World J Gastroenterol. 2011 June 14; 17(22): 2748-2773.
Published online 2011 June 14. doi: 10.3748/wjg.v17.i22..
©2011 Baishideng Publishing Group Co., Limited. All rights reserved.
miRNA studies in in vitro and in vivo activated hepatic stellate cells
Gunter Maubach, Michelle Chin Chia Lim, Lang Zhuo, Institute of Bioengineering and Nanotechnology, 31 Biopolis Way, The Nanos #04-01, Singapore 138669, Singapore
Gunter Maubach, Institute of Experimental Internal Medicine, Leipziger Strasse 44, Magdeburg 39120, Germany
Jinmiao Chen, Henry Yang, Bioinformatics Lab, Singapore Immunology Network, 8A Biomedical Grove, Singapore 138648, Singapore
Received May 19, 2010; Revised September 14, 2010; Accepted September 21, 2010;
AIM: To understand which and how different miRNAs are implicated in the process of hepatic stellate cell (HSC) activation.
METHODS: We used microarrays to examine the differential expression of miRNAs during in vitro activation of primary HSCs (pHSCs). The transcriptome changes upon stable transfection of rno-miR-146a into an HSC cell line were studied using cDNA microarrays. Selected differentially regulated miRNAs were investigated by quantitative real-time polymerase chain reaction during in vivo HSC activation. The effect of miRNA mimics and inhibitor on the in vitro activation of pHSCs was also evaluated.
RESULTS: We found that 16 miRNAs were upregulated and 26 were downregulated significantly in 10-d in vitro activated pHSCs in comparison to quiescent pHSCs. Overexpression of rno-miR-146a was characterized by marked upregulation of tissue inhibitor of metalloproteinase-3, which is implicated in the regulation of tumor necrosis factor-? activity. Differences in the regulation of selected miRNAs were observed comparing in vitro and in vivo HSC activation. Treatment with miR-26a and 29a mimics, and miR-214 inhibitor during in vitro activation of pHSCs induced significant downregulation of collagen type?I?transcription.
CONCLUSION: Our results emphasize the different regulation of miRNAs in in vitro and in vivo activated pHSCs. We also showed that miR-26a, 29a and 214 are involved in the regulation of collagen type I mRNA.
Keywords: Hepatic stellate cells, miRNA, miR-146a, Nuclear factor-?B
Liver fibrosis, characterized by an overproduction of extracellular matrix (ECM), is a common outcome of different chronic liver diseases[1
]. Hepatic stellate cells (HSCs) are one of the major cell types responsible for the production of ECM molecules like collagens, laminin, proteoglycans and fibronectin[2
]. The production of different ECM molecules is increased upon transdifferentiation (activation) of HSCs from a quiescent to an activated myofibroblast-like state[3,4
]. Consequently, the regulation of the complex process of HSC activation is of great interest to the research community. Understanding this process should lead to the discovery of therapeutic strategies for liver fibrosis. Due to the complexity of the activation of HSCs, the number of regulatory steps is expected to be overwhelming[5
], and requires addressing many different targets at the same time, either with different compounds or with one compound that is able to work on many different targets.
miRNAs are small approximately 23-nt non-coding RNAs, which are able to regulate hundreds of different proteins. The versatility of miRNAs is attributed to the imperfect binding (seed region) to the 3’-UTR of mRNAs, which results in, contrary to siRNA, many binding partners. The regulation by miRNAs is also different to siRNAs because it leads to a translational repression and/or mRNA destabilization[6,7
]. That miRNAs fulfill regulatory functions has been established by their involvement in many different processes and diseases[8,9
]. Therefore, it is tempting to use these molecules in order to treat liver fibrosis; a condition that is caused by a deregulation of biological processes. To succeed in this attempt, we need to identify the miRNAs, which are differentially regulated in the normal and diseased liver, and more specifically in the HSCs; one cell type that is responsible for the fibrotic process.
The purpose of this study was to identify differentially regulated miRNAs in in vitro activated HSCs, in order to study them in an in vivo animal model, and finally, to determine their role in the activation process.
MATERIALS AND METHODS
Isolation of rat primary HSCs and cell culture conditions
Wistar rats were used to isolate primary HSCs (pHSCs) according to a published pronase/collagenase in situ
]. The isolation protocol was approved by the Institutional Animal Care and Use Committee under #080389. For in vitro
activation, the cells were seeded into 75-cm2
culture flasks and harvested after 3, 5, 7 or 10 d. Primary cells and the HSC-2 cell line were cultured in high glucose Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10% fetal bovine serum, 100 U/mL penicillin and 100 ?g/mL streptomycin at 37°C in a 5% CO2
HSC-2 is a spontaneous immortalized cell line derived from the pHSCs of a male Wistar rat. The primary cells were passaged several times before clonal selection by limiting dilution[11
The purity of pHSCs from rats on normal and choline-deficient ethionine supplemented (CDE) diet was assessed using vitamin A autofluorescence or real-time polymerase chain reaction (PCR), respectively (Figure ). All cell culture reagents were purchased from Invitrogen (Carlsbad, CA, USA).
Primary hepatic stellate cells and over-expression of miR-146a in hepatic stellate cell-2 cell line. A: Bright-field image of 1 d cultivated primary hepatic stellate cells and the corresponding vitamin A autofluorescence image are shown. Scale bar represents 100 ?m. Real-time polymerase chain reaction (PCR) for in vivo activated hepatic stellate cells from rats on normal (n = 2) and choline-deficient ethionine supplemented (CDE) diet (n = 4). The mean ± SE for each diet model is shown; B: A representative image for the over-expression of miR-146a as visualized by the reporter GFP is shown. Real-time PCR for three independent clones confirmed the expression of miR-146a. The data represent the mean ± SE of triplicate reactions. SMAA: Smooth muscle ?-actin.
In vivo activation of rat HSCs
Six- to eight-week-old male Wistar rats were fed the CDE diet (CDE model) (MP Biomedicals, Solon, OH, USA, #0296021410) for 4 wk (Figure ). Livers were isolated, perfused with PBS and fixed in neutralized formalin (paraffin embedding) or in vivo activated pHSCs were isolated.
Histological and immunohistochemical analysis of livers from rats receiving choline-deficient ethionine supplemented diet for 4 wk. A: HE staining shows the structural changes between control and choline-deficient ethionine supplemented (CDE) diet livers. No severe steatosis is observed; B: The Sirius Red staining depicts the deposition of collagen around the portal area and the whole liver; C: The increase in smooth muscle ?-actin (SMAA) staining reflects the increasing number of myofibroblasts seen in patches throughout the liver. Scale bar represents 200 ?m; D: The Western blotting data confirm the increase in SMAA and ColI.
Isolation of miRNA for microarray and analysis
miRNA was extracted from quiescent (freshly isolated) and 10-d in vitro
-activated pHSCs using the PureLink purification kit (K1570-01; Invitrogen). The miRNA microarray (NCode Multi-Species miRNA microarray V2) was performed according to the manufacturer’s manual (MIRLS-20; Invitrogen). For each experiment, a dye swap was performed. The arrays were scanned using a GenePix 4200AL array scanner. The raw datasets were deposited under #GSE19463 at the Gene Expression Omnibus (GEO) repository[12
]. For two-color miRNA arrays, averaging of dye-swapped arrays was performed to minimize the dye effects prior to normalization using the Cross-Correlation method[13
]. The targets of differentially regulated miRNAs (Table ) were predicted by three different methods, TargetScan 5.1[14
], mirBASE target[15
], and miRNA Viewer[16
] using default parameters. Targets predicted by at least two tools were selected and grouped into upregulated and downregulated miRNAs, respectively. These two groups of targets were subjected to pathway analysis using Ingenuity Pathway Analysis (Ingenuity Systems, Redwood City, CA, USA). A ratio was calculated whereby the number of predicted targets in a given pathway was divided by the total number of molecules in that pathway. The Fisher’s exact test was used by the software to calculate a P
value. This P
value represented the probability that the association between the predicted targets and the pathway could not be explained by chance alone. The P
value cutoff was set at P
? 0.001. The x axis was the negative logarithm of P
value with a base of 10 (-log10 P
Differentially regulated miRNAs as identified by miRNA microarray
The verification of the microarray data and subsequent miRNA assessments were performed for let-7b, let-7c, miR-16, 26a, 29a, 31, 125b, 143, 146a, 150 and 214 by using the respective Taqman MicroRNA assays (P/N 4427975, Applied Biosystems, Foster City, CA, USA). The U6 snRNA assay (ID 001973) served as a normalization control. Total RNA was isolated using the NucleoSpin RNAII kit (Macherey-Nagel, Germany). Total RNA and miRNA were isolated using the same kit but with a small modification. Briefly, the cell lysate was adjusted to contain 35% ethanol and passed through the RNAII column to bind the total RNA. The ethanol concentration of the flow through was then adjusted to > 70% and passed through the same column in order to bind the miRNA. The Cells-to-Ct kit (Invitrogen, P/N 4391848) was used for some experiments to quantify the miRNA expression with the respective miRNA assays. The reverse transcription and real-time PCR were performed according to the assays protocol using the ABI 7500 Fast Real Time PCR System (Applied Biosystems). Taqman assays used were smooth muscle ?-actin (SMAA) (Rn01759928_g1), Col1a1 (Rn01463849_g1), interleukin (IL)-6 (Rn00561420_m1), cyclooxygenase-2 (Cox-2) (Rn00568225_m1), RelA (Rn01502266_m1), CD31 (Rn01467259_m1), Albumin (Rn01413833_m1), CD68 (Rn01495643_g1) and tissue inhibitor of metalloproteinase (TIMP)-3 (Rn00441826_m1).
Nuclear factor-?B siRNA transfection
HSC-2 cells were seeded at a density of 106 per 100 mm cell culture dish and incubated at 37°C. The siRNA was mixed at a final concentration of 10 nmol/L with 1 mL DMEM without serum and 120 ?L HiPerfect transfection reagent (Qiagen, Germany) and incubated for 10 min. The mixture was added drop-wise to the cells and incubated for 48 h. For the mock control, only the HiPerfect reagent was used. The ON-Targetplus nuclear factor (NF)-?B siRNAs used were J-080033-11 and J-080033-12 (Dharmacon, Lafayette, CO, USA). These conditions were tested for transfection efficiency using FITC-labeled siRNA and FACS analysis.
Overexpression of miR-146a in an HSC cell line
The vector was constructed by amplification of a 487-bp fragment containing the rno-miR-146a from rat genomic DNA using the following primer pair: sense 5'-AAGCTTGCCACCAGTCCCATCCTTCACC-3' (HindIII), anti-sense 5'-GGATCCTTCCTCTGTGCTGGGATTACAGGGTG-3' (BamHI). After sub-cloning, the rno-miR-146a was excised using BamHI/EcoRV and cloned into pcDNA6.2/GW EmGFP-miR (Invitrogen). The HSC-2 cells were stably transfected with the construct using Lipofectamine 2000 (Invitrogen) and selected in cell culture medium supplemented with 10 ?g/mL Blasticidin. The clonal selection was achieved using FACS.
Gene expression array and analysis
Total RNA from HSC-2 cells overexpressing miR-146a and control cells (two different passages) were used to study the transcriptome changes using the GeneChip Rat Genome 230 2.0 (Affymetrix, USA). The preparation of the samples was performed according to the technical manual P/N 702232 Rev. 3 (Affymetrix) using one-cycle cDNA and target labeling. The chips were scanned using a Genechip Scanner 3000 (Affymetrix). The raw datasets were deposited under #GSE19463 at the GEO repository[12
The microarray probe set data was summarized using the Robust Multi-Array Average expression measure method, and pre-processed to correct unreliable (small) intensities for each array. The pre-processed data were then normalized using the Cross-Correlation method[13
]. For each gene, a fold change value was calculated for samples vs
control. Differentially expressed genes (DEGs) were selected based on the criterion of fold change > 2. The P
values of DEGs were obtained using one-tailed Student’s t
test. Pathway analysis was carried out on the DEGs using Ingenuity Pathway Analysis (Ingenuity Systems).
Transfection of miRNA mimics and hairpin-inhibitor
Cells were seeded at 20?000 per well in 48-well plates 24 h prior to transfection. The miRNA mimics or hairpin-inhibitor were added at the required final concentration (miR-26a, 146a, controls and quadruple transfection: 50 nmol/L each; miR-29a and 214: 200 nmol/L each) to 750 ?L DMEM without serum, followed by 10 ?L HiPerfect transfection reagent. The mixture was incubated for 10 min. The medium from each well was aspirated and replaced by 250 ?L of the mixture. The transfection was performed in triplicate. Controls were either HiPerfect reagent only (mock) or control miRNAs for the mimic and/or inhibitor.
SDS-PAGE and Western blotting
Cells were lysed in ProteoJet lysis buffer (#K0301; Fermentas, Glen Burnie, MD, USA) and the protein concentration was estimated using the BCA method (Thermo Scientific, USA). The samples were separated in 4%-12% Bis-Tris NuPage gels (Invitrogen) and transferred onto nitrocellulose membranes. The membranes were blocked for 1 h at room temperature using 5% non-fat milk in TBS-Tween (TBS-T). The primary antibodies were applied in the following dilutions: interleukin receptor associated kinase 1 (IRAK1) (sc-7883; Santa Cruz Biotechnology, Santa Cruz, CA, USA,) 1:400; tumor necrosis factor receptor associated factor 6 (TRAF6) (sc-7221; Santa Cruz Biotechnology) 1:400; I?B? (#4814; Cell Signaling, Danvers, MA, USA) 1:1000; pI?B? (#2859; Cell Signaling) 1:750; Cox-2 (sc-1747; Santa Cruz Biotechnology) 1:5000; and ?-actin (ab-8227; Abcam, Cambridge, UK) 1:5000. After three washes in TBS-T, the appropriate HRP-conjugated secondary antibody was given at 1:2000 dilution in blocking solution. After three washes in TBS-T, the membrane was developed using the chemiluminiscence substrate (Millipore, Billerica, MA, USA). Primary and secondary antibodies were incubated at 4°C overnight and 1 h at room temperature, respectively.
Electrophoretic mobility shift assay
Nuclear protein extract from rno-miR-146a-overexpressing clones was obtained using the NE-PER Nuclear and Cytoplasmic Extraction kit (Thermo Scientific). The electrophoretic mobility shift assay (EMSA) was performed using the NF-?B(I) EMSA kit according to its protocol (AY1030; Panomics, USA), as described previously[17
]. The samples were separated in a 6% non-denaturing polyacrylamide gel (Invitrogen) and transferred to a nylon membrane.
Immunohistochemistry and staining of liver sections
Slides were de-paraffinized and the antigen retrieved by heat exposure in the Target Retrieval Solution pH 9 (S2367; Dako, Glostrup, DK) using a 2100-Retriever retrieval steamer for 45 min. The endogenous peroxidase was blocked with 3% H2O2 in methanol for 15 min. Protein was blocked in 10% normal goat serum in PBS for 20 min. The slides were incubated with mouse anti-human SMAA (M0851; Dako) at 1:100 dilution for 1 h, washed and incubated with an anti-mouse HRP-conjugated antibody (K4001; Dako) for 30 min, and developed with DAB (K3468; Dako). All incubations were carried out at room temperature. Nuclei were counter stained with hematoxylin. Hematoxylin and eosin and Sirius Red staining was performed according to standard protocols on paraffin sections. Bright-field images were taken with the LEICA RMB-DM epifluorescence microscope (LEICA, Germany).
All quantitative data were presented as mean ± SE. Experimental data were analyzed using the two-tailed Student’s t test assuming equal variances. P ? 0.05 was considered significant. The time-dependent changes during in vitro HSC activation were tested for significance at the 0.05 level using one-way ANOVA and Bonferroni’s post-hoc test. The array data were normalized and analyzed as described in the respective sections above.
Identification of differentially regulated miRNAs in in vitro activated pHSCs and comparison to in vivo activated pHSCs
In 10-d in vitro activated pHSCs, 16 miRNAs were upregulated and 26 were downregulated significantly in comparison to quiescent pHSCs (Table ). We included miR-29a, although the P value was above the threshold of 0.05, for further studies because of its predicted targets, which consisted of a number of collagens. The microarray data were confirmed for a number of chosen miRNAs (let-7b, 7c, miR-16, 26a, 29a, 31, 125b, 143, 146a, 150 and 214) using real-time PCR in three additional experiments (Figure ). Using isolated in vivo activated pHSCs from rats on CDE diet, we found that only miRNAs let-7b, 7c, miR-31, 143 and 214 showed the same regulation as observed for the in vitro activated pHSCs (Figure ).
Verification of microarray data by real-time polymerase chain reaction of 11 differentially regulated miRNAs and their regulation upon in vivo activation of hepatic stellate cells. A: The graph depicts the changes in the miRNA expression of 11 miRNAs detected by real-time polymerase chain reaction, comparing quiescent with 10-d culture activated primary hepatic stellate cells (pHSCs). The data represent the mean ± SE of three independent experiments (aP ? 0.05, bP ? 0.005); B: The graph illustrates the relative expression levels of miRNAs in isolated in vivo activated pHSCs (n = 4, choline-deficient ethionine supplemented diet) compared to normal diet (n = 2) (aP ? 0.05, bP ? 0.005).
Pathway analysis for differentially regulated miRNAs in in vitro activated pHSCs
We performed a pathway analysis using the predicted targets of the differentially regulated miRNAs. The enrichment of genes in single pathways is shown as the -log of the P value (P ? 0.001). Signaling pathways which were affected include endothelin-1, cyclin-dependent kinase 5, extracellular signal-regulated kinase (ERK)/mitogen-activated protein kinase (MAPK), p70S6K, chemokine, bone morphogenetic protein (BMP) and IL-6 for the upregulated miRNAs, as well as ERK/MAPK, production of NO and reactive oxygen species (ROS), AMP activated protein kinase (AMPK), transforming growth factor (TGF)-?, integrin, cAMP-mediated signaling and phosphatase and tensin homolog (PTEN) for the downregulated miRNAs (Figure ).
Predicted targets of all differentially regulated miRNAs during in vitro activation of primary hepatic stellate cells (Table ) were analyzed. The two charts represent the enrichment of molecules in affected pathways for the upregulated (A) and downregulated (B) miRNAs. Only pathways with P ? 0.001 are shown. IGF-1: Insulin-like growth factor-1; ERK: Extracellular signal-regulated kinase; MAPK: Mitogen-activated protein kinase; BMP: Bone morphogenetic protein; IL: Interleukin; ROS: Reactive oxygen species; AMPK: AMP activated protein kinase; TGF: Transforming growth factor; PTEN: Phosphatase and tensin homolog; LXR: Liver X receptor; RXR: Retinoid X receptor; PPAR: Peroxisome proliferator-activated receptor; HGF: Hepatocyte growth factor.
Overexpression of miR-146a in HSC-2 and transcriptome analysis
Studies have shown that miR-146a is linked to inflammation and the NF-?B pathway through the two known targets IRAK1 and TRAF6[18,19
]. In order to study the function of miR-146a in activated HSCs in vitro
, we overexpressed this miRNA in a HSC cell line HSC-2[11
]. The level of miR-146a in this cell line is very low, making it suitable for the overexpression. The expression of the reporter green fluorescent protein and the real-time PCR validation of the miR-146a expression (Figure ) provided evidence for the successful overexpression of miR-146a in three different clones (S1, S4 and S5).
IRAK1 and TRAF6 are direct targets of miR-146a with two target sites for each mRNA (Figure ). We were able to show downregulation of these proteins in all three clones (Figure ). The functional consequence of this downregulation can be seen by suppression of the phosphorylation of I?B at Ser32 (Figure ). The reduced phosphorylation of I?B in turn should lead to the retention of NF-?B in the cytoplasm. Indeed, our EMSA illustrated that there was reduced nuclear binding activity of NF-?B to an NF-?B probe in all clones (Figure ). One of the genes regulated by NF-?B is Cox-2, which is functionally related to HSCs due to its pro-apoptotic effect on HSCs[20,21
]. Therefore, we investigated the protein level of Cox-2 in the miR-146a-overexpressing clones, and found the expected downregulation (Figure ). Surprisingly, further investigation revealed that the mRNAs of NF-?B and Cox-2 were upregulated (Figure ). In contrast, we observed a significant downregulation of IL-6 mRNA, another target of NF-?B, in the clones S1, S4 and S5 (Figure ). We also found a significant upregulation of SMAA and collagen?I?(ColI) mRNAs, a HSC activation and a fibrotic marker, respectively (Figure ).
Changes during overexpression of rno-miR-146a in the hepatic stellate cell-2 cell line. A: Depicted are two putative binding sites of miR-146a to the 3’-UTR of rat tumor necrosis factor receptor associated factor 6 (TRAF6) and rat interleukin receptor associated kinase 1 (IRAK1), respectively; B: The Western blotting data show the suppression of TRAF6 and IRAK1, resulting in the decreased phosphorylation of I?B, although the expression of I?B remained unchanged. A representative Western blotting for two independent experiments is shown; C: Electrophoretic mobility shift assay (EMSA) results demonstrated a decrease in nuclear factor (NF)-?B DNA binding activity due to the overexpression of miR-146a. TATA binding protein (TBP) showed equal loading of samples. A representative EMSA experiment is shown out of three independent samples for each clone; D: miR-146a-overexpressing clones showed a reduced level of cyclooxygenase-2 (Cox-2) protein. The Western blotting shown is representative of two independent experiments; E: The relative fold change in mRNA expression between hepatic stellate cell (HSC)-2 and miR-146a-overexpressing HSC-2 cells for five different targets [NF-?B (RelA), Cox-2, smooth muscle ?-actin, ColI, interleukin-6] is shown. The data represent the mean ± SE of two independent experiments (aP ? 0.005).
In order to establish a link between the regulation of miR-146a and NF-?B activity, as proposed by Taganov et al[18
], we transfected NF-?B siRNAs into HSC-2 cells. The efficiency of the transfection was shown by the downregulation of NF-?B in total cell lysates and nuclear extracts, which resulted in a decrease in NF-?B DNA binding activity (Figure , respectively). We also found downregulation of miR-146a in NF-?B siRNA-transfected cells, thereby confirming a regulation of miR-146a by NF-?B in HSCs (Figure ). Surprisingly, we noticed an increase in the Cox-2 protein expression (Figure ), which implied a yet unclear involvement of miR-146a in the regulation of this enzyme.
Regulation of miR-146a by nuclear factor-?B. A: Knock-down experiments using nuclear factor (NF)-?B siRNAs showed a reduced level of cellular NF-?B (RelA) protein; B: The nuclear level of NF-?B (RelA) was decreased and showed a diminished DNA binding activity. Depicted is a representative Western blotting and electrophoretic mobility shift assay from three independent experiments; C: Downregulation of NF-?B (RelA) mRNA due to NF-?B siRNA transfection was accompanied by a decrease in miR-146a after 24 h. The data represent the mean ± SE of two independent experiments (aP ? 0.05, bP ? 0.01, cP ? 0.001); D: Cyclooxygenase-2 (Cox-2) protein was upregulated after NF-?B siRNA transfection. Shown are a representative Western blotting and the densitometric analysis of six independent experiments (aP ? 0.05).
The differences in the NF-?B-dependent regulation of Cox-2 and IL-6 have already hinted at the intricacy of the influence of the miR-146a overexpression has on the gene expression in activated HSCs. In order to get an overview of the transcriptome changes, we performed a gene expression analysis of the three miR-146a-overexpressing clones, and compared them with control cells using a cDNA microarray. The analysis yielded 485 up- and 309 downregulated transcripts (Supplementary Tables and ), which satisfied a P value ? 0.05 and at least twofold change. Among the upregulated genes were Lmcd1, CD81, FGF13, Col4a1, Cadherin 11 and BMP-4. The highly downregulated genes included Col15a1, MMP-2, Thy-1, IL-1RL1 and Cadherin 13.
We further analyzed the pathways which were significantly enriched, using a P value ? 0.05 as a threshold. Here, we observed enrichment for signaling pathways like integrin-linked kinase, hepatic fibrosis/HSC activation and caveolar-mediated endocytosis, calcium, cAMP-mediated signaling, integrin, endothelin-1 for the upregulated genes (Figure ), and hepatic fibrosis/HSC activation, lipopolysaccharide (LPS)/IL-1-mediated inhibition of retinoid X receptor (RXR) function and nitrogen metabolism, and liver X receptor/RXR activation for the downregulated genes (Figure ).
Pathway analysis for the differentially expressed genes of miR-146a over-expressing clones. The charts depict the pathways affected by the (A) or (B) of genes upon stable transfection of miR-146a into hepatic stellate cell-2 cells. Only pathways with P ? 0.05 are shown. ILK: Integrin-linked kinase; FGF: Fibroblast growth factor; IL: Interleukin; LPS: Lipopolysaccharide; LXR: Liver X receptor; RXR: Retinoid X receptor.
The most interesting finding was the robust upregulation of TIMP-3 mRNA (Supplementary Tables ), verified by real-time PCR (Figure ), which is an inhibitor of the tumor necrosis factor-? converting enzyme[22
], and has been proposed as a tumor suppressor. Similarly, pHSCs treated with miR-146a mimic also showed induction of TIMP-3 mRNA (Figure ).
Relative expression of tissue inhibitor of metalloproteinase-3 mRNA in rno-miR-146a-overexpressing hepatic stellate cell-2 cells. A: The graph depicts the relative changes in tissue inhibitor of metalloproteinase (TIMP)-3 mRNA of three rno-miR-146a-overexpressing clones detected by real-time polymerase chain reaction (PCR). The data represent the mean ± SE of two different passages for each clone (aP ? 0.005); B: Primary hepatic stellate cells were treated with 50 nmol/L miR-146a mimic, and the expression of TIMP-3 mRNA was analyzed by real-time PCR and expressed as fold change relative to mock controls. The data represent the mean ± SE of three independent experiments (bP ? 0.01). HSC: Hepatic stellate cell.
Time-dependent expression of different miRNAs during in vitro activation of pHSCs
The fact that miR-146a was not downregulated in in vivo activated pHSCs (CDE diet) prompted us to study the time-dependent expression of this miRNA during the in vitro activation of pHSCs, together with miR-26a, 29a and 214. The expression of miR-146a was indeed downregulated at day 3 already, and recovered subsequently until day 10. Although the miR-146a level at day 10 was still lower than that in quiescent pHSCs at day 0, there was still a 10-fold increase between day 3 and 10 (Figure ). In contrast, the expression of miR-26a and 29a did not change as dramatically from day 3 to day 10. We also noticed that miR-214 started to increase only from day 5 onwards (Figure ).
Time-dependent changes in the expression of different miRNAs during in vitro activation and miRNA mimic and inhibitor transfection of primary hepatic stellate cells. A: The relative changes in the expression level after 3, 5, 7 and 10 d of in vitro primary hepatic stellate cells (pHSCs) activation is shown for miR-146a, 26a, 29a and 214 (aP ? 0.05, one-way ANOVA). The data represent one of two independent experiments performed in triplicate; B: pHSCs were transfected with miR-146a, 26a, 29a mimics or miR-214 inhibitor or in combination. The control transfection consisted of control miRNAs for the mimic and/or inhibitor. The smooth muscle ?-actin and ColI?mRNA expression was analyzed as fold change relative to mock controls. The data represent the mean ± SE of two independent experiments, each performed in triplicate (aP ? 0.05, bP ? 0.01, cP ? 0.001). SMAA: Smooth muscle ?-actin.
Regulation of SMAA and ColI?transcripts in pHSCs by different miRNA mimics and inhibitor
In order to study the effect of different miRNA mimics or inhibitor on the in vitro activation process, we transfected 3-d in vitro activated pHSCs for 3 d with miR-146a, 26a, 29a mimics, miR-214 hairpin inhibitor or all combined. The impact on the HSC activation was followed using real-time PCR to study the changes on the mRNA levels of SMAA and ColI. The high efficiency of transfection was demonstrated by real-time PCR (Figure ). We found moderate upregulation of the activation marker SMAA by miR-146a (Figure ); an observation seen also for the miR-146a-overexpressing clones (Figure and cDNA microarray data). In fact, all cells transfected with the mimics, the inhibitor or combined showed an upwards trend for SMAA mRNA compared to the mock control, although the level did not always change significantly (Figure ). For the ColI?expression, we noted again an increase caused by miR-146a mimic (not significant) and a decrease by miR-26a, 29a mimic and miR-214 inhibitor. The quadruple transfection led to a suppression of ColI?mRNA (Figure ).
Transfection of primary hepatic stellate cells with different miRNA mimics and inhibitor. The miRNA expression was analyzed as fold change relative to mock transfected primary hepatic stellate cells (pHSCs). Shown are data for the transfection of miR-146a mimic and miR-214 inhibitor with respective control (A), for the transfection of miR-26a and miR-29a mimic with respective control (B) and for the quadruple transfection of miR-146a, miR-26a, miR-29a mimic and miR-214 inhibitor (C) (aP ? 0.05, bP ? 0.005).
The aim of this study was to gain a deeper insight into the regulation of miRNAs during the activation process of pHSCs, as well as the influence of up- or downregulation of miRNAs on the gene expression and activation of HSCs.
The expression analysis of miRNAs between quiescent and in vitro
activated pHSCs yielded a number of induced and suppressed miRNAs (Table ), some of which (miR-143, 16, 122, 146a, 92b, 126) confirmed the findings of Guo et al[23
]. On the other hand, there were some differences in the regulation of certain miRNAs (miR-328, 207), which could be attributed to the dynamic nature of miRNA regulation and the different use of quiescent pHSCs (day 0 vs
When evaluating the miRNAs expression profile of the in vitro
and in vivo
activated pHSCs, a clear distinction was seen in the expression of miR-16, 26a, 29a, 125b, 146a and 150 (compare Figure ); a phenomenon which could be explained by the distinct HSC activation process. This has been shown at the gene expression level by De Minicis et al[24
]. The in vivo
activation was performed over a period of 4 wk, whereas the in vitro
activation was monitored over 10 d, which could also account for some differences in the miRNA expression, assuming a dynamic regulation.
On the other hand, we found that certain miRNAs (let-7b, 7c and miR-214) were regulated in the same way during in vitro and in vivo activation of pHSCs. It also became clear to us that miR-214 could be a potential candidate for a diagnostic approach, because this miRNA always shows robust upregulation.
Pathway analysis of the miRNA microarray data was performed to obtain information on signaling cascades involving predicted targets of the differentially regulated miRNAs in in vitro
activated pHSCs (Figure ). NO and ROS are known to play a role in the activation process and apoptosis of HSCs[25,26
]. The pathways for AMPK, ERK/MAPK, PTEN and TGF-? are also implicated in HSC activation[27-30
]. We noticed that a number of pathways were present in the charts for both up- and downregulated miRNAs, which could denote the complexity of regulated targets by each single miRNA, and possibly a cooperative effect between up- and downregulated miRNAs.
A number of publications have shown that miR-146a is involved in inflammatory diseases, regulation of the immune response and NF-?B[19,31-33
]. In the early events of liver fibrosis, the activation of HSCs is in part driven by the hepatic inflammatory process, during which different cytokines are secreted by various liver cells, like Kupffer cells, endothelial cells and hepatocytes[34,35
]. Involvement of NF-?B in HSC activation has also been shown in several research papers[36,37
]. Therefore, we overexpressed miR-146a in an HSC cell line and observed changes consistent with the findings from Bhaumik et al[19
]. The detected increase in the NF-?B transcript (Figure ) could be explained by a feedback mechanism to the reduced nuclear activity, which leads to the upregulation of the mRNA.
Cox-2 is inducible in activated HSCs by various stimuli and is thought to regulate proliferation[21
]. Others have shown that the inhibition of this enzyme has a beneficial antifibrotic effect[20,38,39
]. The seemingly discrepant findings of the protein (lower) and transcript (elevated) level for Cox-2 in the miR-146a-overexpressing HSCs (Figure ) hint at independent pathways for the regulation of Cox-2. These pathways have been shown for intestinal myofibroblasts[40
] and during ischemic injury of ileal mucosa[41
]. Lasa et al[42
] and others have shown that the p38 MAPK signaling cascade is able to stabilize the Cox-2 mRNA[43
], which could also explain an elevated transcript level. We also cannot exclude that other mechanisms could be involved in stabilizing the Cox-2 mRNA and/or a regulation of Cox-2 by other miRNAs like miR-26a or 143, which are also present in the cell line HSC-2 and for which Cox-2 is a predicted target.
In contrast, the IL-6 mRNA, another molecule regulated by NF-?B, was downregulated (Figure ). This observation implies that IL-6 regulation in HSCs is more tightly associated with NF-?B than that of Cox-2.
We were also interested to know whether downregulation of the NF-?B DNA binding activity triggered by miR-146a overexpression could facilitate a feedback loop in HSCs; a notion supported by the fact that the promoter region of miR-146a contains a number of NF-?B binding sites[18
]. As expected, a reduction in the NF-?B DNA binding activity (Figure ) leads to a decrease in miR-146a (Figure ). The observed upregulation of Cox-2 protein (Figure ) was somewhat surprising and again substantiated the speculation that other pathways such as p38 MAPK, C-Jun N-terminal kinase and ERK could participate in the regulation of Cox-2 in HSCs[40,44,45
The microarray analysis revealed that the transcriptome changes caused by miR-146a overexpression are complex and numerous pathways are affected (Figure , Supplementary Tables and ). We found that several DEGs coincided with data from earlier publications on HSC activation[24,46
], suggesting that a number of genes affected by miR-146a overexpression are also involved in the activation process. Pathway analysis of the DEGs (Figure ) confirms a link between miR-146a and inflammation (LPS/IL-1 mediated inhibition of RXR function, eicosanoid signaling, nitrogen metabolism and NRF2-mediated oxidative stress response pathways). That the miR-146a overexpression in HSC-2 cells leads to changes in the pathway called hepatic fibrosis/HSC activation emphasizes that these changes are specific for the HSCs. The upregulation of TIMP-3 (Supplementary Table and Figure ) again emphasizes the involvement of miR-146a in inflammatory processes and immunity, by linking it to the TNF? activity[47
We noticed a robust downregulation of miR-146a during in vitro
, but a missing regulation of miR-146a during in vivo
activation of pHSCs (CDE diet). We hypothesized that there is a dynamic component in the regulation of miR-146a. We effectively found that there is a time-dependent regulation of miR-146a over 10 d of in vitro
activation of pHSCs. From an in vivo
perspective, it could be a possibility that miR-146a is decreased following the first insult to the liver, but reaches almost a normal level during the developing fibrosis, as seen for the in vivo
activated pHSCs (CDE diet). The mechanism behind this miR-146a regulation is not clear, but the involvement of different transcription factors [NF-IL6, interferon regulatory factor (IRF 3/7)] binding to its promoter region is conceivable[18
The dynamic nature of miRNA regulation during the in vitro
activation of pHSCs could also partially explain the differences in the expression pattern of the miRNAs in vitro
and in vivo
. The dynamic nature of miRNA expression has been shown for the T-cell development[48
], and it makes sense if we consider the multitude of effects a single miRNA can have due to the imperfect complementarity to its target sequence.
The in vivo
targets of a miRNA treatment are pHSCs, therefore, we assessed the effects of several miRNAs mimics (miR-26a, 29a, 146a) and inhibitor (miR-214) on the activation state of pHSCs. The transfection with a combination of all mimics and inhibitor was performed so as to examine possible cooperative effects between different miRNAs, as a first step to understand the cooperativity of miRNA expression changes during HSC activation. The miR-26a, 29a mimics and miR-214 inhibitor showed a significant suppression of the ColI?mRNA (Figure ). This is somewhat surprising because even though a number of collagens are predicted targets for miR-26a and 29a, none has a perfect binding site, which would explain regulation by mRNA degradation. Therefore, we conclude that the mechanism by which miR-26a, 29a and 214 downregulate the ColI?mRNA is indirect, as also suggested by van Rooij et al[49
] for miR-29a. The downregulation of ColI?by the quadruple transfection shows some synergistic effect between the miRNAs.
Our findings showed the differential regulation of miRNAs in in vitro and in vivo activation of pHSCs, and particularly, the involvement of miR-26a, 29a and 214 in the regulation of ColI?mRNA. Moreover, miR-146a overexpression or treatment with miR-146a mimic upregulates TIMP-3 mRNA, which suggests an association between miR-146a, TNF? activity and inflammation. In conclusion, our observations help build a global picture of the miRNA regulation during HSC activation in vitro and in vivo, and may have important implications when considering a therapeutic approach for treating liver fibrosis using miRNAs.
miRNAs are a relatively new and exciting tool to control the expression of multiple genes. During liver injury and subsequent wound healing involving hepatic stellate cells (HSCs), complex regulatory processes occur and have to be tightly regulated in this cell type. miRNAs could be one tool to control these processes, and therefore, it is of interest to the research community to gain information about the expression of miRNAs during liver fibrosis in HSCs.
Liver fibrosis and subsequently cirrhosis are common outcomes of chronic injuries to the liver. HSCs are involved in liver fibrosis and repair. The tools for the treatment of liver fibrosis are limited and are still under development. In this study, the authors aimed to gain information for the possible role of miRNAs in liver fibrosis and whether they could become a future tool to develop a treatment for liver fibrosis by addressing the changes in HSCs.
Innovations and breakthroughs
Different publications have analyzed the miRNA expression in HSCs in vitro and studied the effect of various differentially regulated miRNAs in HSCs. The authors analyzed the miRNA expression in an in vivo model of hepatic fibrosis, namely choline-deficient ethionine supplemented diet. Furthermore, they studied the transcriptome changes upon overexpression of miR-146a and found that, in particular, tissue inhibitor of metalloproteinase-3 showed robust up-regulation, a hitherto unreported effect, which emphasizes its involvement in inflammation. Another important finding was the dynamics of miRNA regulation during the in vitro activation of HSCs.
miRNAs are becoming a promising tool for the regulation of gene expression. In order to use this tool, it is necessary to understand the role and regulation of the targeted miRNA. In this study, the authors describe the dynamic regulation of specific miRNAs. The results of this study show clearly that the use of miRNAs as target molecules will have to take this dynamic component into consideration. The same is valid for the use of miRNAs as therapeutic agents.
miRNAs are small non-coding RNAs that are about 23 nucleotides long. The versatility of miRNAs depends on the imperfect binding (seed region) to the 3’-UTR of mRNAs. This imperfect binding results in many different binding partners. The regulation by miRNAs leads to a translational repression and/or mRNA destabilization.
The field of miRNA research as well as HSC activation mechanisms are very up to date and important areas of research, in order to find new strategies against liver fibrosis. The methods used are comprehensive and convincing. In all, the study was fairly well conducted and interesting.
Bataller R, Brenner DA. Liver fibrosis. J Clin Invest.
Gressner AM, Weiskirchen R. Modern pathogenetic concepts of liver fibrosis suggest stellate cells and TGF-beta as major players and therapeutic targets. J Cell Mol Med.
Maher JJ, McGuire RF. Extracellular matrix gene expression increases preferentially in rat lipocytes and sinusoidal endothelial cells during hepatic fibrosis in vivo. J Clin Invest.
Stefanovic B, Hellerbrand C, Holcik M, Briendl M, Aliebhaber S, Brenner DA. Posttranscriptional regulation of collagen alpha1(I) mRNA in hepatic stellate cells. Mol Cell Biol.
Mann DA, Smart DE. Transcriptional regulation of hepatic stellate cell activation. Gut.
Bartel DP. MicroRNAs: target recognition and regulatory functions. Cell.
Filipowicz W, Bhattacharyya SN, Sonenberg N. Mechanisms of post-transcriptional regulation by microRNAs: are the answers in sight?. Nat Rev Genet.
O'Hara SP, Mott JL, Splinter PL, Gores GJ, LaRusso NF. MicroRNAs: key modulators of posttranscriptional gene expression. Gastroenterology.
Pauley KM, Cha S, Chan EK. MicroRNA in autoimmunity and autoimmune diseases. J Autoimmun.
Weiskirchen R, Gressner AM. Isolation and culture of hepatic stellate cells. Methods Mol Med.
Maubach G, Lim MC, Zhuo L. Nuclear cathepsin F regulates activation markers in rat hepatic stellate cells. Mol Biol Cell.
Edgar R, Domrachev M, Lash AE. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res.
Chua SW, Vijayakumar P, Nissom PM, Yam CY, Wong VV, Yang H. A novel normalization method for effective removal of systematic variation in microarray data. Nucleic Acids Res.
Lewis BP, Burge CB, Bartel DP. Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets. Cell.
Griffiths-Jones S, Grocock RJ, van Dongen S, Bateman A, Enright AJ. miRBase: microRNA sequences, targets and gene nomenclature. Nucleic Acids Res.
Enright AJ, John B, Gaul U, Tuschl T, Sander C, Marks DS. MicroRNA targets in Drosophila. Genome Biol.
Lim MC, Maubach G, Zhuo L. TGF-beta1 down-regulates connexin 43 expression and gap junction intercellular communication in rat hepatic stellate cells. Eur J Cell Biol.
Taganov KD, Boldin MP, Chang KJ, Baltimore D. NF-kappaB-dependent induction of microRNA miR-146, an inhibitor targeted to signaling proteins of innate immune responses. Proc Natl Acad Sci USA.
Bhaumik D, Scott GK, Schokrpur S, Patil CK, Campisi J, Benz CC. Expression of microRNA-146 suppresses NF-kappaB activity with reduction of metastatic potential in breast cancer cells. Oncogene.
Paik YH, Kim JK, Lee JI, Kang SH, Kim DY, An SH, Lee SJ, Lee DK, Han KH, Chon CY. Celecoxib induces hepatic stellate cell apoptosis through inhibition of Akt activation and suppresses hepatic fibrosis in rats. Gut.
Gallois C, Habib A, Tao J, Moulin S, Maclouf J, Mallat A, Lotersztajn S. Role of NF-kappaB in the antiproliferative effect of endothelin-1 and tumor necrosis factor-alpha in human hepatic stellate cells. Involvement of cyclooxygenase-2. J Biol Chem.
Lee MH, Knäuper V, Becherer JD, Murphy G. Full-length and N-TIMP-3 display equal inhibitory activities toward TNF-alpha convertase. Biochem Biophys Res Commun.
Guo CJ, Pan Q, Cheng T, Jiang B, Chen GY, Li DG. Changes in microRNAs associated with hepatic stellate cell activation status identify signaling pathways. FEBS J.
De Minicis S, Seki E, Uchinami H, Kluwe J, Zhang Y, Brenner DA, Schwabe RF. Gene expression profiles during hepatic stellate cell activation in culture and in vivo. Gastroenterology.
Langer DA, Das A, Semela D, Kang-Decker N, Hendrickson H, Bronk SF, Katusic ZS, Gores GJ, Shah VH. Nitric oxide promotes caspase-independent hepatic stellate cell apoptosis through the generation of reactive oxygen species. Hepatology.
Svegliati-Baroni G, Saccomanno S, van Goor H, Jansen P, Benedetti A, Moshage H. Involvement of reactive oxygen species and nitric oxide radicals in activation and proliferation of rat hepatic stellate cells. Liver.
Hellerbrand C, Stefanovic B, Giordano F, Burchardt ER, Brenner DA. The role of TGFbeta1 in initiating hepatic stellate cell activation in vivo. J Hepatol.
Takashima M, Parsons CJ, Ikejima K, Watanabe S, White ES, Rippe RA. The tumor suppressor protein PTEN inhibits rat hepatic stellate cell activation. J Gastroenterol.
Caligiuri A, Bertolani C, Guerra CT, Aleffi S, Galastri S, Trappoliere M, Vizzutti F, Gelmini S, Laffi G, Pinzani M. Adenosine monophosphate-activated protein kinase modulates the activated phenotype of hepatic stellate cells. Hepatology.
Marra F, Arrighi MC, Fazi M, Caligiuri A, Pinzani M, Romanelli RG, Efsen E, Laffi G, Gentilini P. Extracellular signal-regulated kinase activation differentially regulates platelet-derived growth factor's actions in hepatic stellate cells, and is induced by in vivo liver injury in the rat. Hepatology.
Nakasa T, Miyaki S, Okubo A, Hashimoto M, Nishida K, Ochi M, Asahara H. Expression of microRNA-146 in rheumatoid arthritis synovial tissue. Arthritis Rheum.
Perry MM, Moschos SA, Williams AE, Shepherd NJ, Larner-Svensson HM, Lindsay MA. Rapid changes in microRNA-146a expression negatively regulate the IL-1beta-induced inflammatory response in human lung alveolar epithelial cells. J Immunol.
Williams AE, Perry MM, Moschos SA, Larner-Svensson HM, Lindsay MA. Role of miRNA-146a in the regulation of the innate immune response and cancer. Biochem Soc Trans.
Li JT, Liao ZX, Ping J, Xu D, Wang H. Molecular mechanism of hepatic stellate cell activation and antifibrotic therapeutic strategies. J Gastroenterol.
Friedman SL. Molecular regulation of hepatic fibrosis, an integrated cellular response to tissue injury. J Biol Chem.
Hellerbrand C, Jobin C, Iimuro Y, Licato L, Sartor RB, Brenner DA. Inhibition of NFkappaB in activated rat hepatic stellate cells by proteasome inhibitors and an IkappaB super-repressor. Hepatology.
Lang A, Schoonhoven R, Tuvia S, Brenner DA, Rippe RA. Nuclear factor kappaB in proliferation, activation, and apoptosis in rat hepatic stellate cells. J Hepatol.
Cheng J, Imanishi H, Liu W, Iwasaki A, Ueki N, Nakamura H, Hada T. Inhibition of the expression of alpha-smooth muscle actin in human hepatic stellate cell line, LI90, by a selective cyclooxygenase 2 inhibitor, NS-398. Biochem Biophys Res Commun.
Planagumà A, Clària J, Miquel R, López-Parra M, Titos E, Masferrer JL, Arroyo V, Rodés J. The selective cyclooxygenase-2 inhibitor SC-236 reduces liver fibrosis by mechanisms involving non-parenchymal cell apoptosis and PPARgamma activation. FASEB J.
Mifflin RC, Saada JI, Di Mari JF, Adegboyega PA, Valentich JD, Powell DW. Regulation of COX-2 expression in human intestinal myofibroblasts: mechanisms of IL-1-mediated induction. Am J Physiol Cell Physiol.
Shifflett DE, Jones SL, Moeser AJ, Blikslager AT. Mitogen-activated protein kinases regulate COX-2 and mucosal recovery in ischemic-injured porcine ileum. Am J Physiol Gastrointest Liver Physiol.
Lasa M, Mahtani KR, Finch A, Brewer G, Saklatvala J, Clark AR. Regulation of cyclooxygenase 2 mRNA stability by the mitogen-activated protein kinase p38 signaling cascade. Mol Cell Biol.
Ridley SH, Dean JL, Sarsfield SJ, Brook M, Clark AR, Saklatvala J. A p38 MAP kinase inhibitor regulates stability of interleukin-1-induced cyclooxygenase-2 mRNA. FEBS Lett.
Guan Z, Buckman SY, Miller BW, Springer LD, Morrison AR. Interleukin-1beta-induced cyclooxygenase-2 expression requires activation of both c-Jun NH2-terminal kinase and p38 MAPK signal pathways in rat renal mesangial cells. J Biol Chem.
Guan Z, Buckman SY, Pentland AP, Templeton DJ, Morrison AR. Induction of cyclooxygenase-2 by the activated MEKK1 --> SEK1/MKK4 --> p38 mitogen-activated protein kinase pathway. J Biol Chem.
Jiang F, Parsons CJ, Stefanovic B. Gene expression profile of quiescent and activated rat hepatic stellate cells implicates Wnt signaling pathway in activation. J Hepatol.
Mohammed FF, Smookler DS, Taylor SE, Fingleton B, Kassiri Z, Sanchez OH, English JL, Matrisian LM, Au B, Yeh WC. Abnormal TNF activity in Timp3-/- mice leads to chronic hepatic inflammation and failure of liver regeneration. Nat Genet.
Neilson JR, Zheng GX, Burge CB, Sharp PA. Dynamic regulation of miRNA expression in ordered stages of cellular development. Genes Dev.
van Rooij E, Sutherland LB, Thatcher JE, DiMaio JM, Naseem RH, Marshall WS, Hill JA, Olson EN. Dysregulation of microRNAs after myocardial infarction reveals a role of miR-29 in cardiac fibrosis. Proc Natl Acad Sci USA.