Proteomics study of the antifibrotic effects of α-mangostin in a rat model of renal fibrosis

All procedures involving animals were approved by the Animal Ethics Committee of the Faculty of Medicine, Srinakharinwirot University (Protocol No. U1-01229-2558) and conducted in compliance with The Animals for Scientific Purposes Act, BE 2558 (AD 2015) (Thai Government Gazette, Vol. 132, Part 18 a, 13th March 2015) under a license to use animals for scientific purposes from the Institute of Animals for Scientific Purpose Development (IAD) National Research Council of Thailand following the revised Institute of Laboratory Animal Resources, Commission on Life Sciences, National Research Council “Guide for the Care and Use of Laboratory Animals” Washington, D.C.: National Academy Press; 1996. Male Wistar rats (n = 18, range 180–210 g, 5–8 weeks old) were obtained from the National Laboratory Animal Center, Mahidol University, Thailand. The rats were maintained in an air-conditioned animal facility at 22–25°C under a 12-h light and dark cycle and allowed food and water ad libitum. We obtained 96% pure AM (by high-performance liquid chromatography [HPLC] assay) from Associate Professor Primchanien Mongkarndi, Faculty of Pharmacy, Mahidol University, Thailand [14].

The rats were divided without selection into 3 groups of 6 as follows:

Group I: Vehicle control group. Rats in this group were administered 0.8 mL/kg of dimethyl sulfoxide (DMSO) mixed with phosphate-buffered saline (PBS) by intraperitoneal injection twice a week for 8 weeks.

Group II: TAA-treated group (TAA group). Rats in this group were given 200 mg/kg of TAA dissolved in sterile water 3 times a week via intraperitoneal injection for 8 weeks.

Group III: TAA + AM group. Rats in this group were given 200 mg/kg of TAA as for rats in group II, but were also administered 5 mg/kg of AM dissolved in 0.8 mL/kg of DMSO mixed with PBS twice a week for 8 weeks.

Rats were killed humanely using an overdose of pentobarbital (Nembutal) anesthesia 8 weeks after the first treatment, and kidneys were collected. One kidney from each rat was fixed in 4% paraformaldehyde at 4°C for histology, and the second kidney was stored at –80°C for proteomic analysis.

After fixing the kidney tissues with 4% paraformaldehyde, the tissues were dehydrated in an ascending graded series of ethanol and embedded in paraffin. The tissue blocks were sectioned at 5–7 mm before staining with Sirius Red to assess fibrosis [15]. ImageJ software (http://imagej.nih.gov/ij/docs/index.html) version 1.51r, was used to quantify the relative area of stained collagen fibers [16]. The area quantified contained at least one glomerulus surrounded by renal tubules of the renal cortex; 20 images were captured per group of rats. The quantification comprised 4 main steps repeated for all images from each group: (1) setting an appropriate scale to ensure correct calculation of area; (2) conversion of the stained image to gray scale; (3) isolation of Sirius Red-stained collagen using thresholding; (4) measuring the thresholded area [16]. The relative fibrotic areas were assessed using a one-way analysis of variance (ANOVA) to determine whether the group means differed significantly (P < 0.01) followed by Tukey’s multiple comparison test to compare the differences between each pair of means.

The kidneys from each group were pooled before proteomic analysis. After homogenizing the tissue in liquid nitrogen and dissolving the homogenate in 0.5% sodium dodecyl sulfate (SDS), the proteins were precipitated with acetone at –20°C. The precipitate was centrifuged at 12000 rpm (11700 ×g) for 15 min, and the pellets resuspended in 0.5% SDS. The solutions were then sonicated in an ultrasonic bath and centrifuged at 5000 rpm (2000 ×g) for 5 min to collect the supernatants. The concentration of proteins was determined using a Lowry assay at 690 nm [17].

The proteins were then separated by SDS-polyacrylamide gel electrophoresis (PAGE; 12.5% w/v) at 50 V for the stacking gel and 70 V once the proteins had entered the separating gel, for a total of 4 h. The height of the separating gel is 8 cm from the lower glass border on an Atto AE-6540 apparatus (Japan). The separated proteins were silver stained to visualize the bands [18]. The lane for each group was excised from the gel and separated into 15 smaller pieces (from top to bottom of the lane) that were placed into individual wells of a microtiter plate to digest the proteins in-gel.

Each piece of gel was washed twice with 200 µL of distilled water for 10 min each time, and the solution was then discarded. The washed gel pieces were then soaked in 200 µL acetonitrile for 5 min. After removing acetonitrile, the gel pieces were soaked in 30 µL of 10 mM dithiothreitol (DTT) for 1 h and then incubated in the dark with 100 mM iodoacetamide for 1 h. The solution of DTT and iodoacetamide was then discarded. We then added 10 µL of sequencing grade modified trypsin (10 ng/µL) (Promega) and 20 µL of 30% acetonitrile and incubated the preparation overnight at 37°C. The resulting peptide solutions were extracted with 30 µL of 50% acetonitrile/0.1% formic acid. The extracted peptides were placed into new microtiter plate and incubated at 40°C for 1 day. Finally, the samples were dissolved in 15 µL 0.1% formic acid and examined by liquid chromatography with tandem mass spectrometry (LC-MS/MS) [19].

The peptides extracted from each piece of gel were injected into an UltiMate 3000 LC System (Dionex) equipped with a reversed-phase nanocolumn and the eluted peptides were detected with an HCTultra PTM Discovery System (Bruker Daltonics). The nanocolumn (PepSwift monolithic column 100 µm i.d. × 50 mm). was eluted with a gradient of eluent A (0.1% formic acid) in eluent B (80% acetonitrile in water containing 0.1% formic acid). Peptides were separated with a linear gradient of from 10% to 70% eluent B for 13 min at 300 nL/min, including a column regeneration step at 90% eluent B and an equilibration step at 10% eluent B, producing a run time of 20 min. Mass spectra of the peptide fragments were acquired in a data-dependent auto MS/MS mode with a range 300–1500 m/z using 3 averages, and up to 5 precursor ions selected from an MS scan 50–3000 m/z. LC-MS/MS sample injections were automated and made in triplicate.

To quantify proteins, we used DeCyderMS Differential Analysis software (GE Healthcare) [20]. LC-MS/MS raw data were converted, and the PepDetect module was used for automated peptide detection, charge state assignments, and quantification based on the peptide ion signal intensities in MS mode. The analyzed MS/MS data from DeCyderMS were submitted for searching against the National Center for Biotechnology Information (NCBI) database using Mascot software (version 2.3, Matrix Science). Database interrogation was as follows: taxonomy (Rattus norvegicus); enzyme (trypsin); variable modifications (carbamidomethyl, oxidation of methionine residues); mass values (monoisotopic); protein mass (unrestricted); peptide mass tolerance (1.2 Da); fragment mass tolerance (±0.6 Da), peptide charge state (1+, 2+, and 3+), and maximum missed cleavages (3).

Using a self-organizing tree algorithm (SOTA), Multi-Experiment Viewer (Mev) software (version 4.6.1) was used to determine the proteins whose level of expression followed the trend of the quantified relative area of collagen fibers. The proteins resulting from the Mev analysis, together with proteins and chemicals that were shown to be involved in fibrosis such as transforming growth factor-β (TGF-β), Smads, tissue inhibitors of metalloproteinases (TIMPs), tumor necrosis factor (TNF), folic acid, streptozotocin, and cyclosporine A, were then entered into a Search Tool for Interactions of Chemicals Network (STITCH 5.0; http://stitch.embl.de) to elucidate their cellular functions [21]. Proteins shown to be involved in pathways relevant to the onset of fibrosis, whose information was provided by the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis available as a part of STITCH 5.0, were reevaluated individually with STITCH 5.0 with a limit of 10 functional partners to elucidate their functions. A flow diagram showing the methodological steps of GeLC-MS/MS and bioinformatics analysis is shown in Figure 1.

Figure 1

The relative fibrotic areas were assessed using a one-way analysis of variance (ANOVA) to determine whether the group means differed significantly (P < 0.01) followed by a Tukey multiple comparison test to compare the differences between each pair of means. The PepMatch module of the DeCyder MS 2.0 differential analysis software allowed for mass spectra with an inadequate match to those on the database to be eliminated (t test and ANOVA, where P < 0.05). Mass spectra were compared across each group using a t test and ANOVA to assess their similarity.

Representative graphic output from ImageJ software showing the relative areas of red staining in a given frame across the 3 groups, namely the control group, TAA group, and TAA+AM group, is shown in Figure 2. The accumulation of collagen fibers in Bowman’s capsule and the interstitium was significant in the kidneys of rats in which fibrosis was induced by TAA, when compared with the accumulation in the control group (P < 0.001). Rats in the TAA + AM group had a reduced area of fibrosis (P < 0.001).

Figure 2

Upon interrogation of the NCBI database with mass spectra from the tandem analysis, we identified 2769 proteins, the majority of which were shared among all 3 groups. The Sirius Red staining (mean relative percentages) as shown in Figure 3 were then employed to compare the scores against the proteomic protein profiles (analyzed using Mev software) across different groups, and 68 proteins belonging to R. norvegicus were found to possess similar expression trends to the Sirius Red quantification as summarized in Table 1. These 68 proteins were then analyzed using STITCH 5.0, and 4 of the proteins were found extensively linked to 4 main pathways relevant to fibrosis (as drawn from the KEGG pathway database), namely the apoptotic, TNF-signaling, NF-κB-signaling, and MAPK signaling pathways. The 4 proteins of interest are CASP8 and FADD-like apoptosis regulator (Cflar), Ragulator complex protein LAMTOR3 (Lamtor3), mitogen-activated protein kinase kinase kinase 14 (Map3k14), and C-Jun-aminoterminal kinase-interacting protein 3 (Mapk8ip3), along with the pathways they are involved in as shown in Figures 4–7. The remaining proteins, although shown to have followed the trend of Sirius Red quantification based on Mev software analysis, were discarded due to their nonexistent or weak link to the process of fibrosis as suggested by STITCH 5.0.

Figure 3

Figure 4

In an attempt associate our various findings, we hypothesize that AM exhibits a wide range of effects against fibrosis such as attenuating upstream activators of c-FLIP transcription, such as TNF ligands, interleukins, chemokines, or tumor protein 63, and reducing levels of Lamtor3 to prevent facilitating the MAP kinase signaling involved in fibrotic diseases. Furthermore, AM may attenuate upstream inflammatory responses leading to an overall reduction of noncanonical NF-κB signaling, as suggested by the markedly reduced level of Map3k14 found in the present study. We also propose that AM may reduce signaling by attenuating harmful stimuli that activate the JNK pathway. AM treatment may result in slight alterations in many pathways as shown in the reduction in signal intensities of key actors in the aforementioned pathways leading to an improvement of the overall fibrotic condition as supported by the histology (Figure 8). This is congruent with the pathogenesis of fibrosis, which like many other complex conditions, is caused by a disturbance in not just one biological pathway, but rather as a manifestation of many small disturbances leading to the observed pathogenesis.

Figure 8