Effects of Fucoidans from Five Different Brown Algae on Oxidative Stress and VEGF Interference in Ocular Cells

Background: Fucoidans are interesting for potential usage in ophthalmology, and especially age-related macular degeneration. However, fucoidans from different species may vary in their effects. Here, we compare fucoidans from five algal species in terms of oxidative stress protection and vascular endothelial growth factor (VEGF) interference in ocular cells. Methods: Brown algae (Fucus vesiculosus, Fucus distichus subsp. evanescens, Fucus serratus, Laminaria digitata, Saccharina latissima) were harvested and fucoidans isolated by hot-water extraction. Fucoidans were tested in several concentrations (1, 10, 50, and 100 µg/mL). Effects were measured on a uveal melanoma cell line (OMM-1) (oxidative stress), retinal pigment epithelium (RPE) cell line ARPE19 (oxidative stress and VEGF), and primary RPE cells (VEGF). Oxidative stress was induced by H2O2 or tert-Butyl hydroperoxide (TBHP). Cell viability was investigated with methyl thiazolyl tetrazolium (MTT or MTS) assay, and VEGF secretion with ELISA. Affinity to VEGF was determined by a competitive binding assay. Results: All fucoidans protected OMM-1 from oxidative stress. However, in ARPE19, only fucoidan from Saccharina latissima was protective. The affinity to VEGF of all fucoidans was stronger than that of heparin, and all reduced VEGF secretion in ARPE19. In primary RPE, only the fucoidan from Saccharina latissima was effective. Conclusion: Among the fucoidans from five different species, Saccharina latissima displayed the most promising results concerning oxidative stress protection and reduction of VEGF secretion.


Introduction
Fucoidans are sulfated polysaccharides found in the cell walls of brown algae. Common to all fucoidans is a high amount of L-fucose, yet their structures are complex and variable among different species [1]. Fucoidans have been described as exerting interesting pharmacological activities including, e.g., anti-inflammatory, antitumorigenic, and anti-angiogenic effects [1,2]. In particular, a fucoidan has been found to be potentially beneficial in age-related macular degeneration (AMD), the most common cause of blindness and severe vision loss in the Western world [3,4]. AMD is a disease of the elderly in which photoreceptors and retinal pigment epithelial cells of the macula, the area of high acuity vision, degenerate, and, in the more severe exudative subtype of the disease (wet AMD), vessels grow from the choroid under and into the retina. These immature vessels are leaky and lead to fluid Figure 1. Characterization of the susceptibility of cell lines to oxidative stress. Cell viability was tested in OMM-1 (a) and ARPE19 (b) exposed to H 2 O 2 (a,b) and tert-Butyl hydroperoxide (TBHP) (c). Significance was evaluated with Friedman's ANOVA and Student's t-test, + p < 0.05, ++ p < 0.01, +++ p < 0.001 compared to control (n > 3).
Taken together, the results differ profoundly from those found with OMM-1 cells, with only Saccharina latissima showing protection, Laminaria digitata and Fucus vesiculosus displaying no influence, and Fucus serratus and Fucus distichus subsp. evanescens showing an additional toxic effect on the cells.

VEGF Secretion
The effect of the five different fucoidans on the secretion of VEGF was compared in two different RPE cell types, the human cell line ARPE19 and primary RPE cells derived from porcine eyes. Untreated ARPE19 cells secreted considerably and significantly less VEGF, with a mean of 416.33 pg/mL (±415.27), in a collection time of 24 h compared to primary porcine RPE cells, which secreted a mean of 2386.88 pg/mL (±824.81) (p < 0.001) in a collection time of 4 h, as tested in ANOVA and subsequent t-test.
Based on these results, we investigated VEGF content after 3 days of incubation with fucoidan for all following experiments. VEGF content was normalized to cell viability and displayed in relation to an untreated control.
In ARPE19 cells, all tested fucoidans reduced the amount of secreted VEGF, with the highest concentration (100 µg/mL) generally displaying the strongest effect. The differences between the different fucoidans are not as profound as those found in protective effects and are generally limited to effects seen at certain concentrations. LD fucoidan differs only slightly from FV or FS fucoidan (both 10 µg/mL, p < 0.05) or from FE fucoidan (10 µg/mL and 100 µg/mL both p < 0.05), while there were significant differences compared to SL fucoidan in all concentrations tested (1-10 µg/mL p < 0.001; 50 µg/mL p < 0.05; 100 µg/mL p < 0.01). SL fucoidan significantly differed from FV fucoidan in all concentrations but 10 µg/mL (1 µg/mL p < 0.001; 50 µg/mL p < 0.05; 100 µg/mL p < 0.01), while it differed from FE and FS fucoidan at concentrations of 1 µg/mL and 100 µg/mL (both p < 0.001). FV fucoidan did not display any significant differences from FS fucoidan, and only at 100 µg/mL when compared to FE fucoidan (p < 0.05). Finally, FE and FS fucoidan did not display any significant differences (Table 3). Ranging the effect on VEGF secretion results in LD > FS > FE > FV > SL.
In primary RPE cells, only Saccharina latissima showed a reductive effect on VEGF content in the supernatant. When comparing the different fucoidans, significant differences were only found between SL and LD fucoidan (at a concentration of 10 µg/mL) and between FV and FS fucoidan (at a concentration of 50 µg/mL) ( Table 4). . VEGF content was evaluated in ELISA and normalized to cell viability. Control = 1. In RPE cells, only SL fucoidan reduced the VEGF content of RPE cells (10 µg/mL), while LD, FV, and FE induced a higher signal at concentrations of 50 and 100 µg/mL. Significance was evaluated with Friedman's ANOVA and subsequent Student's t-test, + p < 0.05 reduction compared to the control, * p < 0.05, ** p < 0.01 and *** p < 0.001 (n = 7).

Binding Affinity to VEGF
In the pathogenesis of the exudative form of AMD, VEGF is an important factor [5,6]. VEGF inhibitors are currently used as a therapy standard [11]. As has been known for a long time that heparin and other sulfated polysaccharides have a high affinity for VEGF [20], fucoidans may not only reduce the secretion of VEGF but also directly antagonize its actions. Therefore, we investigated the binding affinity of fucoidans to VEGF in comparison to heparin, and aimed to identify those fucoidans with the strongest binding affinity to VEGF.
In the competitive VEGF binding assay used, the replacement of biotinylated heparin by the test compound was measured. At a concentration of 0.5 µg/mL, maximum binding of biotinylated heparin was achieved (data not shown). By adding increasing concentrations of unlabeled heparin, the detection of biotinylated heparin decreased (data not shown). As expected, incubation with a mixture of 0.5 µg/mL heparin and 0.5 µg/mL biotinylated heparin reduced the signal to 49.96% (±7.28) and thus the binding of biotinylated heparin to VEGF. All five fucoidans showed a significant affinity to VEGF compared to the control, tested in ANOVA and subsequent testing (SL 26.13 ± 13.46%; LD 28.57 ± 13.20%; FS 10.63 ± 6.43%; FV 21.27 ± 4.69% and FE 16.97 ± 9.91%) ( Figure 7). Moreover, the three Fucus fucoidans bound significantly more strongly to VEGF than heparin (p < 0.01) ( Table 5). Although FS and FE showed a slightly higher affinity, there were no significant differences between the five fucoidans (Table 5).

Discussion
Fucoidans are of great interest to biomedical research, especially considering their possible application in age-related macular degeneration [4]. However, the effects of fucoidans can profoundly vary depending on their origin and method of extraction. Therefore, generalized statements about their activities should be avoided.
In this study, we have compared fucoidans from five different brown algae species harvested in summer and obtained from the same supplier (Coastal Research & Management, Kiel, Germany). The preparation of the algal material as well as the extraction and purification of the fucoidans were performed in parallel according to a standardized procedure. Therefore, this study is well equipped to reliably examine the effects of fucoidans from different algal species. The aim was to compare the beneficial properties of fucoidans with regard to age-related macular degeneration.
In our study, we focused on two major factors important for the development of age-related macular degeneration, i.e., oxidative stress and VEGF. While interference with these is without doubt of great interest, we are aware that other activities, e.g., anticomplementary, anti-inflammatory effects, or influence on lipid metabolism, are also potentially beneficial for impairing AMD pathogenesis and are of high interest for further testing.
We have previously shown that OMM-1 cells are protected by commercially available fucoidan [17]. In contrast to uveal melanoma cells, RPE cells are intrinsically strongly protected against oxidative stress [18], so we tested oxidative stress protection both in OMM-1 cell lines, which we know could be protected by fucoidans, having demonstrated this for commercial fucoidan from Fucus vesiculosus [17], as well as the RPE cell line ARPE19. All five fucoidans protected OMM-1 cells from oxidative stress, confirming our previous results. However, fucoidans can differ in their properties depending on the test system or the cell types due to distinct cellular and molecular pathways. A major task of RPE cells is oxidative stress protection [18] and, as mentioned above, they are naturally highly resistant to oxidative stress. ARPE19 cells, as an RPE cell line, consequently behaved differently from OMM-1 cells, not only concerning their susceptibility to oxidative stress itself, but also their reaction to fucoidans. In contrast to OMM-1 cells, only SL fucoidan displayed a protection of cell viability, while FE and FS fucoidan even exacerbated the effect.
Little is known about oxidative stress pathways in uveal melanoma cells, but an increased susceptibility due to the reduction of superoxide dismutase activity in uveal melanoma cells has been described [21]. There is a debate about the direct reactive oxygen species (ROS) scavenging effect of fucoidans. The ROS scavenging potency of fucoidan has been described in several publications [22][23][24][25], but it has recently been demonstrated in cell-free systems that these measured effects are mainly due to co-extracted phenolic and terpenoid compounds in the fucoidan preparations [15,26]. Generally, the scavenging effect of fucoidan against hydrogen peroxide has been described to be rather weak [24,25]. Our results even suggest that the in vitro ROS scavenging activity may be irrelevant as, in contrast to SL and LD, the three Fucus fucoidans showed an ROS scavenging effect (manuscript in preparation) but did not protect ARPE19 cells from oxidative stress.
However, independent of all of the results from simple cell-free assays, fucoidans may exhibit antioxidative activity by cellular effects. Accordingly, fucoidan has been shown to increase the expression of superoxide dismutase (SOD) in several experimental models and activate the transcription factor nuclear factor erythroid 2-like 2 (Nrf2), the "master regulator" of the antioxidative stress response [27][28][29][30][31][32]. Both the overall protective effect of our fucoidans on OMM-1 and the limited protective effect on ARPE19 cells (only found for Saccharina latissima fucoidan) could be explained via these pathways. As mentioned above, uveal melanoma cells have been described to have a reduced SOD activity [21]. Hence, the SOD-inducing effect of fucoidans could protect these cells against an oxidative stress insult. In contrast, RPE cells have a high intrinsic stress response level mediated by Nrf2 [18]. Indeed, knock-out of Nrf2 renders RPE cells highly susceptible to oxidative stress insults [33] and Nrf2 knock-out mice develop AMD-like features at an older age [34]. It is feasible that these protective pathways are already at maximum efficacy, so that any further enhancement by fucoidans may be impossible. Therefore, the different effects of fucoidans on the two cell lines are assumed to be due to distinct impacts on the respective cellular pathways. Further research needs to be conducted to elucidate these pathways. It should be noted, however, that oxidative stress protection in AMD is not only needed for (the rather resistant) RPE cells, but also for the rather fragile photoreceptor cells [35]. Therefore, the effect of fucoidans on oxidative stress-induced photoreceptor cell death should also be evaluated in further studies.
A major contributor to the pathology of wet AMD is the growth factor VEGF, and its inhibition is the only current treatment option for AMD patients. We have previously shown that commercially obtained fucoidan from Fucus vesiculosus additively reduced VEGF expression when co-applied with the VEGF inhibitor bevacizumab [3]. On the molecular level, the interaction of fucoidan with VEGF differs profoundly from that of the current therapeutic anti-VEGF molecules. VEGF antibody-derived compounds bevacizumab and ranibizumab, as well as the fusion protein aflibecept, interact with specific amino acids in the receptor-binding domain of VEGF, causing a steric inhibition of the binding of VEGF to its receptor [36], with differences in affinities between the compounds [37]. The interaction of fucoidan and other heparin-related compounds is complex, however, depending on features such as sulfation and molecular weight [38,39]. Furthermore, fucoidan has been shown to also have a binding affinity to VEGF receptors and to facilitate the internalization of VEGF receptors, blocking the binding and in-vitro functions of VEGF [38,40,41].
In our current study, we were able to demonstrate antagonization as well as a reduction of VEGF secretion in ARPE19 cells, in which all five fucoidans were effective. In primary porcine RPE cells, however, only SL displayed a significant effect.
Fucoidan was found to influence Stat3-regulated promoters, which includes the promoter of VEGF [42]. But it should be noted that we have previously shown that in unchallenged primary RPE cells, Stat3 is not involved in constitutive VEGF expression [43]. Therefore, this mechanism of fucoidan-mediated VEGF reduction is not feasible. We have also previously shown that VEGF is positively regulated in an autocrine way via the VEGFR-2 [43,44], whereby fucoidan has been shown to bind to VEGF165 and to competitively inhibit the interaction of VEGF with VEGFR2 [39,40]. This pathway has been suggested to be involved in VEGF reduction mediated by fucoidan [3,40]. Such an extracellular mode of action for fucoidans is now supported by the binding of the five fucoidans to VEGF. Their affinity was significantly higher than that of heparin, whereas heparan sulfate was not able to reduce the binding of biotinylated heparin to VEGF. Interaction of VEGF with heparin sulfate on the cell surface was found to be involved in effective VEGFR2 activation [45]. Thus, the fucoidans may competitively prevent this interaction and thus attenuate signaling through VEGFR2, resulting in reduced VEGF expression and secretion. In line with this assumed mode of action are the findings that the intraocular injection of heparan sulfate or heparin in mice eyes with aberrant angiogenesis results in reduced neovascularization [46]. Given the even higher affinity to VEGF of fucoidans, this seems promising.
The amount of VEGF secreted by primary RPE cells in this study was much higher than that of ARPE19 cells (596.72 pg/h for primary RPE vs. 17.35 pg/h for ARPE19; factor 34.4). In the presence of such high VEGF concentrations, the VEGF antagonizing mode of action obviously became ineffective, explaining the discrepant results in ARPE19 cells and RPE cells. Therefore, the effect of SL on VEGF secretion in primary RPE cells is even more remarkable.
Our data clearly show a positive effect of the tested fucoidans in terms of oxidative stress protection and VEGF inhibition, with the most promising fucoidan extracted from Saccharina latissima. Among the tested fucoidans, SL had the highest degree of sulfation, the highest molecular weight, and the highest degree of purity (under submission). But these parameters cannot explain its superiority or the ranking of the other fucoidans. Fucoidans from Saccharina latissima have been previously shown to be highly biologically active compared to those from other brown algae species [15,16]. Furthermore, in line with our finding concerning VEGF inhibition, fucoidans from Saccharina latissima have been shown to inhibit angiogenesis in tumor models [47]. However, it seems too early to decide that the other fucoidans are not worth further investigation. Other activities beneficial for AMD therapy should be regarded as well. Further preclinical and clinical research is warranted, but fucoidans may be a potential treatment option for age-related macular degeneration. To develop potential therapeutics from fucoidan, in addition to finding the most suitable source and a sustainable and reliable harvest and extraction method, bioavailability and application forms need to be tested.
In addition, VEGF secretion and oxidative stress are also involved in the pathomechanisms of diabetic retinopathy [48,49]. Therefore, fucoidans may also be of great interest for diabetic patients, especially considering that fucoidan may also reduce blood glucose levels and ameliorate hypertension [4].
In conclusion, we compared fucoidan from five brown algae species in terms of three activities that are considered promising for the treatment of AMD, i.e., their capacity for oxidative stress protection, inhibition of VEGF secretion, and binding affinity to VEGF. Based on these three basic parameters, the fucoidan from Saccharina latissima turned out to be most suitable for further investigations.

Fucoidans
Fucoidans were extracted from dried stocks of the species Saccharina latissima (SL; Atlantic, Funningsfjord, Faroe Island), Fucus vesiculosus (FV; Baltic, Kiel Bay, Germany), Laminaria digitata (LD; Atlantic; Churchbay, Island ), Fucus distichus subsp. evanescens (FE; Baltic, Kiel Cana, Germany), and Fucus serratus (FS; Baltic, Kiel Bay, Germany), all harvested in summer and provided by Coastal Research & Management, Kiel, Germany. The fucoidans were extracted as previously described [53]. Briefly, the pulverized algal material was defatted by Soxhlet extraction with 99% (v/v) ethanol, and was then extracted with aqueous 2% calcium chloride for 2 h at 85 • C under reflux conditions. The supernatants of the raw extracts were concentrated and precipitated with ice-cold ethanol in a final concentration of 60% (w/w). After centrifugation, the sediments were dissolved in demineralized water, dialyzed, and lyophilized. In addition, commercially available fucoidan from Sigma (Sigma-Aldrich, Deisenhofen, Germany, F8190) was used. Fucoidan was solved in Ampuwa bidest (Fresenius, Schweinfurt, Germany), and then further diluted with appropriate cell medium for the cell experiments and phosphate buffered saline (PBS) for the VEGF binding assay, filtered through a 0.2 µm filter (Sarstedt, Nümbrecht, Germany), and applied to the cells in final concentrations of 1, 10, 50, and 100 µg/mL.

OMM-1
OMM-1 cells were treated with hydrogen peroxide (H 2 O 2 ) to induce oxidative stress-related cell death, as previously shown [17]. As OMM-1 is a cancer cell line that may change its characteristics during subculture, we evaluated the appropriate concentrations of H 2 O 2 resulting in approximately 50% cell viability. In order to assess this, OMM-1 cells were treated with different concentrations (100, 200, 400, 1000 mM) of H 2 O 2 for 24 h and cell viability was investigated by MTS assay (see below). To investigate the potential protective effects of the different fucoidans, a concentration of 1 mM H 2 O 2 was chosen. Cells were treated with fucoidan (1, 10, 50, and 100 µg/mL) 30 min prior to the application of H 2 O 2 .

ARPE19
Corresponding experiments to find the appropriate H 2 O 2 concentration for ARPE19 cells revealed that none of the tested concentrations of H 2 O 2 (100 µM, 200 µM, 400 µM, 1000 µM) induced a cell death of about 50% after 24 h, as detected by MTS assay. In addition, the cell death rate was highly variable after H 2 O 2 incubation (see results). Therefore, we tested tert-Butyl hydroperoxide (TBHP), a more stable inducer of oxidative stress in RPE cells [33], at concentrations of 100 µM, 250 µM, and 500 µM, for 24 h and investigated cell viability by MTS assay, as described below. In order to investigate the potential protective effect of the different fucoidans, a concentration of 500 µM TBHP was chosen. Cells were treated with fucoidan 30 min prior to the insult.

Methyl Thiazolyl Tetrazolium (MTT) Assay
MTT assay is a common method in cell research [54] and was conducted as previously described [3]. In brief, after treatment with the fucoidans, the cells were washed and incubated with 0.5 mg/mL MTT (dissolved in DMEM without phenol red). After removal and further washing of the cells, cells were lysed with dimethyl sulfoxide (DMSO) and the absorbance was measured at 550 nm with a spectrometer (Elx800, BioTek, Bad Friedrichshall, Germany).

MTS Assay
The MTS assay is a commercially available viability assay and was used according to the manufacturers' instructions (CellTiter 96 ® AQueous One Solution Cell Proliferation Assay (Promega, Mannheim, Germany)). The cell viability assay was performed in 96 well plates in phenol red-free medium with the same supplements described above. In each well, 20 µL of MTS solution was added for 1 h.

VEGF ELISA
VEGF was detected in the supernatants of ARPE19 and primary RPE cells using commercially available ELISA kits (R&D Systems, Wiesbaden, Germany) according to the manufacturer's instructions. To establish the parameters of VEGF ELISA, we investigated time-dependent VEGF secretion in ARPE19 and primary RPE cells in the presence and absence of commercially available fucoidan from Fucus vesiculosus (Sigma-Aldrich, F8190) for 1 day, 3 days, and 7 days. A cell viability assay (MTT) was conducted after 7 days. According to these results, an incubation time of 3 days was chosen for experiments with the five different fucoidans. The medium was changed 24 h prior in ARPE19 cells and 4 h prior in primary RPE cells and the supernatant collected. Measured VEGF content was normalized for cell survival and is depicted in relation to that of untreated control cells.

Competitive VEGF Binding Assay
The affinity to VEGF of the test compounds was investigated with a competitive VEGF-binding assay using biotinylated heparin. In addition to the fucoidans, heparin (EDQM, no. Y0001282, Strasbourg, France) was tested.
The wells of a 96-well Nunc-Immuno MaxiSorp microplate (Sigma-Aldrich, Deisenhofen, Germany) were coated with 0.1 µg recombinant human VEGF 165 (R&D Systems Cat. 293-VE/CF) dissolved in 100 µL PBS overnight at 4 • C. After washing with PBS, the coated wells were blocked for 90 min at 37 • C with 100 µL of 5 mg/mL bovine serum albumin (BSA, dissolved in PBS) and subsequently washed three times with PBS. During the blocking, 65 µL of 1 µg/mL heparin, biotin conjugate (Merck, Darmstadt, Germany) in PBS, and 65 µL of 1 µg/mL test compounds in PBS were preincubated at 4 • C. For the blank and the 100% binding value, 100 µL PBS and 100 µL of 0.5 µg/mL biotinylated heparin in PBS, respectively, were preincubated at 4 • C. Aliquots of 100 µL of these solutions were pipetted into the coated microplate wells and incubated for 2 h at 37 • C with gentle agitation. After three washing steps, 100 µL of streptavidin alkaline phosphate conjugate (Southern Biotech/, Birmingham, AL, USA, stock solution diluted 1:3000 with PBS) was incubated for 1 h at 37 • C with gentle agitation. The next steps involved three washings with PBS and incubation with 100 µL p-nitrophenyl phosphate substrate system (Sigma-Aldrich, Deisenhofen, Germany) for 30 min in the dark. The reaction was stopped by addition of 25 µL 3 N NaOH, and the absorbance was measured at 405 nm. Blank values in the absence of biotinylated heparin were subtracted from the measured values. The reduction of the binding of biotinylated heparin by the test compounds is indicated as a percentage in relation to the binding of the biotinylated heparin alone.

Statistics
All experiments testing fucoidans were independently repeated at least six times, experiments for establishing oxidative stress response were repeated at least three times, and the VEGF binding experiments were performed in duplicates on three different days. Statistics were calculated using Statistica 7 (Statsoft, Tulsa, OK, USA) and Microsoft Excel (Excel 2010, Microsoft, Redmond, WA, USA). A Friedman's ANOVA was performed, and, if a significant difference between groups was detected, a subsequent Student's t-test was conducted. A p value of <0.05 was considered significant. All bars represent mean and standard deviation.