Integrative â-Omicsâ Analysis in Primary Human Hepatocytes Unravels

Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)

Article

Integrative ‘-omics’ analysis in primary human hepatocytes unravels persistent mechanisms of cyclosporine A-induced cholestasis Jarno E.J. Wolters, Marcel HM van Herwijnen, Daniel H.J. Theunissen, Danyel G.J. Jennen, Wim Fransiscus Petrus Maria Van den Hof, Theo M.C.M. de Kok, Frank G. Schaap, Simone G.J. van Breda, and Jos C.S. Kleinjans Chem. Res. Toxicol., Just Accepted Manuscript • DOI: 10.1021/acs.chemrestox.6b00337 • Publication Date (Web): 21 Nov 2016 Downloaded from http://pubs.acs.org on November 28, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Chemical Research in Toxicology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Chemical Research in Toxicology

Integrative ‘-omics’ analysis in primary human hepatocytes unravels persistent mechanisms of cyclosporine A-induced cholestasis

Jarno E.J. Wolters† *, Marcel H.M. van Herwijnen†, Daniel H.J. Theunissen†, Danyel G.J. Jennen†, Wim Fransiscus Petrus Maria Van den Hof†, Theo M.C.M. de Kok†, Frank G. Schaap, Simone G.J. van Breda†, Jos C.S. Kleinjans†

†

Department of Toxicogenomics, GROW School for Oncology and Developmental Biology,

Maastricht University, P.O. Box 616, 6200 MD, Maastricht, The Netherlands ‡

Department of Surgery, Maastricht University, 6200 MD, Maastricht, The Netherlands

*To whom correspondence should be addressed at the Department of Toxicogenomics, Faculty of Health, Medicine and Life Sciences, Maastricht University, Maastricht, P.O. Box 616, 6200 MD, The Netherlands. Tel.: +31 43 3881090; Fax: +31 43 3884146 E-mail address: [email protected] (J. Wolters)

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 47

Table of Contents Graphic


Page 3 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Abstract Cyclosporine A (CsA), is an undecapeptide with strong immunosuppressant activities and is used a lot after organ transplantation. Furthermore, it may induce cholestasis in the liver. In general, the drug-induced cholestasis (DIC) pathway includes genes involved in the uptake, synthesis, conjugation and secretion of bile acids. However, whether CsAinduced changes in the cholestasis pathway in vitro are persistent for repeated dose toxicity has not yet been investigated. To explore this, primary human hepatocytes (PHH) were exposed to a sub-cytotoxic dose of 30 µM CsA daily for 3 and 5 days. To investigate the persistence of induced changes upon terminating the CsA exposure after 5 days, a subset of PHH was subjected to a washout period (WO-period) of 3 days. Multiple -omics analyses, comprising whole genome analysis of DNA methylation, gene expression and microRNA expression, were performed. The CsA-treatment resulted after 3 and 5 days, respectively in 476 and 20 differentially methylated genes (DMGs), 1353 and 1481 differentially expressed genes (DEGs), and in 22 and 29 differentially expressed microRNAs (DE-miRs). Cholestasis-related pathways appeared induced during CsA-treatment. Interestingly, 828 persistent DEGs and 6 persistent DE-miRs, but no persistent DMGs, were found after the WO-period. These persistent DEGs and DE-miRs showed concordance for 22 genes. Furthermore, 29 persistent DEGs changed into the same direction as observed in livers from cholestasis patients. None of those 29 DEGs which amongst others relate to oxidative stress and lipid metabolism are yet present in the DIC pathway or cholestasis adverse outcome pathway (AOP) thus presenting novel findings. In summary, we have demonstrated for the first time a persistent impact of



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 47

repeated dose administration of CsA on genes and microRNAs related to DIC in the gold standard human liver in vitro model with PHH.

Keywords Cholestasis; Primary Human Hepatocytes; multiple-omics techniques; Cyclosporine A


Page 5 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1. Introduction The liver is the most important organ for the metabolism and excretion of drugs and therefore the liver is vulnerable to adverse reactions known as drug-induced liver injury (DILI). Cholestasis is one of the most severe manifestations of DILI resulting in a high mortality rate and it may account for 10% to 50% of all DILI cases.1-3 Cholestasis is characterized by the accumulation of substances normally excreted via the bile (e.g. cholesterol, bile acids, and drug metabolites) in the liver or in the bile ducts due to inhibition of canalicular transporter function or due to impaired bile flow.2,

4

The

subsequent toxicity, resulting from the intracellular accumulation of bile acids, is induced by the disruption of mitochondrial function which may cause apoptosis or necrosis.5, 6 One of the most important transporter proteins for canalicular secretion of conjugated bile acids in hepatocytes is the Bile Salt Export Pump (BSEP), an ATP-dependent canalicular transporter.2 Vinken et al. (2013) published an Adverse Outcome Pathway (AOP) describing the onset of cholestasis due to the accumulation of bile acids resulting from the inhibition of the ATP binding cassette subfamily B member 11 (ABCB11 (BSEP)).7 In general, the bile acid accumulation will induce an adaptive response, which will decrease the uptake of bile acids into the hepatocytes and increase the export of bile acids from the basolateral side.7, 8 However, not all cholestatic compounds inhibit the BSEP, which indicates that the molecular mechanisms underlying drug-induced cholestasis (DIC) still needs further investigation.9 More specifically, it has been shown that inhibition of the basolateral efflux pump multidrug resistance-associated protein 4 (MRP4; ABCC4), next to BSEP inhibition is associated with DIC.10, 11 In addition, Yuryev et al. (2009) and Vinken et al. (2013)7 published a DIC pathway which includes the nuclear receptors



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 47

Farnesoid X Receptor (FXR (NR1H4)), Pregnane X Receptor (PXR (NR1I2)), and Constitutive Androstane Receptor (CAR (NR1I3)) which act as sensors of intrahepatocyte bile acids.7, 12-17 More interestingly, there is no consensus related to the roles of the nuclear receptor Vitamin D Receptor (VDR), and the genes coding for the conjugation enzyme bile acid-CoA:amino acid N-acyltransferase (BAAT) and the efflux pumps ABCC4, and ATP binding cassette subfamily B member 1 (ABCB1).7, 12 Microarray-based gene expression is a well-established technique which may generate a better insight into the molecular mechanisms of DIC.18, 19 For instance, Van den Hof et al. (2015) showed that an integrated analysis of transcriptomics, microRNA profiling and metabonomics data from cyclosporine A-treated HepG2 cells generated novel insights into response pathways.20 Cyclosporine A (CsA) is considered a prototypical DIC compound.

It is a

neutral highly lipophilic cyclic peptide and a potent

immunosuppressant which has contributed significantly to improvements in organ transplantation but also causes various toxic effects in the kidney and the liver.21 CsA inhibits biliary bile acid secretion while also an inhibition of the secretion of biliary components such as phospholipids, cholesterol and bilirubin has been reported.21-23 CsA, a multidrug resistance-1 protein (MDR1) substrate, may cause cholestatic liver injury through a number of mechanisms,2, 9 namely: I) competitive inhibition of ATP-dependent transporters,24-26 II) inhibition of intrahepatic vesicle transport and targeting of ATPdependent transporters to the canalicular membrane,23, 27, 28 and III) impairment of bile formation partly by increasing canalicular membrane fluidity without affecting the expression of canalicular transporters.21 Other studies suggest that CsA reduces the expression of glutathione synthesizing enzymes and the canalicular glutathione efflux


Page 7 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


system, multidrug resistance-associated protein 2 (MRP2), leading to reduced bile acidindependent bile flow.2 While transcriptomics and microRNA profiling may contribute to further elucidating CsA-related DIC mechanism, the epigenome has not yet been investigated in this context. Interestingly, Rieswijk et al. (2016) have recently, and for the first time, shown that epigenomic effects induced by repeated treatment with the liver carcinogen aflatoxin B1 in vitro were still present after terminating exposure.29 So here, we have undertaken to evaluate the time-dependent contribution of epigenetic changes and microRNA expression modulation to investigate the mechanisms underlying cholestasis development in primary human hepatocytes (PHH) after repeated CsA-treatment. In order to evaluate our hypothesis that persistent epigenomic changes substantiate to CsA-induced cholestasis we particularly focused on cross-omics analysis upon terminating the CsA exposure. For this, we performed whole genome analysis of DNA methylation (using NimbleGen 2.1 deluxe promoter arrays), gene expression (using Affymetrix whole genome gene expression microarrays) and microRNA expression (using Agilent microRNA microarrays) of cell lysates from PHH which were exposed to a non-cytotoxic dose of 30 µM of CsA daily for 5 days. To investigate the persistence of induced changes upon terminating the CsA exposure, a subset of PHH was maintained during a washout period (WO-period) of three days. Detection of differentially methylated genes (DMGs), differentially expressed genes (DEGs), and differentially expressed microRNAs (DEmiRs) was achieved by using several R packages included in the open source software framework BioConductor.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 47

2. Experimental Procedures 2.1 Cell culture and cyclosporine A (CsA) treatment Cryopreserved primary human hepatocytes (PHH) and culture media were purchased from Life Technologies (Bleiswijk, The Netherlands), unless otherwise stated. In order to reduce the influence of inter-donor variability, PHH from 3 donors (lots Hu8119, Hu1591 and Hu1540) with different age, sex, ethnicity, body mass index, and behavioral status concerning alcohol and drug use, and smoking, were pooled. PHH were cultured in precoated multi-well plates in a two-layer collagen sandwich, according to the supplier’s protocol. In short, PHH were thawed for 1 min. at 37°C in a water bath. Three vials were poured in 50 ml Thawing Medium (CHRM CM7000) and centrifuged for 10 min. at 100g at 4°C. After removing the supernatant, the cell pellet was dissolved in CHPM plating medium (CHPM, CM9000), which contain FBS, at a concentration of 1*10E6 cells/ml. After seeding, PHH were incubated in the pre-coated multi-well plates for 4 hours at 37°C. After the incubation, debris was removed by shaking and washing the cells twice with Williams’ Medium E medium. Subsequently, PHH were covered by 150 µL/well with a collagen layer (containing 10*DMEM, 0.2M NaOH, and 1.0mg/mL rat tail Collagen type 1 (BD biosciences, Breda, The Netherlands)) and incubated for 30 min. at 37°C. Finally, culture medium (WME + 20 mL Hepatocyte Supplement Pack (CM4000) substituted with 1% penicillin/streptomycin (Gibco)) was added. For RNA and DNA, 2.1*10E6 cells were seeded in 6 wells of a 24-wells plate (0.35*10E6 cells per well of a 24-wells plate). Before treatment, cells were allowed to acclimatize for 3 days in order to restore an in vivo-like cellular configuration and expression of CYP enzymes. One sub-cytotoxic


Page 9 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


concentration of 30 µM CsA was chosen after a screening experiment examining a dose range from 0 – 60 µM of CsA (using 0.5% DMSO as solvent) in which PHH showed weak cytotoxicity in the MTT assay after daily treatment with 30 µM CsA for 5 days. Crucial liver enzymes such as lactate dehydrogenase (LDH), alanine transaminase (ALT), and aspartate aminotransferase (AST) were also measured after exposure to 0 – 60 µM to verify that toxicity was actually induced in liver cells with a treatment of 30 µM CsA (data not shown). For the main experiment, PHH were exposed to 30 µM of CsA or 0.5% DMSO (solvent control) in a 24 hour repeat-dose testing regime. Medium was changed daily thereby providing a new dose of CsA or DMSO to the PHH. A bi-phasic treatment regimen was applied, combining a 5-day CsA exposure with a subsequent 3-days WOperiod during which the PHH were exposed to medium only. Cell lysates were collected on days 3, 5 and 8, and prepared for DNA methylation, transcriptomic and microRNA analysis. Experiments were carried out in triplicate.

2.2 Methylated DNA immunoprecipitation (MeDIP) chip analysis MeDIP chip analysis was performed according to Deferme et al.,30 Rieswijk et al.,29 and van Breda et al.,31 with minor adjustments. qPCR validation of the MeDIP was successful in all three the studies.29-31

2.2.1 DNA isolation PHH were lysed in 500 µl of digestion buffer (containing 1 mM EDTA; 50 mM Tris–HCl, pH 8.0; 0.5% SDS). Next, 25 µl of proteinase K (20 mg/ml) (Ambion) was added. After incubation of 1 hour at 55°C, the proteinase K was inactivated for 10 minutes at 80°C.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 47

RNAse A (2 µl; 100 mg/ml) (Qiagen) and collagenase (25 µl; 1%) (Sigma) treatment was performed for 1 hour at 37°C. Thereupon, 500 µl of phenol-chloroform-isoamylalcohol (PCI; 25:24:1) (Sigma) was added. The mixture was shaken manually for 5 minutes, and centrifuged for 5 minutes at maximum speed. The upper phase was transferred to a new Eppendorf and the step with PCI was repeated. The upper phase was collected and precipitated using 50 µl of 3M NaAc pH 5.6 and 1250 µl of cold 100% ethanol for 30 min. at -80°C. After centrifugation for 30 min. at maximum speed, the DNA pellet was washed using cold 70% EtOH, dried in a speed-vac and dissolved in 50µl of nuclease free water. The total amount was at least 8 µg DNA, the A260/A280 ratio ranged between 1.7-1.9, and the A260/A230 ratio was higher than 1.6. A total of 18 DNA samples were prepared. DNA was used for MeDIP-Chip analyses.

2.2.2 Methylated DNA immunoprecipitation (MeDIP), whole-genome amplification and methylation enrichment assessment Genomic DNA was sonicated to obtain fragments ranging from 200 bp to 600 bp, cleaned up using silica columns (Zymo Research) and eluted in TE buffer. MeDIP was performed using the MagMeDIP kit (Diagenode, Liège, Belgium) according to the manufacturer’s protocol as was described before.31 Quantitative PCR (qPCR) was used for controlling DNA methylation enrichment. qPCR was performed by measuring the Ct-values of 1 µL of purified DNA sample and 24 µL of qPCR mixture (1 µL of provided primer pair (reverse and forward), 12.5 µL of SYBR Green PCR master mix and 10.5 µL water) using the temperature profile: 95°C for 7 minutes, 40 cycles of 95°C for 15 sec. and 60°C for 1 minute, followed by 1 minute 95°C.


Page 11 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The efficiency of MeDIP was calculated by performing qPCR and using the following formula: %(meDNA-IP/Total input) = 2^[(Ct(10%input)-3.32) – Ct(meDNA-IP)] x 100%.

2.2.3 MeDIP-Chip For analysis of DNA methylation levels, the Human DNA Methylation 2.1M Deluxe Promoter Array (Roche NimbleGen) was used. This array has a density of 2.1 million probes (50-75 oligonucleotides long, median probe spacing 100 bp) that represent all annotated human promoters (~26,210), 27,867 CpG islands, and 750 microRNA promoters per slide. The promoters and CpG islands largely overlap with transcription start sites of well characterized RefSeq genes, covering, on average, 8 kb upstream and 3 kb downstream. Labeling and hybridization of arrays was performed according to the manufactures’ protocol. In brief, 1 µg of input DNA and 1 µg of MeDIP DNA were labeled with Cy3 and Cy5 respectively by random priming using the Dual Color DNA labeling kit (Roche NimbleGen). The labeled samples were precipitated using isopropanol and quantified spectrophotometrically using the NanoDrop 1000 (Thermo Scientific). Thirty-four microgram of Cy3 Input and 34 µg of Cy5 MeDIP DNA were pooled and completely dried by using a speed-vac. The pellet was dissolved in hybridization solution using the NimbleGen Hybridization kit. After denaturing, the probe was hybridized overnight on the 2.1M Deluxe Promoter Arrays using the HX1 mixers and the NimbleGen Hybridization system 4. Slides were washed using the NimbleGen wash buffer kit and scanned using the 2 µm high resolution NimbleGen MS 200 micro array scanner.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 47

2.2.4 DNA methylation data analysis Signal intensity data was extracted from the scanned images of each array using NimbleScan v2.6 software and quantile normalized on a per channel basis. Log2 ratios of the intensities were computed (ratio of MeDIP signal / Input signal) and for each array, centering was performed by subtracting the global array bi-weight mean of the log2 ratios such that the computed log2 ratios were centered around 0. Detection of differential methylation was performed using the Probe Sliding Window-ANOVA algorithm (PSWANOVA). PSW-ANOVA was implemented in the R statistical programming environment (v2.15.3) (http://www.r-project.org) as a custom script and was provided by Roche NimbleGen as described before.32 PSW-ANOVA (sliding window of 750 bp comprising 7 probes, and a False Discovery Rate (FDR) corrected p- value < 0.01) was used to identify differential methylated regions (DMR) which were statistically significantly different between the different conditions tested in the experiment i.e. exposed versus control. Peaks were identified in the DMR by searching for regions containing at least 8 significant consecutive probes (p 0 indicate hyper-methylation and log2 ratios < 0 indicate hypo-methylation.


Page 13 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The DNA methylation microarray data have been deposited in NCBI's Gene Expression Omnibus (GEO),34 and are accessible through accession number GSE84276 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE84276).

2.3 Transcriptomics analysis Transcriptomic analysis was performed according to Deferme et al. (gene expression), Rieswijk et al. (gene and microRNA expression), and van Breda et al. (gene expression), with minor adjustments.29-31

2.3.1 Total RNA isolation At the end of the CsA-treatment, medium was removed and PHH were harvested in Qiazol (Qiagen). Total RNA was isolated using a miRNeasy Mini Kit (Qiagen) according to the manufacturer’s protocol and followed by DNaseI (Qiagen) treatment. Upon purification, RNA concentrations were measured by means of a NanoDrop 1000 spectrophotometer (Thermo Scientific) at 260 and 280 nm. RNA quality and integrity were assessed by using automated gel electrophoresis on an Agilent 2100 Bioanalyzer system (Agilent Technologies). Only RNA samples which showed clear 18S and 28S peaks and demonstrating a RNA-integrity number higher than 6, were used. Samples were stored at -80ºC until RNA hybridization.

2.3.2 Gene expression microarrays High-density oligonucleotide GeneChips from Affymetrix were used to measure gene expression levels (Human Genome U133 Plus 2.0 array (604,258 probes)). Targets for



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 47

these arrays were prepared, hybridized and scanned according to the Affymetrix protocol (Affymetrix). Normalization quality controls appeared to be within acceptable limits for all chips. The Arrayanalysis.org web service was used for quality control and all microarray data were of high quality.35 CEL files were imported into R v2.15.3 (http://www.r-project.org) using the “affy” library within BioConductor (v2.9).36,

37

.

Probe re-annotation, normalization and data filtering was performed as previously described.31 The mRNA microarray data have been deposited in NCBI's GEO,34 and are accessible through

accession

number

GSE83958

(http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE83958).

2.3.3 Selection of differentially expressed genes DEGs were selected using the linear model for microarrays (LIMMA) approach.38 The resulting p-values were FWER-corrected using the FDR method. The following criteria were applied: (1) average expression in at least one of the experimental groups > 6 (log2scale), (2) a FDR-corrected p-value < 0.05 obtained through a moderated t-test, and (3) for the three replicates an average absolute fold change (FC) of 1.5 or higher (i.e., average log2 ratio of < −0.58 or > 0.58).

2.3.4 MicroRNA microarrays and selection of differentially expressed microRNAs MicroRNA expression profiling was performed using Agilent Sureprint G3 Unrestricted Human miRNA V19 8 × 60 K microarrays. The hybridization was performed following standard protocols, after which the microarray slides were washed and scanned using a


Page 15 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


DNA microarray scanner (Agilent Technologies). The scanned images were converted into TXT files using the Feature Extraction Software v10.7.3.1 from Agilent Technologies, which were imported in R 2.15.3 (http://www.r-project.org) for quality control using arrayQC (https://github.com/BiGCAT-UM/arrayQC_Module/). Filtering and normalization was done per time-point and was performed using MagiCMicroRna,39 which applies a LIMMA approach for selection of DE-miRs. Total gene signals were log2-transformed and quantile-normalized. DE-miRs with an absolute FC of 1.5 or higher (i.e., average log2 ratio of < −0.58 or > 0.58) and an FDR-corrected p-value < 0.05 were considered statistically significant. The microRNA microarray data, have been deposited in NCBI's GEO,34 and are accessible

through

accession

number

GSE83954

(http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE83954).

2.3.5 Target information of DE-miRs To understand the biological function of DE-miRs, target information was retrieved using a database containing over 50,000 experimentally validated microRNA-mRNA interactions (miRTarBase – Homo sapiens, release 6.1, miRBase version 20). MiRTarBase (http://mirtarbase.mbc.nctu.edu.tw/).40

2.4 Integrated data analyses 2.4.1 Pathway analyses and network visualization of DEGs, DMGs and DE-miRs The lists of significantly modulated DMGs and DEGs were uploaded onto the web-tool ConsensuspathDB for human pathway over-representation analysis.41,

42

DEGs which



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 47

were present in the significantly modulated pathways (p-value < 0.01, > 2 genes per pathway) were extracted.

2.4.2 Mapping of DEGs, DMGs and microRNAs onto the Adverse Outcome Pathways of cholestasis Vinken et al. (2013) published an Adverse Outcome Pathway (AOP) describing the onset of cholestasis due to the accumulation of bile acids.7 Furthermore, Yuryev et al. (2009) published a DIC pathway which includes proteins involved in bile acid uptake, synthesis, conjugation and secretion.12 In vitro DNA methylation, gene expression, and microRNA expression changes were mapped onto this DIC pathway.

2.4.3 Benchmarking of DEGs, DMGs and microRNAs against a disease signature of cholestasis In order to benchmark our results against the in vivo situation in patients suffering from cholestasis,43 we established a cholestasis disease signature gene set using patient samples (microarray data of cholestatic liver samples and control liver samples) from previously published research. To our knowledge there is no in vivo transcriptomics dataset available from DIC patients, reflecting the difficulties in obtaining medical and ethical justification to procure liver specimens of subjects with DIC. Instead, samples of patients with extrahepatic cholestasis were analyzed to obtain a fingerprint of human cholestasis in vivo since both extrahepatic cholestasis as well CsA-induced cholestasis occur through bile flow impairment (reduced bile acid uptake and increased bile acid export). In short, the protocol of the study was approved by the local Medical Ethical


Page 17 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Committee and the patients gave their informed consent. Samples were derived from 9 perioperative liver biopsies from patients with a pancreatic or periampullary malignancy and subsequent obstructive jaundice (cholestatic group), and from 9 non-jaundiced patients with a pancreatic malignancy or undergoing liver resection (control group). Liver samples were collected in RNAlater and stored at -80°C until RNA isolation. Isolated RNA was hybridized to Agilent SurePrint G3 Human Gene Expression 8x60K v2 Microarrays scanned on an Agilent Technologies G2565CA DNA Microarray Scanner. Raw data on pixel intensities were extracted from the scan images using Agilent Feature Extraction Software. Quality control and data normalization were performed using ArrayQC, an in house developed pipeline in R (https://github.com/BiGCATUM/arrayQC_Module). The arrays were quantile normalized and replicate probes were summarized by taking the median. Probes that were present in 75% of all arrays were included for further processing. The remaining probes were reannotated to EntrezGeneIDs using the Agilent annotation file (date: 2012-06-28). A list of genes differentially expressed between cholestatic patients and controls, was generated using the BioConductor package LIMMA version 3.18.3.38 In brief, a moderated t-test was executed to find DEGs based on a combination of the following criteria: an average FC of 1.5 or higher (i.e. average log2 ratio of < -0.58 or > 0.58), and a P-value 0.05); ** : FDR < 0.05; NE : Not expressed.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 40 of 47

Tables Table 1 3-days daily exposure with 30 µM CsA

DMGs

Direction of effect *

DEGs

3-days washout period after 5days CsA exposure

5-days daily exposure with 30 µM CsA

DE-miRs

DMGs

DEGs

DE-miRs

DMGs

DEGs

DE-miRs

Magnitude; >0 or

Integrative â-Omicsâ Analysis in Primary Human Hepatocytes Unravels

Recommend Documents

Integrative â-Omicsâ Analysis in Primary Human Hepatocytes Unravels