ARTICLE pubs.acs.org/Biomac
In Vitro Evaluation of Antioxidant and Anti-Inflammatory Properties of Genistein-Modified Hemodialysis Membranes Chandrasekaran Neelakandan,†,|| Teng Chang,† Thomas Alexander,‡ Linda Define,‡ Michelle Evancho-Chapman,§ and Thein Kyu*,† †
Department of Polymer Engineering, University of Akron, Akron, Ohio 44325, United States Department of Pathology and Laboratory Medicine and §Falor Division of Surgical Research, Summa Health System, Akron, Ohio 44304, United States
‡
ABSTRACT: Genistein-modified poly(amide):poly(vinyl pyrrolidone) (PA: PVP/G) hemodialysis membranes have been fabricated by coagulation via solvent (dimethyl sulfoxide, DMSO)/nonsolvent (water) exchange. The antioxidant and anti-inflammatory properties of the unmodified PA:PVP membranes were evaluated in vitro using human blood. It was found that these unmodified PA:PVP membranes were noncytotoxic to peripheral blood mononuclear cells (PBMC) but raised intracellular reactive oxygen species (ROS) levels. Pure genistein (in DMSO solution) was not only nontoxic to PBMC, but also suppressed the ROS levels in a manner dependent on genistein dosage. A similar dose-dependent suppression of ROS was found in genistein-modified PA (i.e., PA/G) membranes. However, the PVP addition had little or no effect in the suppression of ROS levels for the ternary PA:PVP/G system; the membrane ROS suppression was largely controlled by the genistein dosage. The levels of tumor necrosis factor-R (TNF-R), interleukin-1β (IL-1β), and interleukin (IL-6) in whole blood were measured by ex vivo stimulation with lipopolysaccharide (LPS). The unmodified PA:PVP membranes drastically increased the level of TNF-R; however, the concentration of IL-1β and IL-6 remained almost the same. The PA/G membranes reduced the concentration of IL-1β and TNF-R even at very low genistein loadings, but it required a higher genistein loading to realize a similar effect in the case of IL-6. Of particular importance is that the genistein-modified blend membranes (PA:PVP/G) showed greater suppression of the concentrations of all three cytokines (TNF-R, IL-1β, and IL-6) in comparison with those of the PA/G membranes, signifying the role of PVP in the enhanced anti-inflammatory properties of these genistein-modified membranes. Ultravioletvisible (UVvis) spectroscopy was employed to quantify any genistein leaching during the in vitro testing.
’ INTRODUCTION Hemodialysis (HD) (hemo: blood; dialysis: purification) is a life-sustaining treatment for end-stage renal disease (ESRD) patients whose physiological rhythm has succumbed to impaired kidney functions.1 In HD, patient blood is pumped for repeated cycles through a dialyzer unit packed with hollow fiber bundles made-up of hydrophobic/hydrophilic polymer blends to reduce urea and uremic toxin concentration in blood to physiologically acceptable levels. Basically, the dialyzer unit implements the filtration function of kidney, and thus it has been regarded as artificial kidney. On the flip side, HD treatment cannot match the kidney’s functions in totality because the synthetic membrane often induces some undesirable alterations in blood chemistry. Typically, an HD patient’s blood is in contact with the synthetic HD membrane for a ∼3 to 4 h/session and three to four sessions/week.2 This prolonged contact of blood with the synthetic polymer surface culminates in two long-term complications, viz. dialysis-induced oxidative stress (DIOS) and membraneinduced inflammation (MII), both of which act in tandem creating cardiovascular problems and contribute to 40% mortality in HD patients. DIOS is initiated when the excess production of oxygen radicals overpowers the natural antioxidant defense r 2011 American Chemical Society
mechanisms of the body.35 MII causes undesirable immune response induced by higher concentration of pro-inflammatory cytokines such as interleukin-1β (IL-1β) and interleukin (IL-6) and tumor necrosis factor-R (TNF-R) in blood.6,7 The bioincompatibility of the polymeric HD membranes has been implicated as the primary source of generation of excessive reactive oxygen species (ROS), which contribute to DIOS. Therefore, there is an urgent demand for functional HD membranes that are capable of suppressing these long-term ill-effects of HD, which is the source of motivation for the present work. Recently, we have engaged in the development of functional HD membranes via modification with phytochemicals. Phytochemicals are plant-derived chemicals that have been shown to possess antioxidant, anti-inflammatory, and other salubrious properties.8 Phytochemicals have garnered the reputation as alternative and complementary medicine by virtue of their natural origin, abundant availability, lower toxicity, and perceived health benefits.810 In this Article, genistein, a soybean-derived Received: January 8, 2011 Revised: June 8, 2011 Published: June 09, 2011 2447
dx.doi.org/10.1021/bm200591q | Biomacromolecules 2011, 12, 2447–2455
Biomacromolecules phytochemical was selected as the modifying agent by virtue of its demonstrated antioxidant11,12 and anti-inflammatory13,14 properties, whereas poly(amide):poly(vinyl pyrrolidone) (PA:PVP) pair was chosen as the polymer matrix because of its extensive use in HD membranes.15 In the present study, we have investigated the in vitro effects of genistein modification of PA:PVP membranes on cytotoxicity, oxidative burst (i.e., generation of ROS), and cytokine response of human blood.
’ EXPERIMENTAL SECTION Materials. TROGAMID T5000 (Mn = 20 000 and Mw = 63 000),16,17 an aromatic poly(amide) was kindly provided by Degussa Corporation (Parsippany, NJ). PVP (Mw = 40 000) was purchased from Sigma Aldrich, (St. Louis, MO). A reagent-grade DMSO (dimethyl sulfoxide) purchased from Sigma-Aldrich was used as a common solvent without further purification. Genistein (>98% purity) was bought from MDidea Exporting Division (YinChuan, China). Reverse osmosis grade water was utilized as nonsolvent for membrane formation. Dihydrorhodamine 123 (DHR) was obtained from Invitrogen (Carlsbad, CA). Phorbol myristate acetate (PMA) was purchased from Sigma Aldrich. DHR and PMA were used for the in vitro compatibility studies to determine ROS levels in blood. Methods. Fabrication of Genistein-Modified PA:PVP Membranes. To prevent moisture absorption, PA pellets were vacuum-dried at 80 °C for 24 h. The dried PA pellets were subsequently dissolved in DMSO at 17.5 wt % concentration. The solution used to cast membranes hereafter is termed as “feed solution”. In the preparation of genistein-modified membranes, appropriate amounts of genistein were added to the feed solution. The solutions were homogenized for 48 h by mechanical stirring, and the entrapped air, if any, was removed under vacuum at room temperature. The homogeneous solutions were then cast in the form of a film of predetermined thickness on a precleaned glass plate and immersed in nonsolvent bath (i.e., reverse osmosis grade water) maintained at 25 °C. The coagulated membranes were peeled-off from the glass plate and rinsed with excess water and then vacuum-dried at room temperature. For the scanning electron microscope (SEM) observations, dried membranes samples were fractured in liquid nitrogen and then sputtered with silver using a sputter coater (Emitech, model K575X) and analyzed with field emission SEM (JEOL-Model JSM-7401F). The sample surfaces were cleaned with compressed air before sputtering. SEM micrographs were taken at three different areas that are representative of the whole sample. Reagents and Blood Samples. The stock solution of DHR (10 mg DHR in 2.0 mL of DMSO) was stored in 50 μL aliquots at 70 °C. For the assay, 20 μL of stock was dissolved in 980 μL of PBS to give a final concentration of 100 μg/mL. PMA was used to activate neutrophils to undergo oxidative burst. PMA stock solution was prepared by dissolving 1 mg PMA in 1 mL of DMSO. This solution was also stored in 50 μL aliquots at 70 °C. Subsequently, 10 μL of stock solution was diluted with 90 μL of PBS to give a final concentration of 5 μg/mL. The first author voluntarily donated his blood for the needed biomedical tests. We collected 10 mL of venous blood in lithium heparin Vacutainer tubes (Becton Dickinson, Rutherford, NJ) and used within 12 h of the collection. Cytotoxicity Studies. Cell viability assay was conducted to determine the effect of unmodified membranes, pure genistein, and genisteinmodified membranes on whole blood. A nucleic acid dye, 7-aminoactinomycin D (7-AAD) (BD-Via-Probe from BD Bioscience, CA), was used as a viability probe, which is based on uptake of 7-AAD. Fresh blood mixed with the dye was used as negative control, whereas blood incubated with membrane, genistein, or both served as the experimental samples. In brief, 100 μL of fresh blood was pipetted into separate polypropylene tubes and mixed with 2 mL of ammonium chloride lysis buffer solution (containing 8.26 g ammonium chloride, 1 g potassium
ARTICLE
bicarbonate, and 0.037 g ethylenediaminetetraacetic acid dissolved in 1 L of deionized water). The tubes were allowed to stand at room temperature for 10 min, during which most of the red blood cells got lysed. The tubes were centrifuged at room temperature for 5 min at 1800 rpm. The supernatant was discarded and the cells were washed twice and centrifuged with 2 mL of phosphate buffered saline (PBS, Baxter, IL, pH 7.4) solution. The supernatant was discarded again and 20 μL of 7-AAD dye was added to the cells and was stored under dark conditions for 10 min to permit cell uptake. The cells were resuspended with 500 μL of PBS, and the cell suspension was analyzed using a flow cytometer (EPICS XL-MCL, Beckman Coulter, Brea, CA) equipped with System II software. 7-AAD is a nucleic acid dye, which is designed in such a way that it cannot enter live cells. The number of cells in each sample was adjusted between 10 000 to 20 000 cells/mL using PBS. The fluorescence signal from the dead cells was measured on FL3 red channel (650 nm). Standard DHR Assay. To establish the reliability of the assay, 100 μL of whole blood samples was pipetted into three separate polypropylene tubes. These were labeled as stimulated, resting, and reagent blank. We added 25 μL of 5 μM PMA solution to the stimulated tubes, and 25 μL of PBS was added to the resting and reagent blank tubes. All tubes were incubated at 37 °C for 15 min. Then, 25 μL of working DHR solution was added to the stimulated and rested samples, and 25 μL of PBS was added to the reagent blank, followed by 15 min of incubation at 37 °C. The tubes were centrifuged at 1800 rpm for 5 min. The supernatant was discarded, and the cells were washed and centrifuged twice. After the second centrifugation, the supernatant was discarded and replaced with 1 mL of immunoprep (Beckman Coulter) solution. Immunoprep reagent is a rapid whole blood lysing solution consisting of three ready-to-use reagents, which lyses the red blood cells, buffers the solution to stop the lysing process, and fixes the cells, respectively. The solutions were filtered through 50 μm filter to remove the cellular aggregates. The samples were analyzed using the flow cytometer described above. A quality control run was executed daily according to the manufacturer’s instructions to ensure proper calibration of the instrument. Before acquiring data, the instrument was set up using a reagent blank sample. The forward and side light scatter profiles were adjusted to ensure that the neutrophil population was well-resolved. Fluorescence was measured on the FL1 green channel (wavelength 530 nm). Data were then collected from the reagent blanks and all resting and stimulated tubes. A total of 10 000 events were collected for each sample. During analysis, the plot of forward scattering versus side scattering was displayed and the neutrophil population was identified by its typical location and selected by gating. Subsequently, a histogram of rhodamine fluorescence (FL1) was obtained for the gated region, and the mean channel fluorescence was recorded. Each experiment was repeated three times.
DHR Assay with Genistein, Unmodified, and Modified Membranes. To determine the effect of membranes, genistein, or both on oxygen radical generation, the standard DHR assay was modified as follows. The assay consisted of one positive control tube, several sample tubes and one negative control tube. The positive control tube contained 100 μL of blood, 25 μL of DHR, and 25 μL of PMA; the total contents were made up to a total volume of 1 mL with PBS. The effect of oxygen radical generation is tested in what follows using the sample tube containing 100 μL of blood, 25 μL of DHR, 25 μL of PMA and the appropriate sample (for example, unmodified membranes or pure genistein or genistein-modified membranes). The membranes were cast and dried under sterile conditions inside a biosafety laminar flow hood. For the assay, two small circular samples having the area of 0.4 cm2 were punched out from the cast films and placed inside the test tubes. After the addition of the membranes, the entire contents were incubated at 37 °C for 3 h (to simulate HD conditions) subjected to continuous shaking. After 3 h of incubation, the membrane samples were 2448
dx.doi.org/10.1021/bm200591q |Biomacromolecules 2011, 12, 2447–2455
Biomacromolecules
ARTICLE
removed from the blood and agitated with 1 mL of PBS to dislodge the cells that were attached to the membrane. In the case of pure genistein, the samples were prepared by dissolving genistein in DMSO over a wide range of concentration (25500 μg/mL). The negative control tube contained 100 μL of blood, 25 μL of DHR, and the sample. The rest of the assay was similar to the standard DHR assay protocol including lysis of red blood cells, fixing of cells with immunoprep solution and followed by flow cytometric analysis. To normalize the data obtained from different set of experiments, the following formula was employed to express ROS level in percentage xz 100 ð1Þ %ROS level ¼ 100 xy where the positive control represent 100% ROS and the negative control represent 0% ROS. The x, y, and z parameters represent the mean fluorescence signal obtained from the positive control, the negative control, and the samples, respectively.
Lipopolysaccharide (LPS) Stimulation of Whole Blood Samples and Multiplex Cytokine Assay. Whole blood samples were collected in ethylenediamine tetraacetic acid (EDTA) containing Vacutainer tubes. The blood was stimulated ex vivo with 100 ng/mL LPS for 24 h at 37 °C. Triton X-100 was then added to a final concentration of 0.5% (v/v) and incubated for 10 min at room temperature to lyse the cells. All samples were centrifuged at 13 200 rpm for 10 min and the serum was separated. Serum levels of cytokines were determined by Bio-Plex Pro Assay (Bio-Rad Laboratories, Hercules, CA). Calibration curves for each cytokine were obtained using the reconstituted standards supplied by the vendor. All assays were carried out directly in a 96-well filtration plate (Millipore, Bedford, MA) at room temperature while protecting the beads from exposure to light. Standards and samples were measured in duplicate, and blank values were subtracted from the readings of all samples. In brief, wells were prewetted with 100 μL of PBS. The beads (∼5000 beads per cytokine) were added along with sample or blank to a final volume of 100 μL in each well and incubated for 30 min at room temperature with continuous shaking. After washing with 100 μL of PBS, 25 μL of biotin-conjugated detection antibody was added to each well and incubated for an additional 30 min. Subsequently, streptavidinphycoerythrin was added to the wells and incubated for 10 min. After washing-off the unbound streptavidin phycoerythrin, the fluorescence intensity of the beads (minimum of 50 beads per cytokine) were analyzed in the Luminex 200 instrument (Austin, TX) equipped with Luminex XYP platform and Luminex xPONENT 3.1 data analysis software. Statistical Analysis. The results are presented as the mean ( SEM (standard error of the mean) and statistical significance between the groups was determined by unpaired Student’s t test; p value less than 0.05 represents statistical significance between the groups.
’ RESULTS AND DISCUSSION Membrane Morphology and Quantification of Actual Amount of Genistein in Membranes. Genistein-modified
PA/G and PA:PVP/G membranes were prepared via coagulation-induced phase separation by immersing the genistein-modified polymer solutions into water bath to form the membranes. Although we have previously demonstrated18 the miscibility enhancement via cross-hydrogen bonding between genistein with PA, it is crucial to know whether such specific interaction would be sufficient to retain, if not all, incorporated genistein in the PA membrane during the coagulation process. To quantify the amount of potential loss of genistein, the nonsolvent bath containing water, any genistein, and DMSO were chilled at 15 °C for 96 h to allow the genistein to crystallize and precipitate out
Figure 1. Actual amounts of genistein retained in the final membrane in relation to the amount of genistein in feed for various polymer blend ratios; neat PA and 75:25 PA:PVP. Dashed line represents the ideal case in which all genistein in feed were retained in the final membrane. The SEM micrographs represent the surface (left) and cross sectional (right) morphology of PA/G membranes.
from the solution mixture, which were filtered, dried, and weighed. On the basis of a simple mass balance calculation, that is, the percent weight loss of genistein = (wt of genistein residue 100)/(wt of PA + wt of genistein in feed), the actual amount of genistein in the final modified membrane was determined.18 As illustrated in Figure 1, at lower genistein loadings (up to ∼30 wt %), most of the genistein added to the feed solution was retained in the final membrane (filled circles). Beyond the 30 wt % genistein loading, the deviation occurs from the dashed straight line representing the ideal situation, where all genistein in feed was retained in the membrane after coagulation in the nonsolvent bath. Upon further increase in genistein concentration (e.g., ∼50 wt %), the genistein loading of the final membranes saturated out at ∼35 wt % of the solid content, indicating substantial weight loss during the membrane formation. A similar experiment was carried for the PA:PVP/G system. Because PVP and genistein are completely miscible, there is a possibility that more genistein might be dragged out by PVP leaching. As expected, there is more genistein leaching out into the nonsolvent bath along with PVP (filled squares, Figure 1), which may be attributed to the profound solubility of PVP in water. The surface and cross-sectional morphologies of unmodified and genistein PA membranes were examined by using SEM (Figure 1). A dense skin layer can be seen on the surface of the unmodified PA membrane, whereas the cascading finger-like channels developed along the cross section (Figure 1a). The size of the cascading fingers progressively increases along the thickness direction from the skin layer (i.e., the top layer), where the solvent escapes and the nonsolvent enters. The formation of cascading finger-like structures is seemingly governed by the changing solvent power at the coagulation front coupled to the hydrodynamic flow of solvent/nonsolvent exchange that eventually creates the asymmetric membrane structure. The effects of genistein modification on membrane morphology are illustrated 2449
dx.doi.org/10.1021/bm200591q |Biomacromolecules 2011, 12, 2447–2455
Biomacromolecules
Figure 2. UVvis spectra of DMSO, genistein dissolved in DMSO, and blood plasma (BP). The signature peaks of genistein are located at 260 and 320 nm, whereas the signature peak of BP appears at 280 nm.
in Figure 1be. The prominent effect of genistein modification is the formation of needle-like genistein crystals that are embedded on the membrane surface even at very low feed compositions of genistein (e.g., 90/10 PA/G). As the genistein loading in feed was increased to 80/20 and 70/30 PA/G, the resulting membranes contained larger seaweed-like crystals (i.e., dense branching morphology) on the membrane surface. Concurrently, the cross-sectional morphology of all modified membranes showed cascading finger-like channels, which were coated with genistein crystals (e.g., Figure 1ce). Increasing the genistein feed composition beyond 30 wt % resulted in the formation of larger incomplete spherulitic crystals on the membrane surface, which are presumably deposited on the surface from the excess genistein (Figure 1e). Moreover, some needle-like crystals are protruded out from the interior surface of the cascading channels. These deposited crystals on the surface can leach out easily upon immersion in nonsolvent, especially at 50 wt % genistein loading. The agreement between the observed crystal morphology at the surface with virtual lack of genistein leaching below 30 wt % implies that these crystals are predominantly segregated to the surface (i.e., the membrane surface as well as the cross-sectional interior surface of the cascading channels) but fully embedded, and thus the crystallization presumably prevents leaching of genistein. A similar morphology development can be confirmed in the PA:PVP/G membranes, except that the PVP makes the surface of the membrane more porous, but smoother (pictures not shown). The role of genistein crystal at the surface on the antioxidant and anti-inflammability properties of genisteinmodified membranes will be discussed later. Determination of Genistein Leaching during Incubation of Membranes in Blood Plasma. To quantify the amount of genistein that can potentially leach into the blood samples during the in vitro blood compatibility tests, a UVvis absorbance technique was employed. Figure 2 illustrates the UVvis spectra of pure DMSO as well as genistein solution in DMSO. Because genistein is insoluble in water, DMSO was chosen as the solvent to obtain the signature absorption peaks of genistein, which are located at 260 and 320 nm. The signature absorption peak of blood plasma (BP) is at 280 nm. Figure 3 shows the UVvis absorption curve for pure BP and BP incubated with two different membranes, viz. 70/30 PA/G
ARTICLE
Figure 3. Overlay of UVvis absorbance spectrum of PA/G mixed with BP exactly matches that of pure BP indicating no leaching of genistein into blood plasma. PA:PVP/G membranes mixed with BP indicate a small peak around 334 nm, suggesting that PVP along with genistein could have leached into BP. However, the amount is too negligible to be quantified.
and 63:7/30 PA:PVP/G compositions. These two membranes were selected as representative from each group that is closer in composition toward the final application. As can be seen in Figure 3, the UVvis absorbance curve of pure BP and BP incubated with PA/G membrane exactly overlay on top of each other indicating that there was virtually no leaching of genistein into the BP. Of particular interest is that the PA:PVP/G membrane incubated with BP showed a slight peak shift in the vicinity of the 280 nm peak and also a minor shoulder in the vicinity of 330 nm, implying that some traces of genistein may have been dragged out by PVP leaching into BP because of the profound affinity of PVP to water. It is reasonable to infer that genistein may not leach out from the PA/G membrane, but there is a possibility that some traces of genistein, although very minor, apparently leach out during the actual in vitro testing of the PA: PVP/G membrane. It should be emphasized that the minor genistein leaching from the PA:PVP/G membrane into the blood may not necessarily compromise its biomedical performance because pure genistein is proven to have excellent antioxidant and anti-inflammatory properties, as will be described in the subsequent sections. Cytotoxicity of Unmodified and Genistein Modified Membranes. Prior to all biocompatibility tests, cytotoxicity test is a minimum requirement to be carried out first. Cytotoxicity test was performed in accordance with the procedure described in the Experimental Section. It was found that the pure PA membranes (i.e., similar to the commercial HD membranes) exhibited >98% cell viability. The incorporation of PVP into PA showed cell viability exceeding 98% implying no harmful effect to blood cells. In principle, any modifying agent intended for medical device applications must be noncytotoxic in its pure form as well as when bound to a substrate (either physically or chemically). Therefore, the cell viability tests were performed on human blood. The pure genistein were found to be noncytotoxic to the PBMC in the entire concentration range of the assay (25500 μg/mL) (data not shown). This finding established the basis for the selection of genistein toward subsequent development of the functional HD membranes for suppressing DIOS and MII. Cytotoxicity studies 2450
dx.doi.org/10.1021/bm200591q |Biomacromolecules 2011, 12, 2447–2455
Biomacromolecules
Figure 4. DHR assay performed with unmodified membranes fabricated with 100:0, 95:5, 90:10, and 75:25 PA:PVP ratios. The membranes were incubated for 3 h at 37 °C to simulate hemodialysis conditions. ROS levels of the membrane samples (blood + PBS + membrane + DHR + PMA) were compared with that of the positive control (blood + PBS + DHR + PMA) to determine statistical significance. Values shown are mean ( SEM obtained from three experiments (n = 3).
were extended to genistein-modified PA/G and PA:PVP/G membranes, which did not exhibit any harmful effects on cell viability (i.e., comparable to the control). On the basis of the nominal area of the samples used for testing, it can be estimated that the genistein loading in these modified membranes (PA/G and PA:PVP/G) ranges between 100 and 200 μg/cm2. Although the cytotoxicity results of pure genistein as well as both PA/G and PA:PVP/G membranes are very promising, it should be cautioned that passing the cytotoxicity test is merely a minimum necessary criterion, but it is by no means sufficient for actual applications. DHR Assay with Unmodified PA:PVP Membranes. The outcome of oxygen radical generation due to blood contact with unmodified PA:PVP membrane is depicted in Figure 4. The unmodified membrane samples used in this experiment were: 100:0, 95:5, 90:10, and 75:25 PA:PVP ratios. To calculate ‘p’ value to establish statistical significance, the mean fluorescence values of the membrane samples were compared with that of the positive control. It can be noticed that the blood contact with unmodified membranes has led to an increase in the ROS level. For example, in the 90:10 PA:PVP membrane, the increase in ROS level was found to be as high as 30% compared with that of the positive control. This result is very important for HD perspective because typical HD patients’ blood comes into contact with the membrane a ∼3 to 4 h/session and three to four sessions/week.1,2 When blood touches the synthetic polymer surface, the neutrophils are activated as a defense mechanism, leading to oxidative burst and the oxidative stress builds up in the patient’s body. The body quickly neutralizes these radicals by natural antioxidant molecules such as superoxide dismutase and glutathione peroxidase. Such a defense mechanism is generally deficient for HD patients, and thus over a period of time, the excessive production of oxygen radicals overpowers the natural antioxidant defense mechanism and thus eventually leading to DIOS. Contrary to the report that the addition of hydrophilic polymer to hydrophobic polymer improves biocompatibility of the membranes,19 the incorporation of PVP to PA shows the rise in ROS with increasing PVP concentration in the membrane (Figure 4). In general, hydrophobic polymer surfaces
ARTICLE
Figure 5. DHR assay executed with pure genistein (25500 μg/mL) clearly demonstrates the antioxidant properties of genistein, as observed by the dose-dependent reduction of ROS level. Values shown are mean ( SEM obtained from three experiments (n = 3). ROS levels of the genistein samples were compared with that of the positive control to determine statistical significance, which is represented by asterisks (*).
are known to be good adhesion promoters with strong protein adsorption capability, but they are poor complement activators. The reverse is true for hydrophilic polymers that show poor protein adsorption but are a strong complement to activation. Therefore, it is generally accepted that an optimum surface for blood contact must be made up of hydrophobic/hydrophilic polymer blends.19 However, in terms of oxygen radical generation, the present study showed that the addition of PVP to PA raises the ROS levels in blood albeit marginally. Antioxidant Properties of Pure Genistein and Genistein Modified Membranes. The results of ROS levels obtained with pure genistein (in DMSO) are plotted in Figure 5. Two important conclusions can be made from Figure 5; that is, genistein was able to suppress ROS levels significantly even at concentrations as low as 25 μg/mL, and the suppression of ROS occurs in a dose-dependent manner; that is, higher concentration of genistein results in greater suppression of ROS levels. Genistein has been reported to inhibit the priming events of high-level ROS production, especially in human polymorphonuclear cells cultured in vitro.20 The reported levels of ROS inhibition by genistein solution (∼50% in the concentration range 15100 μg/mL) agrees well with the present finding. To comprehend the antioxidant properties of genistein, it is critical to examine the mechanism of formation of oxygen radicals. Upon activation of the cells, the membrane-bound NADPH (nicotinamide adenine dinucleotide phosphate) and cytosolic components of the enzyme assemble in the membrane and form the active enzyme. NADPH oxidase catalyzes the reduction of O2 to superoxide anion (O2•), which then rapidly dismutates to hydrogen peroxide (H2O2). This chain of events is referred to as electron transport chain. Subsequently, H2O2 may be converted by the enzyme myeloperoxidase into highly reactive compounds such as hypochlorous acid (HOCl).21 Therefore, for a neutrophil to undergo oxidative burst, a functionally intact NADPH oxidase is crucial. In this case, there might be a possibility that genistein has inhibited the expression of NADPH. Because of such ability of genistein in inhibiting NADPH expression, long-term oral administration of genistein has been recommended to improve the health of endothelial cells.22 Therefore, the excellent cell viability results along with antioxidant properties of genistein displayed in Figure 5 attest to 2451
dx.doi.org/10.1021/bm200591q |Biomacromolecules 2011, 12, 2447–2455
Biomacromolecules
Figure 6. Effect of (a) PA/G membranes and (b) PA:PVP/G membranes on ROS levels measured by DHR assay after 3 h of incubation at 37 °C. Values shown are mean ( SEM obtained from three experiments (n = 3). Asterisks (*) represent statistical significance in comparison with positive control.
not only the safety of genistein to the blood cells but also its efficacy as an antioxidant. The ROS levels of the genistein-modified membranes (i.e., PA/G) are shown in Figure 6a, which shows a dose-dependent suppression of oxygen radicals. As previously explained, the unmodified PA membrane shows an increase in oxygen radicals for a few percent. However, with the addition of genistein in the PA membrane, the ROS level was found to reduce for ∼50% in the feed 70/30 PA/G composition. This 50% reduction of the oxygen radicals is close to that of the pure genistein shown in Figure 5. However, no ROS test was performed with the PA/G above 30% because a usable film cannot be formed above this genistein concentration of 50 wt %; in fact, the film became very brittle and crumbled. It should be kept in mind that most genistein crystals are segregated to the surface of the membrane as well as of the cascading channels at least at the low loading level of up to 30 wt % (Figure 1) without significant leaching. On the basis of the observed suppression of ROS of the 70/30 PA/G membrane comparable to that of the genistein/DMSO solution, it may be concluded that almost all genistein ingredients are accommodated at the membrane surface/pore interface of the PA/G membranes. As previously demonstrated, the surfacesegregated genistein crystal has helped for the improvement of the antioxidant properties of the genistein-modified membranes. Next, the effect of PVP addition on the antioxidant properties of PA:PVP/G membranes is demonstrated in Figure 6b. According to the results of PA:PVP membranes (Figure 4) and pure genistein (Figure 5), PVP and genistein showed diametrically opposing trend in regards to the oxygen radical generation. It is
ARTICLE
Figure 7. Serum levels of IL-1β measured after incubating LPS stimulated whole blood for 24 h with (a) PA/G membranes and (b) PA:PVP/G membranes. Values shown are mean ( SEM obtained from two experiments (n = 2).
therefore of interest to find out how these two ingredients play out in the ternary PA:PVP/G membranes. Because the 70/30 PA/G sample showed the best ROS reduction, we kept the genistein content to be constant at 30%; only the ratio of PA to PVP was varied. All samples show impressive ROS reduction of 4050%. Given the large statistical error, it is fair to infer that the PVP addition has little or no effect in the suppression of ROS levels for the ternary PA:PVP/G system; the membrane ROS suppression is largely controlled by the genistein dosage. Effect of Genistein Modified Membranes on Serum Cytokine Levels. Although the demonstration of genistein modification of the PA:PVP membrane on the cell viability and the ROS suppression is promising, it is imperative to determine the significance of genistein modification on the anti-inflammatory properties. The pro-inflammatory cytokines such as IL-1β, IL-6, and TNF-R in serum were measured by Luminex analyzer after ex vivo stimulation of the whole blood with LPS. These cytokines were chosen because it has been clinically demonstrated that the elevated levels of IL-1β, IL-6, and TNF-R in serum of ESRD patients were closely correlated with increased mortality risk.23 These cytokines have been suggested to cause fever during HD, reduced blood pressure, and anemia in long-term HD patients.24 The effect of PA/G membranes on serum levels of IL-1β is shown in Figure 7a. In this plot, the positive control represents the whole blood stimulated with LPS, sample tubes contained LPS-stimulated blood incubated with various PA/G membranes (two small circular disk of 0.4 cm2 each 100/0, 90/10, 80/20, and 70/30 PA/G), whereas negative control represents the whole blood without the LPS stimulation. As can be noticed in Figure 7a, the unmodified membranes (i.e, 100/0 PA/G) slightly suppressed IL-1β secretion from PBMC. 2452
dx.doi.org/10.1021/bm200591q |Biomacromolecules 2011, 12, 2447–2455
Biomacromolecules
Figure 8. Effect of (a) PA/G membranes and (b) PA:PVP/G membranes on serum levels of IL-6 measured after incubating LPS stimulated whole blood with membranes for 24 h. Values shown are mean ( SEM obtained from two experiments (n = 2).
With the addition of genistein, the suppression of IL-1β secretion from PBMC becomes very pronounced even at very low genistein loading (i.e., 90/10 PA/G). To determine the effect of PVP addition on these PA/G membranes, samples were prepared with a fixed genistein composition of 30 wt %; only the PA:PVP ratio was changed systematically. As shown in Figure 7b, there was a slight increase in IL-1β concentration at a very low PVP concentration (66.5:3.5/30 PA:PVP/G). However, with further increase in PVP content, progressive suppression of IL-1β was realized. In the case of IL-6, the unmodified membranes (100/0 PA/G) virtually produced no change in serum concentration (Figure 8a). Unlike the case of IL-1β, a higher genistein loading (70/30 PA/G) was required to observe even a reasonable decrease in IL-6 concentration. The addition of PVP slightly reduced the IL-6 level at lower PVP compositions, whereas a sudden drop was noticed at a higher PVP composition (e.g., 52.5:17.5/30 PA:PVP/G) (Figure 8b). Among the three cytokines measured, the unmodified membranes produced the largest increase in the concentration of TNF-R as depicted in Figure 9a (i.e., almost four times increase as compared with the positive control). Upon the addition of genistein in the membranes, a dramatic suppression of TNF-R concentration can be noticed even at a very low genistein loading (for example 90/10 PA/G), which indicates the outstanding antiinflammability property of the genistein-modified membranes. Further increase in genistein composition in the PA/G membranes further reduced TNF-R level in serum drastically, that is, close to the negative control. It is striking to notice that with the
ARTICLE
Figure 9. Serum levels of TNF-R measured after incubating LPS stimulated whole blood for 24 h with (a) PA/G membranes and (b) PA:PVP/G membranes. Values shown are mean ( SEM obtained from two experiments (n = 2).
addition of PVP to the PA/G membranes, the TNF-R level is suppressed progressively and approaches that of the negative control (Figure 9b). Cytokines are a family of proteins that are involved in numerous immunological functions including the production and control of other cytokines. They play an important role in the regulation of hematopoiesis, mediating the differentiation and proliferation of diverse type of cells. For example, it has been identified that endotoxins (such as bacterial components) from dialysate induce secretion of IL-1β from neutrophils, which causes fever and low blood pressure during HD.24 IL-1β and TNF-R are known for their autocrine (i.e., induce/regulate its own secretion) and paracrine signaling (induce/regulate other cytokine secretion) functions.25 Clinically, it has been demonstrated that serum concentrations of IL-1β and TNF-R drastically increase during HD in a manner dependent on the choice of the membrane.7 Although a polymeric membrane surface could induce cytokine secretion due to direct contact of PBMC with membrane and endotoxin from dialysate, complement-mediated cytokine secretion has been generally accepted as the common mechanism by which a HD membranes induce inflammation.6,25 In the case of HD, the alternate pathway of complement activation leads to the formation of complement fragments such as C3b, which coats the membrane surface by adsorbing to the membrane surface. C3b molecules together with other soluble complement fragments such as C3a and C5a subsequently stimulate the PBMC triggering the enhanced secretion of pro-inflammatory cytokines. As previously pointed out, hydrophilic polymers that contain charged groups strongly activate complement system, whereas hydrophobic polymers show less complement activation.19 It can 2453
dx.doi.org/10.1021/bm200591q |Biomacromolecules 2011, 12, 2447–2455
Biomacromolecules
’ CONCLUSIONS We have demonstrated how a soybean-derived phytochemical called “genistein” can be incorporated into poly(amide):poly(vinyl pyrrolidone) (PA:PVP) blends for application as functional HD membranes. The PA:PVP membranes were shown to be noncytotoxic to the PBMC. However, these unmodified membranes were shown to increase the concentration of ROS, which lent support to the idea of DIOS. Pure genistein was found to be capable of reducing ROS levels in blood in a manner dependent on dosage, which may be attributed to the NADPH inhibition by genistein. Of particular importance is that genisteinmodified membranes (PA/G membranes) exhibited a dosedependent suppression of ROS levels and showed up to 50% reduction relative to positive control. These membranes also showed significant suppression of all three clinically relevant cytokines, viz. IL-1β, IL-6, and TNF-R although to different extents, which was due to the known ability to genistein to inhibit protein tyrosine kinase and PKC. It may be inferred that the surface crystallization of genistein at the pore interface has contributed to the improvement of antioxidant and anti-inflammability properties of the genistein-modified membranes. Although, the PVP addition has little or no effect in the suppression of ROS levels for the ternary PA:PVP/G system, but it certainly reduces the cytokine levels. Our UVvis spectroscopic experiments
confirmed that there was virtually no genistein leaching from the PA/G membranes. However, negligible traces of genistein were found in the UVvis spectra of the PA:PVP/G membranes, implying that genistein may be possibly dragged out during PVP leaching due to its profound solubility with water.
’ AUTHOR INFORMATION Corresponding Author
*Tel: +1-330-972-6672. Fax: +1-330-258-2339. E-mail: tkyu@ uakron.edu. Present Addresses
)
be argued that the negligible or virtually no increase in IL-1β and IL-6 concentration (in comparison with the positive control) observed with pure PA membranes (i.e., 100/0 PA/G) could be attributed to poor complement activation by PA, which in turn is dictated by the extreme hydrophobicity of these unmodified membranes. However, the dramatic increase in TNF-R by the PA membranes could not be answered solely based on complement activation. It has been shown that blood contact with synthetic polymer surface triggers the secretion of IL-1, which in turn promotes TNF-R secretion by monocytes.7 Therefore, the paracrine signaling function of IL-1 might have contributed to the observed increase in TNF-R concentration by the unmodified PA membranes. It is reasonable to infer that the surface crystallization of genistein at the pore interface has contributed to not only the improvement of antioxidant property but also the anti-inflammability property of the genistein-modified membranes. Mechanistically, it has been demonstrated that cytokine secretion due to LPS stimulation of human monocytes requires the activation of protein tyrosine kinase and protein kinase C (PKC), prior to gene transcription.26 Genistein has been shown to suppress IL-6 production in a dose-dependent manner, which has been attributed to its inhibition of protein tyrosine kinase.27 Similarly, other in vitro experiments have also demonstrated the ability of genistein to suppress significantly serum concentration of IL-1β and TNF-R following LPS stimulation of blood. Therefore, it is reasonable to conclude that the ability of genistein to inhibit protein tyrosine kinase was responsible for the observed suppression of all three cytokines by PA/G membranes. Moreover, the incorporation of PVP (i.e., PA:PVP/G membranes) has resulted in the reduction of all three cytokines. Although PVP is a hydrophilic polymer, it does not contain any charged groups to activate the complement cascade that promotes cytokine secretion. Therefore, the general perception that PVP addition results in improved biocompatibility of the HD membranes might be true in view of lesser complement activation and therefore lesser cytokine secretion.
ARTICLE
Personal Care Division, 3M Company, St. Paul, Minnesota 55144, United States.
’ ACKNOWLEDGMENT Support of this work by the Ohio Soybean Council is gratefully acknowledged. ’ REFERENCES (1) Hamilton, R. W. In Atlas of the Diseases of the Kidney; Henrich, W. L., Bennet, W. L., Eds.; Blackwell Publishing: Malden, MA, 1999; Vol. 5, p 3. (2) Daugridas, J. T.; Blake, P. G.; Ing, T. S. Handbook of Dialysis, 4th ed.; Kippincott Williams and Wilkins: New York, 2006. (3) Galli, F.; Canestarari, F.; Bouncristiani, U. Blood Purif. 1999, 17, 79–94. (4) Turi, S.; Nemath, I.; Torkos, A.; Saghy, L.; Varga, I.; Matkovics, B.; Nagy, J. Free Radical Biol. Med. 1997, 22, 161–168. (5) Canestrari, F.; Buoncristiani, U.; Galli, F.; Giorgini, A.; Albertini, M. C.; Carobi, C.; Pascucci, M.; Bossu, M. Clin. Chim. Acta 1995, 234, 127–136. (6) Nilsson, B; Ekdalh, K. N.; Mollnes, T. E.; Lambris, J. D. Mol. Immunol. 2007, 44, 82–94. (7) Herlebin, A.; Nguyen, A. T.; Zingraff, J.; Urena, P.; DescampsLatscha, B. Kidney Int. 1990, 37, 116–125. (8) Meskin, M. S.; Bidlack, W. R.; Davies, A. J.; Omaye, S. T. Phytochemicals in Nutrition and Health; CRC Press: Boca Raton, FL, 2002. (9) Bolch, A. J. Am. Diet. Assoc. 1995, 95, 493. (10) Bao, Y.; Fenwick, R. Phytochemicals in Health and Disease; Marcel Dekker: Basle, Switzerland, 2004. (11) Wei, H.; Wei, L. H.; Frenkel, K.; Bowen, R.; Barnes, S. Nutr. Cancer. 1993, 20, 1–12. (12) Kapiotis, S. Arterioscler., Thromb., Vasc. Biol. 1997, 17, 2868– 2874. (13) Puente, J.; Salas, M. A.; Canon, C.; Miranda, D.; Wolf, M. E.; Mosnaim, A. D. Int. J. Clin. Pharmacol. Ther. 1996, 34, 212–218. (14) Fiedor, P.; Kozerski, L.; Dobrowolzki, J. C.; Kawecki, R.; Beniecki, K.; Pachecka, J.; Rowinski, W.; Mazurek, A. P. Transplant. Proc. 1998, 30, 537. (15) Gohl, H.; Buck, R.; Strathmann, H. In Polyamide: The Evolution of a Synthetic Membrane for Renal Therapy; Shaldon, S., Koch, K. M., Eds.; Karger: Basle, Switzerland, 1992; Vol. 96, pp 125. (16) Herold, J.; Meyerhoff, G. Makromol. Chem. 1980, 181, 2625– 2636. (17) Herold, J.; Meyerhoff, G. Eur. Polym. J. 1979, 15, 525–532. (18) Neelakandan, C.; Kyu, T. Polymer 2010, 51, 5135–5144. (19) Matsuda, T. Nephrol., Dial., Transplant. 1989, 4, 60–66. (20) Peterson, G. J. Nutr. 1995, 125, 784S. (21) Miesel, R.; Hartung, R.; Kroeger, H. Inflammation 1996, 20, 427–438. (22) Xu, J.; Ikeda, K.; Yamori, Y. Hypertens. Res. 2004, 27, 675–684. 2454
dx.doi.org/10.1021/bm200591q |Biomacromolecules 2011, 12, 2447–2455
Biomacromolecules
ARTICLE
(23) Kimmel, P. L.; Phillips, T. M.; Simmens, S. J.; Peterson, R. A.; Weihs, K. L.; Alleyne, S.; Cruz, I.; Yanovski, J. A.; Veis, J. H. Kidney Int. 1998, 54, 236–244. (24) Sasaki, M. J. Artif. Organs. 2006, 9, 50–60. (25) Schindler, R.; Lonnemann, G.; Shaldon, S.; Koch, K. M.; Dinarello, C. A. Kidney Int. 1990, 37, 85–93. (26) Parikh, A. A.; Salzman, A. L.; Kane, C. D.; Fischer, J. E.; Hasselgren, P. J. Surg. Res. 1997, 69, 139–144. (27) Shapira, L.; Takashiba, S.; Champagne, C.; Amar, S.; Van Dyke, T. E. J. Immunol. 1994, 153, 1818–1824.
2455
dx.doi.org/10.1021/bm200591q |Biomacromolecules 2011, 12, 2447–2455
