Marine organism sulfated polysaccharides exhibiting significant antimalarial activity and inhibition of red blood cell invasion by Plasmodium

The antimalarial activity of heparin, against which there are no resistances known, has not been therapeutically exploited due to its potent anticoagulating activity. Here, we have explored the antiplasmodial capacity of heparin-like sulfated polysaccharides from the sea cucumbers Ludwigothurea grisea and Isostichopus badionotus, from the red alga Botryocladia occidentalis, and from the marine sponge Desmapsamma anchorata. In vitro experiments demonstrated for most compounds significant inhibition of Plasmodium falciparum growth at low-anticoagulant concentrations. This activity was found to operate through inhibition of erythrocyte invasion by Plasmodium, likely mediated by a coating of the parasite similar to that observed for heparin. In vivo four-day suppressive tests showed that several of the sulfated polysaccharides improved the survival of Plasmodium yoelii-infected mice. In one animal treated with I. badionotus fucan parasitemia was reduced from 10.4% to undetectable levels, and Western blot analysis revealed the presence of antibodies against P. yoelii antigens in its plasma. The retarded invasion mediated by sulfated polysaccharides, and the ensuing prolonged exposure of Plasmodium to the immune system, can be explored for the design of new therapeutic approaches against malaria where heparin-related polysaccharides of low anticoagulating activity could play a dual role as drugs and as potentiators of immune responses.


Results
Characterization of sulfated polysaccharide size and integrity. I. badionotus FucCS has simple branches of sulfated α -fucose (Fig. 1a), composed of a single monosaccharide unit, either disulfated at positions 2 and 4 (~90%) or exclusively 4-sulfated (~10%) 35 . L. grisea FucCS has more complex branching structures, mostly composed of disaccharide units of α -fucose, non-sulfated and 3-sulfated at the nonreducing and reducing ends, respectively. This FucCS also has small amounts of branches composed of single α -fucose units, either 2,4-disulfated (~27%) or 2,3-disulfated (~20%) 35 . The linear sulfated fucans from these echinoderms (Fig. 1b) contain repetitive tetrasaccharide sequences, defined by the patterns of sulfation at positions 2 and 4 36,37 , which differ exclusively in the sulfation of the second residue of the tetrasaccharide: 2-sulfated in I. badionotus and non-sulfated in L. grisea. Unlike the majority of sulfated galactans from red algae, that of B. occidentalis (Fig. 1c) has a relatively simple structure, varying only in the sulfation pattern of its units 38 . The sponge glycan used here (Mr ~200 kDa) has had its structure only partially elucidated; preliminary gas chromatography/mass spectroscopy analysis indicated that it is a heteropolysaccharide composed of glucose (75%), fucose (17%) and galactose (8%), with a molar ratio sulfate:total monosaccharide of ~1.5 (data not shown). Integrity of these molecules was analyzed by polyacrylamide gel electrophoresis (Fig. 2), and the result obtained was found to be consistent with the respective approximate molecular masses calculated by size exclusion chromatography after polysaccharide purification from their natural sources, indicating the absence of significant degradation. The bands observed in 6% polyacrylamide gels exhibited a certain degree of size polydispersity, as it is typical of this group of compounds.
Antimalarial and anticoagulating activities in vitro of sulfated polysaccharides. The antimalarial activity of the polysaccharides was analyzed in in vitro cultures of P. falciparum (Fig. 3a), revealing for most of them a significant inhibition of the parasite's growth, with IC50s between 2.3 and 20.3 μg/mL (Table 1). These activities were similar to those found for different heparin batches (between 4 and 18 μg/mL according to our own data) 39 . The sole exception was the D. anchorata glycan, whose antimalarial activity was found to be relatively low, with an IC50 ~66 μg/mL. No correlation was found between Plasmodium growth inhibition and polysaccharide size, since the two best activities were for the largest and smallest structures corresponding, respectively, to the B. occidentalis galactan (~700 kDa) and the L. grisea FucCS (~30 kDa). The sulfated polysaccharides from marine organisms assessed here, especially the fucosylated chondroitin sulfates and sulfated fucans from sea cucumbers, have been showing remarkably homogenous structures, as demonstrated by their coincident NMR spectra obtained from different preparations 29,35,40 . Therefore, if these polysaccharides are extracted and purified properly, no significant variations in structure, and thus in pharmacological properties, are expected between batches.
According to the activated partial thromboplastin time (APTT) determined in vitro for the sulfated polysaccharides used in this work (Fig. 3b and Supplementary Table 1), the three best antimalarial compounds (both FucCSs and the galactan) were the most anticoagulant and the three polymers more innocuous for Plasmodium (both fucans and the sponge glycan) exhibited the worst anticlotting activities. Nevertheless, all compounds had comparatively small anticoagulant activities never above 16% of that from heparin, as demonstrated by their significantly higher doses necessary to double the control APTT (Supplementary Table 1), which suggests that anticoagulating and antimalarial activities are not directly related.

Sulfated polysaccharides inhibit Plasmodium invasion of red blood cells.
Since the antimalarial mechanism of heparin and related polysaccharides had been described to operate through inhibition of the invasion of RBCs by Plasmodium, we proceeded to investigate the invasion inhibition activity of the marine Scientific RepoRts | 6:24368 | DOI: 10.1038/srep24368 sulfated polysaccharides. Late-stage pRBC cultures that had been treated with the different structures revealed upon microscopic examination at 20 h post-treatment a clear decrease in ring stages relative to untreated samples (Fig. 4). Polysaccharide-treated samples thus showed a delay in P. falciparum development, as evidenced by the presence at 40 h within the intraerythrocytic cycle of a significant fraction of ring stages relative to untreated controls which contained, as expected, trophozoites and schizonts only. Quantitative microscopic counts evidenced a clear decrease in the invasion rate of all polysaccharide-treated samples ( Table 2). Maturation rates, on the other hand, were not negatively affected, indicating that if rings are formed, their progression towards trophozoites and schizonts seems to proceed normally. The observation that some of the samples had maturation rates larger than the untreated control suggests that these cases reflected the presence of a significant number of parasites that either completed their invasion or started their differentiation into identifiable rings after the count of ring stages was made. A clear example of this is represented by L. grisea FucCS-treated samples, whose parasitemias at 40 h post-treatment were higher than those expected from the low invasion rate reported for this polysaccharide in Table 2. Likely, the reduced ring numbers observed in microscopic counts of in vitro assays indicates a slower invasion process of otherwise viable merozoites. These retarded invasions (therefore not detected as ring stages) eventually develop into trophozoites, which results in an apparently high maturation rate from rings to late forms. Indeed, when multiplying invasion by maturation rates, which gives an approximate comparative estimation of parasite viability, the values obtained (not shown) are in good agreement with the respective inhibitory effects on parasite growth (Fig. 3), with L. grisea FucCS exhibiting the highest antiplasmodial activity when all polysaccharides are tested at 4 μg/mL. Flow cytometry analyses of P. falciparum cultures treated at late stage with sulfated polysaccharides confirmed a clear dose-dependent invasion inhibition at their respective IC50 (Fig. 5a) and IC90 (Fig. 5b). The accumulated experimental evidence indicates that invasion inhibition is the antimalarial mechanism through which sulfated polysaccharides operate. However, the short time that free merozoites are present in the blood circulation suggests that the process of parasite binding might occur, at least in part, inside pRBCs. Following previously established protocols 39 , we added fluorescein-labeled heparin to live pRBC cultures and after 30 min of incubation the samples were processed for confocal fluorescence microscopy analysis. The resulting data show that heparin added to living pRBC cultures not only specifically targeted pRBCs vs. RBCs in vitro, but it entered live pRBCs and bound intraerythrocytic developing merozoites ( Supplementary Fig. 1).
In vivo antimalarial activity analysis of sulfated polysaccharides. P. yoelii-infected mice were treated iv with polysaccharide doses selected after consideration of their in vitro antimalarial activity, anticoagulation capacity, and unspecific cytotoxicity. Although a significant toxicity in endothelial cell cultures was found at some of the administered doses for most marine sulfated polysaccharides when compared to the same heparin concentrations ( Fig. 6a and Supplementary Table 2), no adverse effects were observed in the animals during the first week of the assay apart from the symptoms characteristic of a malaria infection. Except for D. anchorata glycan and L. grisea fucan, all compounds reduced parasitemia when compared to untreated controls (Table 3). I. badionotus fucan provided the best improvement in mice survival (Fig. 6b), and in one animal treated with this compound parasitemia was reduced from 10.4% at day 4 to undetectable levels. Western blot analysis revealed the presence of antibodies against P. yoelii antigens in the plasma of surviving mice (Fig. 6c). To explore if the observed increased antibody titers were consequence of an immune response against the parasite, the surviving animals were re-inoculated with P. yoelii 73 days after the initial infection; all mice, including that treated with the I. badionotus fucan, survived the new infection without any treatment (Fig. 6b). Microscopic observation of blood smears confirmed the infection of the latter animal, which at day 4 had 18.3% parasitemia (Fig. 6d), and the progressive reduction of parasitemia until its complete elimination (Fig. 6e). All surviving animals were alive and without symptoms of disease at day 42 after the second, untreated infection.

Discussion
Previous work had demonstrated that the presence of sulfate groups was paramount for the binding of L. grisea FucCS to human lung endothelial cells and placenta cryosections under static and flow conditions 14 , and that sulfated FucCS was capable of inhibiting pRBC cytoadherence in these cell models. Because pRBC sequestration in the microvasculature of vital organs plays a key role in the pathogenesis of cerebral and pregnancy malaria, L. grisea FucCS has been proposed for the treatment of severe disease. The crucial role of sulfate groups in the context of malaria was further evidenced by the ability of L. grisea FucCS to disrupt P. falciparum rosettes, which was significantly lost upon desulfation 14 . Other evidences illustrating the physiological importance of sulfate groups came from reports showing that their removal abolished the antithrombotic and anticoagulant effects of FucCS 33 , and that their presence was essential for preserving the inhibitory effects of the polysaccharide in interactions mediated by P-and L-selectin 41 . The in vitro data reported here show that polysaccharides containing α -fucose as internal units are less active as antimalarials than polymers having α -fucose as branches. The sulfated fucan from I. badionotus was found to have a slightly higher in vitro antimalarial activity than that of L. grisea, probably because of the additional 2-sulfation. The observation that both FucCSs have similar in vitro antimalarial effect despite the marked differences in their α -fucose-containing branches suggests that, beyond a minimal threshold, the presence of additional 2,4-disulfated fucose units does not result in higher antiplasmodial potency, as it was similarly reported for the anticoagulant activity of FucCS 35 . Compared to heparin, marine sulfated polysaccharides exert their antimalarial activity in vitro at concentrations where their anticoagulant activity is low. The semi-synthetic heparin-like polysaccharide K5-NSOS-H also showed high antiplasmodial activity despite of its low anticoagulant capacity 16 , related to its lack of iduronic acid units which are essential to promote the interactions with antithrombin and heparin co-factor II 42 . However, to trace a parallel between the antimalarial and anticoagulant activities of heparin-like molecules and fucosylated chondroitin sulfates and sulfated galactan is difficult because these polysaccharides from marine organisms present serpin-independent anticoagulant properties 31,32 . The polysaccharides used in this work do not require fractionation and/or chemical modification after purification 31 and, unlike heparin, are not derived from mammals, thus reducing the risk of contamination by human-affecting pathogens. These compounds are present at high concentrations in marine organisms and can be isolated with relatively high yields of at least ~1% dry weight. The synthesis of sulfated polysaccharides from the marine organisms used here is unfeasible because the enzymatic machinery for their synthesis is still unknown. However, sea cucumbers are already mass cultivated in several countries, especially in China, where they are used as food 43 . Several species of seaweeds are also commercially farmed 44 , and particularly the alga B. occidentalis is abundant in the northeastern coast of Brazil; almost 50% of its dry weight is sulfated galactan 45 , making its harvesting a feasible strategy.
Other already described antimalarial compounds like pentosan polysulfate, curdlan sulfate and dextran sulfate are obtained via chemical sulfation of neutral polysaccharides and show serious side effects such as thrombocytopenia, intracerebral hemorrhage and colitis 46,47 . Curdlan sulfate, which has been proposed as adjunct medication to conventional therapy in patients with severe malaria, has been described to possess as adverse effect an increase in APTT 48 . Fucoidan, also reported to have an inhibitory effect on Plasmodium growth 49 , had a certain level of toxicity for a murine macrophage cell line and was described to occasionally cause eye hemorrhages and death of the animals 49 . Although at concentrations close to their in vitro IC50s marine sulfated polysaccharides exhibited significant toxicity in endothelial cell cultures, higher in vivo amounts did not trigger observable adverse effects in mice during the first week of treatment. Consistently, FucCS can be satisfactorily administered orally 50 , without toxic or cumulative effects in tissue observed after daily doses to animals of 50 mg/kg for 30 days (Mourão, unpublished data). Nevertheless, the potential unspecific toxicity of sulfated polysaccharides in future antimalarial clinical applications can possibly be averted by encapsulating them in pRBC-targeted nanocapsules as it has been reported for other antimalarial agents [51][52][53] . In the case of heparin, the polysaccharide itself has been demonstrated to be capable of acting as targeting molecule of drug-loaded nanocarriers 39 , thus adding to its own antiparasitic action and potentiating therapeutic activity. Some of the pernicious effects of sulfated polysaccharides, such as the anticoagulant activity of heparin, are significantly reduced when immobilized on a substrate 54 . Surface plasmon resonance biosensor studies showed that covalent binding through its carboxyl groups dramatically reduced the interaction of heparin with antithrombin III 55 . Conjugation to nanoparticles can thus be explored as an interesting approach to reduce potential toxic side-effects.
We have observed that formation of ring-stage parasites is clearly reduced in the presence of sulfated polysaccharides, in agreement with preexisting data indicating that their antimalarial activity unfolds by inhibition of merozoite invasion 8,9,[13][14][15][16]49,[56][57][58] . The mechanism through which this invasion blocking proceeds has not been elucidated yet, although the finding that sulfation patterns are crucial for the inhibitory effect of heparin and similar compounds 16 , suggests that it is the result not only of nonspecific ionic interactions but also of particular conformations of anions present in the polysaccharides 8 . Whereas binding of heparin to merozoites has been described to be mediated by multiple protein receptors 17,18 , GAG-pRBC associations are mainly based on interactions with the parasite-derived adhesin, P. falciparum erythrocyte membrane protein 1, PfEMP1 59 . The subsequent internalization of heparin into pRBCs might be an unspecific uptake through the tubulovesicular network induced by Plasmodium during its intraerythrocytic growth 60 . Such entry into pRBCs and coating of developing merozoites before they egress, permits the invasion inhibition activity of heparin to be manifested since the first moment when Plasmodium cells are free in the blood circulation. This is important regarding possible future therapeutic applications of sulfated polysaccharides; if the observed activity were only exerted upon binding to free, extraerythrocytic merozoites, their rapid invasion of RBCs 61 would severely compromise clinical applications. Because heparin is capable of penetrating live pRBCs and of binding intracellular merozoites 39 , heparin-based antimalarial therapies can be administered during the wide time frame when late stages are present in clinical malaria. Smaller fragments of marine polysaccharides might also have this behavior, although the finding that the ~700 kDa B. occidentalis galactan has antimalarial activity similar to that of the ~30-40 kDa FucCSs from L. grisea and I. badionotus suggests that pRBC internalization of the polymers might not be essential for their capacity to inhibit Plasmodium growth.
Experimental evidence presented here and elsewhere 7-16 has shown that pRBCs and merozoites are targeted by different sulfated polysaccharides, and that heparin targets pRBCs and merozoites from widely diverging malarias (e.g. human-infecting P. falciparum and the murine malaria parasite P. yoelii 39 ). The widespread pathogen resistance against virtually all currently used drugs 6 and the difficulties in selecting heparin-resistant parasites 16 , places antimalarial sulfated polysaccharides as interesting candidate molecules deserving careful exploration.   Extraction and purification of sulfated polysaccharides. Samples of the marine organisms were cut into 1 mm 3 pieces, immersed three times in acetone and dried at 60 °C. Sulfated polysaccharides were extracted from 10 g of the desiccated tissues by extensive papain digestion, and the extracts were partially purified by cetylpyridinium and ethanol precipitations as described 62 . Approximately 100 mg dry weight of crude polysaccharide extract was obtained from each species. Extracts were applied to a high-performance liquid chromatography system-linked Mono Q column (GE Healthcare, UK), equilibrated with 5 mM ethylenediaminetetraacetic acid, 20 mM Tris-HCl, pH 7.0. The polysaccharides were eluted from the column using a 0-3 M NaCl linear gradient at a flow rate of 1 mL/min. 0.5 mL fractions were collected and checked by metachromatic assay using 1,9-dimethylmethylene blue 63 , and by measuring conductivity to estimate NaCl concentration. The fractions containing sulfated polysaccharides were pooled, dialyzed against distilled water and lyophilized, and the corresponding structures confirmed by nuclear magnetic resonance analysis as described 35,37,38 . P. falciparum cell culture and growth inhibition assays. The P. falciparum 3D7 strain was grown in vitro in group B human RBCs using previously described conditions 51 . Briefly, parasites (thawed from glycerol stocks) were cultured at 37 °C in Petri dishes containing RBCs in Roswell Park Memorial Institute (RPMI) complete medium under a gas mixture of 92% N 2 , 5% CO 2 , and 3% O 2 . Synchronized cultures were obtained by 5% sorbitol lysis, and the medium was changed every 2 days keeping 3% hematocrit. For culture maintenance, parasitemias were kept below 5% late forms by dilution with washed RBCs prepared as described elsewhere 51 . For growth inhibition assays, parasitemia was adjusted to 1.5% with more than 90% of parasites at ring stage after sorbitol synchronization. 150 μL of this Plasmodium culture was plated in 96-well plates and incubated in the presence of polysaccharides for 48 h in the conditions described above. Parasitemia was determined by flow cytometry, after staining pRBC DNA with the nucleic acid dye Syto 11, added 10 min before analysis. Samples were analyzed using a BD FACSCalibur ™ flow cytometer and parasitemia was expressed as the number of parasitized cells per 100 erythrocytes. Acquisition was configured to stop after recording 50,000 events within the RBC population. IC50 and IC90 were derived from non-linear fit dose response curves (Log doses versus normalized inhibitions).

Methods
Merozoite invasion inhibition assay. Synchronized cultures of P. falciparum 3D7 were enriched using Percoll (GE Healthcare) purification to obtain late trophozoites and early schizonts, and diluted to ~1% initial parasitemia and 3% hematocrit. Assays were performed in 24-well, flat-bottomed microculture plates where 1 mL of culture was incubated in RPMI supplemented with different amounts of each polysaccharide in study, for 20 h as described above. After incubation, smears were prepared by fixing cells in methanol for a few seconds and then staining them for 10 min with Giemsa (Merck Chemicals, Germany) diluted 1:10 in Sorenson's Buffer, pH 7.2.
Plates were incubated for another 20 h before preparing a new set of smears. Slides were observed with an optical microscope Nikon Eclipse 50i (Japan) and pictures were taken with a Nikon Digital Sight DS-U2 camera. For quantitative determinations, the cultures were analyzed by flow cytometry as described above. To assess invasion and maturation rates, respectively, the following formulae were applied:  Table 2. Invasion and maturation rates corresponding to the invasion inhibition assay from Fig. 4.
The data are derived from microscopic counting at, respectively, 20 and 40 h post-treatment of the samples incubated with 4 μg/mL of sulfated polysaccharides. The results are shown as the means of three independent experiments ± standard deviation, with 900 cells counted for each sample (300 × 3 replicates). Significant differences relative to non-treated control as determined by t-tests are indicated by asterisks (*p< 0.05, **p< 0.005, ***p< 0.001).
Scientific RepoRts | 6:24368 | DOI: 10.1038/srep24368  Table 1. The results are shown as the means of three independent experiments; the error bars represent standard deviations. Significant differences in the numbers of rings and schizonts relative to the respective non-treated controls as determined by t-tests are indicated by asterisks (*p< 0.05, **p< 0.005, ***p< 0.001).  Antimalarial activity assay in vivo. The in vivo antimalarial activity of sulfated polysaccharides was studied in a 4-day blood suppressive test as previously described 64 . Briefly, Balb/C female mice (n = 5/sample; Janvier Laboratories) were inoculated intraperitoneally with 2 × 10 6 RBCs extracted from an animal infected by the Plasmodium yoelii yoelii 17 XL lethal strain. Treatment with the antimalarial drug chloroquine (5 mg kg −1 day −1 ) 65 or polysaccharides dissolved in PBS started 2 h later (day 0) with a 200-μL single dose administered intravenously, followed by identical dose administration for the next 3 days. To obtain the desired final in vivo doses, the compounds were prepared 10× concentrated, assuming a volume of 2 mL of circulating mouse blood. A control untreated group received PBS. Activity was determined by microscopic counting at day 4 of blood smears stained with Wright's solution (Merck Chemicals). Mice were fed a commercial diet ad libitum and treated with humane care, being euthanized if reaching a 20% weight loss for two consecutive days. The sacrifice method was exposure to 95% CO 2 following anesthesia with 5% isoflurane vaporized in O 2 . On day 73 after the initial inoculation those surviving animals were re-inoculated as described above. The mice did not receive any treatment after this second infection and parasitemia was monitored by blood smear counting. P. yoelii protein extraction from infected whole blood. Protein lysates were extracted from the whole blood of infected Balb/C female mice having >50% parasitemia. Blood was collected in Microvette ® CB 300 tubes (Sarstedt, Germany) and kept at −80 °C until protein extraction. RBC lysis was performed by adding 10× vol of saponin 0.1% (w/v) in PBS. After washing twice with cold PBS, the pellet was treated with 2 vol of extraction buffer consisting of 50 mM NaCl, 0.5% Mega 10, 3% CHAPS, and 50 mM Tris-HCl, pH 8.0, supplemented with a protease inhibitor cocktail (Roche). The samples were subjected to four freeze-thaw cycles, and the lysates were finally centrifuged at 20,000 g for 30 min (4 °C). Protein concentration was determined by the DC protein assay (Bio-Rad), and P. yoelii total protein samples were stored at −20 °C until use.
Western blot. 10 μg of P. yoelii total protein extract were fractionated in a reducing 10% SDS-PAGE (Bio-Rad), transferred to polyvinylidene difluoride membranes and blocked with Z buffer (100 mM MgCl 2 , 0.5% Tween 20, 1% Triton X-100, 1% BSA,100 mM Tris-HCl, pH 7.4, supplemented with 5% FCS). Membranes were then incubated at 4 °C overnight with 1:10,000 dilutions of sera from the mice that survived the infection, followed by a 1-h incubation with secondary horseradish peroxidase-labeled anti-mouse IgG (Amersham Biosciences) at a 1:10,000 dilution.  Table 3. Estimated compound concentrations in blood and determined day 4 parasitemias for the in vivo assay from Fig. 6b. These daily administered doses assume 2 mL of blood in a mouse into which were injected 200 μL of 10× concentrated compound solutions. The results are shown as the means of five independent experiments ± standard deviation.
Scientific RepoRts | 6:24368 | DOI: 10.1038/srep24368 Statistical analysis. Data are presented as the mean ± standard deviation of at least three independent experiments, and the corresponding standard deviations in histograms are represented by error bars. Cell counts from Giemsa-stained slides were done using the Plasmoscore programme. The parametric Student's t-test was used to compare two independent groups when data followed a Gaussian distribution, and differences were considered significant when p ≤ 0.05. Percentages of viability were obtained using non-treated cells as control of survival and IC 50 values were calculated by nonlinear regression with an inhibitory dose-response model using GraphPad Prism5 software (95% confidence interval). Concentrations were transformed using natural log for linear regression. Regression models were adjusted for replicates and assay data. In anticoagulant assays, the polysaccharide concentrations (mean ± standard deviation) necessary to double the control (T0) APTT 66 were compared via one-way ANOVA with Tukey´s test using the software Origin-Pro 8.0 (OriginLab).