Multiple Cellular Transport and Binding Processes of Unesterified Docosahexaenoic Acid in Outer Blood–Retinal Barrier Retinal Pigment Epithelial Cells

Abstract

Docosahexaenoic acid (DHA, 22 : 6) is an essential omega-3 long-chain polyunsaturated fatty acid that plays a pivotal role in vision. The purpose of this study was to clarify the cellular uptake and binding processes of free and protein-bound unesterified DHA in retinal pigment epithelial cell (RPE) line ARPE-19 as a model of the human outer blood–retinal barrier and isolated porcine RPE cell fractions. Uptake of free [¹⁴C]DHA by ARPE-19 cells was saturable with a Michaelis–Menten constant of 283 µM, and was significantly inhibited by eicosapentaenoic acid, arachidonic acid, and linoleic acid, but not by oleic acid. Further, the uptakes of [¹⁴C]DHA associated with retinol-binding protein ([¹⁴C]DHA-RBP), [¹⁴C]DHA associated with low-density lipoprotein ([¹⁴C]DHA-LDL) and [¹⁴C]DHA associated with bovine serum albumin ([¹⁴C]DHA-BSA) in ARPE-19 cells increased time-dependently at 37°C, and were significantly reduced at 4°C, suggesting the involvement of energy-dependent transport processes. [¹⁴C]DHA-LDL uptake by ARPE-19 cells was significantly inhibited by excess unlabeled LDL, but not by an inhibitor of scavenger receptor B type I. Fatty acid transport protein (FATP) 2 and 4 mRNAs were expressed in ARPE-19 cells, and [¹⁴C]DHA uptake was observed in FATP2- and FATP4-expressing Xenopus oocytes. Photo-reactive crosslinking and mass spectrometry analyses identified 65-kDa retinal pigment epithelium-specific protein (RPE65) as a DHA-binding protein in porcine RPE cell membrane fractions. Thus, RPE cells possess multiple cellular transport/binding processes for unesterified DHA, involving at least partly FATP2, FATP4, LDL, RBP, and RPE65.

Docosahexaenoic acid (DHA, 22 : 6) is an omega-3 long-chain polyunsaturated fatty acid (PUFA) that plays an essential role in visual function,¹⁾ serving to regulate photo-signal transduction via rhodopsin.²⁾ It has also been reported that 15-lipoxgenase-1 converts DHA to neuroprotectin D1, which functions to protect retinal pigment epithelial (RPE) cells from oxidative stress.³⁾ Therefore, tight control of the DHA level is essential in retinal tissue, especially in photoreceptor cells.

DHA is enriched in phospholipids of photoreceptor outer segment disk membranes in rats^4,5) and accounts for 80% of PUFAs in photoreceptor outer segments and 30–40% of total fatty acids in the retina.⁶⁾ [¹⁴C]DHA was hardly detected in the retina after oral administration of [¹⁴C]α-linolenic acid in adult rats,⁷⁾ although RPE cells can synthesize DHA from α-linolenic acid⁸⁾ in the frog. This strongly suggests that DHA is delivered from the circulating blood to the retina to meet the high demand in the photoreceptor outer segments in rats. In support of this notion, several reports have demonstrated that (i) generational dietary DHA deficiency resulted in eye DHA deficiency in rodents,⁹⁾ (ii) lack of DHA supplementation caused anatomical and functional abnormalities of vision, and (iii) DHA supplementation ameliorated certain visual processing deficits.^10,11) It is thus conceivable that the retina possesses a specific transport system(s) for DHA, which may be in part distinct from the system(s) for other PUFAs.

The outer blood–retinal barrier (outer BRB), which consists of RPE cells linked by tight junctions, expresses various transport systems that regulate the supply of nutrients and drugs from the choroidal blood flow to the photoreceptor cells.¹²⁾ The major facilitator superfamily domain-containing protein 2a (Mfsd2a)¹³⁾ and the adiponectin receptor 1 (ADIPOR1)¹⁴⁾ have been proposed to mediate this transport. It was confirmed that Mfsd2a mediates the transport of esterified DHA as well as other long-chain fatty acids, e.g., oleic acid (OA) and palmitic acid (PA), as a chemical form of lysophosphatidylcholine (LPC), but does not transport unesterified fatty acids.¹⁵⁾ Further, it functions as the primary transporter in RPE cells, which supply DHA to the photoreceptor outer segments.¹³⁾ On the other hand, DHA is not completely lost in the eyes of Mfsd2a-knockout mice,¹³⁾ implying that there may be alternative mechanisms for DHA uptake into the retina. Since unesterified DHA and phospholipid-esterified DHA in the plasma account for 1.6 and 2.2%, respectively, of total plasma fatty acids,¹⁶⁾ it seems plausible that there would be a transport system(s) for unesterified DHA at the outer BRB.

Cellular transport of unesterified PUFAs involves fatty acid-binding protein plasma membrane (FABPpm), intracellular fatty acid binding proteins (FABPs), fatty acid translocase (FAT/CD36), and fatty acid transport proteins (FATPs) in mammalian cells.¹⁷⁾ We have reported that FATP1 transports unesterified DHA in human brain microvessel endothelial cells (hCMEC/D3 cells), an in vitro blood–brain barrier (BBB) model, and the transport is competitively inhibited by OA.¹⁸⁾ It has also been suggested that FATP1 and/or FATP4 are involved in placental transfer of long-chain PUFAs.¹⁹⁾ Furthermore, more than 90% of unesterified fatty acids in the plasma are bound with albumin and incorporated into high-density lipoprotein (HDL) and low-density lipoprotein (LDL).²⁰⁾ We previously showed that HDL-associated α-tocopherol (vitamin E) is transported by scavenger receptor class B type I, which is likely to recognize HDL.²¹⁾ It has also been reported that vitamin A associated with retinol-binding protein (RBP) is transported by the RBP receptor stimulated by retinoic acid 6 (STRA6).²²⁾ These results suggest the presence of the multiple transport and binding pathways at the outer BRB for the retinal transfer of DHA as either free unesterified DHA or protein-associated DHA.

The purpose of this study was to investigate the characteristics of the multiple DHA transport/binding processes at the outer BRB using a cultured human RPE cell line (ARPE-19) as an in vitro model and freshly isolated porcine RPE cell membrane fractions.

MATERIALS AND METHODS

Reagents

4,7,10,13,16,19-[1-¹⁴C]Docosahexaenoic acid ([¹⁴C]DHA,51 mCi/mmol) was obtained from Moravek Biochemicals (Brea, CA, U.S.A.). Albumin-, LDL-, or RBP-associated [¹⁴C]DHA was generated by the incubation of ethanolic DHA with bovine serum albumin (BSA, Sigma-Aldrich, St. Louis, MO, U.S.A.), human plasma-derived LDL (Merck, Darmstadt, Germany), or human urine-derived RBP (Sigma-Aldrich), respectively, in extracellular fluid (ECF) buffer containing NaCl (122 mM), NaHCO₃ (25 mM), KCl (3 mM), CaCl₂ (1.4 mM), MgSO₄ (1.2 mM), K₂HPO₄ (0.4 mM), and N-(2-hydroxyethyl)piperazine-N-2-ethanesulfonic acid (HEPES) (10 mM), pH 7.4, at 37°C for 3 h. The protein-associated [¹⁴C]DHA was purified by size-exclusion chromatography (PD-10 column, GE Healthcare, Little Chalfont, U.K.).

Uptake Study of Free [¹⁴C]DHA, and BSA-, LDL-, or RBP-Associated [¹⁴C]DHA in ARPE-19 Cells

ARPE-19 cells (American Type Culture Collection, Manassas, VA, U.S.A.), a human retinal pigment epithelial cell line, were cultured as previously reported.²³⁾ Uptake of each compound by ARPE-19 cells and inhibitory effects on the uptake were evaluated according to the previously reported methods.²³⁾ To examine the concentration dependence of the uptake of free DHA in ARPE-19 cells, DHA was dissolved in the ECF buffer (containing no protein) at concentrations of 40–1000 µM at 37°C. According to the product information (Cayman Chemical, Ann Arbor, MI, U.S.A.), the solubility of DHA is 1 mg/mL (ca. 3 mM) in a solution of 0.15 M Tris–HCl (pH 8.5). Thus, it seems likely that DHA can be dissolved in the aqueous ECF uptake buffer to at least the maximal concentration of 1 mM used in the present study.

Kinetic Analyses

The following equation was used to estimate kinetic parameters of free DHA uptake by ARPE-19 cells:   

where V and V_max are the initial velocity and maximum velocity of DHA uptake, respectively, S is the DHA concentration in the medium, and K_m is the Michaelis–Menten constant. The data were fitted by iterative non-linear least-squares regression analysis using the program MULTI.²⁴⁾

RT-PCR Analysis

RT-PCR analysis was performed as reported previously²³⁾ using specific primers for human FATP1-6 and β-actin (Table 1). The program for PCR was 30 cycles of 94°C for 30 s, 60°C for 30 s, 72°C for 1 min, and a final period of 72°C for 10 min.

Table 1. Oligonucleotide Primers for PCR Amplification of Human FATP1-6 and β-Actin cDNAs

	Upstream primer (5′ to 3′)	Downstream primer (5′ to 3′)	Accession Number	Expected size of PCR product (bp)
FATP1	accatcccgcgcatctttcaggc	cgccgccaccatttctcctccaa	NM_198580	336 bp
FATP2	cgccagacgccacacaagccttt	cagggacttcgcgcggatcttgt	NM_003645	240 bp
FATP3	cattggggccacagtggtgct	aagcgccgcacaaaacgctccca	NM_024330	216 bp
FATP4	tgctgcatggcatgacggtggtg	tggggtatgtggaagcggctgga	NM_005094	232 bp
FATP5	tgcggcaatcgggcgacgtttac	tagctgcacagcagccatgccca	NM_012254	236 bp
FATP6	ggctgcgtggtggcctttctcaa	acaacatggtggctgcgtggca	NM_014031	251 bp
β-Actin	tcatgaagtgtgacgtggacatccgc	cctagaagcatttgcggtggacgatg	NM_001101	285 bp

Uptake Study of BSA- or RBP-Associated [¹⁴C]DHA in FATP2-, or FATP4-Expressing Oocytes

T7 RNA polymerase was used to transcribe the capped cRNA from NotI-linearized pGEM-HEN into which an open reading frame of FATP2 (GenBank accession number NM_017206) and FATP4 (GenBank accession number NM_021594) cDNA had been inserted. The open reading frames were amplified by RT-PCR from ARPE-19 cells using the sense primers 5′-cgggatccatgctttccgccatctacacagtc-3′ for FATP2 and 5′-cgggatccatgctgcttggagcctctctggtgg-3′ for FATP4 (BamHI site underlined) and the antisense primers 5′-cggaattccctcctgggaatattcagagtttcagg-3′ for FATP2 and 5′-ggaattccacagcttctcctcgcctgcctggatg-3′ for FATP4 (XhoI site underlined). The uptake of BSA- or RBP-associated [¹⁴C]DHA (14 nM) by the capped cRNA- or water-injected Xenopus laevis oocytes was examined as reported previously.²⁵⁾ The uptake amount of [¹⁴C]DHA in the oocytes was evaluated as the oocyte-to-medium (oocyte/medium) ratio (µL/oocyte) as reported previously.²⁵⁾

Chemical Synthesis of Photoreactive DHA Analogue Bearing a Diazirine Moiety as a Photophore (Fig. 1)

Fig. 1. Structure of Photoreactive DHA Probe

The probe consists of DHA, a diazirine moiety as a photophore, and a biotin tag for protein purification.

Chemical synthesis of a photoreactive DHA analogue was performed according to the previously reported method, using DHA in place of palmitic acid.²⁶⁾ In brief, a phenylalanine analogue was used to introduce a diazirine photophore at the carboxyl terminus of DHA as shown in Supplementary Materials and Methods. The coupling reaction with DHA and the subsequent biotinylation were performed as reported previously.²⁶⁾

Photoaffinity Crosslinking of DHA with RBP or Membrane Fraction of Porcine RPE Cells

The photoreactive DHA probe was dissolved in CHCl₃ to prepare a stock solution (200 µM). The solvent was removed under a nitrogen gas stream and replaced with binding buffer containing 50 mM Tris–HCl (pH 7.5), 150 mM NaCl, 1 mM CaCl₂, and 0.5% ethanol. Porcine eye balls were purchased from the meat processing plant. After removing the retinas, RPE cell layers were peeled out from the residual eye bolls and the cell membrane fractions were prepared as reported previously.²⁷⁾ RBP or the cell membrane fraction was incubated with the binding solution containing 20 µM DHA probe (200 µL) at 4°C for 3 h in the presence or absence of unlabeled DHA, retinol, or OA (1 mM). The samples were UV-irradiated (30 W/m² at 360 nm) for 30 min at 4°C, and mixed with the buffer containing 200 µM Tris–HCl, 9% weight per volume (w/v) sodium dodecyl sulfate (SDS), 30% volume per volume (v/v) glycerol, 0.15%w/v bromophenol blue, and 1 mM dithiothreitol. The labeled proteins were separated by SDS-polyacrylamide gel electrophoresis (PAGE) and detected with streptavidin-conjugated horseradish peroxidase.

Identification of Photoaffinity-Labeled Protein in Membrane Fraction of Porcine RPE Cells by Matrix Assisted Laser Desorption/Ionization-Time-of-Flight Mass Spectrometer (MALDI-TOF/MS) Analysis

After UV irradiation as described above, the solution containing the membrane fraction of porcine RPE cells was centrifuged at 100000×g for 1 h at 4°C. The pellet was dissolved in the solubilization buffer containing 50 mM Tris–HCl (pH 7.5), 1 mM CaCl₂, 150 mM NaCl, 1% Triton X, and centrifuged at 100000×g for 1 h at 4°C. The supernatant was incubated with binding buffer-equilibrated avidin agarose at 4°C for 12 h. The avidin agarose was washed with solubilization buffer and wash buffer containing 50 mM Tris–HCl, 500 mM NaCl, 1 mM CaCl₂, 1% Triton X, and 1% NP-40. The avidin agarose-associated proteins were eluted by incubation with sample buffer containing 67 µM Tris–HCl (pH 7.5), 6% SDS, 20% glycerol, 0.1% bromophenol blue, and 1 mM dithiothreitol at 95°C for 5 min. The purified proteins were separated by SDS-PAGE and visualized by CBB staining. The bands were excised from the running gel and incubated with de-staining solution containing 50 mM NH₄HCO₃, 30% acetonitrile three times, each for 20 min. The gel was vacuum-dried and incubated with reduction solution containing 10 mM ethylenediaminetetraacetic acid (EDTA), 10 mM dithiothreitol, 50 mM NH₄HCO₃ at 65°C for 1 h, followed by incubation with alkylation solution containing 10 mM EDTA, 40 mM iodoacetamide, 50 mM NH₄HCO₃ at 18°C for 45 min. The gel was vacuum-dried and subjected to in gel digestion with sequencing-grade modified trypsin dissolved in 50 mM NH₄HCO₃ at 37°C for 16 h. The digested peptides were eluted from the gel by incubation with the elution solution containing 0.1% trifluoroacetic acid and 50% acetonitrile for 30 min, followed by sonication for 10 min. The peptide solution was concentrated by vacuum-drying and desalted with Zip TipC18 (Millipore). The peptides were eluted in 33% acetonitrile containing 0.06% trifluoroacetic acid. The samples were mixed with the matrix, α-cyano-4-hydroxycinnamic acid (CCA), in 33.3% acetonitrile containing 0.06% trifluoroacetic acid, and analyzed by MALDI-TOF/MS (Autoflex-T1; Bruker Daltonics, Billerica, MA, U.S.A.). The peak data were analyzed using Flexcontrol (Bruker Daltonics) and the proteins were identified by MASCOT (Matrix Science, London, U.K.).

Statistical Analysis

The kinetic parameters are shown as mean±standard deviation (S.D.), and other data are shown as the mean±standard error of the mean (S.E.M.) Statistical significance of differences between two groups was evaluated by use of an unpaired two-tailed Student’s t-test. One-way ANOVA followed by the modified Fisher’s least-squares difference method was used to determine statistical significance among more than two groups.

RESULTS

Characterization of Unbound [¹⁴C]DHA Uptake by ARPE-19 Cells

To characterize [¹⁴C]DHA transport in RPE cells, we first evaluated the uptake of unbound [¹⁴C]DHA by ARPE-19 cells. As shown in Fig. 2A, [¹⁴C]DHA uptake by ARPE-19 cells increased linearly for 10 min at 37°C, and the initial uptake rate was 17.6 µL/(min·mg protein). The [¹⁴C]DHA uptake by ARPE-19 cells was saturable with a Km of 283±35 µM and a Vmax of 10.6±0.5 nmol/(min·mg protein) (Fig. 2B). As shown in Fig. 2C, the uptake was inhibited by more than 33% by long-chain PUFAs such as DHA, eicosapentaenoic acid (EPA), arachidonic acid (AA), and linoleic acid (LA) at the concentration of 500 µM. In contrast, OA had no significant effect on [¹⁴C]DHA uptake.

Fig. 2. Characteristics of Free [¹⁴C]DHA Uptake by ARPE-19 Cells

(A) Time course of free [¹⁴C]DHA uptake by ARPE-19 cells. Free [¹⁴C]DHA (0.05 µCi, 4.4 µM) uptake was measured at 37°C at the designated time points. Each point is the mean±S.E.M. (n=3–4). (B) Concentration dependence of free [¹⁴C]DHA uptake by ARPE-19 cells. Free DHA (0.05 µCi) uptake was examined at 37°C for 10 min over the concentration range of 40–1000 µM. Data were subjected to Michaelis–Menten and Eadie–Scatchard analyses (inset). Each point is the mean±S.E.M. (n=4). (C) Effect of several long-chain unsaturated fatty acids on the uptake of free [¹⁴C]DHA by ARPE-19 cells. Free [¹⁴C]DHA (0.05 µCi, 4.4 µM) uptake was examined in the presence or absence (control) of extracellularly applied 500 µM fatty acids (DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; AA, arachidonic acid; LA, linoleic acid; OA, oleic acid) with 1% ethanol at 37°C for 10 min. Each column indicates the mean±S.E.M. (n=4). * p<0.05 and ** p<0.01, significantly different from the control.

Characterization of RBP-, LDL-, and Albumin-Associated [³H]DHA Uptake by ARPE-19 Cells

More than 90% of DHA in the circulating blood is present in bound form with plasma proteins such as albumin and lipoprotein,²⁰⁾ and DHA also shows affinity for intraphotoreceptor retinoid-binding protein (IRBP).²⁸⁾ This raised the possibility that DHA could bind to retinoid-binding protein(s). Since RBP functions as a carrier protein of vitamin A,²²⁾ we therefore examined the binding of DHA with RBP by means of photoaffinity cross-linking. As shown in Fig. 3, a band at ca. 19 kDa corresponding to RBP was detected in the UV-irradiated mixture of the photoreactive DHA probe and RBP. The intensity of the band was lower in the presence of excess unlabeled DHA and retinol (vitamin A, VA). In contrast, the presence of excess OA had no effect on the intensity of the RBP-derived band. This result indicated that RBP could bind unesterified DHA. In size-exclusion chromatography, RBP-associated [¹⁴C]DHA ([¹⁴C]DHA-RBP), LDL-associated [¹⁴C]DHA ([¹⁴C]DHA-LDL), and BSA-associated [¹⁴C]DHA ([¹⁴C]DHA-BSA) were eluted in fraction numbers 10-20, although unbound [¹⁴C]DHA was not detected in any fraction (Fig. 4). These results confirmed that [¹⁴C]DHA binds to RBP, LDL, and albumin. The uptakes of RBP-, LDL-, and BSA-associated [¹⁴C]DHA by ARPE-19 cells showed a time-dependent increase for 20 min at 37°C with linear uptake rates of 37.4±1.3, 21.2±1.0, and 9.98±1.01 µL/(min·mg protein), respectively, which are 0.6 to 2.1 times that of free [¹⁴C]DHA [17.6 µL/(min·mg protein)] (Fig. 5A). The uptakes of RBP-, LDL-, and BSA-associated [¹⁴C]DHA in 10 min were significantly reduced by 79.2, 77.0, and 90.6%, respectively, at 4°C compared to those at 37°C (Fig. 5B). Ten-times excess LDL and 300 µM dansylcadaverine (DC), an inhibitor of clathrin-dependent endocytosis, significantly inhibited the LDL-associated [¹⁴C]DHA uptake by 78.7 and 27.0%, respectively (Fig. 5C). In contrast, BLT-1 (100 nM), an inhibitor of scavenger receptor class B type I-mediated HDL transport, which was used as a negative control for the LDL-related uptake, did not inhibit the uptake (Fig. 5C).

Fig. 3. Binding of Photoreactive DHA Probe with Retinol-Binding Protein (RBP)

RBP was incubated with the DHA probe (5 µM) in solution containing 0.5% ethanol for 1 h at 37°C in the presence or absence of 250 µM unlabeled DHA, oleic acid (OA), and vitamin A (VA, retinol). After UV irradiation (+), the protein sample was separated on a 12% SDS-polyacrylamide gel. The labeled protein (arrowhead) was detected with streptavidin-horseradish peroxidase.

Fig. 4. Size-Exclusion Chromatography of Protein-Bound [¹⁴C]DHA

(A) [¹⁴C]DHA-RBP, (B) [¹⁴C]DHA-LDL, and (C) [¹⁴C]DHA-BSA. The numbers are the fraction numbers.

Fig. 5. Characteristics of Protein-Bound [¹⁴C]DHA Uptake by ARPE-19 Cells

(A) Comparison of initial uptake rates of protein-bound [¹⁴C]DHA and free [¹⁴C]DHA by ARPE-19 cells. [¹⁴C]DHA-RBP (open circle), [¹⁴C]DHA-LDL (open triangle), [¹⁴C]DHA-BSA (open square) and free [¹⁴C]DHA (closed circle) uptake was measured at 37°C. All protein-bound [¹⁴C]DHAs were used at the concentration of 0.05 µCi (as [¹⁴C]DHA). Each point is the mean±S.E.M. (n=3–4). (B) Temperature dependence of protein-bound [¹⁴C]DHA uptake by ARPE-19 cells. [¹⁴C]DHA-BSA, [¹⁴C]DHA-LDL, or [¹⁴C]DHA-RBP (0.05 µCi, 4.4 µM as [¹⁴C]DHA) uptake was measured at 37°C (open column) or 4°C (closed column) Each column indicates the mean±S.E.M. (n=4). ** p<0.01, significantly different from the value at 37°C. (C) Effect of several compounds on [¹⁴C]DHA-LDL uptake by ARPE-19 cells. [¹⁴C]DHA-LDL (0.05 µCi, 4.4 µM as [¹⁴C]DHA) uptake was measured in the presence or absence (control) of extracellularly applied LDL (10-fold excess) (10 x LDL), BLT-1 (an inhibitor of lipid transport) (100 nM) or dansylcadaverine (DC) (300 µM) at 37°C. The inhibition studies with BLT-1 and DC was performed in the presence of 0.5% DMSO. Each value is mean±S.E.M. (n=4). ** p<0.01, significantly different from the control.

Expression of FATP mRNAs in ARPE-19 Cells

The expression of FATP1–6 and β-actin (positive control) mRNAs in ARPE-19 cells was evaluated by RT-PCR analysis. Bands of the expected size (Table 1) for FATP2 and FATP4 as well as β-actin were amplified (Fig. 6). The nucleotide sequences of the amplified products were identical to the reported sequences of human FATP2 and FATP4, respectively, in GenBank. These results indicate that FATP2 and FATP4 are predominantly expressed in ARPE-19 cells.

Fig. 6. RT-PCR Analysis of Fatty Acid Transport Protein 1–6 (FATP1-6) and β-Actin mRNAs in ARPE-19 Cells

RT-PCR was performed in the presence (+) or absence (−) of reverse transcriptase (RT). β-Actin is used as a positive control.

[¹⁴C]DHA uptake in Human FATP2- and FATP4-Expressing Xenopus Oocytes

Several sequential steps have been proposed to be involved in the import of free fatty acids into cells i.e., dissociation of free fatty acids from albumin in the plasma followed by plasma membrane binding and transport across the plasma membranes mediated by FATPs.²⁹⁾ To determine whether FATP2 and FATP4 mediate the import of [¹⁴C]DHA into the cells, we measured the uptakes of [¹⁴C]DHA-BSA and [¹⁴C]DHA-RBP by human FATP2- and FATP4-expressing oocytes (FATP2/oocytes and FATP4/oocytes, respectively) (Fig. 7). The uptake of [¹⁴C]DHA-BSA by FATP4/oocytes was significantly (1.5-fold) greater than that by water-injected oocytes, whereas there was no significant difference of [¹⁴C]DHA-BSA uptake by FATP2/oocytes compared with water-injected oocytes (Fig. 7A). The uptakes of [¹⁴C]DHA-RBP by FATP2/oocytes and FATP4/oocytes were significantly (1.6- and 1.3-fold, respectively) greater than that by water-injected oocytes (Fig. 7B). These results indicated that human FATP2 and FATP4 can mediate DHA uptake.

Fig. 7. Uptake of [¹⁴C]DHA-BSA (A) and [¹⁴C]DHA-RBP (B) by Human FATP2- or FATP4-Expressing Oocytes

[¹⁴C]DHA-BSA and [¹⁴C]DHA-RBP (0.1 µCi, 8.8 µM as [¹⁴C]DHA) uptake by human FATP2-expressing (gray column), FATP4-expressing (closed column) or water-injected (open column: control) oocytes was measured at 20°C for 60 min. Each column indicates the mean±S.E.M. (n=16). ** p<0.01, significantly different from the control.

Photo-affinity Crosslinking of DHA with Target Proteins

To identify membrane-associated binding or target protein(s) of unbound DHA in RPE cells, we conducted photo-affinity crosslinking between the DHA probe and the membrane protein fraction of isolated porcine RPE cells (Fig. 8). The intensities of protein bands at ca. 75 and ca. 65 kDa were increased after UV irradiation, as determined by streptavidin-horseradish peroxidase detection. To evaluate the binding specificity, we examined the competitive effects of unlabeled DHA and OA on the photoaffinity labeling of these proteins. The band intensities were decreased in the presence of unlabeled DHA, but not unlabeled OA. These results indicated that the photoreactive probe-labeled proteins of ca. 75 and ca. 65 kDa preferentially bind DHA. Alignments of the peptide fragments obtained from the ca. 75- and ca. 65-kDa proteins are shown in Supplemental Fig. 1. MALDI-TOF MS analysis and MASCOT search revealed that the peptide fragments could be assigned to peroxisomal hydroxysteroid17β-dehydrogenase 4 (asterisk, Fig. 8) and 65-kDa retinal pigment epithelium specific protein (RPE65; arrowhead, Fig. 8).

Fig. 8. Photoaffinity Labeling of Porcine RPE Cell Membrane Fractions

The photoreactive DHA probe was incubated with the cell membrane fractions for 3 h at 4°C in the presence or absence (−) of unlabeled DHA or OA each at a concentration of 1 mM. After UV irradiation (UV; +) at 360 nm for 30 min at 4°C or without UV irradiation (UV; −), labeled proteins were detected with streptavidin-horseradish peroxidase. The asterisk and arrowhead indicate the ca. 75  and ca. 65 kDa proteins labeled with the photoaffinity probe, respectively.

DISCUSSION

Our present findings indicate that transport of unesterified DHA in RPE cells involves multiple transporters/binding proteins, including FATP2, FATP4, LDL, RBP, and RPE65.

Firstly, we observed concentrative and saturable (K_m value: 283 µM) uptake of free DHA, which was inhibited by long-chain PUFAs, such as DHA, EPA, AA, and LA, but not by OA. These findings support the belief that the outer BRB possesses a carrier-mediated process for uptake of free unesterified DHA, which can also take up EPA, AA, and LA, but not OA. We reported previously that FATP1-mediated uptake of free unesterified DHA was significantly inhibited by OA at the extracellular concentration of 1 mM in human microvessel endothelial cells (hCMEC/D3 cells).¹⁸⁾ In this regard, the DHA-preferential transport system in RPE cells appears to be different from that at the BBB. It has been proposed that import of free fatty acids into cells involves several sequential steps²⁹⁾: (i) dissociation of free fatty acids from albumin in the plasma, (ii) plasma membrane binding and transport across the plasma membranes by FABPpm, FAT/CD36, and FATPs, and (iii) intracellular binding and/or metabolism in the cytoplasm by FABPs, acyl-CoA hydrolase and other proteins. Indeed, FATP4 has dual functions as a transporter and acyl-CoA synthetase for free fatty acids,³⁰⁾ and FATPs-mediated acylation would decrease the intracellular pool of free fatty acids, facilitating the apparent transport of free fatty acids across the plasma membrane.²⁹⁾ Therefore, the discrepancy between ARPE-19 cells and hCMEC/D3 cells might be explained by differences in substrate specificity between FATP1 and other FATPs, or differences in plasma membrane binding of DHA.

In this regard, the RT-PCR results indicate that FATP2 and FATP4 mRNAs are expressed in ARPE19 cells, whereas FATP1 mRNA is not. The expression of FATP4 is consistent with previous RT-PCR, immunoblot, and immunohistochemistry findings, which showed that FATP4 is most highly expressed among FATP1-6 in mouse RPE cells.³¹⁾ FATP1 appears to have no marked preference among long-chain PUFAs, as it mediates the internalization of free PA, OA, and AA in 3T3-L1 cells.³²⁾ It has also been reported that overexpression of FATP2 or FATP4 protein resulted in increased transport of free OA and fluorescent fatty acid analogs.^30,33) Therefore, the absence of a significant inhibitory effect of OA on free [¹⁴C]DHA uptake by ARPE-19 cells seems inconsistent with the substrate specificity of FATP2 or FATP4 in DHA uptake. On the other hand, preferential binding of DHA over OA on human placental membranes has been reported,³⁴⁾ and the in vivo placental transfer rate of DHA was greater than that of LA, PA or OA at 4 h after oral administration in humans.³⁵⁾ These findings seem broadly consistent with the present results. However, FABPpm has not yet been molecularly identified in RPE cells. Further studies are needed to examine whether the preferential recognition of DHA over OA by ARPE-19 cells is due to distinctive features of plasma membrane fatty acid binding in RPE cells. As regards the uptake of protein-bound DHA, we found that the uptakes of [¹⁴C]DHA-RBP, [¹⁴C]DHA-LDL, and [¹⁴C]DHA-BSA by ARPE-19 cells were significantly reduced at 4°C compared to 37°C. This suggests that energy-dependent transport processes are involved. We found that FATP2 specifically mediates cellular uptake [¹⁴C]DHA from RBP-DHA, whereas FATP4 mediates uptake from both [¹⁴C]BSA-DHA and [¹⁴C]RBP-DHA. One possible explanation of this discrepancy is that there may be a difference in the transfer efficacy of free [¹⁴C]DHA from albumin/RBP to FATP2/FATP4. Furthermore, our previous results have shown that free [¹⁴C]DHA can also be taken up into FATP1-overexpressing HEK293 cells without any binding partner in the uptake buffer.¹⁸⁾ This at least suggests that FATP1 does not necessarily require binding protein(s) for the DHA transport. It thus seems plausible that FATP2 and/or FATP4 might be involved, at least in part, as common carriers in the uptakes of free DHA and protein-bound DHA, though further work will be needed to confirm this.

It has been reported that 33 and 26% of [³H]DHA is present as LDL- and HDL-associated forms, respectively, in the circulating blood at 6 h after oral administration.³⁶⁾ Considering that the distribution of [³H]DHA in the retina reaches a peak at 12 h after oral administration,³⁶⁾ it is plausible that HDL- or LDL-associated [³H]DHA is involved in the uptake by RPE cells. This would be consistent with reports showing that intravenously injected rhodamine-labeled LDL was predominantly taken up by rat RPE cells at 24 h after injection, and LDL receptor and FAT/CD36 proteins (potent receptors of LDL) are co-localized with the internalized LDL and oxidized LDL particles, respectively.³⁷⁾ We also found here that LDL-associated [¹⁴C]DHA uptake in ARPE-19 cells was reduced by an excess amount of unlabeled LDL, but not by a scavenger receptor class B type I inhibitor, further supporting a role of LDL-mediated [¹⁴C]DHA uptake in the delivery of circulating DHA to RPE cells. Interestingly, the DHA levels in brain and retina were not changed in LDL receptor-knockout mice.³⁸⁾ Since FAT/CD36 is expressed in ARPE-19 cells at both the transcript and protein levels,³⁹⁾ FAT/CD36 might be the predominant carrier for [¹⁴C]DHA-LDL in ARPE-19 cells. It will be interesting to pursue the molecular contribution of HDL-associated DHA to DHA transport across the outer BRB in a future study.

The present study indicates that [¹⁴C]DHA-RBP showed the greatest uptake rate in ARPE-19 cells among free and protein-bound [¹⁴C]DHA species. Considering that [¹⁴C]DHA-RBP was taken by human FATP2 and FATP4-expressing oocytes, RBP may facilitate the transfer of DHA to FATP2 or FATP4. Furthermore, it has been reported that STRA6 plays a major role in the uptake of RBP-vitamin A complex and that the uptake is significantly enhanced by overexpression of lecithin retinol acyltransferase, which converts retinol into retinyl ester and is involved in vitamin A storage.²²⁾ Our size exclusion and photo-affinity crosslinking analyses indicate that RBP also binds with DHA. It would therefore be intriguing in future studies to clarify the possible involvement of an intracellular DHA acceptor protein, as well as the possible involvement of STRA6 in the cellular uptake of RBP-DHA complex in RPE cells.

FATP4 has been identified as a negative regulator of RPE65 isomerase, which play a role in the isomerization from all-trans retinyl fatty acid esters to 11-cis retinol in the visual cycle, from bovine RPE cDNA library.³¹⁾ Indeed, FATP4-knockout mice accumulate cytotoxic all-trans retinaldehyde in the retina and show hyper-susceptibility to light-induced photoreceptor degradation.³¹⁾ The role of FATP4 as a negative regulator of RPE65 isomerase in RPE cells could not fully be explained by competition between FATP4 and RPE65 for their common substrates, or by inhibition of RPE65 isomerase by the product (lignoceroyl C24:0-CoA) of FATP4′s action as a long-chain fatty acid acyl-CoA synthetase. Interestingly, the present photo-affinity crosslinking and MALDI-TOF/MS analyses identified RPE65 isomerase and hydroxysteroid17β-dehydrogenase 4 (an RPE cell peroxisomal membrane protein involved in β-oxidation)⁴⁰⁾ as binding partners of DHA in the membrane fraction of porcine RPE cells. Thus, DHA transported by FATP4 in RPE cells may be targeted to RPE65 isomerase as a potential modulator of the visual cycle in photoreceptor cells. Alternatively, it seems plausible that RPE65 may facilitate DHA transport by receiving its free form at the cytosolic side of the plasma membrane. So far, there is no direct evidence for plasma membrane localization of RPE65 in RPE cells, but RPE65 protein has been identified in other membrane fractions, including endoplasmic reticulum (ER) membranes.⁴¹⁾ The physiological contribution of RPE65 to DHA transport, and the binding properties and functional interactions among DHA, FATP4 and RPE65 remain to be addressed in future studies.

In conclusion, DHA undergoes active influx transport into ARPE-19 cells, both as free DHA and as protein-bound DHA via multiple mechanisms, including FATP2-, FATP4-, and LDL-mediated pathways. DHA binds with cell membrane components such as RPE65 and hydroxysteroid17β-dehydrogenase 4 in RPE cells. These findings should be helpful to increase our understanding of the regulation of essential PUFAs in the retina.

Acknowledgments

We thank Dr. T. Abe for supplying the pGEM-HEN vector. This study was supported in part by a Grant-in-Aid for Research (C) (JP16K08364) and a Grant-in-Aid for Research (B) (JP16H05110) from the Japan Society for the Promotion of Science (JSPS), and Terumo Foundation for Life Sciences and Arts, Japan.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

The online version of this article contains supplementary materials.

REFERENCES

• 1) Lauritzen L, Hansen HS, Jorgensen MH, Michaelsen KF. The essentiality of long chain n-3 fatty acids in relation to development and function of the brain and retina. Prog. Lipid Res., 40, 1–94 (2001).
• 2) Litman BJ, Niu SL, Polozova A, Mitchell DC. The role of docosahexaenoic acid containing phospholipids in modulating G protein-coupled signaling pathways: visual transduction. J. Mol. Neurosci., 16, 237–242, discussion, 279–284 (2001).
• 3) Mukherjee PK, Marcheselli VL, Serhan CN, Bazan NG. Neuroprotectin D1: a docosahexaenoic acid-derived docosatriene protects human retinal pigment epithelial cells from oxidative stress. Proc. Natl. Acad. Sci. U.S.A., 101, 8491–8496 (2004).
• 4) Fliesler SJ, Anderson RE. Chemistry and metabolism of lipids in the vertebrate retina. Prog. Lipid Res., 22, 79–131 (1983).
• 5) SanGiovanni JP, Chew EY. The role of omega-3 long-chain polyunsaturated fatty acids in health and disease of the retina. Prog. Retin. Eye Res., 24, 87–138 (2005).
• 6) Léveillard T, Sahel JA. Metabolic and redox signaling in the retina. Cell. Mol. Life Sci., 74, 3649–3665 (2017).
• 7) Li J, Wetzel MG, O’Brien PJ. Transport of n-3 fatty acids from the intestine to the retina in rats. J. Lipid Res., 33, 539–548 (1992).
• 8) Wang N, Anderson RE. Synthesis of docosahexaenoic acid by retina and retinal pigment epithelium. Biochemistry, 32, 13703–13709 (1993).
• 9) Weisinger HS, Armitage JA, Jeffrey BG, Mitchell DC, Moriguchi T, Sinclair AJ, Weisinger RS, Salem N Jr. Retinal sensitivity loss in third-generation n-3 PUFA-deficient rats. Lipids, 37, 759–765 (2002).
• 10) Futterman S, Downer JL, Hendrickson A. Effect of essential fatty acid deficiency on the fatty acid composition, morphology, and electroretinographic response of the retina. Invest. Ophthalmol., 10, 151–156 (1971).
• 11) Uauy R, Mena P, Rojas C. Essential fatty acids in early life: structural and functional role. Proc. Nutr. Soc., 59, 3–15 (2000).
• 12) Hosoya K, Tomi M, Tachikawa M. Strategies for therapy of retinal diseases using systemic drug delivery: relevance of transporters at the blood–retinal barrier. Expert Opin. Drug Deliv., 8, 1571–1587 (2011).
• 13) Wong BH, Chan JP, Cazenave-Gassiot A, Poh RW, Foo JC, Galam DL, Ghosh S, Nguyen LN, Barathi VA, Yeo SW, Luu CD, Wenk MR, Silver DL. Mfsd2a is a transporter for the essential omega-3 fatty acid docosahexaenoic acid (DHA) in eye and is important for photoreceptor cell development. J. Biol. Chem., 291, 10501–10514 (2016).
• 14) Rice DS, Calandria JM, Gordon WC, Jun B, Zhou Y, Gelfman CM, Li S, Jin M, Knott EJ, Chang B, Abuin A, Issa T, Potter D, Platt KA, Bazan NG. Adiponectin receptor 1 conserves docosahexaenoic acid and promotes photoreceptor cell survival. Nat. Commun., 6, 6228 (2015).
• 15) Nguyen LN, Ma D, Shui G, Wong P, Cazenave-Gassiot A, Zhang X, Wenk MR, Goh EL, Silver DL. Mfsd2a is a transporter for the essential omega-3 fatty acid docosahexaenoic acid. Nature, 509, 503–506 (2014).
• 16) Spector AA. Plasma free fatty acid and lipoproteins as sources of polyunsaturated fatty acid for the brain. J. Mol. Neurosci., 16, 159–165, discussion, 215–121 (2001).
• 17) Abumrad N, Coburn C, Ibrahimi A. Membrane proteins implicated in long-chain fatty acid uptake by mammalian cells: CD36, FATP and FABPm. Biochim. Biophys. Acta, 1441, 4–13 (1999).
• 18) Ochiai Y, Uchida Y, Ohtsuki S, Tachikawa M, Aizawa S, Terasaki T. The blood–brain barrier fatty acid transport protein 1 (FATP1/SLC27A1) supplies docosahexaenoic acid to the brain, and insulin facilitates transport. J. Neurochem., 141, 400–412 (2017).
• 19) Larqué E, Krauss-Etschmann S, Campoy C, Hartl D, Linde J, Klingler M, Demmelmair H, Cano A, Gil A, Bondy B, Koletzko B. Docosahexaenoic acid supply in pregnancy affects placental expression of fatty acid transport proteins. Am. J. Clin. Nutr., 84, 853–861 (2006).
• 20) Richieri GV, Kleinfeld AM. Unbound free fatty acid levels in human serum. J. Lipid Res., 36, 229–240 (1995).
• 21) Tachikawa M, Okayasu S, Hosoya K. Functional involvement of scavenger receptor class B, type I, in the uptake of alpha-tocopherol using cultured rat retinal capillary endothelial cells. Mol. Vis., 13, 2041–2047 (2007).
• 22) Kawaguchi R, Yu J, Honda J, Hu J, Whitelegge J, Ping P, Wiita P, Bok D, Sun H. A membrane receptor for retinol binding protein mediates cellular uptake of vitamin A. Science, 315, 820–825 (2007).
• 23) Akanuma S, Yamamoto A, Okayasu S, Tachikawa M, Hosoya K. High-density lipoprotein-associated alpha-tocopherol uptake by human retinal pigment epithelial cells (ARPE-19 Cells): the irrelevance of scavenger receptor class B, type I. Biol. Pharm. Bull., 32, 1131–1134 (2009).
• 24) Yamaoka K, Tanaka H, Okumura K, Yasuhara M, Hori R. An analysis program MULTI(ELS) based on extended nonlinear least squares method for microcomputers. J. Pharmacobiodyn., 9, 161–173 (1986).
• 25) Tachikawa M, Kasai Y, Yokoyama R, Fujinawa J, Ganapathy V, Terasaki T, Hosoya K. The blood–brain barrier transport and cerebral distribution of guanidinoacetate in rats: involvement of creatine and taurine transporters. J. Neurochem., 111, 499–509 (2009).
• 26) Kashiwayama Y, Tomohiro T, Narita K, Suzumura M, Glumoff T, Hiltunen JK, Van Veldhoven PP, Hatanaka Y, Imanaka T. Identification of a substrate-binding site in a peroxisomal beta-oxidation enzyme by photoaffinity labeling with a novel palmitoyl derivative. J. Biol. Chem., 285, 26315–26325 (2010).
• 27) Akanuma SI, Higashi H, Maruyama S, Murakami K, Tachikawa M, Kubo Y, Hosoya KI. Expression and function of connexin 43 protein in mouse and human retinal pigment epithelial cells as hemichannels and gap junction proteins. Exp. Eye Res., 168, 128–137 (2018).
• 28) Chen Y, Saari JC, Noy N. Interactions of all-trans-retinol and long-chain fatty acids with interphotoreceptor retinoid-binding protein. Biochemistry, 32, 11311–11318 (1993).
• 29) Doege H, Stahl A. Protein-mediated fatty acid uptake: novel insights from in vivo models. Physiology (Bethesda), 21, 259–268 (2006).
• 30) Milger K, Herrmann T, Becker C, Gotthardt D, Zickwolf J, Ehehalt R, Watkins PA, Stremmel W, Fullekrug J. Cellular uptake of fatty acids driven by the ER-localized acyl-CoA synthetase FATP4. J. Cell Sci., 119, 4678–4688 (2006).
• 31) Li S, Lee J, Zhou Y, Gordon WC, Hill JM, Bazan NG, Miner JH, Jin M. Fatty acid transport protein 4 (FATP4) prevents light-induced degeneration of cone and rod photoreceptors by inhibiting RPE65 isomerase. J. Neurosci., 33, 3178–3189 (2013).
• 32) Schaffer JE, Lodish HF. Expression cloning and characterization of a novel adipocyte long chain fatty acid transport protein. Cell, 79, 427–436 (1994).
• 33) Falcon A, Doege H, Fluitt A, Tsang B, Watson N, Kay MA, Stahl A. FATP2 is a hepatic fatty acid transporter and peroxisomal very long-chain acyl-CoA synthetase. Am. J. Physiol. Endocrinol. Metab., 299, E384–E393 (2010).
• 34) Campbell FM, Gordon MJ, Dutta-Roy AK. Placental membrane fatty acid-binding protein preferentially binds arachidonic and docosahexaenoic acids. Life Sci., 63, 235–240 (1998).
• 35) Larqué E, Demmelmair H, Berger B, Hasbargen U, Koletzko B. In vivo investigation of the placental transfer of (13)C-labeled fatty acids in humans. J. Lipid Res., 44, 49–55 (2003).
• 36) Wang N, Anderson RE. Transport of 22:6n-3 in the plasma and uptake into retinal pigment epithelium and retina. Exp. Eye Res., 57, 225–233 (1993).
• 37) Gordiyenko N, Campos M, Lee JW, Fariss RN, Sztein J, Rodriguez IR. RPE cells internalize low-density lipoprotein (LDL) and oxidized LDL (oxLDL) in large quantities in vitro and in vivo. Invest. Ophthalmol. Vis. Sci., 45, 2822–2829 (2004).
• 38) Bretillon L, Acar N, Seeliger MW, Santos M, Maire MA, Juaneda P, Martine L, Gregoire S, Joffre C, Bron AM, Creuzot-Garcher C. ApoB100,LDLR−/− mice exhibit reduced electroretinographic response and cholesteryl esters deposits in the retina. Invest. Ophthalmol. Vis. Sci., 49, 1307–1314 (2008).
• 39) Gnanaguru G, Choi AR, Amarnani D, D’Amore PA. Oxidized lipoprotein uptake through the cd36 receptor activates the NLRP3 inflammasome in human retinal pigment epithelial cells. Invest. Ophthalmol. Vis. Sci., 57, 4704–4712 (2016).
• 40) Kobayashi K, Kobayashi H, Ueda M, Honda Y. Expression of 17 beta-hydroxysteroid dehydrogenase type IV in chick retinal pigment epithelium. Exp. Eye Res., 64, 719–726 (1997).
• 41) Takahashi Y, Moiseyev G, Ablonczy Z, Chen Y, Crouch RK, Ma JX. Identification of a novel palmitylation site essential for membrane association and isomerohydrolase activity of RPE65. J. Biol. Chem., 284, 3211–3218 (2009).