Differential inhibition of host cell cholesterol de novo ...

Parasitol Res DOI 10.1007/s00436-014-4092-5

ORIGINAL PAPER

Differential inhibition of host cell cholesterol de novo biosynthesis and processing abrogates Eimeria bovis intracellular development Penny H. Hamid & Jörg Hirzmann & Carlos Hermosilla & Anja Taubert

Received: 29 July 2014 / Accepted: 22 August 2014 # Springer-Verlag Berlin Heidelberg 2014

Abstract Eimeria bovis macromeront formation in bovine endothelial host cells is an energy- and nutrient-demanding process. Obligate intracellular replicating coccidians are generally considered as auxotrophic for cholesterol synthesis and scavenge cholesterol from the host cell by either enhancing the uptake of extracellular cholesterol sources or by upregulating the host cellular de novo biosynthesis. We here focused on the latter mechanism and analyzed the effects of several inhibitors targeting the host cellular mevalonate biosynthesis pathway and cholesterol processing. The following inhibitors were used: lovastatin, squalestatin, CI976 and C75 targeting HMGCoA reductase, squalene synthase, acyl-CoA:cholesterol acyltransferase, and fatty acid synthase, respectively. In summary, all inhibitors significantly interfered with E. bovis meront formation and merozoite production in a dose-dependent manner. Dose effect responses identified lovastatin as the most effective compound, followed by CI976, C75, and squalestatin, respectively. Overall, merozoite production was inhibited by 99.6, 99.7, 84.6, and 70.2 % via lovastatin (1 μM), CI976, C75, and squalestatin (all 5 μM) treatments, respectively. Concerning macromeront formation, both the rate and size of developing meronts were affected by inhibitor treatments. The effects were characterized by developmental arrest and meront degradation. In the case of CI976 treatment, we additionally observed detrimental effects on host cellular lipid droplet formation leading to meront developmental arrest irrespective of the time point of treatment onset. These analyses clearly indicate that successful E. bovis intracellular development strictly depends on the host cellular de novo biosynthesis of cholesterol and on the adequate subsequent processing thereof. P. H. Hamid (*) : J. Hirzmann : C. Hermosilla : A. Taubert Institute of Parasitology, Biomedical Research Centre Seltersberg, Justus Liebig University Giessen, Schubertstr 81, 35392 Giessen, Germany e-mail: [email protected]

Keywords Cholesterol metabolism . Eimeria . Coccidia . Statins . Mevalonate biosynthesis

Introduction Eimeria bovis is an obligate intracellular apicomplexan parasite which represents one of the most pathogenic species in cattle coccidiosis, causing severe hemorrhagic typhlocolitis in calves and high economic losses worldwide (Daugschies et al. 1998; Daugschies and Najdrowski 2005). In common with other relevant pathogenic eimerian species infecting ruminants, the life cycle of E. bovis includes the formation of macromeronts of up to 400 μm in size which develop in endothelial cells of the lacteals (Nyberg and Hammond 1965). This long-lasting process (14–18 days) is associated with a significant enlargement and reorganization of the host cell cytoskeleton (Hermosilla et al. 2008) and with the production of >120,000 merozoites I as offspring (Hammond et al. 1966; Hammond and Fayer 1968). In consequence, E. bovis multiplying in macromeronts have a high demand for molecules being involved in membrane biogenesis, such as cholesterol and fatty acids. However, E. bovis belongs to the coccidia, a subclass of protozoa which are generally considered as auxotrophic for cholesterol synthesis (Ehrenman et al. 2013; Wenk 2006). From other coccidian parasites, such as Toxoplasma gondii or Cryptosporidium parvum, it is known that they scavenge cholesterol from their host cells to successfully implement requirements for merozoite replication (Coppens et al. 2000; Ehrenman et al. 2013). In general, two main alternative pathways for cholesterol acquisition exist for host endothelial cells: Cholesterol may be synthesized de novo via the mevalonate biosynthesis pathway or be internalized in terms of LDL uptake from the cellular environment. However, the exploitation of these two pathways apparently differs with the infecting apicomplexan species and the host

Parasitol Res

cell type. By this means, optimal infections with the cholesterol auxotrophic coccidian parasite T. gondii depend on LDL uptake in Chinese hamster ovary (CHO) cells and seem independent of cholesterol de novo synthesis in this cell type (Coppens et al. 2000). Vice versa, in macrophages, T. gondii infections depend on the cholesterol de novo synthesis via the mevalonate biosynthesis pathway while they did not appear to subvert the LDL trafficking pathway (Nishikawa et al. 2011). In agreement with this biochemical requirement, T. gondii replication in macrophages was significantly inhibited by statins which interfere with an early time point of cholesterol de novo synthesis (Cortez et al. 2009; Nishikawa et al. 2011). In addition, Martins-Duarte et al. (2006) have reported on the antiproliferative effects of squalene synthase inhibitors in T. gondii-infected epithelial cells. Accordingly, Blader et al. (2001) showed several molecules being involved in the mevalonate biosynthesis to be transcriptionally upregulated in T. gondii-infected fibroblasts. In C. parvum-infected intestinal epithelial cells, cholesterol requirements were mainly fulfilled via parasite-triggered LDL uptake, but in addition, a modest contribution by de novo synthesis was estimated since treatments with squalestatin (determining the switch toward sterol biosynthesis within the mevalonate pathway) at least marginally influenced parasite proliferation (Ehrenman et al. 2013). In the case of the erythrocytic stages of the apicomplexan parasites Babesia divergens and Plasmodium falciparum, the latter of which was also proven as auxotrophic for cholesterol synthesis (Mbaya et al. 1990), blockage of the mevalonate biosynthesis via different statins also impeded parasite development (Grellier et al. 1994). It is noteworthy that in hepatic stages, Plasmodium spp. in principle scavenge both types of cholesterol, i.e., cholesterol that was internalized by LDL uptake and de novo synthesized one, but neither of these appeared to be necessary for successful parasite replication indicating a marginal need of the liver stages (Labaied et al. 2011). Recent transcriptomic data on E. bovis-infected primary host endothelial cells indicated an infection-triggered, simultaneous induction of both cholesterol acquisition mechanisms since receptors for LDL and molecules being involved in de novo synthesis [e.g., cytochrome P450, family 51 (CYP51), cholesterol 25-hydroxylase (CH25H), hydroxymethylglutaryl-CoA synthase 1 (HMGCS1), acyl-CoA thioesterase 7 (ACOT7), acetyl-CoA C-acetyltransferase 2 (ACAT2)] were found to be upregulated during first merogony (Taubert et al. 2010). However, detailed experiments estimating the effect of these mechanisms on successful parasite replication are missing so far. Excess cellular cholesterol is esterified by sterol O-acyltransferase (SOAT; alternatively named acyl-CoA cholesterolO-acyltransferase, ACAT) and stored in cytoplasmatic lipid droplets, which are metabolically active inclusions (reviewed in Sandager et al. 2002). Several authors (Coppens and Joiner 2003; Coppens et al. 2000; Ehrenman et al. 2013; Sonda et al.

2001) reported on lipid droplet-like inclusions in the cytoplasm of T. gondii and C. parvum intracellular stages. Interestingly, two parasite-specific SOAT-like enzymes have been described for T. gondii (Lige et al. 2013; Nishikawa et al. 2005) indicating the importance of cholesterol esterification and storage for this parasite. Indeed, treatments of T. gondiiinfected host cells with SOAT inhibitors resulted in detrimental effects on parasite replication confirming the key role of SOAT and cholesteryl esters for optimal intracellular parasite development (Sonda et al. 2001). To date, no data are available on the actual role of the de novo biosynthesis and subsequent processing of cholesterol for E. bovis. We here demonstrate that inhibitors targeting different steps of the mevalonate biosynthesis as well as cholesterol esterification and fatty acid synthesis exhibit detrimental effects on E. bovis proliferation as measured on the level of macromeront development and merozoite I production. These data clearly indicate that E. bovis meront I development in endothelial host cells significantly depends on host cellular cholesterol de novo biosynthesis, confirming previous results on changes in host cell gene transcription of mostly metabolism/biosynthesis-related molecules of lipids (Taubert et al. 2010).

Materials and methods Parasites The E. bovis strain H used in the present study was initially isolated from the field in Northern Germany (Fiege et al. 1992) and maintained by passages in parasite-free male Holstein Friesian calves according to Hermosilla et al. (1999). All animal procedures were performed according to the Justus Liebig University Animal Care Committee guidelines, approved by the Ethic Commission for Experimental Animal Studies of the State of Hesse (Regierungspräsidium Giessen, GI 18/10 No A37/2011) and in accordance to the current German Animal Protection Laws. For oocyst production, calves were infected orally at the age of 10 weeks with 3×105 sporulated E. bovis oocysts. Then, excreted oocysts were isolated from the feces beginning 18 days p.i. according to the method of Jackson (1964). The sporulation of oocysts was achieved by the incubation in a 2 % (w/v) potassium dichromate (Merck) solution at room temperature (RT). Sporulated oocysts were stored in this 2 % (w/v) potassium dichromate solution at 4 °C until further use. Sporozoites were excysted from sporulated oocysts as previously described by Hermosilla et al. (2002). Free sporozoites were washed three times in single strength phosphate-buffered solution (PBS) and counted in a Neubauer hemocytometer as described elsewhere (Hermosilla et al. 2002).

Parasitol Res

Host cells and Eimeria bovis host cell infections Primary bovine umbilical vein endothelial cells (BUVECs) were isolated from umbilical cord veins, according to Jaffe et al. (1973). Umbilical cords were collected under aseptic conditions from animals born by section caesaria and kept at 4 °C in 0.9 % HBSS-HEPES buffer (pH 7.4, Gibco) supplemented with 1 % penicillin (500 U/ml, Sigma-Aldrich) and streptomycin (50 μg/ml, Sigma-Aldrich) until use. For the isolation of endothelial cells, 0.025 % collagenase type II (Worthington Biochemical Corporation) suspended in Pucks solution containing 8 g/l NaCl, 0.4 g/l KCl, 0.012 g/l CaCl2, 0.154 g/l MgSO4·7H2O, 0.39 g/l NaH2PO4, 0.15 g/l KH2PO4, 1.1 g/l glucose was infused into the lumen of the ligated umbilical vein and incubated for 20 min at 37 °C in 5 % CO2 atmosphere. After gently massaging the umbilical vein, collagenase cell suspension was collected and supplemented with 1 ml fetal calf serum (FCS, Gibco) in order to inactivate the collagenase type II. After two washings (400×g, 10 min, 4 °C), cells were resuspended in complete endothelial cell growth medium (ECGM, PromoCell), plated in 25-cm2 tissue plastic culture flasks (Greiner), and kept at 37 °C in 5 % CO2 atmosphere. BUVECs were fed with modified ECGM (EGCM, PromoCell, diluted 0.3× in M199 medium, Sigma) 1 day after isolation and thereafter every 2–3 days. BUVEC cell layers were used for infection after 1–2 passages in vitro. BUVEC cell layers were cultured in 24-well formats and infected with 2×104 freshly excysted E. bovis sporozoites per well at 80–90 % confluency. Culture medium was changed 24 h after parasite infection and thereafter every third day.

Host cell toxicity assay To test for toxic effects of lovastatin (solved in aceton), squalestatin (syn. zaragozic acid; solved in ethanol), CI976 and C75 (solved in DMSO; all Sigma-Aldrich) on BUVEC viability, these compounds were applied at the following final concentrations (diluted in cell culture medium): 200, 100, 50, 25, 12.5, 6.25 μM. The effects of lovastatin were additionally tested in 1.58- and 3.125-μM concentrations for long-term cultivation. Therefore, lovastatin was continuously applied every third days for a period of 30 days of cell culture. Toxicity experiments revealed high mortality rates of BUVEC at compound concentrations ranging from 200 to 12.5 μM irrespective of the type of inhibitor. While a concentration of 6.25 μM of zaragozic acid, CI976, and C75 treatments induced little cell death in BUVEC, lovastatin concentrations had to be lowered to 1.58 μM to achieve comparable effects. Considering these results, the following maximum inhibitor concentrations were chosen for E. bovis-related inhibition experiments: 1 μM for lovastatin and 5 μM for zaragozic acid, CI976 and C75 treatments.

The effects of treatment were assessed by three measures: cell morphology (microscopic examination applying 40-fold magnification), cell viability (MTT test, Sigma-Aldrich), and cell numbers (cell counting). For MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (Sylvester 2011; van Meerloo et al. 2011)] tests, BUVEC cultivated in 96-well formats (Greiner) were incubated in 200 μl of the respective compound concentrations for 72 h (five wells for each compound concentration). Thereafter, 20 μl of MTT working solution (5 mg/ml, SigmaAldrich) was added, and samples were incubated for 4 h (37 °C, 5 % CO2 atmosphere). The medium was removed, and 150-μl acidic isopropanol (0.04 N HCl in isopropanol) was added. After 1 h of incubation (37 °C), the resulting formazan product was estimated via optical density (OD) measurements at 590 nm and reference filter 620-nm wavelength using VarioskanTM Flash Multimode Reader (Thermo Scientific). BUVEC treated with the solvent of the compounds only and nontreated BUVEC were used as negative controls. The cell numbers of treated and nontreated BUVEC were estimated after trypsinization by direct counting the cells using a Neubauer chamber after 7, 10, 13, 16, and 19 days of cultivation. Lovastatin, squalestatin, CI976, and C75 treatments of Eimeria bovis-infected host cells Based on the results of the host cell toxicity assays, we chose the following final concentrations of the compounds for inhibition assays: lovastatin, 1 μM; squalestatin, CI976, and C75, all 5 μM. Infected cells were treated from day 1 until 30 days p.i. In case of CI976, additional assays were performed starting the treatment with day 10 p.i. The effects of the inhibitors were estimated by three parameters throughout in vitro development of E. bovis macromeronts: (i) proportion of host cells carrying meronts, (ii) size of meronts (n=10), and (iii) production of merozoites. The infection rates were calculated as follows: infection rate (%)=(infected cells/total counted cells)×100 % and determined by microscopic examination in at least 10 randomly selected vision power fields. The size of meronts was estimated microscopically at different time points p.i. (days 10, 14, 18, 22, 26, 30 p. i) using an IX81 inverted microscope (Olympus) equipped with a software for size measurements (cellSens 1.7, Olympus). The rate of host cells carrying developing parasites was estimated microscopically by analyzing the numbers of sporozoite-infected cells on days 2 and 6 p.i. From day 10 p.i. onward, only cells carrying developing meronts were counted while cells with nondeveloping sporozoites were not incorporated into the calculation. The estimation of merozoite numbers was performed via a quantitative PCR (see “Quantitative Eimeria bovis microneme protein 4 (EbMIC4) real-time PCR assay”). Therefore, cell culture supernatants containing released merozoites were collected at

Parasitol Res

different time points after infection (20, 23, 26, 29, 30 days p.i.), washed in PBS (600×g, 15 min, RT) and pelleted (600×g, 15 min, RT). In addition, BUVEC layers after collection of free merozoites were trypsinized at 30 days p.i., washed with PBS (600×g, 15 min, RT), and pelleted (400×g, 10 min, RT). DNA isolation was performed by addition of 100-μl cell lysis buffer containing 0.32 M sucrose, 1 % Triton X-100, 0.01 M Tris-HCl (pH 7.5), and 5 mM MgCl2 (all Sigma-Aldrich) to the pelleted samples, followed by resuspension and subsequent centrifugation (9,520×g, 10 min). The supernatant was removed, and the pellet was washed once more with the cell lysis buffer. The merozoite pellet was resuspended in 100-μl 1× PCR buffer (PeqLab, Erlangen, Germany) plus 20 μl proteinase K (20 mg/ml Qiagen, Hilden, Germany) and incubated for 1 h at 56 °C. To inactivate the proteinase K, samples were heated at 95 °C for 10 min. Five microliters of each sample was used for PCR analysis. Quantitative Eimeria bovis microneme protein 4 (EbMIC4) real-time PCR assay For merozoite quantification, we established a real-time PCR assay based on the single-copy gene of E. bovis microneme protein 4 (EbMIC4: Lutz 2008). Primers and probes for EbMic4 were designed using the Beacon Designer software (Premier Biosoft): forward primer 5′-CACAGAAAGCAAAA GACA-3′, reverse primer 5′-GACCATTCTCCAAATTCC-3′, and probe 5′-FAM-CGCAGTCAGTCTTCTCCTTCC-BHQ13′. For the efficiency analysis of the EbMIC4-PCR assay, the amplicon was cloned in pDrive vector (Qiagen), and a dilution series of the recombinant plasmid linearized with restriction endonuclease BamHI (6,250,000-6 plasmid molecules) was used as target. PCR amplification was performed on a RotorGene Q cycler (Qiagen). All samples were analyzed in duplicate. For PCR, 5 μl DNA was used in a 20-μl PCR reaction mixture containing 10-μl PerfeCTa FastMix (Quanta, MD, USA), 400 μM of each primer, and 200 μM of probe. The PCR conditions were initial denaturation at 95 °C for 5 min followed by 45 cycles at 94 °C for 15 s, 60 °C for 45 s, and 72 °C for 20 s, with acquisition in green channel. In each quantitative PCR experiment, DNA from a dilution series of known merozoite numbers (16; 160; 1,600; 16,000; 160,000; 1,600,000) was included allowing for absolute quantification of merozoite counts in the samples. The standard curve was generated by plotting EbMIC4-Ct values against the logarithm of the number of merozoites Semiquantitative assessment of neutral lipid/lipid droplet deposition We here assessed the degree of lipid droplets/neutral lipid deposition during E. bovis in vitro macromeront formation

via Nile red (Cayman Chemical, USA) staining on a singlecell level. Since the inhibitor CI976 acts on cholesterol esterification and since cholesteryl esters are mainly stored in lipid droplets, we additionally investigated the effect of CI976 treatments on lipid droplets/neutral lipid deposition in E. bovis-infected cells. Therefore, BUVEC were grown to confluence on glass coverslips in the 12-well plates and infected with freshly excysted sporozoites (3×104 sporozoites/ sample). CI976 treatments (5 μM) were performed from day 1 or 10 p.i. onward. At different time points after infection (2, 6, 10, 14, and 18 days p.i.), the cells were washed twice with PBS, fixed (4 % paraformaldehyde, 15 min), and stained with Nile red fluorescence dye (Cayman Chemical, USA). Nile red suspended in DMSO (Cayman Chemical) was diluted 1:1,000 in PBS for staining (for 10–15 min, RT). After two washings in PBS, the samples were mounted in aqueous mounting medium (PBS). Fluorescence intensities on single-cell level (n=10) were analyzed using an IX81 fluorescence microscope (Olympus). Corrected single-cell fluorescence intensities were calculated using the software NIH’s ImageJ (Burgess et al. 2010; Gavet and Pines 2010) as follows: corrected total cell fluorescence (CTCF)=integrated intensity–(area of selected cell×mean fluorescence of background readings). Statistical analysis To determine median inhibition of the compounds, a nonlinear regression was performed using the software GraphPad Prism 6.02 to generate sigmoidal model of dose-response curve based on a four-parameter fit (Motulsky and Christopoulos 2003). The relative inhibition was calculated as a response of treatment as follows: (mean value of control value of test sample)/(mean value of control) represented in percent compared to control (see Ehrenman et al. 2013; Labaied et al. 2011). Statistical analysis was performed by using paired Student’s t tests to compare treated and nontreated groups, while multiple variable was tested using one-way ANOVA.

Results EbMIC4 qPCR Titration assays of merozoite DNA and EbMIC4 plasmid DNA revealed comparable efficiencies of 0.93–1.03 and 0.97–0.99 with a correlation value (r2) of 0.98–0.99 and 0.99–1.11, respectively. The sensitivity of EbMIC4 qPCR by using merozoite dilution series was two EbMIC4 molecules (corresponding to two merozoites since EbMIC4 is a single-copy gene) in a 20-μl PCR total reaction volume applying 5 μl of sample.

Parasitol Res

Host cellular de novo synthesis via the mevalonate biosynthesis pathway was blocked on two different levels. While

lovastatin inhibits cholesterol synthesis with HMG-CoAreductase at an early step of this pathway affecting the total cellular isoprenoid/steroid synthesis, squalestatin treatments block cholesterol synthesis more specific by affecting the squalene synthase step. Lovastatin treatments exhibited significant effects on both the rate and size of developing macromeronts (Fig. 1a). As such, the rate of cells carrying developing macromeronts was significantly lower in treated cultures compared to nontreated ones at most time points tested (14 days p.i. p