In-vitro effects of the antimicrobial peptide Ala8,13,18-magainin II amide on isolated human first trimester villous trophoblast cells

Background Research on antimicrobial cationic peptides (AMPs) has gained pace toward using their potential to replace conventional antibiotics. These peptides preferentially interact with negatively charged membrane lipids typically seen in bacteria and thereby lead to membrane perturbations and membrane dysfunction. However, one possible disadvantage of AMP drugs is their potential for toxicity, especially to those cells which display externalization of negatively charged moieties to the outer leaflet of the plasma membrane during the process of syncytialization. Human placental villous trophoblast is one such cell type. Indeed, intra-vaginal administration of an antimicrobial cationic peptide Ala8,13,18-magainin II amide (AMA) which is a synthetic analogue of magainin 2 derived from Xenopus frog has been observed to result in inhibition of pregnancy establishment in monkeys. However, only little is known about the cellular behavior of early placental cytotrophoblasts (CTB) in the presence of cationic antimicrobial peptides. It is believed that suitable cell culture approaches using AMA as a representative alpha-helical AMP may yield tangible knowledge in this regard. Methods Immunocytochemical (ICC) analyses using confocal microscopy (n = 6 for each treatment sub-group) and Western blot (WB) method (n = 5 for each treatment sub-group) of CTB differentiation based on synthesis of beta-hCG and hPL, and apoptosis based on apoptosis-associated cytokeratin 18 neo-epitope (CK18f) were performed for CTB isolated from human first trimester placental villi and grown in serum-free primary culture for 24 h, 48 h and 96 h on rat-tail collagen with and without AMA (1000 ng/ml). Moreover, secretion of beta-hCG and hPL into conditioned media from isolated CTB grown in vitro for 24 h, 48 h and 96 h (n = 6/each sub-group) with and without AMA was examined using enzyme immunoassays. Furthermore, TUNEL assay, and cell viability based on LDH leakage into medium (n = 6/each sub-group) were assessed to examine the phenomenon of cell death with time and administration of AMA. Results CTB in serum-free primary culture showed increased (P < 0.05) level of synthesis and secretion of beta-hCG and hPL with time, and higher (P < 0.05) level of cellular cytokeratin 18 neo-epitope and number of TUNEL-positive cells, and LDH activity in conditioned medium at 96 h of culture. Exposure of CTB to AMA resulted in lower (P < 0.05) level of synthesis and secretion of beta-hCG and hPL, as well as, an increase (P < 0.05) of cellular cytokeratin 18 neo-epitope and number of TUNEL-positive cells, and LDH activity in conditioned medium at 96 h as compared to the control treatment. Conclusions Administration of AMA resulted in attenuation of differentiation, enhancement in apoptosis and loss of viability in early placental villi trophoblast cells in primary culture. Thus, it appears that administration of alpha-helical AMP may adversely affect the process of placentation and pregnancy outcome.


Background
One major challenge of medicine today is the growing number of bacterial strains resistant to conventional antibiotic therapies. Hence, the need for new antibiotics or even alternative compounds has stimulated research in the field of antimicrobial peptides to be used as human therapeutics [1,2]. To this effect, research on gene-encoded cationic antimicrobial peptides (AMPs) has gained pace in the recent time [3]. AMPs can be defined as being short peptides (10-50 amino acids) with an overall positive charge (+2 to +9) and a substantial proportion of (>30%) of hydrophobic residues [3][4][5]. These chemical properties in AMPs generally result in folds into amphiphilic structures, especially upon contact with membranes and give rise to formation of separate patches rich in positively charged and hydrophobic amino acids [3][4][5]. There are four broad structural groups of AMPs: α-helical peptides (for example, cercopin B, magainins, LL37), extended structures rich in glycine, proline, tryptophan, arginine, histidine (for example, indolicidin and histatin 1), peptides with one disulfide bond (for example, bactenecin and esculentin A), and β-sheet peptides stabilized by two or more disulfide bridges (for example, human defensins and protegrins) [3][4][5][6]. These peptides preferentially interact with negatively charged lipids, which are major components of bacterial cell membranes resulting in membrane perturbations such as pore formation, alterations of the curvature strain and induction of lipid-peptide domain formation [7]. Such perturbations may alter the micro-environment of membrane proteins resulting in membrane dysfunction. In mammalian cell membranes however negatively charged lipids such as phosphatidylserine are mostly located in the inner leaflet of the membrane and thus are not exposed to the outer surface of the cell. However, during pregnancy such peptides may jeopardize placental villous trophoblasts because these cells show externalization of negatively charged phosphatidylserine moieties to the outer leaflet of the plasma membrane during the process of syncytialization [8]. This may make them vulnerable to AMPs [9]. In fact, one possible disadvantage AMP drugs, especially for α-helical AMPs, is their potential for toxicity [10], although the issue of AMP drugs mediated toxicity on placental trophoblasts has hitherto been little addressed [3].
Ala 8,13,18 -magainin II amide (AMA) is a synthetic AMP [11,12] belonging to α-helical peptide group of AMPs and modified from natural magainin 2 obtained from African frog Xenopus leavis [13,14]. Replacement of three amino acid residues (Ser 8 , Gly 13 , Gly 18 ) of magainin 2 peptide with alanine enhances its antimicrobial activity, and amidation of its terminal amino acid renders stabilization to its α-helical conformation [12]. It shows chemical properties that are very similar to that of human α-helical AMP, cathelicidin peptide LL-37 [3,6] and is commercially available. In a previous study, we have reported that intra-vaginal administration of AMA resulted in the inhibition of blastocyst implantation in the rhesus monkey [9]. Further study in the same species indicated that the anti-nidatory effect of AMA might be resulted from its inhibitory effect on trophoblast cell differentiation [15]. However, little is known about the cellular behavior of early placental trophoblast cells in the presence of cationic antimicrobial peptides. It is assumed that suitable cell culture approaches using AMA as an AMP of non-defensins group may yield tangible knowledge in this regard.
In the present study, we aimed to examine the in vitro effects of AMA on cellular processes like differentiation and apoptosis in primary human placental villous cytotrophoblasts (CTB) isolated from first trimester placental tissues. To this end, we have employed well known markers of cellular phenotype for trophoblast differentiation, apoptosis and viability in primary cell culture system [16]. Table 1 provides a summary of the target proteins examined in the present study. The assessment of CTB differentiation in culture has been performed by examining synthesis and secretion of human chorionic gonadotropin beta (βhCG) and human placental lactogen (hPL) [16][17][18]. The assessment of apoptosis was done by examining the level of caspase-mediated cleavage of cytokeratin 18 yielding a specific neo-epitope (cytokeratin 18 neo-epitope, CK18f), and nuclear DNA fragmentation using terminal deoxynucleotidyl transferase (TdT) enzyme based TUNEL (TdT dUTP nick end labeling) assay in trophoblast cells. Both methods are based on well characterized phenomena of apoptosis at cytoplasm and nucleus, respectively; these are considered robust approaches in studying apoptosis [19,20]. The loss of cell viability in culture with and without AMA at different time points was assessed from lactate dehydrogenase (LDH) activity in the conditioned media. This approach is based on the fact that the loss of intracellular LDH into the culture medium is a reliable indicator of irreversible cell membrane damage and cell death [21].

Tissue samples and chemicals
Human placental samples (N = 47) were obtained from women (age: 23-37 years) undergoing elective surgical termination of singleton pregnancy between 6 and 8 weeks of gestation (timed from last menstrual period) without undergoing any prior medication. All women provided their written informed consent. The Ethics Committee of the All India Institute of Medical Sciences approved the research study. The study also complied with the Helsinki declaration. Placental samples were collected in sterile ice-cold phosphate buffered saline (PBS, pH 7.4) and transported on ice to the laboratory within 15 min after collection for further processing. All the chemicals were obtained from Sigma Chemical Co.

Cell culture
The methodological details of CTB culture have been detailed previously [24][25][26]. Briefly, isolated mononucleated CTB were plated at a density of 1 × 10 5 /cm 2 on collagen I and cultured at 37°C in a humidified air atmosphere of 5% CO 2 and in complete medium [DMEM: F12 (1:1), 10% (v/v) fetal calf serum, penicillin (100 IU/ml), streptomycin (100 μg/ml), amphotericin B (2.5 μg/ml)] for 24 h to allow for their attachment to collagen. Subsequently, the cells were maintained by daily feeding in serum-free medium with antibiotics and antimycotics as described above and supplemented with insulin (5 μg/ml), transferrin (5 μg/ml), selenium (5 ng/ml) and hydrocortisone (0.5 μg/ml) based on the initial optimization observation that addition of these factors render more number of target cells attached, viable and aggregated in serum-free primary culture on rat-tail collagen. It has also been earlier reported that collagen and insulin supported CTB differentiation in primary culture [27]. The same pools of cells were treated with or without AMA (1000 ng/ml) in triplicates to harvest at 24 h, 48 h and 96 h. The concentration of AMA in the present study was selected based on the previously reported dose finding study on viability, hCG secretion and invasion efficiency for isolated CTB indicating that AMA at 1000 ng/ml affected trophoblast invasion into collagen coated membrane with 8 μm pores without any marked effect on hCG secretion and mitochondrial viability at 48 h in presence of serum containing medium [25]. In initial experiments, method optimization for the methods employed in the present study was performed at three time points as mentioned above in at least three successive satisfactory experiments.

TUNEL assay
After termination of cultures at 24 h (n = 6), 48 h (n = 6) and 96 h (n = 6) in triplicates, cells treated with and without AMA (1000 ng/ml) were subjected to in situ detection of nuclear fragmentation according to the methods described elsewhere [28,29]. In brief, cells were Samples along with positive control with nuclease treatment and negative control without enzyme were run as per the method protocol provided by the supplier. The labeled cells were counterstained with nuclear fast red and cell numbers were estimated from nuclear counts. All samples were analysed in one run.

Immunofluorescent staining and image analysis
After termination of culture at 24 h, 48 h, and 96 h, cells treated with and without AMA (1000 ng/ml) were subjected to immunocytochemical (ICC) staining using various antibodies as described in Table 1. Immunofluorescent staining was performed on six samples in duplicates for each treatment group using a method described earlier [30]. Appropriate fluorochrome-conjugated secondary antibodies (Molecular Probes, Grand Island, NY, USA) were used for visualization. Specificity of the antibody binding was assessed by omitting primary antibodies, immunoadsorption of primary antibodies with target antigens, replacing primary antibodies with unrelated IgG from same species and other species, omitting secondary antibodies, and replacing secondary antibodies with unrelated IgGs from same or other species in parallel cultures. Non-overlapping images (at least eight) from entire area of confluence of each culture after ICC were examined using a Confocal laser scanning microscope (Leica Microsystems, Wetzlar GmbH, Germany). The cells were detected using an interactive planimeter analyzer only in cases where discernibility was distinct. The immunopositive cells were identified by detecting positive profiles in digitized images. Following image grabbing and segmentation, digitalized images were subjected to image analysis for estimation of immunopositive area fraction (in per cent) based on manual outlining method and using an optimized grey level threshold yielding normal distribution of image pixels throughout the entire gray scale and shading correction. Pixel calibration was done against the standard provided by the manufacturer. The per cent immunostained areas was measured using a precalibrated computer-assisted digital image analysis system (Leica QWIN DC 200, Cambridge, UK) as described elsewhere [30]. The nuclei were counterstained with diamidino-6-phenylindole (DAPI) (Molecular Probes), and cell numbers were estimated from DAPI positive nuclear counts.

Immunoblots
After termination of cultures at 24 h (n = 5), 48 h (n = 5) and 96 h (n = 5) in duplicate with and without AMA (1000 ng/ml), cells were subjected to Western blot (WB) analysis. Profiles of the candidate proteins in cell homogenates were assessed in samples of 25 μg protein content along with pre-stained molecular weight markers based on SDS-PAGE/Western immunoblotting method on nitrocellulose membrane using electrophoresis and trans-blot equipment, and chemicals obtained from Bio-Rad (Hercules, CA, USA) as described elsewhere [30]. Final visualization was achieved by using Vectastain ABC immunoperoxidase kits (Vector Laboratories). Respective primary antibody and secondary antibody controls were run simultaneously to examine the specificity of procedure. The molecular weights and semi-quantitative densitometric analysis of bands were determined using a densitometric equipment (Pharos FX Plus Molecular Imager, Bio-Rad, Hercules, CA, USA) and an optimized densitometric analysis software (PD Quest Advanced, Bio-Rad). The integrated measures of optical densities for individual antigens were calculated from log of transmittance for each of the target antigen and normalized with that of β-actin.

Analysis of bhCG, hPL and LDH in conditioned media
After termination of cultures at 24 h (n = 6), 48 h (n = 6) and 96 h (n = 6) in triplicates with and without AMA (1000 ng/ml), supernatant from each cell culture well was collected, centrifuged and stored at -80°C for estimation of secreted bhCG and hPL by enzyme immunoassays (25) and lactate dehydrogenase (LDH) activity using standardized methods. The immunoassay kits for βhCG and hPL were purchased from DRG International Inc. (New Jersey, NJ, USA). The ranges of detectable concentrations in the assays were 5 -300 mIU/ml for βhCG and 1 -20 mIU/ml for hPL. All samples were analysed in one run. The intra-assay coefficients of variation for both enzyme immunoassays were less than 9%.
The potential cytotoxic effect of AMA (1000 ng/ml) to isolated CTB in culture with or without AMA at different time points was assessed by lactate dehydrogenase (LDH) leakage into the conditioned culture medium using a LDH kit obtained from Sigma. The assay is based on the conversion of lactate to pyruvate in the presence of LDH with parallel reduction of NAD estimated at 340 nm in a Benchmark Plus Spectrophotometer (Bio-Rad, Hercules, CA, USA) with necessary corrections [31]. All samples were analysed in one run. The intra-assay coefficients of variation for both assays were less than 6%.

Statistical analysis
Statistical analysis of data was performed using Kruskal-Wallis test followed by rank sums Wilcoxon test using SPSS v17 (Chicago, IL, USA). The probability level of P = 0.05 was taken as the limit of significance.

Effect of time course
Isolated villous trophoblast cells grown on rat-tail type I collagen in serum-free medium yielded adherent spheroid cells at 24 h (Figure 2A), showed aggregations at 48 h ( Figure 2B), and formed distinct aggregates of fused cells at 96 h ( Figure 2C).
Immunocytochemical analyses revealed that isolated primary villous cytokeratin 7-and βhCG-positive and vitronectin receptor (CD51)-negative CTB in culture synthesized an increasing amount of βhCG (P < 0.05) and hPL (P < 0.05) at 48 h as compared to 24 h, and at 96 h as compared to 24 h and 48 h ( Table 2; Figures 3A-C, G-I). A concomitant increase (P < 0.05) in concentrations of βhCG and hPL was also detected in the conditioned media with time ( Table 3). As shown in Figure 4 and Table 2 Table 2).

Effect of AMA
Administration of AMA at a concentration of 1000 ng/ ml resulted in reduced numbers of cells and cell aggregates ( Figures 2D-F) Tables 2 and 3) and the number of TUNEL-positive cells along with higher degree of LDH leakage into the conditioned medium with marked (P < 0.05) effect being observed at 96 h ( Table 3).

Discussion
In the present study we could obtain an enriched population of cytokeratin 7 and βhCG positive and CD51 negative CTB from first trimester placental villi based on a method of sequential enzymatic treatment and plating the isolated cells on collagen I. It has been demonstrated earlier that cytokeratin 7-and βhCG-positive and CD51 (vitronectin receptor)-negative CTB were primarily villous in source, while cytokeratin 7-, βhCGand CD51-positive CTB were derived from extravillous in source and invasive in nature [24,32]. Villous CTB in the present primary culture model displayed functional differentiation marked by increasing synthesis and secretion of hCG and hPL over time. In the present study, trophoblasts retained their biological functionality and formed aggregates of fused cells in primary culture on collagen [27]. However, we have not tested for syncytialization using staining for membrane markers such as Ecadherin or fodrin [33,34]; so we could not assess the numbers of syncytia in our cultures.
Previously, several dynamic processes related to CTB differentiation have been studied using primary cultures of human placental villous trophoblast cells as a model system [16,27,35]. It was earlier demonstrated that mononucleated CTB isolated from placental villi by trypsinization aggregated in a timed manner and fused to form syncytiotrophoblast (STB) that synthesized and secreted hCG and hPL [36,37]. Typically, CTB retrieved from first trimester placental villi have been shown to aggregate and start synthesizing hCG and hPL in culture by 3-days [27,35]. Similar results were obtained in the present study. A relatively high level of secretion of both hormones by 48 h and more markedly at 96 h along with marked degree of cell aggregration might be indicative of positive influence of hCG on CTB differentiation as well [38]. It is also apparent from the results of the present study that a higher number of cells underwent apoptosis and showed loss of viability as evident from higher level of cytokeratin 18 neo-epitope in cells, higher number of TUNEL-positive cells, and higher concentration of LDH leaked into the conditioned medium at 96 h as compared to 24 h and 48 h cultures. It is generally believed that these three parameters relate well with the apoptosis process and loss of cell viability [39]. The additional observation that the propidium iodide positive cells showing binding to annexin V increased at 96 h as compared with that at 24 h and 48 h (data not shown) also substantiated the conclusion [40].
We also observed that exposure of villous CTB to the cationic peptide AMA resulted in decreased production and secretion of βhCG and hPL along with increased degree cell apoptosis and loss of viability assessed from cytokeratin 18 neo-epitope level and TUNEL-positive nuclei containing cells, and LDH leakage into medium, respectively. These actions of AMA collectively might have resulted in observed pregnancy inhibition in monkeys [9,15].
The effect of AMA in attenuating the process of functional differentiation in terms of βhCG and hPL synthesis and secretion and that on cell viability in primary culture as observed in the present study should be considered against the background that villous trophoblasts follow two tightly regulated pathways of differentiation during placental development giving rise to differentiated CTB in anchoring villi and floating villi. In the anchoring villi, CTB aggregate into cell columns and invade maternal decidua, while in the floating villi, CTB differentiate into STB [41]. Thus, the present results document that AMA inhibited only the pathway of CTB differentiation into STB in vitro.
In a previous dose finding study, we observed that first trimester human placental villous CTB maintained in three dimensional culture subjected to AMA (1000 ng/ ml) treatment in the presence of serum (10%, v/v) did    not result in any significant change in their mitochondrial viability and hCG secretion, however, with detectable reduction in invasion efficiency at 48 h as compared to the control treatment (25). In the present study, same concentration of AMA in serum-free medium resulted in attenuation of differentiation, enhancement in apoptosis and loss of viability in early placental villi trophoblast cells at 96 h in vitro. Thus, it appears that serum factors might render a potential protection to cytotoxic action of AMA on early placental villous CTB. The possibility that serum factors may provide protection to mammalian cells from AMA mediated cytotoxicity has also been indicated earlier by others [3,42].
The underlying mechanism of the observed AMA mediated intervention of trophoblast function in the present study is not known, however, it appears that two potential mechanisms may be examined in this regard. Firstly, AMA might engage anionic domain in the cell membrane of CTB during the process of syncytialization and thereby resulted in functional inadequacy and loss of viability as we proposed in the background section based on previous reports [8,43]. It would indeed be of interest to directly examine the involvement of AMA with CTB cell membrane using high resolution time lapse photomicrography with suitable fluorochrome labels. The model used in the present study however did not allow this type of investigation. Secondly, AMA might act through calcium mediated process as suggested in studies of magainin 2 mediated histamine release [44] and caspase-independent apoptosis [45]. Indeed there is evidence to support that STB of first trimester placenta express transient receptor potential (TRPC) homologues that are involved in store-operated calcium entry [46]. Therefore it will be interesting to study calcium imaging in CTB treated with and without AMA in vitro. In this context, it is notable that our pilot experiments with isolated human endometrial stromal cells grown on collagen biomatrix failed to    document any marked change in the cellular behavior and viability following AMA administration, unlike isolated CTB in vitro as observed in the present study. Thus, it indicates that AMA mediated a specific effect on the differentiation dynamics and viability of first trimester placental villous CTB maintained in primary culture on collagen. Further experiments are needed to investigate the anti-nidatory risk factors of AMP, vis-àvis, their potential application to combat reproductive tract infection saving pregnancy based on new therapeutic antibiotic approach [47]. It is interesting to note in this context that AMA is an effective broad-spectrum, non-specific antimicrobial agent that also inhibits Herpes simplex virus types 1 and 2 also [2][3][4][5][6]48].

Conclusions
The antimicrobial cationic peptide AMA attenuates differentiation and enhances apoptosis in primary villous CTB isolated from first trimester placenta. Further studies to delineate the effect of AMA on early stage placental trophoblast cells may give important leads towards understanding the anti-implantation risk factors in new therapeutic antibiotic approaches.