Proteins Modified by Malondialdehyde, 4-Hydroxynonenal, or Advanced Glycation End Products in Lipofuscin of Human Retinal Pigment Epithelium

Abstract

purpose. Lipofuscin (LF) accumulation in the retinal pigment epithelium (RPE) is associated with age and various retinal diseases. Toxic LF compounds may interfere with normal RPE function. Oxidative modification of proteins was determined in LF granules from human eyes.

methods. LF was isolated from the RPE-choroid complex of 10 pairs of donor eyes by gradient ultracentrifugation. Protein compounds were separated by two-dimensional (2-D) gel electrophoresis and screened by Western blot analysis for lipid peroxidation- or glucoxidation-induced damage—in particular, by malondialdehyde (MDA), 4-hydroxynonenal (HNE), and advanced glycation end products (AGEs). Identity of the immunostained proteins was revealed using 2-D software for comparison of the spot position with Coomassie-stained 2-D gels of the same samples.

results. By comparing the results taken from the authors’ previous proteome analysis of RPE LF with an immunoblot analysis of the same samples, this study shows that a variety of LF-associated proteins were damaged by aberrant covalent modifications of MDA, 4-HNE, and AGEs. Several proteins were altered by two or three different modification types. Modified mitochondrial proteins indicated that autophagy of altered proteins also contributed to lipofuscin formation.

conclusions. The identification of lipid peroxidation and glucoxidation products in proteinaceous LF components in human RPE supports the hypothesis that these compounds are involved in lipofuscinogenesis and may contribute to the cytotoxic effects of LF in retinal diseases such as age-related macular degeneration and Stargardt disease. Their identification may help to identify potential future treatment targets.

Accumulation of lipofuscin granules in postmitotic retinal pigment epithelial (RPE) cells is associated with age and various retinal diseases, including age-related macular degeneration (AMD) and monogenetic juvenile disorders, such as Best or Stargardt disease. ^{1 2 3} Several lines of evidence indicate that RPE lipofuscin interferes with normal RPE cell function through various molecular mechanisms including inhibition of lysosomal degradation and reduction of antioxidant capacity. ^{3 4 5} Lipofuscin is an enzymatically nondegradable heterogenous mixture of numerous biomolecules. Its biochemical analysis is a prerequisite for the elucidation of pathomechanisms of lipofuscin formation and lipofuscin-induced RPE dysfunction. There is evidence to suggest that lipofuscin arises as a byproduct of constant phagocytosis of membranous discs shed from distal photoreceptor outer segments. ⁶ A recent proteome analysis of lipofuscin granules showed that most lipofuscin proteins are cellular housekeeping proteins derived from photoreceptors and RPE cells but would not be expected to cause cell damage. ⁷ However, we hypothesized that posttranslational modifications may occur, and, thus, undegradable and/or toxic compounds would form. Lifelong UV-light exposure, along with high ocular oxygen levels, trigger permanent peroxidation of polyunsaturated lipids in the membrane system of ocular photoreceptor membranes leading to lipid peroxidation–derived protein modifications. ⁸ The resultant end products, malondialdehyde (MDA) lysin protein adducts or 4-hydroxynonenal (HNE) cysteine adducts are well-known markers in the pathologic molecular process in oxidative stress. ^{9 10} Enhanced oxidative stress is also assumed to participate in the formation of nonenzymatically glycated proteins and their advanced stages, the advanced glycation end products (AGEs), which have been detected in long-lived proteins and protein deposits in human and animal tissues. They are thought to contribute to normal aging phenomena in various tissues as well as to the pathogenesis of diseases including diabetic complications and Alzheimer’s disease. ^{11 12} There is also strong evidence to suggest that the lipofuscin is derived from oxidatively damaged proteins. ^{13 14} A recently published drusen proteome analysis confirmed that oxidative modification of proteins originated from isolated drusen. ¹⁵ Drusen as extracellular deposits are known risk factors for development of the late stages of AMD. The authors argue that oxidative protein modification such as protein cross-links, AGEs, and lipid-derived modifications may be a trigger in the formation of drusen. ¹⁵ Because intracellular lipofuscin accumulation in RPE cells is a hallmark of aging and various forms of AMD, ^{16 17 18 19} it is speculated that the aberrant modifications of proteins just described may contribute to lipofuscin formation and have an impact on RPE dysfunction, ultimately resulting in cell death. ^{17 20} Lipofuscin itself provides an additional oxidative burden on the RPE. ⁴ Rozanowska et al. ²¹ were able to show that aerobic illumination of human lipofuscin isolated from aged donors leads to formation of hydroperoxides and MDA. To substantiate this hypothesis further, we screened our previously determined lipofuscin proteome for MDA, HNE, and AGE modifications. ⁷  

Materials and Methods

Lipofuscin Isolation

Lipofuscin was isolated from the RPE-choroid complex of 10 pairs of human donor eyes for corneal transplantation with no known ocular disease, as described previously. ⁷ Permission had been given to use the eyes of all organ/eye donors for research purposes by their relatives. All globes from white donors aged 65 to 96 (median, 78) years were obtained within 4 to 6 hours after death due to extraocular malignancies or cerebral insults. Eyes were constantly cooled and transported at +4°C. The cornea was trephined, and each posterior pole was examined under a macroscope (Zeiss, Jena, Germany) to screen the posterior pole for macroscopically visible drusen, RPE atrophy, choroidal neovascularization, and subretinal fibrous scars. None of the eyes exhibited such changes. Histopathologic sections were not obtained. The posterior pole of the eye was cooled and stored in balanced saline solution until preparation of the RPE-choroid complex, which occurred 24 to 48 hours after death. Because of minimal oxygen exposure and constant cooling, we assumed only minimal additional oxidation of lipofuscin components including proteins had occurred. Isolated RPE-choroid tissue was frozen at −20°C and pooled. After RPE-choroid acquisition of 10 pairs of human donor eyes we started isolating the lipofuscin. Purity of the isolated lipofuscin was checked by recording its excitation and emission spectra, as well as by light microscopic examination. ⁷ The samples were stored at −20°C for further studies. 

Two-Dimensional Gel Electrophoresis of Lipofuscin

Isoelectric focusing (IEF) and SDS-PAGE were performed according to previously published protocols. ⁷ Briefly, lipofuscin proteins were extracted in 7 M urea, 2 M thiourea, 4% Triton X-100, 65 mM dithiothreitol, and 0.8% Pharmalite (Sigma-Aldrich, Deisenhofen, Germany) and separated in an immobilized pH gradient (IPG) strip (Immobiline DryStrip pH 3–10, 11 cm, Amersham Pharmacia Biotech, Freiburg, Germany) using an IEF system (IPGphor; Amersham Pharmacia Biotech). For subsequent SDS-PAGE, the IPG strip was applied to a precast SDS gradient gel (8-18 ExcelGel; Amersham Pharmacia Biotech), and the separation performed on an electrophoretic transfer unit (Multiphor II; Amersham Pharmacia Biotech). 

Western Blot Analysis and Immunostaining

For immunoblot analysis, proteins of the two-dimensional (2-D) separation were transferred to a nitrocellulose membrane (Hybond ECL; Amersham Pharmacia Biotech) with the electrophoretic transfer unit (Multiphor II NovaBlot; Amersham Pharmacia Biotech). Electrical settings were 10 V, 0.8 mA/cm² of membrane, and 5 W for 75 minutes After electroblot transfer, the nitrocellulose membrane was equilibrated for 10 minutes in TBS buffer (20 mM Tris-HCl [pH 7.5], and 0.5 M NaCl) followed by 45 minutes in TBS buffer and 5% dry milk (Blotting Grade Blocker; Bio-Rad, Munich, Germany). Then the membrane was washed twice for 15 minutes in TTBS buffer (20 mM Tris-HCl [pH 7.5], 0.5 M NaCl, and 0.05% Tween 20) and then incubated for 1 hour at room temperature with the following primary antibodies: rabbit anti-MDA antiserum (MDA11-S; Alpha Diagnostics, San Antonio, TX), rabbit anti-HNE antiserum (HNE11-S; Alpha Diagnostics) and monoclonal anti-AGE IgG antibody (Clone No.6D12; Wako Chemicals, Richmond, VA). Dilutions of primary antibodies were always 1:500 in TTBS buffer and 1% dry milk. Visualization was performed with alkaline phosphatase-conjugated secondary antibodies (1:500 dilution) in combination with 5-bromo-4-chloro-3-indoyl phosphate/nitroblue tetrazolium (BCIP/NBT; Sigma Fast tablets; Sigma-Aldrich). Finally, the membrane was washed with distilled water, air dried, and documented by scanning. Identity of the immunostained proteins was revealed by use of a computer program (TDS PDQuest 2-D; Bio-Rad) for comparison of the spot positions with Coomassie-stained 2-D gels of the same samples, in which protein identification was achieved by using matrix-assisted laser desorption/ionization mass spectrometry and HPLC-coupled electrospray tandem mass spectrometry. ⁷  

MDA- or HNE-modified ovalbumin were tested as positive controls in the immunoblot experiments. To detect nonspecific binding, an excess of MDA- or HNE-modified ovalbumin was included in the solution of primary antibody in parallel experiments. All spots shown in Figure 1 were quenched in these control experiments. Therefore all spots can be considered specifically stained for MDA or HNE modification. For the detection of AGE modification, such specificity control appeared unnecessary, because a specific monoclonal antibody was used. 

Results

Proteome analysis of lipofuscin isolated from human RPE revealed that a large proportion of the identified proteins represent abundant cellular proteins, including cytoskeleton proteins, proteins of phototransduction, enzymes of metabolism, proteins of the mitochondrial respiratory chain, ion channel proteins, and chaperones, whose native structure neither readily explains their deposition in lipofuscin nor any toxic effects. ⁷ Clusters of spots containing the same protein located at different gel positions were suspected of causing oxidative damage. By comparing the results taken from this proteome analysis of RPE lipofuscin with an immunoblot analysis of the same samples, we show that a variety of these lipofuscin-associated proteins were damaged by aberrant covalent modifications. Thus, in 32 of the 81 spots found by Coomassie staining, an MDA modification was detected by immunoblot analysis (Fig. 1A) . Furthermore 15 spots gave a possible signal for HNE (Fig. 1B) and 4 spots a signal for AGEs (Fig. 1C) . The results are summarized in Table 1 1 . Pairs of spots (34,36; 37–39; or 55,56) contained the same protein at different gel positions, indicating different modifications within the molecules. Spot 20 (gel position at 40 kDa; molecular weight 53.3 kDa) revealed a lower molecular weight (MW) compared with theoretical data representing a probably partial enzymatically degraded protein, whereas spot 49 (gel position at 30 kDa; MW 23.5 kDa) showed a higher molecular weight, indicating a possible cross-linking to other molecules. A couple of proteins showed combined MDA- and HNE-modifications (spots 17, 21, 23, 24, 36, and 64). Spot 61 was altered by all three modification types. Multiple mitochondrial immunostained proteins (spots 18, 20, 23, 24, 27, 33, and 57) were detected in the lipofuscin, indicating that autophagy of modified proteins within the RPE contributes to lipofuscinogenesis. The presented series of experiments were performed twice. 

Discussion

Evidence has been brought forward to suggest that lipid peroxidation plays a role in the pathogenesis of various age-related diseases, including atherosclerosis, ²² Alzheimer’s disease, ²³ Parkinson’s disease, ²⁴ diabetic retinopathy, ²⁵ and AMD. ^{8 15 20 26} Lipid peroxidation, a complex process that occurs in all cellular membranes and involves the interaction of oxygen-derived free radicals with polyunsaturated fatty acids (PUFAs), finally results in a variety of reactive aldehydes. ^{9 27} Thus, MDA, the most abundant breakdown product of lipid peroxidation, originates from the peroxidation of unsaturated bonds in PUFAs and subsequent nonenzymatic degradation of the lipid peroxides to MDA. HNE, an especially cytotoxic aldehyde, ^{10 28} is created similarly by peroxidation of lipids, such as arachidonic and linoleic acid esters. The very reactive electrophilic aldehydes are capable of easily attaching covalently to proteins by forming adducts with cysteine, lysine, or histidine residues, and thereby damaging the protein structure. Such damage may affect the protein function and its own catabolism. There is also convincing evidence linking the formation of lipofuscin to this lipoperoxidation damage. ²⁹ Aging and the progression of several degenerative diseases are accompanied by accumulation of intracellular lipofuscin within granules composed in part of damaged protein. ^{17 30} In the eye, lipofuscin accumulates within the RPE throughout life, eventually occupying up to 19% of cytoplasmic volume by 80 years of age, ³¹ and it is likely that condensation between lipoxidation-derived reactive aldehydes and protein groups may represent a process common to the formation of lipofuscin. Diminished susceptibility to proteolysis after protein modification by MDA and HNE was suggested to be a causative factor in lipofuscin generation. ^{32 33} From our present study it is evident that a major part of the proteins detectable in human RPE lipofuscin are modified by MDA or HNE adducts. Photoreceptor outer segments contain high concentrations of PUFAs (up to 70%) that can be easily peroxidized in the presence of high retinal oxygen concentrations and lifelong UV irradiation. Shed photoreceptor outer segments are constantly phagocytosed and degraded by RPE cells. Therefore, phagocytosis of rod outer segments (ROS) constituents damaged by lipid peroxidation and subsequent formation of MDA- or HNE-adducts is likely to be a source of material resistant to lysosomal degradation, finally resulting in deposition of lipofuscin. ^{4 34} Although ROS lack mitochondria, we detected mitochondrial proteins that had MDA or HNE modifications, indicating that autophagy of damaged proteins may also contribute to lipofuscin formation. This is in line with the assumption that autooxidative damage, particularly in mitochondrial membranes, generates precursors of lipofuscin. ^{26 29} It has also been observed that lipofuscin may act as a photosensitizer, thereby producing lipid peroxidation within the granule. ^{26 35} Thus, the protein modifications observed in the present study may also at least partially arise from such intragranular or intralysosomal reactions. Our results are supported by recent findings of Crabb et al., ¹⁵ who observed a variety of cross-linked species in common drusen proteins generated from lipoxidation and hypothesized that this process may have impact on drusen formation. 

The described alterations of proteins detected in the lipofuscin proteome would require a series of events including oxidative reactions outside the RPE (e.g., in the photoreceptors) and transport (phagocytosis) of the reaction products to the lysosomal compartment of the RPE cells. Also in the case of autophagy, active transport processes are required. Such transport does not occur at temperatures below 20°C. All steps in the preparation of lipofuscin in our study were performed below 4°C and were protected from direct light; therefore, a significant alteration of the results by processes occurring during the isolation of lipofuscin appears unlikely. 

Enhanced oxidative-carbonyl stress is also thought to participate in the formation of AGEs, which may also contribute to lipofuscinogenesis. ^{11 13 17} The detection of AGE-modified proteins in ocular lipofuscin shown in our present study supports this hypothesis. The glycoxidation product N^ε-(carboxymethyl)lysin (CML) was found in soft macular drusen, basal laminar deposits, and subfoveal membranes of patients with AMD and colocalized with the receptor for advanced glycation end products (RAGE). ^{36 37}  

Besides the effect of forming nondegradable constituents of lipofuscin deposits that physically impair cell structure and function or act as photosensitizers, ^{35 38} MDA-, HNE- and AGE-modified proteins may actively contribute to inflammatory reactions associated with the development of advanced AMD. ³⁹ Thus, cellular remnants and debris derived from degenerated RPE cells are considered a chronic inflammatory stimulus that is involved in the pathogenesis of AMD. ^{15 36 37 40 41}  

The identification of lipid peroxidation and glucoxidation products in proteinaceous lipofuscin components in human RPE cells may indicate the participation of such biomolecules not only in lipofuscinogenesis but also in the pathogenesis of complex and monogenetic retinal diseases such as AMD and Stargardt disease. Their identification may help to identify potential future treatment targets. AGE inhibitors (e.g., pimagedine or pyridoxamine) have been shown to reduce the severity of diseases involving advanced glycosylation and could offer potential treatment targets for currently untreatable blinding retinal diseases. ^{42 43}  

 Supported by German Research Foundation, Bonn, Germany, Grants Schu 1388/2-1; Ho 1926/2-1, DFG-Research-Priority Program and the State of Baden-Wuerttemberg Research Fund Grant 500/2000.

 Submitted for publication February 19, 2003; revised March 23, 2003; accepted April 14, 2003.

 Disclosure: F. Schutt, None; M. Bergmann, None; F.G. Holz, None; J. Kopitz, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Detection of MDA, HNE, and AGE modifications in human ocular lipofuscin. Lipofuscin was isolated from the RPE-choroid complex of human donors. Proteins were extracted from the pure lipofuscin fraction, separated by 2-D-gel electrophoresis, and MDA (A), HNE (B), and AGE (C) modifications detected by Western blot analysis.

Table 1.

 

Identification of MDA, HNE, and AGE Modifications in Human Ocular Lipofuscin

Spot	Protein	NCBI	MW (kDa)	Modifications
								MDA	HNE	AGEs
1	Hypothetical protein (alpha-spectrin homolog)	7512790	152.4
2	Not identified
3	Tumor rejection antigen (gp96)	4507677	92.7
4	Tumor rejection antigen (gp96)	4507677	92.7
5	Not identified
6	Hexokinase 1	15991833	102.2
7	Not identified			X	X
8	Motor protein	5803115	84.0		X
9	Valosin-containing protein	6005942	90.0
10	Heat shock protein HSP 90-alpha (HSP 86)	123678	85.0		X
11	Heat shock 70 kD protein 5	16507237	72.3
12	Annexin VI (lipocortin VI)	113962	75.8
13	Annexin VI (lipocortin VI)	113962	75.8
14	Annexin VI (lipocortin VI)	113962	75.8
14a	Heat shock 70 kDa protein 8	5729877	70.8
15	Calreticulin precursor; sicca syndrome antigen A	4757900	48.1			X
16	Tubulin, beta chain	5174739	49.6	X
	Prolyl-4-hydroxylase beta-subunit	2507460	57.0
17	Tubulin, alpha-1 chain	135395	50.8	X	X
17a	Tubulin beta chain	135448	50.2
18	ATP synthase beta chain	1145449	56.5	X
19	Enolase 2, gamma	5803011	47.6
20	Ubiquinol-cytochrome c reductase core protein I	4507841	53.3	X
21	Pyruvate kinase, muscle	14750405	58.5	X	X
22	Pyruvate kinase, muscle	14750405	58.5	X
23	ATP synthase, alpha subunit	4757810	59.7	X	X
24	ATP synthase, alpha subunit	4757810	59.7	X	X
25	Enolase 1, alpha	4503571	47.5	X
26	Enolase 1, alpha	4503571	47.5	X
27	S-arrestin	14737493	45.3	X
	NADH dehydrogenase Fe-S protein 2	4758786	52.5	X
28	Glutamate-ammonia ligase	15297214	42.7			X
	Tu translation elongation factor	4507733	49.5			X
	NADH dehydrogenase Fe-S protein 2	4758786	52.5			X
29	Serum albumin	28592	71.3
30	Creatin kinase	180570	42.9
30a	Beta actin	4501885	42.1
31	Vimentin	14742600	53.7
32	Guanine nucleotide-binding protein (0), alpha subunit 2	232134	40.6
33	Guanine nucleotide-binding protein, beta polypeptide 2	4885283	38.1	X
	Pyruvate dehydrogenase	2144337	39.2	X
34	Annexin A2	16306978	38.8		X
35	Annexin A2	16306978	38.8
36	Annexin A2	16306978	38.8	X	X
37	Glyceraldehyde-3-phosphate dehydrogenase	31645	36.2	X
38	Glyceraldehyde-3-phosphate dehydrogenase	31645	36.2	X
39	Glyceraldehyde-3-phosphate dehydrogenase	31645	36.2	X
40	Ubiquinol-cytochrome c reductase core protein II	14775827	48.6
41	Porin 31HM (anion channel 1)	238427	30.7	X
42	Voltage-dependent anion channel 2	4507881	32.1
43	Porin 31HM (anion channel 1)	238427	30.7	X
44	Not identified
45	Voltage-dependent anion channel 2	4507881	32.1	X
46	Glyceraldehyde-3-phosphate dehydrogenase	31645	36.2
47	H119n carbonic anhydrase Ii	2554664	29.1
48	H119n carbonic anhydrase Ii	2554664	29.1	X
48a	Phosphoglycerate mutase I	4505753	28.9
49	Crystallin, beta B2	4503063	23.5		X
50	Annexin IV	14738103	27.2
51	Annexin V (Lipocortin V)	999937	35.8	X
	Retinaldehyde binding protein 1	4506541	36.5
52	Tyrosine 3-monooxygenase/tryptophan 5-monoox. activation protein, epsilon	5803225	29.3
53	Tyrosine 3-monooxygenase/tryptophan 5-monoox. activation protein, theta	5803225	29.3		X
	Tyrosine 3-monooxygenase/tryptophan 5-monoox. activation protein, beta	4507949	28.1		X
	Tyrosine 3-monooxygenase/tryptophan 5-monoox. activation protein, zeta	4507953	27.7		X
54	Cathepsin D	4503143	45.0
55	Prohibitin	4505773	29.8	X
						(continues)

Table 1A.

 

Table 1 (continued). Identification of MDA, HNE, and AGE Modifications in Human Ocular Lipofuscin

56	Prohibitin	4505773	29.8	X
56a	Cathepsin D	4503143	45.0
57	ATP synthase, subunit d	5453559	18.5	X
58	Recoverin	4506459	23.2
59	Calmodulin 2 (phosphorylase kinase, delta)	14250065	16.8	X
60	ATP synthase, H+transporting	4502297	17.4
61	Cytochrome c oxidase subunit Va precursor	4758038	16.9	X	X	X
62	Peroxiredoxin 2	13631440	22.0	X
63	Crystallin, alpha A	4503055	20.0	X
64	Crystallin, alpha B	4503057	20.1	X		X
65	Crystallin, alpha B	4503057	20.1
66	Neuropolypetide h3	9131159	21.0
67	Ubiquinol-cytochrome c reductase bindidng protein	5454152	13.5
	Hemoglobin alpha 2	4504345	15.3
68	Delta globin	4504351	16.2
69	Hemoglobin alpha 1 globin chain	13195586	10.7
70	Chain A, P11 (S100a10), ligand of annexin Ii	3212355	11.2
71	Mutant hemoglobin beta chain	18418633	16.1
72	Mutant hemoglobin beta chain	18418633	16.1
73	Mutant hemoglobin beta chain	18418633	16.1
74	Cellular retinoic acid binding protein 1	18314500	15.7
75	Cytochrome c oxidase subunit Vib	4502985	10.4		X
76	Glial fibrillary acidic protein	17479453	42.2