell. Biol. 2014, 24, 400–406. [CrossRef][PubMed] 63. El-Hattab, A.W. Serine biosynthesis and transport defects. Mol. Genet. Metab. 2016, 118, 153–159. [CrossRef] [PubMed] 64. Marsillach, J.; Suzuki, S.M.; Richter, R.J.; McDonald, M.G.; Rademacher, P.M.; MacCoss, M.J.; Hsieh, E.J.; Rettie, A.E.; Furlong, C.E. Human valacyclovir hydrolase/biphenyl hydrolase-like protein is a highly efficient homocysteine thiolactonase. PLoS ONE 2014, 9, e110054. [CrossRef][PubMed] 65. Shayakul, C.; Clemencon, B.; Hediger, M.A. The urea transporter family (slc14): Physiological, pathological and structural aspects. Mol. Aspects Med. 2013, 34, 313–322. [CrossRef][PubMed] 66. Pugacheva, E.N.; Jablonski, S.A.; Hartman, T.R.; Henske, E.P.; Golemis, E.A. Hef1-dependent aurora a activation induces disassembly of the primary cilium. Cell 2007, 129, 1351–1363. [CrossRef][PubMed] 67. Ran, J.; Yang, Y.; Li, D.; Liu, M.; Zhou, J. Deacetylation of alpha-tubulin and cortactin is required for hdac6 to trigger ciliary disassembly. Sci. Rep. 2015, 5, 12917. [CrossRef] 68. Wang, J.; Morita, Y.; Mazelova, J.; Deretic, D. The arf gap asap1 provides a platform to regulate arf4- and rab11-rab8-mediated ciliary receptor targeting. EMBO J. 2012, 31, 4057–4071. [CrossRef] 69. Wang, J.; Deretic, D. The arf and rab11 effector fip3 acts synergistically with asap1 to direct rabin8 in ciliary receptor targeting. J. Cell. Sci. 2015, 128, 1375–1385. [CrossRef] 70. Chen, Z.S.; Wong, A.K.Y.; Cheng, T.C.; Koon, A.C.; Chan, H.Y.E.Fipoq/fbxo33, a cullin-1-based ubiquitin ligase complex component modulates ubiquitination and solubility of polyglutamine disease protein. J. Neurochem. 2019, 149, 781–798. [CrossRef] 71. Balakrishnan, S.S.; Basu, U.; Shinde, D.; Thakur, R.; Jaiswal, M.; Raghu, P. Regulation of pi4p levels by pi4kiiialpha during g-protein-coupled plc signaling in drosophila photoreceptors. J. Cell. Sci. 2018, 131. [CrossRef][PubMed] 72. Weixel, K.M.; Blumental-Perry, A.; Watkins, S.C.; Aridor, M.; Weisz, O.A. Distinct golgi populations of phosphatidylinositol 4-phosphate regulated by phosphatidylinositol 4-kinases. J. Biol. Chem. 2005, 280, 10501–10508. [CrossRef][PubMed] 73. Al-Ramahi, I.; Lam, Y.C.; Chen, H.K.; de Gouyon, B.; Zhang, M.; Perez, A.M.; Branco, J.; de Haro, M.; Patterson, C.; Zoghbi, H.Y.; et al. Chip protects from the neurotoxicity of expanded and wild-type ataxin-1 and promotes their ubiquitination and degradation. J. Biol. Chem. 2006, 281, 26714–26724. [CrossRef] [PubMed] 74. Kowalski, J.R.; Juo, P. The role of deubiquitinating enzymes in synaptic function and nervous system diseases. Neural Plast. 2012, 2012, 892749. [CrossRef][PubMed] 75. Haider, N.B.; Mollema, N.; Gaule, M.; Yuan, Y.; Sachs, A.J.; Nystuen, A.M.; Naggert, J.K.; Nishina, P.M. Nr2e3-directed transcriptional regulation of genes involved in photoreceptor development and cell-type specific phototransduction. Exp. Eye Res. 2009, 89, 365–372. [CrossRef][PubMed] 76. Jia, L.; Oh, E.C.; Ng, L.; Srinivas, M.; Brooks, M.; Swaroop, A.; Forrest, D. Retinoid-related orphan nuclear receptor rorbeta is an early-acting factor in rod photoreceptor development. Proc. Natl. Acad. Sci. USA 2009, 106, 17534–17539. [CrossRef][PubMed] 77. Formoso, K.; Garcia, M.D.; Frasch, A.C.; Scorticati, C. Evidence for a role of glycoprotein m6a in dendritic spine formation and synaptogenesis. Mol. Cell. Neurosci. 2016, 77, 95–104. [CrossRef] 78. Kang, H.J.; Voleti, B.; Hajszan, T.; Rajkowska, G.; Stockmeier, C.A.; Licznerski, P.; Lepack, A.; Majik, M.S.; Jeong, L.S.; Banasr, M.; et al. Decreased expression of synapse-related genes and loss of synapses in major depressive disorder. Nat. Med. 2012, 18, 1413–1417. [CrossRef] 79. Rajalu, M.; Fritzius, T.; Adelfinger, L.; Jacquier, V.; Besseyrias, V.; Gassmann, M.; Bettler, B. Pharmacological characterization of gabab receptor subtypes assembled with auxiliary kctd subunits. Neuropharmacology 2015, 88, 145–154. [CrossRef] 80. Holahan, M.R. A shift from a pivotal to supporting role for the growth-associated protein (gap-43) in the coordination of axonal structural and functional plasticity. Front. Cell. Neurosci. 2017, 11, 266. [CrossRef] 81. Dabo, S.; Maillard, P.; Collados Rodriguez, M.; Hansen, M.D.; Mazouz, S.; Bigot, D.J.; Tible, M.; Janvier, G.; Helynck, O.; Cassonnet, P.; et al. Inhibition of the inflammatory response to stress by targeting interaction between pkr and its cellular activator pact. Sci. Rep. 2017, 7, 16129. [CrossRef][PubMed] 82. Hu, Y.H.; Zhang, Y.; Jiang, L.Q.; Wang, S.; Lei, C.Q.; Sun, M.S.; Shu, H.B.; Liu, Y. Wdfy1 mediates tlr3/4 signaling by recruiting trif. EMBO Rep. 2015, 16, 447–455. [CrossRef][PubMed] Biomolecules 2019, 9, 867 22 of 22

83. Zhou, T.; Huang, Z.; Zhu, X.; Sun, X.; Liu, Y.; Cheng, B.; Li, M.; Liu, Y.; He, C.; Liu, X. Alpha-1 antitrypsin attenuates m1 microglia-mediated neuroinflammation in retinal degeneration. Front. Immunol. 2018, 9, 1202. [CrossRef][PubMed] 84. Sun, W.; Kimura, H.; Hattori, N.; Tanaka, S.; Matsuyama, S.; Shiota, K. Proliferation related acidic leucine-rich protein pal31 functions as a caspase-3 inhibitor. Biochem. Biophys. Res. Commun. 2006, 342, 817–823. [CrossRef][PubMed] 85. Chemnitz, J.; Pieper, D.; Stich, L.; Schumacher, U.; Balabanov, S.; Spohn, M.; Grundhoff, A.; Steinkasserer, A.; Hauber, J.; Zinser, E. The acidic protein rich in leucines anp32b is an immunomodulator of inflammation in mice. Sci. Rep. 2019, 9, 4853. [CrossRef] 86. Yu, F.P.S.; Sajdak, B.S.; Sikora, J.; Salmon, A.E.; Nagree, M.S.; Gurka, J.; Kassem, I.S.; Lipinski, D.M.; Carroll, J.; Medin, J.A. Acid ceramidase deficiency in mice leads to severe ocular pathology and visual impairment. Am. J. Pathol. 2019, 189, 320–338. [CrossRef] 87. Herrmann, P.; Cowing, J.A.; Cristante, E.; Liyanage, S.E.; Ribeiro, J.; Duran, Y.; Abelleira Hervas, L.; Carvalho, L.S.; Bainbridge, J.W.; Luhmann, U.F.; et al. Cd59a deficiency in mice leads to preferential innate immune activation in the retinal pigment epithelium-choroid with age. Neurobiol. Aging 2015, 36, 2637–2648. [CrossRef] 88. Campbell, A.M.; Zagon, I.S.; McLaughlin, P.J. Astrocyte proliferation is regulated by the ogf-ogfr axis in vitro and in experimental autoimmune encephalomyelitis. Brain Res. Bull. 2013, 90, 43–51. [CrossRef] 89. Mashima, R.; Saeki, K.; Aki, D.; Minoda, Y.; Takaki, H.; Sanada, T.; Kobayashi, T.; Aburatani, H.; Yamanashi, Y.; Yoshimura, A. Fln29, a novel interferon- and lps-inducible gene acting as a negative regulator of toll-like receptor signaling. J. Biol. Chem. 2005, 280, 41289–41297. [CrossRef] 90. Calkin, A.C.; Tontonoz, P. Transcriptional integration of metabolism by the nuclear sterol-activated receptors lxr and fxr. Nat. Rev. Mol. Cell. Biol. 2012, 13, 213–224. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).