METHODS
Animals and ethic statement. The experiments reported in this study followed the ethical guidelines for investigations of experimental animals approved by the Animal Care and Use board and the Ethical Committee of Spanish Council of Scientific Research and performed in accordance with the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes. Animal studies are reported in compliance with the ARRIVE v.2.0 guidelines (27) and with the recommendations made by the British Journal of Pharmacology . Mice lacking the gene for cortistatin (CST-/-) were a generous gift of Dr. Luis de Lecea (Stanford University, La Jolla, CA, USA) and were generated in a C57BL/6 background and backcrossed with C57BL/6 mice for ten generations as previously described (28). Mice heterozygous (CST+/-) for cortistatin were generated by crossing female CST-/- and male CST+/+ mice. CST+/- breeding pairs were used to generate a littermate colony of wild-type (CST+/+), heterozygous (CST+/-) and knockout (CST-/-) mice for cortistatin. Both male and female mice (20-24g body weight, 8-10 weeks-old) were used in all experiments described in this study, and no differences were found between sexes. All animals were housed in a controlled-temperature/humidity environment (22±1°C, 60-70% relative humidity) in individual cages (10 mice per cage, with wood shaving bedding and nesting material), with a 12 h light/dark cycle (lights on at 7:00 a.m.) and fed with rodent chow (Global Diet 2018, Harlan) and tap water ad libitum. Mice were allowed to acclimatize to the experimental room for one hour before experiments. Mice were randomly assigned to the different experimental groups. Experiments were designed to make sample sizes relatively equal. However, this was not possible in some experiments due to the differential mortality rates occurring between genotypes and response to liver fibrosis. None of the animals were excluded from the study. Power calculations were performed using the software G*Power (www.gpower.hhu.de) to ensure that adequate group sizes were used for the studies detailed below. For in vivoanimal models, we calculated a minimum size of five to eight mice per group in order to have a power >0.95 of detecting approximately a 30% change, assuming a standard deviation of 30% at a significance level of p<0.05, expecting an effect size of 1.82 for ANOVA tests. In primary cell cultures, for effect sizes between 3.1 and 4, experiments were repeated at least four times to obtain p<0.05 and a power >0.95.
Induction of experimental liver fibrosis. To investigate the effect of cortistatin deficiency in severity of toxic hepatic fibrosis,CST+/+ , CST+/- andCST-/- mice were injected i.p. with a low-dose of CCl4 (0.5 µl/g body weight, dissolved at 1:9 in olive oil, twice/week, for six weeks). Mice injected with olive oil were used as controls of reference. To evaluate the therapeutic effect of cortistatin, hepatic fibrosis was induced inCST+/+ mice by i.p. injections of high-dose of CCl4 (0.5 µl/g, dissolved at 1:3 in olive oil, twice/week, for six weeks) and then treated i.p. with PBS or mouse cortistatin-29 (1 nmol/mouse, three times weekly, from Bachem, Bubendorf, Switzerland), starting 5 or 14 days after the first CCl4 injection. Moreover, hepatotoxic fibrosis was induced in CST+/- andCST-/- mice with low-dose CCl4and immediately treated with cortistatin-29 (1 nmol/mouse, three times weekly).
To study the effect of cortistatin deficiency in cholestasis-induced hepatic fibrosis, anesthetized CST+/+ ,CST+/- and CST-/- mice were subjected to BDL as previously described (29). To investigate the therapeutic effect of cortistatin in BDL-induced fibrosis,CST+/+ and CST+/- mice were injected i.p. with PBS or cortistatin-29 (1 nmol/mouse), three times weekly, starting one or five days after surgery, respectively.
In both models, survival was monitored daily, and on different times after BDL or initiation of CCl4 injection, animals were sacrificed by CO2 affixation, liver lobes were collected and analyzed for histopathological signs, immunofluorescence and fibrotic gene expression as described below. Collagen content in liver lobes was measured using the hydroxyproline assay as previously described (30), and bilirubin levels in serum were determined by using a colorimetric MAK126 assay kit (Sigma-Aldrich, San Louis, MO, USA).
Isolation and culture of HSCs. HSCs were isolated fromCST+/+ and CST-/- mice as described (29,31) using in vivo enzymatic digestion of liver by pronase-collagenase perfusion through portal vein followed by enrichment by centrifugation of non-parenchymal cell suspension through an 8% Nycodenz density gradient (Merk/Sigma). Enriched HSCs (pooled from six mice) were cultured in complete-DMEM (DMEM supplemented with 10% fetal bovine serum, 100 U/ml penicillin/streptomycin and 2 mM L-glutamine, all from Gibco/Thermo-Fisher, Waltham, MA, USA) in 75 cm2-Nunc flasks at 37ºC/5% CO2. After six days in culture, HSCs were either subjected to RNA isolation for real-time qPCR and next-generation transcriptome sequencing (RNAseq) or seeded secondarily (at 104) in glass-coverslips inserted in 24-well plates and incubated for 24h or 96h in complete-DMEM for immunofluorescence analysis as described below. Cortistatin-29 was added (10 nM, every other day, for one week) during the primary 6 day-culture to study its capacity to reverse the observed phenotype.
Analysis of gene expression of cortistatin in human fibrotic livers. The levels of cortistatin gene expression were obtained from the National Cancer for Biotechnology Information Gene Expression Omnibus database based on the gene chips of fibrotic liver tissues associated to hepatitis B (HBV-GSE84044, ref. 24), hepatitis C-induced hepatocarcinoma (HCV-GSE14323, ref. 25), and nonalcoholic fatty liver disease (NAFLD-GSE49541, ref. 26). In HBV, 10 healthy samples (Scheuer stage S0/G0) were compared with 10 age-matched liver samples showing high fibrosis signs (Scheuer’s stage S4). In HCV, 19 healthy samples were compared with 41 pre-malignant cirrhotic samples. In NAFLD, 40 samples showing mild stage were compared with 32 samples showing advanced stage. The raw data (Affymetrix U133A array) were processed together by correcting the background and normalizing the expression values through the rma function of the affy Bioconductor package (32).
Histopathological analysis of liver fibrosis. For histopathologic evaluation, freshly collected liver lobes were fixed in 10% buffered-formalin, paraffin-embedded and sectioned. Cross-sections (4-µm) were stained with hematoxylin/eosin (H&E) or Picrosirius Red using standard techniques. Images were acquired in an Axio Scope.A1 microscope (Carl Zeiss, Germany). Histopathological analysis was performed in a blinded manner in whole liver sections. The extent of fibrosis was calculated as a percentage of Sirius red-positive area of the total section area (excluding the Glisson capsule from quantification) by using the thresholding method in the green channel of Fiji-ImageJ software (http://imagej.net/Fiji). Toxic-induced hepatic fibrosis was also quantified using a semi-quantitative Ishak-modified scale from 0 (no fibrosis) to 4 (fibrous expansion of portal areas with marked portal-to-portal and portal-central bridging) as described (33,34). BDL-induced hepatic damage was also determined by quantifying the percentage of necrotic area in H&E-stained liver sections using Fiji-ImageJ software. Results show the mean value of 10 randomly selected areas per section and four sections per mouse.
Immunofluorescence analysis. Formalin-fixed liver sections were antigen-retrieved with 10 nM sodium citrate/0.05% Tween-20 (20 min/95ºC), blocked with 10% goat serum/1% bovine serum albumin (BSA, 2h/20°C) and marked with primary mouse anti-mouse αSMA antibody (dilution 1:1000 in PBS/1% BSA, 8h/4°C), and subsequently with secondary Alexa Fluor 488-goat anti-mouse IgG antibody (1:1000, 1h/20°C). See table S6 for antibodies’ information. Nuclei were DAPI-counterstained (Sigma-Aldrich, 1:1000, 5 min/20ºC). Samples in which primary antibody was omitted were used as negative controls. Sections were examined in an Olympus iX81 fluorescence microscope (Olympus Life Science, Hamburg, Germany) and the images were acquired (Olympus CellSens Imaging software) using the same parameters between samples. The percentage of αSMA-positive area was quantified using the Fiji-ImageJ software (10 areas/section, two sections/mouse).
HSCs cultured in coverslips as above were fixed with 4% paraformaldehyde/2% glucose (15 min/20ºC), blocked with 30 mM glycine (5 min/20ºC) and PBS/5% BSA/0.3% Triton X-100 (1h/20°C) and marked with primary mouse anti-mouse αSMA (1:1000, in PBS/1% BSA/0.3% Triton X-100, 8h/4ºC) and rabbit anti-mouse glial fibrillary acidic protein (GFAP, 1:300, 8h/4ºC) antibodies, and then with secondary Alexa Fluor 488-donkey anti-rabbit IgG and Alexa Fluor 594-goat anti-mouse IgG antibodies (1:1000 in PBS/1% BSA/0.1% Triton X-100, 1h/20°C). Nuclei were DAPI-counterstained and images were acquired as described above.
Gene expression analysis by real-time PCR. RNA was isolated from liver by homogenization in TriPure (Roche, Basilea, Switzerland) and from HSCs by lysis in EZNA HP Total RNA Kit (Omega Bio-Tek, Norcross, GA, USA) and treated with DNase 1 (Sigma-Aldrich) before reverse transcription (RevertAid First Strand cDNA Synthesis Kit, Thermo-Fisher) using random hexamers. SYBER green quantitative PCR (SensiFast Sybr No-Rox mix, Bioline, Germany) was performed on thermocycler (CFX96, Bio-Rad, Hercules, CA, USA) using the following conditions: 5 min/94°C followed by 40 cycles at 94°C/30 sec, 60ºC/30 sec and 72°C/30 sec (primers’ sequences are listed in table S7). The expression of each gene was normalized against the housekeeping gene RPLP0 in every PCR reaction and fold-change expression was estimated with Delta-Delta Ct method.
Next-generation transcriptome sequencing (RNAseq). RNA (1μg) from primary CST-/- andCST+/+ HSC cultures (two independent biological replicated experiments each, with RNA Integrity Number coefficients >9, using Bioanalyzer RNA 6000 Nano-chip, from Agilent, Santa Clara, CA, USA) was used to prepare mRNA libraries with TruSeq Stranded mRNA Library Prep Kit (Illumina, San Diego, CA, USA). Quality and size distributions of indexed mRNA libraries were validated by Bioanalyzer High Sensitivity DNA assay (Agilent) and the final libraries were pooled equimolecularly and diluted/denatured as recommended. The 40x2nt paired-end sequencing was conducted on an Illumina NextSeq-500 sequencer (highest output mode), producing 46,850,000 raw paired-reads on average. We used the miARma-Seq pipeline to analyze and calculate the differential expressed genes (DEGs) (35), with Mus musculusGencode version M25 genome-build (mmGRCm38.p6) for alignment of reads. Differential expression analysis was conducted by using edgeR package (36) and genes were normalized by trimmed mean of M-values method (37). We calculated reads per kb per million mapped reads and log2-counts per million per gene in each sample (36). To infer the replicability/similarity of RNA-sequencing samples, we used Principal Component analysis and Hierarchical Clustering of normalized samples (38,39). Genes giving False Discovery Rate (FDR) values <0.05 were marked as DEGs and Log2-FC was used to calculate the fold-change expression of each gene betweenCST-/- HSCs and CST+/+HSCs. To identify the effects of DEGs, functional enrichment study was carried out using the clusterProfiler Bioconductor package (40), in which DEGs were compared against all expressed genes in the RNA-seq assay. Gene Ontology (GO) terms (Biological process, Molecular function, Cellular components) were obtained from the Bioconductor Mus musculus database and associated to Entrez gene identifiers in anorgDB R object through the AnnotationForge package (with clusterProfiler).
LX2 cell culture. The human HSC line LX2 was obtained from Merck Millipore (Burlington, MA, USA, Cat. #SCC064) and maintained in complete DMEM medium (DMEM supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin, all from Gibco/Thermo-Fisher) at 37ºC and 5% CO2.
To determine the effect of cortistatin in protein and gene expression of fibrotic markers, 0.7-1 x 105 LX2 cells were cultured in 25 cm2-Nunc flasks (for protein expression) or in 6-well-Nunc plates (for gene expression) in complete DMEM medium for 24h (for protein) or until reach confluence of 70-80% (for gene). Cells were synchronized overnight with incomplete serum-free DMEM (supplemented with 1% penicillin/streptomycin) and then cultured in complete DMEM in the absence (unstimulated) or presence of TGFβ1 (5 ng/ml, from PeproTech, London, UK), with or without human cortistatin-17 (100 nM, added every 48h, from Bachem). When indicated, the pan-antagonist for sstr1-5 cyclosomatostatin (at 1 µM, Sigma-Aldrich), the specific GHSR1-antagonist GHRP6 (at 1 µM, Sigma-Aldrich) and the protein kinase A inhibitor H89 (at 100 nM, Sigma-Aldrich) were added to cultures simultaneously to cortistatin, and the Gαi-inhibitor pertussis toxin (at 250 ng/ml, Sigma-Aldrich) was administered during the phase ofLX2 synchronization. After 24h or 7d of culture, cells were lysed and protein extracts and analyzed as described below and total RNA were isolated with TriPure and analyzed by real-time qPCR as described for liver and HSCs.
To determine the effect of cortistatin in the presence of intracellular αSMA in activated LX2 , 5x104 cells were plated on glass-coverslips inserted in 24-well-Nunc plates and incubated in complete DMEM until reach 70% confluence. LX2 cells were synchronized overnight and then stimulated in complete DMEM with TGFβ1 (5 ng/ml) in the absence or presence of human cortistatin-17 (100 nM, added every 48h). After 7 days of culture, immunofluorescence analysis was conducted as described below.
Western blot analysis of LX2 cultures. LX2 cells were cultured and activated as described above and then lysed by incubation with lysis buffer containing 50 mM Tris-HCl pH 7.8, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 1% sodium deoxycholic acid, 0.1% SDS, 10 µg/ml-1 protease inhibitor cocktail (Sigma-Aldrich, cat#P8465) and phosphatase inhibitor (PhosSTOP, from Roche) for 2h at 4ºC and shaking. Lysates were centrifuged (21,000xg, 15 min, 4ºC) and supernatants containing proteins extracts (20 µg) were separated on 12% SDS-polyacrylamide gels and blotted onto polyvinylidene difluoride membranes (ImmMobilon-FL PVDF, Millipore) using a semidry system (transfer buffer: 25 mM pH 8.3, 192 mM glycine, 20% methanol). Membranes were blocked with TBS-T buffer (10 mM Tris, 150 mM NaCl, pH 7.5, 0.1% Tween-20) and 5% BSA for 1h at 20ºC and subsequently probed overnight at 4°C with pairs of primary rabbit anti-human CTGF or rabbit anti-human αSMA antibodies (both diluted at 1:1,000 in TBS-T/2% BSA) and mouse anti-human α-tubulin antibody (diluted at 1:8,000). Immunodetection of primary antibodies was performed by incubation with secondary antibodies labelled to the near-infrared fluorophores: goat anti-rabbit IRDye 800CW (green dye, for CTGF and αSMA) or goat anti-mouse IRDye 680RD (red dye, for α-tubulin) diluted at 1:20,000 in TBS-T/2% BSA/0.02% SDS for 1h at 20ºC. Images of blots were acquired in an Odyssey CLX (LI-COR Biosciences, Lincoln, NE, USA) and fluorescence intensities of specific bands corresponding to CTGF, αSMA or α-tubulin (used to normalize protein expression) were quantified using Fiji-ImageJ software.
Immunofluorescence analysis of LX2 cultures. LX2cells were cultured in coverslips as described above and then fixed with 4% paraformaldehyde/2% glucose during 15 min at 20ºC. After extensive washing with PBS, cells were incubated with 30 mM glycine for 5 min (to reduce autofluorescence), permeabilized with 0.1% Triton X-100 (15 min, 20ºC), and blocked with PBS/5% FBS/0.3% Triton X-100 (60 min, 20°C). Cell were then incubated with primary anti-human αSMA monoclonal antibody diluted at 1:1,000 in PBS/1% BSA/0.3% Triton X-100 at 4°C for 8h. After extensive washing with PBS/0.025% Triton X-100, samples were incubated with secondary Alexa Fluor 488-conjugated goat anti-mouse IgG antibody (60 min, 20°C, diluted at 1:1,000 in PBS/1% BSA/0.1% Triton X-100). Nuclei were DAPI-counterstained (1:500 in PBS, 5 min, 20ºC) and were mounted in Mowiol. Samples in which we omitted the primary antibodies were used as negative controls, showing in all cases lack of fluorescence signal. Samples were examined in an Olympus IX81 fluorescence microscope and the images acquired at 400X magnification (Olympus CellSens Imaging software) using the same parameters and ROI of at least eight independent experiments and fluorescence intensity and area were determined using the Fiji-ImageJ software.
Data and statistical analysis. All experiments are randomized and blinded. All data are expressed as mean±SEM. To control for unwanted sources of variation between individual experiments, data obtained from qPCR and western blot analysis of LX2 cell cultures were normalized to the mean of unstimulated cells. No data were excluded and outliers were included in data analysis and presentation. Group size is the number of independent animals or cell cultures, and statistical analysis was performed using these independent values. The data and statistical analysis comply with the recommendations on experimental design and analysis in pharmacology (41). In accordance with journal policy, statistical analysis was performed only when a minimum ofn  = 5 independent samples was acquired. We analysed data for statistical differences between groups using the unpaired Student’s t-test or the non-parametric Mann-Whitney U-test and, if appropriate, by Kruskal-Wallis analysis of variance test. Survival curves were analysed by the Kaplan-Meier log-rank test. All analyses were performed using GraphPad Prism v5.0 software (La Jolla, CA, USA). We considered P-values < 0.05 (two-tailed) as significant.