Proteomic Identification of IPSE/alpha-1 as a Major Hepatotoxin Secreted by Schistosoma mansoni Eggs

Background Eggs deposited in the liver of the mammalian host by the blood fluke parasite, Schistosoma mansoni, normally drive a T-helper-2 (Th2)-mediated granulomatous response in immune-competent mice. By contrast, in mice deprived of T-cells and incapable of producing granulomata, egg-secreted proteins (ESP) induce acute hepatic injury and death. Previous work has shown that one such ESP, the T2 ribonuclease known as omega-1, is hepatotoxic in vivo in that specific antisera to omega-1 prevent hepatocyte damage. Methodology/Principal Findings Using an in vitro culture system employing mouse primary hepatocytes and alanine transaminase (ALT) activity as a marker of heptocyte injury, we demonstrated that S. mansoni eggs, egg-secreted proteins (ESP), soluble-egg antigen (SEA), and omega-1 are directly hepatotoxic and in a dose-dependent manner. Depletion of omega-1 using a monoclonal antibody abolished the toxicity of pure omega-1 and diminished the toxicity in ESP and SEA by 47 and 33%, respectively. Anion exchange chromatography of ESP yielded one predominant hepatotoxic fraction. Proteomics of that fraction identified the presence of IPSE/alpha-1 (IL-4 inducing principle from S. mansoni eggs), a known activator of basophils and inducer of Th2-type responses. Pure recombinant IPSE/alpha-1 also displayed a dose-dependent hepatotoxicity in vitro. Monoclonal antibody depletion of IPSE/alpha-1 abolished the latter's toxicity and diminished the total toxicity of ESP and SEA by 32 and 35%, respectively. Combined depletion of omega-1 and IPSE/alpha-1 diminished hepatotoxicity of ESP and SEA by 60 and 58% respectively. Conclusions We identified IPSE/alpha-1 as a novel hepatotoxin and conclude that both IPSE/alpha-1 and omega-1 account for the majority of the hepatotoxicity secreted by S. mansoni eggs.


Introduction
Schistosomiasis is a chronic parasitic disease that affects more than 200 million people worldwide [1]. The central pathological characteristic during chronic infection is a granulomatous reaction around trapped parasite eggs in the host's liver, bladder, or intestine [2]. Granulomatous inflammation in the liver may result in fibrosis, scarring, portal hypertension, and in the worst cases hemorrhaging and death [3].
Schistosoma mansoni infection in mice is the most common experimental model employed. Approximately five-to-six weeks post-infection, parasite eggs deposited by adult worms induce a Thelper-2 (Th2)-type-polarized immune response [4]. A number of the immunodominant molecules in eggs have been described [5,6,7,8], in addition to the two molecules central to this report (see below). The ability of S. mansoni eggs to induce Th2-type differentiation during infection is underscored by the observation that eggs alone, or soluble egg antigen (SEA) released by the eggs through pores in the shell, are sufficient to drive Th2 polarization in naïve uninfected mice [9,10,11].
Research from the late 1960s and 1970s has documented that mice lacking T-cells due to genetic loss or surgical removal of the thymus [12,13,14], or administration of specific immunosuppressives [15,16,17], do not develop a typical granulomatous response to trapped parasite eggs. In these T-cell depleted mice, infection was associated with extensive hepatic parenchymal damage suggestive of a cytotoxic egg product(s) diffusing into hepatic tissue [18]. Histopathology of livers from schistosome-infected immunocompromised mice displayed microvesicular hepatocyte steatosis [18,19,20], nuclear degeneration, and hepatocyte apoptosis [21]. Coincident with hepatocyte injury, there is an increase in liver cell transaminases in the plasma [19]. Immunosuppressed mice also have higher mortality once egg deposition in the liver begins [17,18,19,22]. In immunologically intact mice, circumoval granulomata, and antibody responses to released S. mansoni egg components, likely protect against hepatocyte damage. Also, hepatotoxicity is prevented in infected T cell-deprived mice by transfer of serum from intact mice immunized with S. mansoni eggs or egg homogenate, whilst antisera against other lifecycle stages do not prolong survival [19]. Egg-induced hepatotoxicity appears to be specific to S. mansoni; it is not observed during infection of T-cell deprived mice with S. haematobium or S. bovis [23]. Finally, transfer of sera from S. japonicum-infected mice failed to alleviate hepatotoxicity induced by S. mansoni eggs in T-cell deprived mice [24].
Research in the 1980s identified several proteins in S. mansoni egg antigen preparations based on their electrophoretic mobility and their recognition by sera from mice with chronic infection [18,25]. Two of these proteins, termed omega-1 and alpha-1, were isolated from S. mansoni egg homogenates (SmAE) by cation exchange chromatography in a single salt-eluted fraction that was termed cationic egg fraction 6 (CEF6) [26].
Omega-1 is a 31 kDa monomeric glycoprotein [26] released from S. mansoni eggs [27] that has previously been reported to be hepatotoxic [18]. Monospecific antisera against omega-1, which is a highly immunoreactive egg antigen, were protective against hepatocyte damage in T-cell deprived mice [18,23]. Immunochemical characterization of omega-1 using sera from humans and mice infected with different schistosome species suggested that the antigen is specific to S. mansoni [26]. Omega-1 is a functional T2 ribonuclease (RNase) [28]. Cytotoxic RNases (which includes T2 family RNase members) have been found in a wide range of species from bacteria to mammals. A range of biological roles has been suggested, including serving as extra or intracellular cytotoxins and modulating host immune responses [29]. Omega-1 was also reported to drive Th2 polarization in human monocyte-derived dendritic cells (DCs), whereas SEA depleted of omega-1 loses this ability [30]. Omega-1 directly affects both DC morphology and the ability of these antigenpresenting cells to interact physically with CD4 T-lymphocytes [31]. Furthermore, when injected into IL-4 dual reporter mice, omega-1 is a potent inducer of the Th2 response in vivo [30].
A second major protein in S. mansoni egg homogenates, originally termed alpha-1 [25], was recently reported to be identical to IPSE (IL-4 inducing principle from S. mansoni eggs) [32]. IPSE/alpa-1 is a glycoprotein [33] that has been crystallized [34], occurs naturally as a dimer (33)(34)(35), and is enriched in the sub-shell area of S. mansoni eggs from where it is secreted into the surrounding tissue [32]. It is not found in the miracidium residing within the egg [35]. Various proteomics analyses have identified IPSE/alpha-1 as an abundant protein in egg secretions [5,36,37,38]. IPSE/alpha-1 binds immunoglobulin and activates naïve basophils, leading to histamine release and facilitating the production of Th2-type cytokines [32]. In vivo, IPSE/alpha-1 induces interleukin (IL)-4 secretion from murine basophils in an IgE-dependent but antigen-independent manner [39]. Most recently, IPSE/alpha-1 has been shown to contain a functional C-terminal, monopartite, nuclear localization sequence that binds DNA such that it may alter gene expression in the host cell [40].
We employed a primary hepatocyte in vitro culture system to identify and measure direct toxicity of S. mansoni eggs and their derived fractions, including pure proteins. Hepatotoxicity of omega-1 was confirmed, and IPSE/alpha-1 was identified as a novel hepatotoxin. Both proteins together account for more than half of the egg-derived toxicity measured.

Ethics statement
These studies were performed in accordance with the recommendations by the University of California San Francisco Institutional Animal Care and Use Committee. The protocol was approved by the Committee on the Ethics of Animal Experiments of the University of California San Francisco (Permit Number: AN080237-03). All surgery was performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering. The protocol followed these guidelines in the study: All U.S. Federal Policy and Guidelines governing the use of laboratory animals, Public Health Service Policy on Humane Care and Use of Laboratory Animals, Guide for the Care and Use of Laboratory Animals, National Academy Press, USDA Animal Welfare Act and Regulations, and NIH Office of Laboratory Animal Welfare Guidelines.
In vitro maintenance of S. mansoni and collection of Egg-Secreted Proteins (ESP) S. mansoni eggs were isolated from the livers of female golden hamsters six weeks following infection with 500 cercariae, as previously described [41]. Approximately 0.5 million eggs were washed twice in serum-free RPMI-1640 supplemented with 100 mg/ml streptomycin. Eggs were then resuspended in 2 ml RPMI-1640, and 500 ml aliquots were placed in 12-well culture plates (Costar). Cultures were checked daily by microscopy to ensure sterility. Medium was harvested after 72 h, and centrifuged for 10 min at 2006g to remove eggs. ESP, usually containing approximately 0.5 mg/ml protein by Bradford assay [42], was stored at 280uC. After collection of ESP, egg viability was confirmed by hatching of miracidia; normally .85% of the eggs hatched. Hatching of eggs during the collection period was ,1%. SEA was prepared, as described previously [43].

Fractionation of ESP by anion exchange chromatography
ESP (2 mg) were added to 2 ml 30 mM Tris-HCl, pH 8.0, and centrifuged at 50006g for 10 min at 4 C. The supernatant was loaded onto an Hr 5/5 Mono Q column (GE Healthcare), and

Author Summary
The flatworm disease, schistosomiasis, is a major public health problem in sub-Saharan Africa, South America and East Asia. A hallmark of infection with Schistosoma mansoni is the immune response to parasite eggs trapped in the liver and other organs. This response involves an infiltration of cells that surround the parasite egg forming a ''granuloma.'' In mice deprived of T-cells, this granulomatous response is lacking, and toxic products released by eggs quickly cause liver damage and death. Thus the granulomata protect the host from toxic egg products. Only one hepatotoxic molecule, omega-1, has been described to date. We set out to identify other S. mansoni egg hepatotoxins using liver cells grown in culture. We first showed that live eggs, their secretions, and pure omega-1 are toxic. Using a physical separation technique to prepare fractions from whole egg secretions, we identified the presence of IPSE/alpha-1, a protein that is known to strongly influence the immune system. We showed that IPSE/alpha-1 is also hepatotoxic, and that toxicity of both omega-1 and IPSE/alpha-1 can be prevented by first mixing the proteins with specific neutralizing antibodies. Both proteins constitute the majority of hepatotoxicity released by eggs. equilibrated with the same buffer and elute by a 0 to 1 M linear NaCl gradient in six column volumes of Tris-NaCl buffer. Flowthrough and eluted fractions (1 ml) were stored at 280uC prior to testing for toxicity with cultured primary hepatocytes (applied volume 100 ml of each fraction to 2610 5 hepatocytes/0.5 ml).
Primary hepatocyte preparation, culture and exposure to egg-derived material Hepatocytes were isolated from C57/BL6 mice by in situ perfusion of liver with collagenase, as described previously [44]. The portal vein was severed to permit outflow followed by cannulation of the inferior vena cava with a 22-guage catheter. The liver was then flushed with a calcium-chelating buffer (liver perfusion medium) for 3 to 5 min, followed by perfusion with collagenase (liver digest medium) for an additional 6-8 min. At the end of the digestion, the liver was removed to a sterile dish and minced thoroughly with a scissors. This crude liver cell isolate was suspended in 25 ml of Dulbecco's modified Eagle's medium/ Ham's F-12, filtered through sterile gauze, centrifuged at 706g for 2 min, and resuspended in Dulbecco's modified Eagle's medium/ Ham's F-12. After an additional round of centrifugation and resuspension, hepatocytes were isolated by centrifugation using a 50% Percoll gradient. Hepatocytes were cultured at a density of 2610 5 /0.5 ml per well in a 12-well culture plate (Costar), previously coated with a 5 mm layer of matrigel (BD Bioscience) [45], which is a tumor biomatrix prepared from the Engelbroth-Holm-Swarm mouse sarcoma [46]. Hepatocytes were allowed to attach for 1 hour at 37uC. Culture plates were gently swirled and the medium containing unattached cells and debris was aspirated. Cultures were then incubated for 72 h in a final volume of 0.5 ml RPMI medium containing 50, 100, or 200 eggs, unfractionated ESP (10 mg/ml) or 100 ml of chromatography fractions. Supernatants were collected and stored at 220uC and analyzed for alanine transaminase (ALT), a serum marker of hepatoxicity [47]) using a Beckman Chemical Analyzer in the Clinical Chemistry Laboratory of the San Francisco Veterans Affairs Medical Center (VAMC).

Proteomic Analysis of ESP
ESP was fractionated by SDS-PAGE then silver stained [48,49], and 40 evenly spaced protein bands were sliced out of the gel ( Fig  S1). The gel slices were then diced into small cubes, reduced and alkylated with dithiothreitol and iodoacetamide, and in-gel digested with trypsin [50,51]. The resulting peptides were extracted and analyzed by on-line liquid chromatography/mass spectrometry, using an Eksigent nanoflow pump and a Famos autosampler, which were coupled to quadrupole-orthogonalacceleration-time-of-flight hybrid mass spectrometer (QStar XL or Pulsar, Applied Biosystems). Peptides were fractionated on a reversed-phase column (C18, 0.756150 mm), and a ''5-50% B gradient-in-gradient'' was developed in 35 min at a 350-nl/min flow-rate. Solvent A was 0.1% formic acid in water and solvent B was 0.1% formic acid in acetonitrile. Data were acquired in information-dependent acquisition mode: 1 sec MS surveys were followed by 3 sec CID experiments on computer-selected multiply charged precursor ions. Peak lists were generated using Analyst 2.0 software (Applied Biosystems) with the Mascot script 1.6b20 (Matrix Science). Database searches were performed using ProteinProspector v. 5.1.7 (http://prospector2.ucsf.edu) [52]. Searches were performed first on the SwissProt databank (December 16, 2008, 405,506 entries) to evaluate sample purity, followed by searching in the S. mansoni database SchistoDB v. 4.0 (www.schistodb.net; 13,174 entries downloaded July 2009). Batch-Tag settings were selected for samples prepared with trypsin allowing a maximum of one missed cleavage and no nonspecific cleavages. Peptide modifications searched for included carbamidomethyl (Cys) as the only fixed modification, and up to two variable modifications from among the following: oxidation (Met), acetyl (N-term), oxidized acetyl (N-term), pyroglutamate (Gln) and Met-loss (N-term). The mass accuracy considered was 200 ppm, and 300 ppm for the precursor and fragment ions, respectively. The following acceptance scores for database matches were required: a minimum protein score of 22, a minimum peptide score of 15, and a maximum expectation value of 0.02 required for both peptide and protein identification. These criteria resulted in an approximate 2% false determination rate. Protein identifications are reported with a minimum of two peptide matches per protein. For the analysis of ESP anion exchange fraction #11, the maximum expectation value was changed to 0.05, and no decoy proteins were identified using these acceptance criteria.

S. mansoni eggs and their secretions are hepatotoxic in vitro
To measure hepatotoxicity caused by parasite eggs ESP, and their chromatographic fractions, we employed an in vitro system involving murine primary hepatocytes cultured on matrigel. ALT was employed as a hepatoxicity biomarker. With parasite eggs, measurements were taken 24, 48 and 72 h. No alteration in ALT levels was seen at 24 or 48 h (not shown); however, by 72 h, ALT had increased markedly and in a dose-dependent manner ( Figure 1A). Dose-dependent hepatocellular injury elicited by ESP was also measured after 72 h ( Figure 1B).
To aid identification of those ESP constituents responsible for hepatotoxicity, ESP was fractionated by Mono Q anion exchange chromatography. The flow-through (component not bound to the column), and each of the eluted fractions (100 mL), was co-incubated with cultured hepatocytes. At 72 h, fraction #11 induced an approximate two-fold greater release of ALT relative to the other eluted fractions and flow-through, such that it was the equivalent of 10 mg/ml unfractionated ESP ( Figure 2). Proteomic analysis of ESP and hepatotoxic fraction #11: identification of IPSE/alpha-1 SDS-PAGE and tryptic digestion followed by mass spectrometry of ESP and hepatotoxic fraction #11 identified 99 and nine proteins, respectively (Tables 1 and 2). Previously, total proteomic analysis of ESP identified 188 proteins [36], and many are common between this and the present dataset (Table 1). Among the nine proteins identified in fraction #11 were metabolic enzymes involved in glucose metabolism, glycogen storage, in addition to chaperones. IPSE/alpha-1 was also identified; and, given its potent immunomodulatory properties [32,34], was of immediate interest in discovering of whether or not it was hepatotoxic.

In vitro confirmation that Omega-1 is hepatotoxic
Omega-1 is an egg-secreted glycoprotein with RNase activity and in vivo-demonstrated hepatotoxicity [18,28]. To measure direct toxicity to primary cultured hepatocytes, purified native omega-1 was co-incubated with primary hepatocytes. At 72 h, a dose-dependent release of ALT was measured ( Figure 3A). Likewise, Aspergillus oryzae T2 RNase (25 U/ml) (Invitrogen, # 18031-013 Carlsbad, CA) was toxic. Importantly, pre-incubation of pure omega-1 with a monoclonal anti-omega-1 antibody bound to Protein G Sepharose abolished cytotoxicity ( Figure 3B). Depleting ESP and SEA with the same antibody decreased toxicity by 47 and 33%, respectively. All reductions in hepatoxicity were statistically significant ( Figure 3B).

IPSE/alpha-1 is also a hepatotoxin in vitro
IPSE/alpha-1 was an abundant protein in the hepatotoxic fraction #11 from anion exchange chromatography (Table 2). Recombinant IPSE/alpha-1 was added to hepatocyte cultures, and a dose-dependent    toxicity was measured at 72 h ALT levels that was significantly elevated relative to negative controls ( Figure 4A). Similar to that found for omega-1, specific neutralization of rIPSE/alpha-1 with an anti-rIPSE/alpha-1 monoclonal antibody abolished activity and decreased the cytotoxicity of ESP and SEA by 32 and 35%, respectively ( Figure 4B). All reductions in hepatoxicity were statistically significant. To measure the combined contributions of omega-1 and IPSE/ alpha-1 to the hepatotoxicity of ESP and SEA in vitro, both eggderived preparations were depleted of both omega-1 and IPSE/ alpha-1 with specific monoclonal antibodies prior to incubation with hepatocytes. The combination of both antibodies diminished hepatotoxicity of ESP and SEA by 60 and 58%, respectively ( Figure 5).

Discussion
The pathogenesis of hepatic schistosomiasis is due to the host's granulomatous response to eggs deposited in the liver [2]. The initial cellular granuloma is characterized by the presence of activated macrophages, lymphocytes, and eosinophils, as reviewed in both Agnew and Pearce [23,53]. Over time, granulomata become fibrotic, and their accumulation in periportal areas, as is the case in chronic S.mansoni infection, can lead to portal hypertension, hemorrhaging, and death [3]. Ironically, in the  absence of a granulomatous response, experimental hepatic schistosomiasis in mice leads to a more acute and lethal disease [17,18,19,20,22,54]. The understanding from such observations is that schistosome eggs release hepatotoxins, toxins that are normally prevented from diffusing by circumoval granulomata. To date, the only hepatotoxin characterized in S. mansoni eggs is omega-1 [18] which is RNaseT2 [28], that also induces a Th2 response [30].
We established an in vitro primary hepatocyte culture system using ALT as the metric for cell injury to identify egg components with direct hepatotoxicity. We first confirmed the toxicity of S. mansoni eggs and their derivatives, ESP and SEA, and then showed that pure native omega-1 is hepatotoxic in vitro, consistent with previous in vivo observations. Based on the present system, omega-1 is a major toxin released by S. mansoni eggs, as depletion of ESP or SEA with a specific monoclonal antibody decreased ALT levels by 47 and 33%, respectively.
To search for additional hepatotoxins, we combined anion exchange chromatography of ESP with proteomics. A single hepatotoxic fraction (#11) was identified which contained a short list of nine proteins. IPSE/alpha-1 stood out as a molecule of interest given its potent immunomodulatory properties [39,55]. Subsequent characterization of pure rIPSE/alpha-1 demonstrated that the molecule is indeed directly hepatotoxic. The finding was confirmed using a specific monoclonal antibody that essentially neutralized IPSE/alpha-1 toxicity while decreasing the cell injury produced by both ESP and SEA by approximately one-third. Further depletion of ESP and SEA with a combination of monoclonal antibodies targeting both omega-1 and IPSE/alpha-1 indicated that approximately 60% of the toxicity of the egg-derived material is due to these two proteins. This leaves room for additional hepatotoxins to be identified, perhaps by different chemical and physical separation approaches. We also note that although both omega-1 and IPSE/alpha-1 were identified in the total ESP proteome, only IPSE/alpha-1 was subsequently found in the single hepatotoxic fraction #11. This suggests that omega-1 was below the mass spectrometry detection limits used to identify proteins.
Recently, IPSE/alpha-1 was reported to be internalized by Chinese hamster ovary cells (CHO) and primary monocyte-derived dendritic cells, but not by peripheral blood basophils [40]; and in each case without apparent toxicity. This suggests that host cell-specific factors determine how cells interact with and respond to IPSE/alpha-1. Such factors might explain why IPSE/alpha-1 is directly toxic to hepatocytes. Studies to understand the mechanism of hepatoxicity induced by IPSE/alpha-1, and other hepatotoxins such as omega-1, can now be undertaken with the present in vitro system. Ribonuclease activity is often associated with cytotoxicity, and Steinfelder et al noted that omega-1 was initially characterized as a hepatotoxic agent from S. mansoni [28]. Nevertheless, the Th2-promoting activity of omega-1 cannot be explained by a cytotoxic effect, as the molecule failed to induce a detectable reduction in dendritic cell viability. The exact mechanism(s) by which the ribonuclease activity of omega-1 may promote Th2 responses is currently under investigation [31].
The results presented here underscore the paradox of the granulomatous response in hepatic schistosomiasis. Though detrimental to the host in the longer term due to its contribution to disease sequelae such as portal hypertension, it nevertheless protects against more acute hepatocyte injury resulting from toxins released by the schistosome egg. Figure S1 SDS-PAGE preparation of ESP. ESP (20 mg) was loaded into SDS-PAGE, the gel was silver stained, then sliced into 40 bands as indicated for in gel trypsin digestion and peptide sequencing by LC-MS/MS. The picture shows Fluorescence image for SyproRuby stained gel. (DOC) Figure 5. Omega-1 and IPSE are major hepatotoxins in ESP and SEA. Hepatocytes were co-incubated with 10 mg/ml ESP or SEA, or with ESP or SEA first depleted of both omega-1 and IPSE with a mixture of anti-IPSE and anti-omega-1 monoclonal antibodies (each 5 mg/ml). After 72 h, the presence of ALT in the medium was measured. A combination of 5 mM D-galactosamine hydrochloride (D-gal) and 1 mg/ml rTNF-a (Dgal/TNF-a) was employed as a known hepatotoxic control and negative control cultures used PBS. Depletion of both IPSE and omega-1 diminished the toxicity of ESP and SEA by 60% and 58%, respectively. Data are displayed as the mean 6 S.D. from two experiments each in duplicate. * P,0.05 for values significantly different from ESP and SEA is based on paired analysis (one sided paired Student's t-test). doi:10.1371/journal.pntd.0001368.g005 IPSE-alpha-1 Is a Major Hepatotoxin www.plosntds.org